Guide

Riemann Sums: Area To Definite Integrals

Was this lesson helpful?

Dive Deeper

Generate a follow-up sub-lesson on any aspect of this topic

Guide complete

Riemann Sums: Area To Definite Integrals

Learn to approximate area under a curve with rectangles, predict whether your estimate is high or low, and see how shrinking rectangle widths turns a sum into the definite integral notation you use in calculus.

You want the area under a curve, but the shape is not a neat rectangle or triangle. Riemann sums handle that by chopping the interval into thin slices and pretending each slice is a rectangle whose height comes from the function. The trick is that you are not trying to be perfect at first. You are trying to be systematic, so you can improve the estimate whenever you want and understand what the definite integral is really measuring.

See what these rectangle estimates look like on the same curve.

What a Riemann sum is

A Riemann sum is an area estimate made by adding up areas of rectangles under (or over) the graph of $y=f(x)$ on an interval $[a,b]$ .

You choose some rectangles, compute each rectangle’s area as base times height, then add. The base is a small width along the $x$ -axis. The height is the function value at a chosen point in that slice. If the curve is above the axis, you are estimating positive area. If parts dip below, the rectangles contribute negative area, which is why integrals are often described as signed area.

Rectangles first
The power of Riemann sums is not the rectangles. It is the repeatable recipe that lets you make them thinner and watch the estimate stabilize.

Build a Riemann sum step by step

Start with the interval $[a,b]$ and split it into $n$ equal pieces. That split is called a partition. Each piece has width

\Delta x=\frac{b-a}{n}.

Now each rectangle has base $\Delta x$ . The only freedom left is the height.

A sample point $x_i^*$ is the point you pick in the $i$ th subinterval to decide the rectangle’s height $f(x_i^*)$ . The star reminds you it is chosen by a rule, not automatically the left endpoint.

Once you have sample points, the sum looks like

\sum_{i=1}^{n} f(x_i^*)\Delta x.

Try constructing the pieces and watching the total build.

Left right and midpoint sums

Three common sample point rules give three common Riemann sums.

Left sum uses the left endpoint of each subinterval for $x_i^*$
Right sum uses the right endpoint of each subinterval for $x_i^*$
Midpoint sum uses the midpoint of each subinterval for $x_i^*$

They differ because a slanted curve cannot be captured by a single rectangle per slice. If $f$ is increasing, left rectangles tend to sit below the curve and right rectangles tend to stick above it. If $f$ is decreasing, that flips.

Explore how the methods line up and when each tends to overestimate.

Bias clue
If the function rises left to right, left sums usually underestimate and right sums usually overestimate. For decreasing functions, swap the conclusion.

Refining the partition

Using more rectangles means using a smaller $\Delta x$ . Visually, each rectangle hugs the curve more closely. Numerically, each rectangle’s error shrinks, and there are more rectangles, but the total error often shrinks overall because the curve is not changing much within a tiny slice.

The key idea is convergence. As $n$ grows, the Riemann sum can settle toward a single number, even though you are still using rectangles. That settling value is what the definite integral is.

Move $n$ up and down and watch the estimate approach a limit.

From sums to the definite integral

The definite integral is the number you get when you push a Riemann sum to its limit with infinitely many infinitely thin rectangles.

Formally, for a function that behaves nicely on $[a,b]$ ,

\int_a^b f(x)\,dx=\lim_{n\to\infty}\sum_{i=1}^{n} f(x_i^*)\Delta x.

The symbol $\int$ is a stretched S for sum. The $dx$ tells you the variable and hints at the width pieces. The bounds $a$ and $b$ tell you where you start and stop along the $x$ axis.

The subtle part is that the exact choice of sample points stops mattering in the limit for well behaved functions. Left, right, and midpoint sums may differ for small $n$ , but they can converge to the same value as $n\to\infty$ .

See how the sigma notation turns into the integral notation.

Worked examples to build intuition

Example 1 constant function

Let $f(x)=3$ on $[0,2]$ . Every rectangle has height 3 no matter where you sample. With $n$ rectangles, $\Delta x=\frac{2-0}{n}=\frac{2}{n}$ , so each rectangle’s area is $3\cdot\frac{2}{n}$ . Adding $n$ identical rectangles gives

\sum_{i=1}^{n}3\Delta x=n\cdot 3\cdot\frac{2}{n}=6.

No approximation happened. The curve was already flat.

Example 2 linear function

Let $f(x)=x$ on $[0,1]$ . With $n$ rectangles, $\Delta x=\frac{1}{n}$ .

Left endpoints are $x_{i-1}=\frac{i-1}{n}$ , so the left sum is

L_n=\sum_{i=1}^{n}\left(\frac{i-1}{n}\right)\frac{1}{n} =\frac{1}{n^2}\sum_{i=1}^{n}(i-1) =\frac{1}{n^2}\cdot\frac{(n-1)n}{2} =\frac{n-1}{2n}.

Right endpoints are $x_i=\frac{i}{n}$ , so the right sum is

R_n=\sum_{i=1}^{n}\left(\frac{i}{n}\right)\frac{1}{n} =\frac{1}{n^2}\sum_{i=1}^{n} i =\frac{1}{n^2}\cdot\frac{n(n+1)}{2} =\frac{n+1}{2n}.

So $L_n<\frac{1}{2}<R_n$ and both squeeze toward $\frac{1}{2}$ as $n\to\infty$ . That matches $\int_0^1 x\,dx=\frac{1}{2}$ .

Example 3 curved function

Let $f(x)=x^2$ on $[0,1]$ . The curve bends upward, so midpoint sums usually beat left and right for the same $n$ because the midpoint height often tracks the slice’s average height better.

You can still write a sum, for example the right sum

R_n=\sum_{i=1}^{n}\left(\frac{i}{n}\right)^2\frac{1}{n} =\frac{1}{n^3}\sum_{i=1}^{n} i^2.

The exact limit uses the known formula $\sum i^2=\frac{n(n+1)(2n+1)}{6}$ , which leads to $R_n\to\frac{1}{3}$ . So $\int_0^1 x^2\,dx=\frac{1}{3}$ .

Work through a few choices and notice what changes when you change the method or $n$ .

Curvature matters
When the graph bends, endpoints can miss the slice’s typical height. Midpoints often reduce that miss because they sample where the slice is centered.

A practical mental checklist

Before you compute, decide what you need from the estimate.

Goal Decide if you want a quick estimate, a guaranteed bound, or the exact integral value.
Method For monotone functions, left and right give natural low and high bounds. Midpoint is often a strong default for accuracy.
Width Compute $\Delta x=\frac{b-a}{n}$ first and write it once. Most mistakes are really $\Delta x$ mistakes.
Units Area has units of $y$ times units of $x$ . If that feels wrong, something is off.
Error sense If $f$ is increasing, expect $L_n\le\int_a^b f(x)\,dx\le R_n$ . If decreasing, reverse it.

When you have both a left and right sum, the true integral is trapped between them, so $R_n-L_n$ is a simple way to gauge how tight your estimate is. If you need it tighter, increase $n$ until that gap is small enough for your purpose.

Pick one function you can sketch, like $f(x)=x^2$ on $[0,1]$ , and compute $L_4$ and $R_4$ by hand. If your numbers do not bracket $\frac{1}{3}$ , the issue is almost always an endpoint or $\Delta x$ slip.

Was this lesson helpful?

Dive Deeper

Generate a follow-up sub-lesson on any aspect of this topic

Guide complete

Riemann Sums: Area To Definite Integrals

What a Riemann sum is

A Riemann sum is an area estimate made by adding up areas of rectangles under (or over) the graph of $y=f(x)$ on an interval $[a,b]$ .

Rectangles first
The power of Riemann sums is not the rectangles. It is the repeatable recipe that lets you make them thinner and watch the estimate stabilize.

Build a Riemann sum step by step

Start with the interval $[a,b]$ and split it into $n$ equal pieces. That split is called a partition. Each piece has width

\Delta x=\frac{b-a}{n}.

Now each rectangle has base $\Delta x$ . The only freedom left is the height.

Once you have sample points, the sum looks like

\sum_{i=1}^{n} f(x_i^*)\Delta x.

Try constructing the pieces and watching the total build.

Left right and midpoint sums

Three common sample point rules give three common Riemann sums.

Left sum uses the left endpoint of each subinterval for $x_i^*$
Right sum uses the right endpoint of each subinterval for $x_i^*$
Midpoint sum uses the midpoint of each subinterval for $x_i^*$

Explore how the methods line up and when each tends to overestimate.

Bias clue
If the function rises left to right, left sums usually underestimate and right sums usually overestimate. For decreasing functions, swap the conclusion.

Refining the partition

The key idea is convergence. As $n$ grows, the Riemann sum can settle toward a single number, even though you are still using rectangles. That settling value is what the definite integral is.

Move $n$ up and down and watch the estimate approach a limit.

From sums to the definite integral

The definite integral is the number you get when you push a Riemann sum to its limit with infinitely many infinitely thin rectangles.

Formally, for a function that behaves nicely on $[a,b]$ ,

\int_a^b f(x)\,dx=\lim_{n\to\infty}\sum_{i=1}^{n} f(x_i^*)\Delta x.

The symbol $\int$ is a stretched S for sum. The $dx$ tells you the variable and hints at the width pieces. The bounds $a$ and $b$ tell you where you start and stop along the $x$ axis.

See how the sigma notation turns into the integral notation.

Worked examples to build intuition

Example 1 constant function

\sum_{i=1}^{n}3\Delta x=n\cdot 3\cdot\frac{2}{n}=6.

No approximation happened. The curve was already flat.

Example 2 linear function

Let $f(x)=x$ on $[0,1]$ . With $n$ rectangles, $\Delta x=\frac{1}{n}$ .

Left endpoints are $x_{i-1}=\frac{i-1}{n}$ , so the left sum is

L_n=\sum_{i=1}^{n}\left(\frac{i-1}{n}\right)\frac{1}{n} =\frac{1}{n^2}\sum_{i=1}^{n}(i-1) =\frac{1}{n^2}\cdot\frac{(n-1)n}{2} =\frac{n-1}{2n}.

Right endpoints are $x_i=\frac{i}{n}$ , so the right sum is

R_n=\sum_{i=1}^{n}\left(\frac{i}{n}\right)\frac{1}{n} =\frac{1}{n^2}\sum_{i=1}^{n} i =\frac{1}{n^2}\cdot\frac{n(n+1)}{2} =\frac{n+1}{2n}.

So $L_n<\frac{1}{2}<R_n$ and both squeeze toward $\frac{1}{2}$ as $n\to\infty$ . That matches $\int_0^1 x\,dx=\frac{1}{2}$ .

Example 3 curved function

Let $f(x)=x^2$ on $[0,1]$ . The curve bends upward, so midpoint sums usually beat left and right for the same $n$ because the midpoint height often tracks the slice’s average height better.

You can still write a sum, for example the right sum

R_n=\sum_{i=1}^{n}\left(\frac{i}{n}\right)^2\frac{1}{n} =\frac{1}{n^3}\sum_{i=1}^{n} i^2.

The exact limit uses the known formula $\sum i^2=\frac{n(n+1)(2n+1)}{6}$ , which leads to $R_n\to\frac{1}{3}$ . So $\int_0^1 x^2\,dx=\frac{1}{3}$ .

Work through a few choices and notice what changes when you change the method or $n$ .

Curvature matters
When the graph bends, endpoints can miss the slice’s typical height. Midpoints often reduce that miss because they sample where the slice is centered.

A practical mental checklist

Before you compute, decide what you need from the estimate.

Goal Decide if you want a quick estimate, a guaranteed bound, or the exact integral value.
Method For monotone functions, left and right give natural low and high bounds. Midpoint is often a strong default for accuracy.
Width Compute $\Delta x=\frac{b-a}{n}$ first and write it once. Most mistakes are really $\Delta x$ mistakes.
Units Area has units of $y$ times units of $x$ . If that feels wrong, something is off.
Error sense If $f$ is increasing, expect $L_n\le\int_a^b f(x)\,dx\le R_n$ . If decreasing, reverse it.

Guide complete

Related content

Riemann Sums: Area To Definite Integrals

What a Riemann sum is

Build a Riemann sum step by step

Left right and midpoint sums

Refining the partition

From sums to the definite integral

Worked examples to build intuition

Example 1 constant function

Example 2 linear function

Example 3 curved function

A practical mental checklist

Guide complete

Related content

Riemann Sums: Area To Definite Integrals

What a Riemann sum is

Build a Riemann sum step by step

Left right and midpoint sums

Refining the partition

From sums to the definite integral

Worked examples to build intuition

Example 1 constant function

Example 2 linear function

Example 3 curved function

A practical mental checklist