Calculus › Section E · Integrals › Topic E.2
Area and sigma notation
How do you find the area of a region with a curved edge? Slice it into thin rectangles, add their areas, and let the slices get infinitely thin. Sigma notation is the bookkeeping that makes the limit computable.
Two problems, one strategy
Integral calculus starts with a pair of questions that turn out to be twins. What is the area under a curve \(y = f(x)\) between two points? And: if a car's speed keeps changing, how far does it travel? At a constant 50 km/h for 3 hours a car covers \(50 \times 3 = 150\) km, which is the area of the rectangle under its flat velocity graph. In general, on a graph of velocity against time, distance is area under the curve, so both questions submit to one strategy: slice the interval into narrow strips, pretend the function is constant on each strip, add up the pieces, and shrink the strips toward zero width. This topic builds the machinery of the first three verbs, slice, pretend, add, and introduces the notation that makes the fourth, shrink, exact.
Building the rectangles
Split \([a, b]\) into \(n\) strips of equal width \(\Delta x = \frac{b - a}{n}\). On each strip, erect a rectangle; its height is the function value at a chosen sample point of the strip. Three popular choices: the left endpoint, the right endpoint, or the midpoint, giving the left, right and midpoint sums. On strip \(i\) the right endpoint is \(x_i = a + i\,\Delta x\) and the left endpoint is \(a + (i - 1)\,\Delta x\). The sum is always the same shape: heights times widths, added up, \(\sum f(x_i^*)\,\Delta x\). For a curved graph it is only an estimate; the exact answer arrives in the limit as the strips get infinitely thin.
Worked example: bracketing the area under x²
Estimate the area under \(f(x) = x^2\) on \([0, 3]\) with \(n = 3\) rectangles, so \(\Delta x = 1\). Left endpoints \(0, 1, 2\) give heights \(0, 1, 4\), and the left sum is \(5\). Right endpoints \(1, 2, 3\) give heights \(1, 4, 9\), and the right sum is \(14\). Because \(x^2\) is increasing here, the rectangles of the left sum sit under the curve (underestimate) and those of the right sum poke above it (overestimate), and the true area, which we will compute exactly below, lies strictly between \(5\) and \(14\). Even this coarse picture already traps the answer.
Why the bracketing works
For an increasing function, the left endpoint of each strip is its lowest point and the right endpoint its highest, so left sums lie below the true area and right sums above it; decreasing functions swap the roles. Either way, monotone functions get bracketed between two computable numbers, and the gap between them is a built-in error bound. Midpoint sums usually do much better than either side, because on each strip the overshoot on one half roughly cancels the undershoot on the other.
Sigma notation
Long sums demand a shorthand with the Greek capital sigma: \(\sum_{i=1}^{n} a_i\) means "compute \(a_i\) for each \(i\) from \(1\) to \(n\), then add". The index \(i\) is a dummy variable, and \(\sum_{i=1}^{3} i^2 = 1 + 4 + 9 = 14\) is the whole idea in one line. Two algebraic facts let you compress sums before adding: sums split over addition, and a constant factor pulls out front, \(\sum c a_i = c \sum a_i\). These mirror the corresponding rules you will use for integrals later, because an integral is, at root, a limit of exactly these sums.
The power-sum formulas
A few sums have closed forms worth owning:
The first says a constant added \(n\) times. The second is Gauss's schoolroom trick, pairing first and last terms. The third is proved by induction in an algebra course. These turn a right-sum expression for \(x^2\), which contains \(\sum i^2\), into a closed formula in \(n\), and once the formula is closed, the limit as \(n \to \infty\) is ordinary Section-D algebra.
Worked example: exact area as a limit
Compute the exact area under \(x^2\) on \([0, 3]\) as the limit of right sums. With \(n\) strips, \(\Delta x = \frac{3}{n}\) and the right endpoints are \(x_i = \frac{3i}{n}\). The sum is
Now take \(n \to \infty\): the ratio \(n(n+1)(2n+1)/n^3\) tends to \(2\), the ratio of the leading terms, so the area is \(\frac{27}{6} \cdot 2 = 9\). Exactly \(9\), inside our bracket \(5 < 9 < 14\) from the coarse estimate, as it must be. E.4 will reach this same 9 in one line.
What the limit buys
The limit did two things at once. It produced an exact number, \(9\), rather than an estimate, and it showed that the rectangle sums really do settle on a single value. But notice the cost: we needed the closed form for \(\sum i^2\) and a careful limit. For \(x^3\) you would need \(\sum i^3\); for \(\sin x\), a trig sum you do not have. The method is exact yet scales terribly, and that is precisely the problem the fundamental theorem of calculus solves in E.4 by replacing all of this with antiderivatives.
Areas in the field
Surveyor's assistants and earthwork estimators really do this. To price the fill needed for a building plot with a curved boundary, measure the depth of the ground profile at equal spacings along a line; each reading times the spacing is one rectangle of cross-section, and their sum estimates the cross-sectional area, with finer spacings tightening the answer. The same strip-sum logic estimates rainfall accumulation from hourly rate readings and energy use from a sampled power meter. Wherever a quantity accumulates under a curve you can only sample, sigma bookkeeping is how the estimate is made, and how its error is reasoned about.
Common mistakes to avoid
The usual slips: using \(\Delta x = \frac{1}{n}\) on an interval of length other than \(1\), forgetting the \(\frac{b-a}{n}\); including both endpoints in a left sum, which adds a phantom rectangle; forgetting the \(\Delta x\) factor, so the sum reports heights instead of areas; mishandling the index start, since \(\sum_{i=1}^{n}\) with right endpoints and \(\sum_{i=0}^{n-1}\) with left endpoints are easy to swap; and stopping with a closed sum in \(n\) when the question asks for the limit. One more: with non-monotone functions, left and right sums do not reliably bracket, so do not claim a sandwich you do not have.
Key takeaways
Curved-edge area, and its twin, distance travelled at varying speed, come from one strategy: slice \([a, b]\) into \(n\) strips of width \(\Delta x = \frac{b-a}{n}\), sample a height per strip, add, and take the limit. Left and right sums bracket the answer for monotone functions; midpoints usually hug it. Sigma notation plus the power-sum formulas turn brute-force addition into closed expressions, and the limit delivers exactness, like the \(9\) hiding under \(x^2\) on \([0, 3]\). These limits-of-sums get an official name and symbol next: the definite integral, and the famous theorem that computes them in one line.
Flashcards
Flashcard
Slice, pretend constant, add, shrink
Test your understanding
Question 1 of 5
For the interval [1, 5] cut into n = 4 strips, what is Δx?
Question 2 of 5
What is the left sum for f(x) = x² on [0, 3] with n = 3?
Question 3 of 5
What is Σ from i = 1 to 10 of i?
Question 4 of 5
For an increasing function, the left sum…
Question 5 of 5
Taking the limit of right sums for x² on [0, 3] as n → ∞ gives which area?
Standards: 6
