CALCULUS 2 • INTEGRATION FOUNDATIONS

Riemann Sums & Notation — Riemann Sums, Summation Notation, and Definite Integral Notation

How finite rectangular approximations evolve into the precise language of definite integrals.

Historical Context & Motivation

The problem of computing areas bounded by curves is among the oldest in mathematics, predating formal calculus by nearly two millennia. Ancient mathematicians recognized that curved regions could be approximated by collections of simpler shapes—polygons, triangles, and eventually rectangles—and that increasing the number of such shapes would yield increasingly accurate estimates. This intuition, refined across centuries of mathematical thought, culminated in the rigorous framework of the Riemann sum, which provides the formal bridge between discrete approximation and the continuous accumulation captured by the definite integral. Understanding this historical arc reveals why mathematicians needed precise notation for sums and limits, and why the integral sign itself carries deep geometric meaning.

~250 BCE
Archimedes' Method of Exhaustion
Archimedes approximated the area under a parabolic segment by inscribing triangles of decreasing size, effectively performing a geometric series summation. His method anticipated integration by using successive polygonal approximations that exhausted the region between the polygon and the curve.
1635
Cavalieri's Indivisibles
Bonaventura Cavalieri proposed that areas could be computed by summing infinitely many infinitesimally thin slices (indivisibles). Though lacking rigor by modern standards, his principle established the conceptual groundwork for thinking of area as an accumulation of thin strips.
1665–1686
Newton & Leibniz Formalize Calculus
Newton developed fluents and fluxions while Leibniz introduced the integral sign ∫ (an elongated 'S' for summa) and differential notation dx. Leibniz's notation became the standard because of its suggestive geometric meaning: the integral as a sum of infinitesimal rectangles f(x) dx.
1854
Riemann's Rigorous Definition
Bernhard Riemann, in his habilitation lecture, provided the first rigorous definition of the integral using tagged partitions. By defining integrability in terms of the convergence of finite sums as the partition norm tends to zero, Riemann placed the integral on solid logical foundations, replacing intuitive appeals to infinitesimals.
1902
Lebesgue Generalizes Integration
Henri Lebesgue extended Riemann's framework by partitioning the range instead of the domain, enabling integration of a far broader class of functions. Nevertheless, the Riemann integral remains the pedagogical cornerstone of introductory calculus and suffices for most applications in science and engineering.

The central question that Riemann sums address is deceptively simple: given a function f defined on a closed interval [a, b], how can we assign a precise numerical value to the signed area between the graph of f and the x-axis? The answer requires three ingredients—a systematic way to partition the interval, a compact notation to express the resulting sums, and a limiting process that transforms finite approximations into exact values. These are precisely the topics of this lesson.

Core Principles & Definitions

The Riemann sum framework rests on a small number of foundational ideas that connect finite arithmetic to the abstract concept of continuous accumulation. Before writing any formulas, it is essential to understand what a partition is, how sample points are selected, and why the width of the widest subinterval—the mesh or norm of the partition—governs the quality of the approximation. These definitions underpin everything that follows, from summation notation through the formal definition of the definite integral.

1

Partition of [a, b]

A partition P is a finite set of points a = x₀ < x₁ < x₂ < ⋯ < xₙ = b that divides [a, b] into n subintervals [xᵢ₋₁, xᵢ]. The width of the i-th subinterval is Δxᵢ = xᵢ − xᵢ₋₁. When all widths are equal we call it a regular partition with Δx = (b − a)/n.
2

Sample Points & Tags

For each subinterval [xᵢ₋₁, xᵢ], we choose a sample point (or tag) xᵢ* ∈ [xᵢ₋₁, xᵢ]. Common choices include the left endpoint (xᵢ* = xᵢ₋₁), right endpoint (xᵢ* = xᵢ), or midpoint (xᵢ* = (xᵢ₋₁ + xᵢ)/2). The function value f(xᵢ*) determines the height of the approximating rectangle.
3

Riemann Sum

The Riemann sum associated with a tagged partition is Sₙ = Σᵢ₌₁ⁿ f(xᵢ*)Δxᵢ. Each term f(xᵢ*)Δxᵢ represents the signed area of one rectangle; the full sum approximates the total signed area under the curve over [a, b].
4

Norm (Mesh) of a Partition

The norm ‖P‖ = max{Δxᵢ : 1 ≤ i ≤ n} measures the width of the largest subinterval. As ‖P‖ → 0 (which forces n → ∞ for bounded intervals), the Riemann sum converges to the definite integral for any choice of sample points, provided f is Riemann integrable.
5

Definite Integral as a Limit

The definite integral ∫ₐᵇ f(x) dx is defined as the limit of Riemann sums: lim_{‖P‖→0} Σᵢ₌₁ⁿ f(xᵢ*)Δxᵢ. When this limit exists and is independent of partition and tag choices, f is Riemann integrable on [a, b]. All continuous functions on closed bounded intervals are Riemann integrable.
KEY TAKEAWAY
Think of a Riemann sum the way a surveyor estimates the area of an irregularly shaped plot of land by laying down a grid of rectangular tiles. Each tile has a known width (Δxᵢ) and a height determined by measuring the terrain at some point within the tile (f(xᵢ*)). The sum of tile areas gives an estimate; using narrower tiles makes the estimate better. The definite integral is the exact area you would obtain if the tiles became infinitely narrow—the conceptual limit of 'perfect tiling.'

Visual Explanation — Left, Right, and Midpoint Sums

The following diagram illustrates a continuous function f(x) = x² + 1 on the interval [0, 4], partitioned into four equal subintervals. Three versions of the Riemann sum—left-endpoint, right-endpoint, and midpoint—are shown side by side so that you can visually compare how the choice of sample point affects the accuracy of the approximation. Notice that for an increasing function, left sums underestimate and right sums overestimate the true area, while midpoint sums tend to fall closer to the exact value.

Three Riemann sums for f(x) = x² + 1 on [0, 4] with n = 4 subintervals. The left sum L₄ = 24 underestimates, the right sum R₄ = 44 overestimates, and the midpoint sum M₄ = 33 is closest to the exact value of 76/3 ≈ 25.33. The dots indicate the sample points used for each rectangle's height.

Observe the interplay between sample-point choice and error. In the left-endpoint panel, each rectangle's height is determined by the function value at the left edge of its subinterval, so the tops of the rectangles consistently lie below the curve for this increasing function; the total area of the rectangles falls short of the true area. The right-endpoint panel exhibits the opposite behavior, with rectangle tops extending above the curve, producing an overestimate. The midpoint panel strikes a balance: the sample point sits at the center of each subinterval, and the resulting rectangles straddle the curve more closely. As the number of subintervals n increases, all three methods converge to the same value—the definite integral—but the midpoint rule typically converges fastest among these three elementary choices.

Mathematical Framework

We now formalize the three notational systems that form the backbone of integration theory: summation (sigma) notation for expressing finite sums compactly, the Riemann sum formula itself, and finally the definite integral as the limit of Riemann sums. Mastering these notations is essential not only for computing integrals but also for reading and writing proofs throughout analysis and applied mathematics.

Summation (Sigma) Notation

SIGMA NOTATION
Σᵢ₌ₘⁿ aᵢ = aₘ + aₘ₊₁ + aₘ₊₂ + ⋯ + aₙ
Here i is the index of summation, m is the lower limit, n is the upper limit, and aᵢ is the general term. The index i is a dummy variable—renaming it to j or k does not change the sum.

Essential Summation Identities

LINEARITY OF SUMMATION
Σᵢ₌₁ⁿ (caᵢ + dbᵢ) = c Σᵢ₌₁ⁿ aᵢ + d Σᵢ₌₁ⁿ bᵢ
Constants c and d factor out of the sum, and sums distribute over addition. This property is critical when expanding Riemann sums of polynomial functions.
CLOSED-FORM POWER SUMS
Σᵢ₌₁ⁿ i = n(n+1)/2 , Σᵢ₌₁ⁿ i² = n(n+1)(2n+1)/6 , Σᵢ₌₁ⁿ i³ = [n(n+1)/2]²
These identities allow us to evaluate Riemann sums in closed form before taking limits. The first formula is attributed to Gauss; the others follow from telescoping or induction arguments.

The Riemann Sum Formula (Regular Partition)

RIEMANN SUM — REGULAR PARTITION
Sₙ = Σᵢ₌₁ⁿ f(xᵢ*) · Δx , where Δx = (b − a)/n
For a regular partition, all subintervals have equal width Δx. The i-th partition point is xᵢ = a + iΔx. Common sample point choices are: left endpoint xᵢ* = a + (i−1)Δx, right endpoint xᵢ* = a + iΔx, midpoint xᵢ* = a + (i − ½)Δx.

The Definite Integral

DEFINITE INTEGRAL
∫ₐᵇ f(x) dx = lim_{n→∞} Σᵢ₌₁ⁿ f(xᵢ*) Δx
The integral sign ∫ is Leibniz's elongated S for summa. The lower limit a and upper limit b specify the interval. The integrand f(x) corresponds to the function being integrated, and dx indicates the variable of integration and reminds us of the infinitesimal width Δx → dx. The result is a number (not a function) representing the net signed area.
🔗 Notation Correspondence
Every piece of Riemann sum notation maps directly to integral notation: the sum Σ becomes ∫, the index step Δx becomes the differential dx, the general term f(xᵢ*) becomes the integrand f(x), the starting index i = 1 corresponds to the lower limit a, and the ending index i = n corresponds to the upper limit b. Recognizing these correspondences demystifies integral notation—it is simply the limiting case of sigma notation applied to infinitely many infinitesimally thin rectangles.

How Riemann Sums Converge to the Integral

A natural question arises: how quickly does a Riemann sum approach the definite integral as n increases, and does the choice of sample points matter asymptotically? The diagram below illustrates the convergence process for a single function using an increasing number of subintervals. Alongside the visual, the table quantifies the approximation error for each method at various values of n, making the convergence rate concrete.

Top panels: right Riemann sums for f(x) = x² on [0, 2] with n = 4 (amber) and n = 10 (emerald) subintervals. Bottom: convergence summary showing that the error decreases as 1/n, confirming that the right Riemann sum converges to the exact value ∫₀² x² dx = 8/3.
Convergence of three Riemann sum methods for f(x) = x² on [0, 2]
n (subintervals)Left Sum LₙRight Sum RₙMidpoint Sum MₙExact ∫₀² x² dx
41.7503.7502.6252.667
102.2803.0802.6602.667
502.5872.7472.66642.667
1002.6272.7072.666672.667
n → ∞8/38/38/38/3

The table makes a crucial point explicit: regardless of whether left, right, or midpoint sample points are used, all three methods converge to the same limiting value 8/3 as n → ∞. The midpoint rule exhibits visibly smaller errors at each stage because its error is proportional to 1/n² (second-order convergence), whereas left and right sums have error proportional to 1/n (first-order convergence). This distinction becomes practically important in numerical integration: for the same computational effort, the midpoint rule delivers significantly better accuracy than endpoint rules.

Worked Example — Evaluating a Definite Integral via Riemann Sums

We now compute ∫₁³ (2x + 3) dx by constructing a right Riemann sum with n subintervals, expressing the result using sigma notation and closed-form identities, and then taking the limit as n → ∞. This example demonstrates the full pipeline from partition to integral.

Evaluate ∫₁³ (2x + 3) dx Using a Right Riemann Sum
1
Step 1 — Set Up the Regular PartitionThe interval is [a, b] = [1, 3], so the subinterval width is Δx = (b − a)/n = (3 − 1)/n = 2/n. The partition points are xᵢ = a + iΔx = 1 + 2i/n for i = 0, 1, 2, …, n.
Δx = 2/n, xᵢ = 1 + 2i/n
2
Step 2 — Identify the Sample PointsUsing right endpoints, the sample point for the i-th subinterval is xᵢ* = xᵢ = 1 + 2i/n. We evaluate the integrand at this point: f(xᵢ*) = 2(1 + 2i/n) + 3 = 2 + 4i/n + 3 = 5 + 4i/n.
f(xᵢ*) = 5 + 4i/n
3
Step 3 — Write the Riemann Sum in Sigma NotationSubstituting into the Riemann sum formula: Rₙ = Σᵢ₌₁ⁿ f(xᵢ*)Δx = Σᵢ₌₁ⁿ (5 + 4i/n)(2/n) = (2/n) Σᵢ₌₁ⁿ (5 + 4i/n). Distributing: Rₙ = (2/n)[Σᵢ₌₁ⁿ 5 + (4/n) Σᵢ₌₁ⁿ i].
Rₙ = (2/n)[5n + (4/n) · n(n+1)/2]
4
Step 4 — Apply Closed-Form Summation IdentitiesWe use Σᵢ₌₁ⁿ 5 = 5n and Σᵢ₌₁ⁿ i = n(n+1)/2. Substituting: Rₙ = (2/n)[5n + (4/n) · n(n+1)/2] = (2/n)[5n + 2(n+1)] = (2/n)[5n + 2n + 2] = (2/n)(7n + 2) = 14 + 4/n.
Rₙ = 14 + 4/n
5
Step 5 — Take the Limit as n → ∞As n → ∞, the term 4/n → 0, so lim_{n→∞} Rₙ = lim_{n→∞} (14 + 4/n) = 14. Therefore ∫₁³ (2x + 3) dx = 14.
∫₁³ (2x + 3) dx = 14
6
Step 6 — Verification via GeometryThe function f(x) = 2x + 3 is linear, so the region under its graph from x = 1 to x = 3 is a trapezoid. The parallel sides have lengths f(1) = 5 and f(3) = 9, and the width is 2. The trapezoid area is (1/2)(5 + 9)(2) = 14, confirming the Riemann sum result.
Trapezoid area = 14 ✓

Comparing Riemann Sum Methods

While all Riemann sum methods converge to the same definite integral for integrable functions, they differ substantially in practical accuracy for a given number of subintervals, computational complexity, and the ease with which they lend themselves to error analysis. The following table compares the four most common Riemann sum variants along these dimensions, helping you choose the appropriate method for a given context—whether that context is a hand calculation, a numerical simulation, or a theoretical proof.

Comparison of Riemann sum methods by sample point, error rate, and practical considerations
MethodSample Point xᵢ*Error OrderStrengthsLimitations
Left Endpointxᵢ₋₁ = a + (i−1)ΔxO(1/n)Simple to implement; provides lower bound for increasing functionsSlow convergence; systematic bias for monotonic functions
Right Endpointxᵢ = a + iΔxO(1/n)Simple to implement; provides upper bound for increasing functionsSlow convergence; opposite bias to left sums
Midpointa + (i − ½)ΔxO(1/n²)Cancellation of first-order errors; good accuracy per function evaluationSlightly more complex; no guaranteed bound direction
Trapezoidal(Lₙ + Rₙ)/2O(1/n²)Averages left and right; same order as midpointRequires two endpoint evaluations per subinterval
KEY TAKEAWAY
Think of the different Riemann sum methods as different GPS sampling strategies for mapping terrain elevation. Left and right endpoint sampling is like measuring elevation only at the edges of each grid cell—cheap but biased. Midpoint sampling is like measuring at the center of each cell—more representative with no extra data points. The trapezoidal rule averages the two edge measurements, achieving similar accuracy to midpoint. In both surveying and integration, the choice of sampling strategy determines how quickly the map converges to the true landscape.

Connection to the Fundamental Theorem & Advanced Integration

Evaluating definite integrals via Riemann sums is conceptually illuminating but computationally demanding for all but the simplest functions. The Fundamental Theorem of Calculus (FTC) provides a vastly more efficient alternative by linking the definite integral to antiderivatives. Nevertheless, understanding Riemann sums remains essential: the FTC's proof relies on them, numerical methods generalize them, and they form the basis for defining integrals in higher dimensions and for abstract measure-theoretic settings.

Riemann sum definition vs. Fundamental Theorem evaluation
AspectRiemann Sum ApproachFTC / Antiderivative Approach
Definition∫ₐᵇ f(x) dx = lim Σ f(xᵢ*)Δx∫ₐᵇ f(x) dx = F(b) − F(a), where F′ = f
When to UseBuilding intuition; numerical approximation; proving the FTCWhen an antiderivative F is known in closed form
Computational CostRequires n function evaluations and a limit; algebraically intensiveTypically one antidifferentiation and two evaluations
GeneralizabilityExtends to higher dimensions, numerical quadrature, Lebesgue integralsLimited to functions with elementary antiderivatives
Error ControlError bounds via remainder terms; adaptive refinement possibleExact (no approximation error); symbolic computation

Looking ahead in your study of Calculus 2, Riemann sums reappear in several important contexts. The integral definition of arc length arises as the limit of sums of chord lengths. Volumes of revolution via the disk, washer, and shell methods all begin by summing the volumes of thin cylindrical slices—essentially Riemann sums in three dimensions. Improper integrals extend the framework to unbounded intervals or integrands, requiring an additional limiting process. Even probability theory relies on the Riemann integral (or its Lebesgue generalization) to define continuous probability distributions. In each case, the conceptual template is the same: approximate, sum, take a limit.

Practice Problems

PROBLEM 1CONCEPTUAL
Explain why, for a continuous function f on [a, b], the definite integral ∫ₐᵇ f(x) dx is independent of the choice of sample points xᵢ* in the Riemann sum definition. What mathematical property of continuous functions on closed intervals guarantees this?
PROBLEM 2BASIC CALCULATION
Using sigma notation and the closed-form identity Σᵢ₌₁ⁿ i = n(n+1)/2, evaluate the right Riemann sum Rₙ = Σᵢ₌₁ⁿ f(xᵢ)Δx for f(x) = 3x on [0, 4] with n subintervals. Then take the limit as n → ∞ to find ∫₀⁴ 3x dx.
PROBLEM 3INTERMEDIATE
Compute ∫₀² (x² − 1) dx using a left Riemann sum with n subintervals. Express the sum in sigma notation, apply the identities Σᵢ₌₁ⁿ 1 = n and Σᵢ₌₁ⁿ i² = n(n+1)(2n+1)/6, carefully handle the index shift for left endpoints, and evaluate the limit.
PROBLEM 4APPLIED
A chemical reactor's temperature T(t) (in °C) over a 6-hour window [0, 6] is recorded at hourly intervals: T(0) = 80, T(1) = 85, T(2) = 92, T(3) = 96, T(4) = 91, T(5) = 84, T(6) = 78. Using the trapezoidal rule (the average of left and right Riemann sums), estimate ∫₀⁶ T(t) dt. Interpret the physical meaning of this integral and specify its units.
PROBLEM 5CRITICAL THINKING
Prove directly from the Riemann sum definition that the definite integral is linear: if f and g are integrable on [a, b] and c₁, c₂ are constants, then ∫ₐᵇ [c₁f(x) + c₂g(x)] dx = c₁∫ₐᵇ f(x) dx + c₂∫ₐᵇ g(x) dx. State clearly where you use properties of limits and of finite sums.

Lesson Summary

A Riemann sum approximates the area under a curve by partitioning the interval [a, b] into n subintervals, selecting a sample point xᵢ* in each subinterval, and summing the products f(xᵢ*)Δxᵢ. Summation (sigma) notation Σᵢ₌₁ⁿ provides the compact language for expressing these sums, and closed-form identities for power sums enable algebraic evaluation before taking limits. Common sample point choices—left endpoint, right endpoint, and midpoint—all converge to the same limit for integrable functions, though they differ in convergence rate.

The definite integral ∫ₐᵇ f(x) dx is defined as the limit of Riemann sums as the partition norm ‖P‖ → 0. Leibniz's integral notation directly mirrors sigma notation: Σ becomes ∫, Δx becomes dx, and the index bounds become the limits of integration a and b. While the Fundamental Theorem of Calculus provides a far more efficient evaluation technique via antiderivatives, the Riemann sum definition remains foundational—it underpins the FTC's proof, numerical quadrature methods, and the generalization of integration to higher dimensions and abstract measure spaces.

Varsity Tutors • Calculus 2 • Riemann Sums & Notation