1. Limits and Continuity of Functions
1.1 Limits of Functions
A function limit describes what values of a function approach near a point, independently of whether the function is defined at that point. Start from a general neighborhood definition: let
A finite real limit at a finite point can be written in epsilon-delta form:
The exclusion
Useful prerequisites:
| Concept | Meaning |
|---|---|
| Neighborhood of | A set of points close to |
| Cluster point of | Every neighborhood of |
| Punctured neighborhood | A neighborhood of |
| One-sided neighborhood | A neighborhood restricted to the left or right side of |
The special cases of limits are best remembered as the same neighborhood idea with the target or source allowed to be infinite.
| Case | Definition pattern |
|---|---|
| Finite limit at finite point | For every small target error |
| Infinite limit at finite point | For every large bound |
| Finite limit at infinity | For every |
| Infinite limit at infinity | For every |
| One-sided finite limit | The same finite-point definition, but using only |
For a two-sided limit at a finite point, both one-sided limits must exist and be equal:
The sequential characterization says that a function limit can be tested by sequences. If
if and only if for every sequence
Examples:
| Function | Point | Reason |
|---|---|---|
| Has finite limit | ||
| Goes to | ||
| No two-sided infinite limit in the signed sense: right side goes to | ||
| No limit: sequences can force values toward different accumulation values. |
Limits interact well with algebraic operations. If
The same principle applies to powers and roots under the usual domain restrictions. These rules are not a substitute for checking denominator limits: a quotient rule is valid only when the denominator limit is nonzero. Forms such as
What to Emphasize in an Oral Answer
- Define the limit at a cluster point using the punctured-neighborhood or epsilon-delta condition; emphasize that
is irrelevant. - Treat finite, infinite, one-sided, and
limits as the same neighborhood idea with changed source or target neighborhoods. - State that a two-sided finite limit exists exactly when the two one-sided limits exist and are equal.
- Include the sequential/Heine characterization, especially its use in disproving limits with two approaching sequences.
- Mention algebraic limit laws, but also the caveat about nonzero denominator limits and indeterminate forms.
Suggested answer
A limit of a function describes the value approached by
The same definition gives the special cases by changing the source or target neighborhoods. If the limit is infinite at a finite point, then near
The sequential characterization is another way to say the same thing:
Finally, limits are compatible with algebraic operations. Sums, scalar multiples, products, and quotients have the expected limits, as long as the denominator limit is nonzero and the expression is defined near the point. When a direct operation produces an indeterminate form such as
1.2 Continuity of Functions
Continuity at a point is defined directly by epsilon-delta language. A function
If
The sequential characterization of continuity follows the limit version. A function is continuous at
Continuity is preserved by algebraic operations. If
The preservation of sign theorem is important for exam answers. If
Continuity at endpoints is understood one-sidedly. For a function on
Discontinuities are points where the function is not continuous. The main types are:
| Type | What happens | Example pattern |
|---|---|---|
| Removable discontinuity | The finite limit exists, but | |
| Jump discontinuity | Both one-sided finite limits exist, but they are different. | Step function at the jump point. |
| Infinite discontinuity | At least one one-sided limit is infinite. | |
| Oscillatory / essential discontinuity | The limiting behavior does not settle to a finite or infinite one-sided value. | |
| Dirichlet-type discontinuity | Values depend on dense subsets in incompatible ways. |
A function can often be made continuous at a removable discontinuity by assigning the missing limit value. A jump, infinite, or oscillatory discontinuity cannot be repaired by changing only the value at the point.
What to Emphasize in an Oral Answer
- Define continuity at
by epsilon-delta language and, at cluster points, by . - Contrast continuity with a plain limit: for continuity the value at
matters and must equal the limiting value. - State the sequential characterization: every sequence in the domain converging to
must have images converging to . - Mention stability under algebraic operations, quotients with nonzero denominator, and composition, plus local sign preservation.
- Distinguish endpoint continuity as one-sided continuity when the domain is an interval.
- Classify discontinuities by failure mode: removable, jump, infinite, oscillatory/essential; only removable ones are repaired by redefining
.
Suggested answer
Continuity at a point means that small changes in the input produce small changes in the function value. Formally,
There is also a sequential characterization:
Continuous functions are stable under the usual algebraic operations. Sums, differences, products, scalar multiples, and quotients by a nonzero continuous denominator are continuous, and compositions of continuous functions are continuous. A useful local consequence is preservation of sign: if
Discontinuities are classified by what goes wrong. If the finite limit exists but the value is missing or wrong, the discontinuity is removable. If the two one-sided finite limits exist but differ, it is a jump discontinuity. If the function tends to infinity near the point, it is an infinite discontinuity. If the values oscillate without approaching any limit, it is an essential or oscillatory discontinuity.
1.3 Continuous Functions on Compact Sets
Sequential compactness means that every sequence in the set has a convergent subsequence whose limit is still in the set. In the real line, compact sets are exactly closed and bounded sets.
The core compact-set properties of continuous functions are:
| Theorem | Statement | Consequence |
|---|---|---|
| Boundedness theorem | If | A continuous function on |
| Weierstrass theorem | If | There exist |
| Bolzano-Darboux / intermediate value theorem | If | Continuous functions cannot jump over intermediate values. |
| Bolzano zero theorem | If | A sign change forces a zero. |
| Heine theorem | If | The same |
The important statements are Weierstrass, Bolzano's zero theorem, Heine's theorem, the boundedness theorem, and the Bolzano-Darboux theorem.
Uniform continuity is stronger than pointwise continuity. Pointwise continuity allows
The bisection method is the numerical version of Bolzano's zero theorem. Suppose
- Set
, . - Let
. - If
, stop. - Otherwise choose the half-interval where the sign change remains:
- if
, set ; - otherwise set
.
- if
- Repeat until the interval is short enough.
The invariant is that each interval contains a root and has half the length of the previous interval. After
What to Emphasize in an Oral Answer
- Define compactness on the real line as closed and bounded, equivalently sequential compactness.
- State boundedness and Weierstrass: a continuous function on a compact set is bounded and attains its minimum and maximum.
- State the intermediate value/Bolzano-Darboux theorem and the zero theorem as the sign-change corollary.
- Include Heine's theorem and contrast pointwise continuity with uniform continuity through the dependence of
. - Explain bisection as the numerical use of the zero theorem: preserve a sign-change bracket and halve the error interval each step.
Suggested answer
On the real line, compactness means closed and boundedness, and equivalently every sequence has a convergent subsequence with limit still in the set. Continuous functions behave especially well on compact sets. First, a continuous function on a compact set is bounded. Second, by the Weierstrass theorem it attains its bounds, so it has both a minimum and a maximum. Third, by the Bolzano-Darboux or intermediate value theorem, a continuous function on an interval takes every value between two of its values; in particular, if
These theorems express the same geometric idea: continuity prevents jumps, and compactness prevents escape to infinity or to a missing endpoint. The Heine theorem adds that a continuous function on a compact set is uniformly continuous, so one input tolerance can be chosen for the whole domain.
The bisection method uses the zero theorem numerically. Starting with an interval where a continuous function changes sign, take the midpoint. If the midpoint is not a root, one of the two half-intervals still has opposite signs at its endpoints. Keep that half and repeat. The root remains inside every chosen interval, and the interval length halves at each step, giving a simple guaranteed error bound.
1.4 Power Series
A power series is defined from a center
For each fixed
The Cauchy-Hadamard formula is:
with the conventions
| Limit of | Radius |
|---|---|
| positive finite value | |
A useful shortcut, when it exists, is the ratio-test form:
This is not more general than Cauchy-Hadamard, but it is often easier for factorial-type coefficients.
Analytic functions are defined through power series with nonzero radius of convergence. On the interior of the convergence interval, the sum function of a power series is continuous. In standard calculus, it is also differentiable and integrable term by term inside the radius of convergence. For continuity, the key point is uniform convergence on every smaller closed interval
Important analytic functions:
The same list includes the hyperbolic functions:
For all five listed examples, the root-test limit is
Taylor series are useful context for analytic functions: a Taylor series at
When the Taylor series converges to the function in a neighborhood of
What to Emphasize in an Oral Answer
- Define a power series
by center, coefficients, and pointwise convergence as a numerical series for each . - State Cauchy-Hadamard:
, with absolute convergence inside and divergence outside. - Emphasize that real boundary points
require separate tests; the radius formula does not decide them. - Mention the ratio-test shortcut when
exists. - Explain why the sum is continuous inside the radius: uniform convergence on every smaller closed interval.
- Give standard analytic examples such as
, , and , all with infinite radius.
Suggested answer
A power series is a function series of the form
The Cauchy-Hadamard theorem gives the radius of convergence. In general,
When the ratio limit exists, the radius can also be computed as
Inside the radius of convergence, a power series defines a continuous sum function. The reason is that on every smaller closed interval inside the radius, the series converges uniformly and its partial sums are polynomials, hence continuous. Analytic functions are precisely functions represented locally by such convergent power series.
The standard examples are the exponential, sine, and cosine series. The coefficients contain factorials, so the root-test limit is zero and the radius of convergence is infinite. Thus the series define continuous functions on the whole real line:
