Why Substitution Can Evaluate a Limit
A limit asks what output values a function approaches as its input approaches a specified point. When a function is continuous at that point, its nearby output values approach the value the function actually takes there. This lets us evaluate many limits with a direct calculation: substitute the input value into the function.
The method is simple, but its condition matters. Substitution evaluates a limit directly only when the function is continuous at the point being approached, with the relevant domain taken into account. If the function is undefined there or discontinuous there, substitution may fail—even if the limit itself exists.
The Direct-Substitution Rule
Let \(f:D\to\mathbb{R}\), where \(D\) is the domain of \(f\), and let \(c\) be a point in \(D\). For a two-sided limit, assume that points of \(D\) other than \(c\) can approach \(c\). The function \(f\) is continuous at \(c\) when the limit as \(x\) approaches \(c\), with \(x\) restricted to the domain, exists and equals \(f(c)\).
The equation is not permission to substitute into every expression without checking. It is a consequence of continuity, so first make sure the function is defined and continuous at the point. When \(c\) is an endpoint of the domain, the relevant limit is one-sided, and continuity is understood from within the domain.
Why the Rule Works
The result follows directly from the definition of continuity. In the usual epsilon-delta definition, continuity of \(f\) at \(c\) means that for every \(\varepsilon>0\), there is a \(\delta>0\) such that, for \(x\) in the domain,
This says that inputs sufficiently close to \(c\) produce outputs within any prescribed distance \(\varepsilon\) of \(f(c)\). The definition of the limit \(\lim_{x\to c}f(x)=L\) asks for the same kind of control, but compares \(f(x)\) with \(L\). Here the continuity condition supplies that control with \(L=f(c)\). Therefore the limit is \(f(c)\).
Let \(\varepsilon>0\) be the desired maximum distance between \(f(x)\) and \(f(c)\).
Continuity supplies a \(\delta>0\) such that domain inputs satisfying \(|x-c|<\delta\) also satisfy \(|f(x)-f(c)|<\varepsilon\).
The condition in Step 2 is exactly the limit condition for \(f(x)\) to approach \(f(c)\). Hence \(\lim_{x\to c}f(x)=f(c)\).
This justification explains why direct substitution is reliable when continuity is known. In practice, continuity of standard functions and the rules for sums, products, quotients, and compositions let us establish the required condition without returning to the epsilon-delta definition each time.
Worked Examples
Worked Example: Substituting into a Polynomial
Evaluate \(\lim_{x\to -2}(3x^2-2x+1)\).
A polynomial is continuous at every real input, so the direct-substitution rule applies at \(x=-2\). Substitute \(-2\) for \(x\):
Check: Evaluating the polynomial itself at \(-2\) gives \(3(-2)^2-2(-2)+1=12+4+1=17\), matching the limit. Continuity guarantees that nearby polynomial values approach this value.
Worked Example: A Rational Function with a Nonzero Denominator
Evaluate \(\lim_{x\to 1}\frac{x^2+3x+1}{x+2}\).
The numerator and denominator are polynomials, hence continuous. The quotient is continuous wherever its denominator is nonzero. At \(x=1\), the denominator is \(1+2=3\), which is nonzero, so substitution is valid:
Check: The original function is defined at \(1\), and its value there is \(\frac{1+3+1}{3}=\frac{5}{3}\). Since it is continuous there, the limit agrees with that value.
Worked Example: Substitution into a Square-Root Expression
Evaluate \(\lim_{x\to 1}\sqrt{2x+7}\).
The expression under the square root at \(x=1\) is \(2(1)+7=9\), which is positive. The square-root function is continuous on its domain, and \(2x+7\) is continuous. Their composition is therefore continuous at \(1\). Substitute:
Check: The function's value at \(1\) is \(\sqrt{9}=3\). The input is in the domain, and continuity at \(1\) means the limit is that same value.
Check the Domain Before Substituting
A formula may contain operations that are not defined for every real input. For a quotient, the denominator must be nonzero at the point for the quotient rule for continuity to apply there. For a real square root, the radicand must be nonnegative at inputs in the domain. These checks tell you whether the formula defines a continuous function at the point in question.
If the function is defined only on part of the real line, consider inputs approaching \(c\) through that domain. At an endpoint, this is a one-sided approach. For example, \(\sqrt{x}\) is defined for \(x\geq 0\), and its limit as \(x\) approaches \(0\) through its domain is \(0\). This is a one-sided limit, consistent with continuity at the endpoint relative to the domain.
A failed domain check does not automatically tell you what a limit is. It tells you that the direct-substitution theorem cannot be used as stated. The limit may fail to exist, or it may exist even though the function is undefined or discontinuous at the point.
Worked Example: A Limit Exists Even Though Substitution Fails
Define \(g\) by \(g(x)=\frac{x^2-1}{x-1}\) for \(x\ne 1\), and \(g(1)=4\). Determine whether direct substitution evaluates \(\lim_{x\to 1}g(x)\).
The quotient formula is not defined at \(x=1\), because its denominator would be \(1-1=0\). Also, the separately specified function value is \(g(1)=4\). Thus the direct-substitution rule cannot be applied unless continuity at \(1\) is established; in this case, the function is not continuous there.
For \(x\ne 1\), factor the numerator and cancel the common factor, which is nonzero for these nearby inputs:
Therefore the limit can be evaluated using the expression that equals \(g(x)\) for \(x\ne 1\):
Check: For \(x\ne 1\), \(|g(x)-2|=|(x+1)-2|=|x-1|\), so inputs close to \(1\) make \(g(x)\) arbitrarily close to \(2\). But \(g(1)=4\), not \(2\), so the function is not continuous at \(1\). The limit is \(2\), but substituting into the original quotient produces \(0/0\), not the limit value.
A Common Mistake: Treating an Undefined Substitution as an Answer
A frequent error is to substitute first and treat whatever appears as the limit. In the example above, someone might write
The conclusion is wrong. The expression \(0/0\) is undefined; it is not the number zero and does not determine a limit. Direct substitution only gives the limit when continuity at the point justifies it. Here, rewriting the expression for \(x\ne 1\) shows that the limit is \(2\), while the function's assigned value is \(4\).
The practical correction is to pause when substitution gives an undefined expression or violates a domain restriction. Do not assign a limit value from that substitution. Instead, determine whether continuity is present; if it is not, use an appropriate limit method or conclude that the substitution rule does not apply.
When the Rule Applies—and When It Does Not
| Situation at the point | What substitution tells you |
|---|---|
| The function is defined and continuous there. | The limit equals the function value at the point. |
| A quotient has denominator zero there. | The quotient continuity rule does not apply there; investigate separately. |
| The point is outside the function's domain. | The function value is unavailable, so direct substitution is not justified. |
| The function is defined there but discontinuous. | The limit need not equal the function value; check the limit separately. |
The essential distinction is between the value \(f(c)\) and the limiting behavior of \(f(x)\) as \(x\) approaches \(c\). Continuity connects them. Without that connection, they must not be assumed to agree.
Check Your Understanding
For each question, identify whether continuity justifies direct substitution before calculating.
- Evaluate \(\lim_{x\to 2}(x^3-4x+6)\). Why is substitution valid?
- For \(\lim_{x\to 3}\frac{x+1}{x-1}\), check the denominator at the input and evaluate the limit.
- For \(\lim_{x\to 0}\sqrt{x+4}\), check the domain near the input and evaluate the limit.
- If \(h(2)\) is defined to be \(7\), but \(\lim_{x\to 2}h(x)=5\), is \(h\) continuous at \(2\)? Does substitution give the limit?
- Why does obtaining \(0/0\) after substitution fail to determine a limit?
