Why Cancellation Can Reveal an Infinite Sum
A series is defined through the limit of its partial sums. That means a useful first question is not simply whether its terms look small, but whether the finite sums have a structure that can be simplified. In a telescoping series, consecutive terms cancel in those finite sums, often leaving only a few terms at the beginning and end.
The surviving terms determine what happens as the number of terms increases. They may approach a finite number, giving a convergent series, or fail to approach a finite number, giving a divergent series. The key is to write and inspect the partial sum before drawing a conclusion.
Difference Form and the Basic Identity
The most common telescoping form is a difference of consecutive values of a sequence. Suppose the terms are \(a_n-a_{n+1}\). The \(N\)th partial sum is the sum of the first \(N\) terms:
Write out the terms instead of trying to cancel mentally. This displays precisely which occurrences have opposite signs:
Each intermediate term \(a_2,\ldots,a_N\) appears once with each sign and cancels. The endpoints \(a_1\) and \(-a_{N+1}\) remain. Thus the finite sum is \(a_1-a_{N+1}\), not zero.
This identity explains why a limit of the uncancelled terms decides the series. A series \(\sum_{n=1}^{\infty}(a_n-a_{n+1})\) converges to a finite real number exactly when the sequence \(a_{N+1}\) has a finite limit. If \(a_n\to L\), then the sum is \(a_1-L\).
Worked Examples
Worked Example: A Rational Difference with a Finite Sum
Determine whether \(\displaystyle\sum_{n=1}^{\infty}\frac{1}{n(n+1)}\) converges, and find its sum if it does.
First rewrite one term as a difference. Since \(\frac{1}{n}-\frac{1}{n+1}=\frac{(n+1)-n}{n(n+1)}=\frac{1}{n(n+1)}\), the series has the form of consecutive differences. Its \(N\)th partial sum is
The intermediate terms \(1/2,\ldots,1/N\) cancel. Therefore \(S_N=1-\frac{1}{N+1}\). As \(N\to\infty\), \(\frac{1}{N+1}\to0\), so \(S_N\to1\). The series converges to \(1\).
Check: The proposed sum agrees with the partial-sum limit: \(\lim_{N\to\infty}(1-\frac{1}{N+1})=1\). Also, the first partial sum is \(S_1=1-\frac12=\frac12\), which agrees with the original first term \(\frac{1}{1(2)}=\frac12\).
Worked Example: A Difference with a Two-Term Shift
Find the sum of \(\displaystyle\sum_{n=1}^{\infty}\left(\frac{1}{n+1}-\frac{1}{n+3}\right)\), if it converges.
Write out the first terms of the partial sum. The two denominator patterns are offset by two places:
For \(N\ge2\), the overlapping terms with denominators \(4,\ldots,N+1\) cancel. The first two positive terms and the last two negative terms remain. This gives \(S_N=\frac12+\frac13-\frac{1}{N+2}-\frac{1}{N+3}\). This endpoint formula also gives the correct value for \(N=1\): its \(+\frac13\) and \(-\frac13\) cancel, leaving \(\frac12-\frac14\).
Both final fractions approach zero as \(N\to\infty\). Thus \(S_N\to\frac56\), and the series converges to \(\frac56\).
Check: At \(N=2\), the original partial sum is \((\frac12-\frac14)+(\frac13-\frac15)=\frac14+\frac{2}{15}=\frac{23}{60}\). The endpoint formula gives \(\frac56-\frac14-\frac15=\frac{50}{60}-\frac{15}{60}-\frac{12}{60}=\frac{23}{60}\), as required.
Worked Example: Cancellation Does Not Guarantee Convergence
Determine the behavior of \(\displaystyle\sum_{n=1}^{\infty}(\sqrt{n+1}-\sqrt n)\).
The partial sum expands as
Every intermediate square root appears once with each sign. The endpoints are \(\sqrt{N+1}\) and \(-\sqrt1=-1\). As \(N\to\infty\), \(\sqrt{N+1}-1\to\infty\), so the partial sums do not approach a finite real number. The series diverges.
Check: For \(N=3\), the original partial sum is \((\sqrt2-1)+(\sqrt3-\sqrt2)+(2-\sqrt3)=2-1=1\). The formula gives \(\sqrt4-1=2-1=1\). The partial-sum identity is consistent.
Why the Partial-Sum Argument Works
To define the sum of an infinite series, we take the limit of its finite partial sums. This is why telescoping cancellation is justified: it happens inside each finite sum, where the number of terms is fixed and the cancellations are exact. Only after simplifying \(S_N\) do we take the limit as \(N\to\infty\).
For the basic difference form, the argument is a direct proof. Expanding the first \(N\) terms gives \(S_N=(a_1-a_2)+(a_2-a_3)+\cdots+(a_N-a_{N+1})\). For every index \(k\) with \(2\le k\le N\), the term \(-a_k\) from one pair and \(+a_k\) from the next pair sum to zero. The first term \(a_1\) has no preceding negative copy, and \(-a_{N+1}\) has no following positive copy. Therefore \(S_N=a_1-a_{N+1}\). If \(a_{N+1}\to L\), the limit of these partial sums is \(a_1-L\), which proves the theorem.
The same reasoning handles a fixed shift larger than one. For example, if a term is \(a_n-a_{n+2}\), then the partial sum is \((a_1-a_3)+(a_2-a_4)+\cdots+(a_N-a_{N+2})\). The cancellations occur separately along the odd-indexed and even-indexed terms. This illustrates an important habit: expand enough terms to see the pattern, then identify all the endpoints rather than assuming that every term disappears.
Common Mistakes and How to Avoid Them
A frequent error is to say that every term cancels and therefore the sum is zero. For instance, for the first three terms of the basic pattern,
not zero. The intermediate terms cancel, but the endpoints remain. A reliable correction is to mark the first positive term and the final negative term before simplifying.
Another mistake is to use a cancellation formula without checking small values of \(N\). With shifted patterns, the ranges of positive and negative terms can overlap at the endpoints. In the second worked example, \(N=1\) produces a \(+\frac13\) and a \(-\frac13\) in the endpoint formula; they cancel each other. They should not be described as two distinct surviving terms. Substituting \(N=1\), and then checking another small value such as \(N=2\), is a quick way to catch indexing errors.
Finally, seeing cancellation does not by itself show convergence. In the square-root example, almost all terms cancel in every partial sum, but the remaining endpoint grows without bound. The conclusion must come from the limit of the actual partial-sum expression, not from the visual amount of cancellation.
How to Approach a Telescoping Series
Look for a form such as \(a_n-a_{n+1}\), or a difference with a fixed shift. Verify the algebra before summing.
Expand several consecutive terms so the positive and negative copies are visible.
Keep track of the beginning and end of each index pattern. Check a small value of \(N\) if endpoint ranges overlap.
A finite limit gives the value of the series. If the partial sums do not approach a finite real number, the series diverges.
Why Telescoping Matters
Telescoping is one of the clearest ways to connect the structure of a term with the behavior of its partial sums. It can turn a long finite sum into a short expression, and it makes the definition of an infinite sum visible: the series converges precisely when its sequence of partial sums approaches a finite limit.
It also emphasizes a useful distinction. A series may have terms that cancel substantially, yet its remaining endpoints may still grow. Conversely, a complicated-looking term may become simple once it is expressed as a difference. The decisive calculation is always the finite partial sum followed by its limit.
Check Your Understanding
For each question, write enough of the partial sum to make the cancellation pattern clear.
- Rewrite \(\frac{1}{n(n+2)}\) as a difference of fractions. What are the first four terms of the resulting series?
- For \(\sum_{n=1}^{\infty}(a_n-a_{n+1})\), state the formula for \(S_N\) and the condition on \(a_n\) that gives a finite sum.
- Find the \(N\)th partial sum of \(\sum_{n=1}^{\infty}\left(\frac{1}{n}-\frac{1}{n+2}\right)\), taking care to check \(N=1\).
- Explain why the series \(\sum_{n=1}^{\infty}(\sqrt{n+1}-\sqrt n)\) diverges even though its partial sums involve cancellation.
- In a telescoping partial sum, which terms cancel, and why do the endpoint terms require separate attention?
