Calculus › Section J · Infinite series › Topic J.2
Series, convergence, and geometric series
Can you add infinitely many numbers and get a finite answer? Sometimes yes, and the idea powers everything from repeating decimals to how calculators compute functions.
Adding infinitely many numbers
Start halfway across a room, then cover half of what remains, then half again: \(\frac{1}{2} + \frac{1}{4} + \frac{1}{8} + \cdots\). Each step crosses only a fraction of the gap, yet the total distance approaches exactly 1. An infinite sum can land on a finite number. But not every infinite sum behaves: \(1 + 1 + 1 + \cdots\) plainly grows without bound, and even \(1 + \frac{1}{2} + \frac{1}{3} + \cdots\), whose terms shrink, eventually passes every whole number. A series is an infinite sum, and the first job of this topic is to say precisely when such a sum makes sense.
Series and partial sums
A series is the sum of the terms of a sequence:
To define the value, look at what the running totals do. The \(N\)th partial sum is \(S_N = a_1 + a_2 + \cdots + a_N\). If the sequence of partial sums \(S_1, S_2, S_3, \dots\) converges to a number \(S\), the series converges and its sum is \(S\); otherwise the series diverges. This converts a question about sums into a question about sequences, which you already know how to analyse. A series can diverge by shooting off to infinity or by oscillating, as \(1 - 1 + 1 - 1 + \cdots\) does, with partial sums \(1, 0, 1, 0, \dots\) that never settle.
Geometric series
A geometric series starts with a first term \(a\) and multiplies by a fixed ratio \(r\) each step: \(a + ar + ar^2 + ar^3 + \cdots\). Its behaviour depends entirely on \(r\):
and the series diverges when \(|r| \ge 1\). The room-crossing example is geometric with \(a = \frac{1}{2}\) and \(r = \frac{1}{2}\), so its sum is \(\frac{1/2}{1 - 1/2} = 1\), confirming the picture. When \(|r| \ge 1\) the terms refuse to die away, so the running totals cannot approach anything: multiplying by 2 each time runs to infinity, and multiplying by \(-1\) oscillates forever.
Why the geometric formula works
The derivation is a two-line classic. The partial sum is \(S_N = a + ar + ar^2 + \cdots + ar^{N-1}\). Multiply through by \(r\): \(rS_N = ar + ar^2 + \cdots + ar^N\). Almost every term of \(rS_N\) also appears in \(S_N\), so subtracting cancels the whole middle: \(S_N - rS_N = a - ar^N\), giving \(S_N = \frac{a(1 - r^N)}{1 - r}\). Now let \(N\) grow. When \(|r| < 1\), the power \(r^N\) shrinks to 0, leaving \(\frac{a}{1 - r}\). When \(|r| \ge 1\), the piece \(r^N\) never settles, and neither do the partial sums. The formula is not magic; it is one subtraction plus one limit.
Worked example: repeating decimals
What number is \(0.777\ldots\)? Each digit contributes 7 tenths, 7 hundredths, 7 thousandths, and so on:
This is geometric with first term \(a = \frac{7}{10}\) and ratio \(r = \frac{1}{10}\), both plain from one term to the next. The sum is \(\frac{7/10}{1 - 1/10} = \frac{7/10}{9/10} = \frac{7}{9}\). Check it from the other direction: \(\frac{7}{9} = 7 \div 9\), and long division produces exactly the digit 7 repeating. Every repeating decimal is secretly a geometric series, which is why fractions have repeating expansions while numbers like \(\sqrt{2}\) never fall into a repeating pattern.
Telescoping series
Some sums collapse on their own. Take \(\sum_{n=1}^{\infty} \frac{1}{n(n+1)}\). Partial fractions split each term: \(\frac{1}{n(n+1)} = \frac{1}{n} - \frac{1}{n+1}\). Written out, the series is \(\left(1 - \frac{1}{2}\right) + \left(\frac{1}{2} - \frac{1}{3}\right) + \left(\frac{1}{3} - \frac{1}{4}\right) + \cdots\). In the partial sum \(S_N\), every middle number appears once positive and once negative, so all that survives is the first piece of the first term and the last piece of the last term: \(S_N = 1 - \frac{1}{N+1}\). As \(N \to \infty\) this approaches 1, so the series converges to 1. Spotting telescopes is a partial-fractions skill: whenever the terms can be written as a difference, try writing out three or four terms and watch for the cancellations.
The divergence test
If a series converges, its terms must shrink to 0: the partial sums settle down, and adding a term that stays away from 0 would keep jostling them. This gives a one-way test: if \(\lim_{n \to \infty} a_n \ne 0\) (including when the limit does not exist), then \(\sum a_n\) diverges. For instance \(\sum \frac{n}{n+1}\) has terms approaching 1, so it diverges immediately. But be very careful with the direction: terms approaching 0 does not prove convergence. The test can only ever prove divergence, never convergence. When terms do approach 0, you need a real convergence test, and the next sections build a toolkit for exactly that.
The harmonic series diverges
The series \(\sum \frac{1}{n} = 1 + \frac{1}{2} + \frac{1}{3} + \frac{1}{4} + \cdots\) is the great counterexample: its terms approach 0, yet it diverges. Group the terms as \(1 + \frac{1}{2} + \left(\frac{1}{3} + \frac{1}{4}\right) + \left(\frac{1}{5} + \cdots + \frac{1}{8}\right) + \cdots\). The third group is \(\frac{1}{3} + \frac{1}{4} > \frac{1}{2}\); the next four terms together exceed \(4 \times \frac{1}{8} = \frac{1}{2}\); and every doubling block, running from \(\frac{1}{2^k + 1}\) to \(\frac{1}{2^{k+1}}\), clears \(\frac{1}{2}\) the same way. Infinitely many half-sized blocks add up without bound. The divergence is slow beyond intuition: the partial sums need more than 12,000 terms just to pass 10, yet pass 10 they do, and every number after it.
Worked example: a drug reaching steady state
A patient takes a 250 mg dose of a drug every 6 hours, and after each 6-hour interval the body has eliminated 80% of the drug present, keeping 20%. Assume the timing and the elimination rate are exactly regular. How much drug is in the body right after a dose, once the routine has run for a long time? Right after the latest dose the body holds that 250 mg plus 20% of the previous total, which itself was 250 mg plus 20% of the one before, and so on back: \(250 + 250(0.2) + 250(0.2)^2 + \cdots\). This is geometric with \(a = 250\) and \(r = 0.2\), so the level approaches \(\frac{250}{1 - 0.2} = \frac{250}{0.8} = 312.5\) mg. The same reasoning explains the steady hum of any repeated dose: repeated doses level off at dose divided by one minus the retention fraction.
Common mistakes to avoid
First, keep sequences and series separate: a sequence is a list, a series is the running total of a list, and a sequence can converge while its series diverges, as \(\frac{1}{n}\) shows. Second, never apply \(\frac{a}{1-r}\) when \(|r| \ge 1\); the formula answers a question that does not even exist there. Third, read off \(a\) as the actual first term of the series you face: for \(\sum_{n=1}^{\infty} \frac{1}{2^n}\) the first term is \(\frac{1}{2}\), not 1, so the sum is 1. Fourth, do not use the divergence test to prove convergence: terms shrinking to 0 is a necessary condition, not a sufficient one. Finally, remember that series can diverge by oscillating, with \(1 - 1 + 1 - 1 + \cdots\) the standard example.
Key takeaways
A series converges when its sequence of partial sums converges, and its sum is that limit. Geometric series with first term \(a\) and ratio \(r\) sum to \(\frac{a}{1-r}\) exactly when \(|r| < 1\); the formula comes from one subtraction and the limit of \(r^N\). Repeating decimals and steady-state repeated doses are geometric series in disguise. Telescoping series collapse through partial fractions to a first and last piece. The divergence test is one-way: terms not approaching 0 proves divergence, but terms approaching 0 proves nothing, as the ever-growing harmonic series shows.
Flashcards
Flashcard
The sum of a sequence's terms: a₁ + a₂ + a₃ + …
Test your understanding
Question 1 of 5
What is the sum 12 + 3 + 3/4 + 3/16 + … ?
Question 2 of 5
What fraction is the repeating decimal 0.555… ?
Question 3 of 5
What is the sum of 1/(n(n+1)) from n = 1 to infinity?
Question 4 of 5
What happens with the series n/(2n + 3) from n = 1 to infinity?
Question 5 of 5
Why does the harmonic series diverge even though its terms approach 0?
Standards: 10 (BC)
