← back to MATH 102
Week 2Stewart §11.272 min full read
6 concepts19 worked examples25 exercises3 exam-level4 figures
What are you here for?

02Series: partial sums, the geometric and telescoping series, and the Test for Divergence

Drop a ball from a height of $2$ m. Each time it lands it comes back to exactly three quarters of the height it fell from, then falls again, and it keeps doing this forever. A tape measure that adds up every rise and every fall records a longer and longer number as the bounces pile up. The bounces never stop, so it is tempting to say the tape reading runs off to infinity — yet the ball plainly comes to rest in a few seconds over a patch of floor you could cover with your hand.

By the end of this section you can add those infinitely many bounces to one exact number, $14$ m, and state precisely when an endless sum settles on a value and when it genuinely blows up.

In 60 seconds

An infinite series adds up to the number its running totals close in on: name the type first (geometric, telescoping, or neither), apply that type's method to the partial sum $s_n$, and take $n\to\infty$.

Series as a limit of partial sums
$$\sum_{n=1}^{\infty} a_n = \lim_{n\to\infty} s_n,\quad s_n = a_1+a_2+\cdots+a_n$$

the definition behind every convergence question; the sum is the limit of the running totals, nothing more

Geometric series
$$\sum_{n=1}^{\infty} ar^{\,n-1} = \frac{a}{1-r}\quad(|r|<1),\ \text{diverges for }|r|\ge 1$$

each term is a fixed multiple of the one before it; repeating decimals; bouncing distances

Telescoping series
$$a_n=b_n-b_{n+1}\ \Rightarrow\ s_n=b_1-b_{n+1}$$

a term splits by partial fractions into a difference whose middle pieces cancel

Test for Divergence
$$\lim_{n\to\infty} a_n \ne 0\ \Rightarrow\ \sum a_n\ \text{diverges}$$

always the first check; if the terms do not shrink to zero the sum cannot settle

Three most common mistakes
  1. Reading $a_n\to 0$ as a guarantee of convergence. It is only a necessary condition; the harmonic series $\sum 1/n$ has $a_n\to 0$ and still diverges.

  2. Using $\dfrac{a}{1-r}$ without checking $|r|<1$. When $|r|\ge 1$ that fraction is a meaningless number for a series that has no sum at all.

  3. Taking the first term $a$ to be the wrong value when the sum starts at $n=0$ instead of $n=1$, or when the terms are written as $ar^{n}$ instead of $ar^{n-1}$.

On this course the Quiz and Homework components carry $10\%$ each. The three moves in this section — sum a geometric series, collapse a telescoping one, and reject a series by the Test for Divergence — are exactly the short single-method questions those two components are built from.

How much time do you have?
10 minutes

The two moves that clear most quiz questions: the geometric sum with its |r|<1 check, and the one-line test that rejects a series whose terms do not go to zero.

card, c-geometric-series, c-divergence-test, formula card
45 minutes

Every type you can be handed: what a sum even means, the geometric and telescoping methods, the one series everybody trips over (harmonic), the rejection test, and one full ladder from a worked solution to a bare one.

card, c-series-as-partial-sums, c-geometric-series, c-telescoping, c-harmonic, c-divergence-test, faded ladder, formula card
full read

Where an infinite sum comes from, why each formula is allowed, and enough mixed practice that you decide the method yourself instead of being told it.

hook, prerequisites, c-series-as-partial-sums, c-geometric-series, c-telescoping, c-harmonic, c-divergence-test, c-algebra-of-series, contrast pairs, faded ladder, exam example, practice A to D
By the end of this section
  1. Define the sum of an infinite series as the limit of its sequence of partial sums, and read off convergence or divergence from that limit.

  2. Sum a geometric series with the formula $a/(1-r)$ after checking $|r|<1$, and turn a repeating decimal or a bouncing distance into such a series.

  3. Collapse a telescoping series by splitting its term into a difference and cancelling the middle, then take the limit of what survives.

  4. Explain why the harmonic series diverges even though its terms shrink to zero, using the grouping argument.

  5. Apply the Test for Divergence to reject a series whose terms do not tend to zero, and state exactly why the test says nothing when they do.

  6. Combine convergent series with the linearity rules, and recognise when a sum of a convergent and a divergent series must diverge.

Syllabus coverage
11.2

Series: partial sums and convergence, the geometric series, telescoping sums, the harmonic series, the Test for Divergence, and the algebra of convergent series

The single book section for this week is spread over six blocks below: the definition through partial sums, the geometric series, telescoping sums, the harmonic series, the Test for Divergence, and the linearity rules for combining convergent series.

covered
Recall first
Limit of a sequence

$\lim_{n\to\infty} c_n = L$ means the terms $c_n$ close in on $L$ as $n$ grows; if no single such $L$ exists the sequence diverges.

The sum of a series is defined as one particular sequence limit, so every convergence question here is a sequence limit in disguise.

Limit of a power

$\lim_{n\to\infty} r^{\,n} = 0$ when $|r|<1$, while $|r^{\,n}|\to\infty$ when $|r|>1$.

This single fact is what makes the geometric series for small ratios and blow up for large ones.

Partial fractions

A quotient like $\dfrac{1}{n(n+1)}$ can be rewritten as a difference $\dfrac{1}{n}-\dfrac{1}{n+1}$ by solving for the unknown numerators.

Splitting a term into a difference is the whole trick that makes a telescoping series collapse.

Sequence versus its running totals

From a sequence $a_n$ we build a second sequence $s_n=a_1+\cdots+a_n$; these are different objects with possibly different fates.

The previous section studied $a_n$; this section studies $s_n$, and confusing the two is the single most common error below.

Try it yourself first (3 questions)
1§02.0 — do shrinking terms force a finite sum?●○○○○

Before the section, an honest guess. A classmate argues: even though the terms of $1+\tfrac12+\tfrac13+\tfrac14+\cdots$ get as small as we like, the running totals still climb past every bound, so the sum is infinite.

Given
  • Claim: even though $\dfrac1n\to 0$, the running totals of $\displaystyle\sum_{n=1}^{\infty}\dfrac1n$ grow past every bound, so the sum is infinite.

Find
  1. (a) True or false? A guess is fine; the section settles it.

Hint 1/4

Ask whether small terms alone can stop a running total from growing.

Hint 2/4

Terms going to zero is necessary for a finite sum, but the question is whether it is enough.

Hint 3/4

Think of adding many terms each near, say, $1/1000$: a thousand of them still add to about $1$.

Hint 4/4

Shrinking terms do not by themselves force a finite total; the harmonic totals do run off, so the claim is true.

Show solution
Test the intuition against grouping
$$\tfrac13+\tfrac14 > \tfrac14+\tfrac14=\tfrac12$$

two terms already beat one half, and later blocks do the same

Answer $$\boxed{\text{True}}$$
Check

The full grouping argument is developed in the harmonic-series block below.

2§02.0 — a running total you can already read●○○○○

You have met sequence limits already. Here is a sequence of running totals with a formula, so you can find the sum without knowing the individual terms.

Given
  • The running total of the first $n$ terms is $s_n=\dfrac{2n}{n+1}$.

Find
  1. (a) What number do the running totals approach as n grows?

Hint 1/4

The sum of the series is just the limit of the running totals you are given.

Hint 2/4

For a quotient of polynomials of equal degree, the limit is the ratio of leading coefficients.

Hint 3/4

Here $s_n=\dfrac{2n}{n+1}$, and dividing top and bottom by $n$ gives $\dfrac{2}{1+1/n}$.

Hint 4/4

As $n\to\infty$ the total approaches $2$.

Show solution
Take the limit of the running totals
$$\lim_{n\to\infty}\frac{2n}{n+1}=\lim_{n\to\infty}\frac{2}{1+1/n}=2$$

divide numerator and denominator by the highest power, then let 1/n vanish

Answer $$\boxed{2}$$
Check

Whenever you are handed a formula for the running totals, the sum is one limit away.

3§02.0 — is a geometric-looking sum always finite?●○○○○

One more guess. A geometric-looking sum multiplies each term by a fixed number to get the next. Does such a sum always land on a finite value?

Given
  • Claim: $\displaystyle\sum_{n=1}^{\infty} 2^{\,n-1}=1+2+4+8+\cdots$ adds up to a finite number.

Find
  1. (a) True or false?

Hint 1/4

Ask whether the running totals here settle or keep growing.

Hint 2/4

A geometric series only settles when the fixed multiplier has size below $1$.

Hint 3/4

Here the multiplier is $2$, so each term is bigger than the last: $1,2,4,8,\dots$

Hint 4/4

The totals $1,3,7,15,\dots$ run off to infinity, so the sum is not finite.

Show solution
Look at the running totals
$$s_1=1,\ s_2=3,\ s_3=7,\ s_4=15$$

each total is one less than the next power of two, so they increase without bound

Answer $$\boxed{\text{False}}$$
Check

Geometric sums settle only when the ratio is small; this is the content of the geometric block.

Notation
symbolreads asmeanswatch out
$\displaystyle\sum_{n=1}^{\infty} a_n$

the sum from n equals 1 to infinity of a sub n

the limit of the running totals of the terms a_1, a_2, a_3 and so on

It is a single number (or the word diverges), not a list of terms; the letter $n$ is a dummy and never appears in the answer.

$s_n$

s sub n, the n-th partial sum

the finite total of just the first n terms, a_1 through a_n

This is an ordinary finite sum; the infinite series is $\lim_{n\to\infty} s_n$, a different object from $s_n$ itself.

$a_n$

a sub n, the n-th term

one single term of the series, the thing being added at step n

Do not confuse $a_n$ (a term) with $s_n$ (a running total); a whole section of mistakes lives in that confusion.

$r$

r, the

the fixed number each term of a geometric series is multiplied by to reach the next

Compute it as $r=a_{n+1}/a_n$ and confirm it is the same for every $n$; if it is not constant the series is not geometric.

$\text{converges}$ / $\text{diverges}$

converges or diverges

the running totals do, or do not, close in on a single finite number

Diverges covers three different behaviours: running off to $\pm\infty$, or oscillating with no limit at all.

Conventions used here
What convergence of a series means on this page.

A series converges when its sequence of partial sums $s_n$ has a finite limit, and that limit is called the sum. It diverges otherwise. We never treat $\sum a_n$ as a literal act of adding infinitely many numbers; it is always shorthand for $\lim_{n\to\infty} s_n$. So the question is never can you add them all up, but does the running total settle.

The whole subject is built on one definition, and half of the errors below come from reasoning about the terms $a_n$ when the definition only ever talks about the totals $s_n$.

How we report a divergent series.

When the partial sums run off in one direction we write $\sum a_n=\infty$ or $-\infty$, because that says more than the bare word diverges. When they oscillate with no limit, as for $\sum(-1)^n$, we simply say the series diverges. Either way there is no finite sum, and writing a number down for a divergent series is wrong even if the number is $\infty$ treated as an ordinary value.

Both phrasings are standard, and mixing them inside one answer is what loses the mark that the divergence discussion is meant to earn.

Where the index starts.

A geometric series can be written $\sum_{n=1}^{\infty} ar^{\,n-1}$ (first term $a$) or $\sum_{n=0}^{\infty} ar^{\,n}$ (also first term $a$) or $\sum_{n=1}^{\infty} ar^{\,n}$ (first term $ar$). The letter $a$ in the formula $a/(1-r)$ is always the actual first term you are adding, not the constant sitting in front of the power. Always read off the first term by plugging in the starting index.

The single most common geometric-series slip is using the wrong $a$ because the starting index or the exponent was read carelessly.

Adding or dropping finitely many terms.

Changing, inserting or deleting a finite number of terms never changes whether a series converges or diverges, though it does change the sum. So $\sum_{n=1}^{\infty} a_n$ and $\sum_{n=5}^{\infty} a_n$ live or die together; only the value differs by the finitely many terms you dropped.

Convergence is a statement about the tail, and the tail is untouched by any finite edit at the front.

2.1A series is the limit of its partial sums

Turns an endless sum into a sequence limit by watching the running totals $s_n$ instead of adding all the terms at once.

The previous section handed us a way to take the limit of a sequence; now we point that same tool at the running totals of an infinite list of numbers.

Solvable with what we have
  • Add any finite list: $3+1+4+1+5=14$.

  • Take the limit of a sequence, from the previous section.

Not solvable yet
  • Literally add infinitely many numbers, one after another, because that process never ends.

  • Decide by eye whether $1+\tfrac12+\tfrac14+\cdots$ is a real number or nonsense.

Try to just keep adding. For $1+\tfrac12+\tfrac14+\tfrac18+\cdots$ the running totals are $1,\ 1.5,\ 1.75,\ 1.875,\ 1.9375,\dots$ and they press up toward $2$. For $1+2+3+4+\cdots$ the running totals $1,3,6,10,\dots$ just keep climbing. The adding never finishes in either case, so how can one have an answer and the other not?

Why it fails

There is no last term to reach, so adding is the wrong picture. What separates the two examples is not the adding but the running totals: one sequence of totals has a limit and the other does not. That observation is the whole definition.

DefinitionDefinition: sum of an infinite series
Conditions
  • the terms $a_1,a_2,a_3,\dots$ are given

  • the partial sums are $s_n=a_1+a_2+\cdots+a_n$

$$\boxed{\ \sum_{n=1}^{\infty} a_n=\lim_{n\to\infty} s_n\quad\text{when this limit exists (finite).}\ }$$

The sum of the whole series is the number the running totals close in on. If those running totals settle on a finite value the series converges to it; if they run off or oscillate, the series diverges and has no sum.

Looks like this, but is not

The series $1-1+1-1+1-\cdots$ has terms that stay small and bounded, so it looks like it ought to have a sum, perhaps $0$ by pairing, or $\tfrac12$ by a clever average.

The partial sums are $1,0,1,0,1,0,\dots$, which never close in on a single number. Boundedness is not convergence; the definition demands a limit of the totals, and this sequence of totals has none, so the series diverges.

Summing 1/2 + 1/4 + 1/8 + ... by a closed form for the totals

Show that $\displaystyle\sum_{n=1}^{\infty}\frac1{2^{\,n}}$ converges and find its sum.

Given
  • $a_n=\dfrac1{2^{\,n}}$, so the terms are $\tfrac12,\tfrac14,\tfrac18,\dots$

Find

The sum, as a limit of partial sums.

Solution
Find a closed form for the running total
$$s_n=\tfrac12+\tfrac14+\cdots+\tfrac1{2^{\,n}}=1-\tfrac1{2^{\,n}}$$

each new term fills half the gap left to 1, so the total is always exactly 1 minus the last piece

Take the limit of the totals
$$\lim_{n\to\infty}\Big(1-\tfrac1{2^{\,n}}\Big)=1-0=1$$

the leftover piece 1/2^n tends to 0, which is the only moving part

Answer $$\boxed{\displaystyle\sum_{n=1}^{\infty}\frac1{2^{\,n}}=1}$$
Check

Sanity check against the square in the figure: the pieces of area 1/2, 1/4, 1/8, ... exactly fill a square of area 1.

The sum came from a formula for s_n, not from adding forever. Get the totals in closed form and the rest is a sequence limit.

A series whose totals march off: the terms do not shrink

Decide whether $\displaystyle\sum_{n=1}^{\infty} 1$ converges.

Given
  • Every term is $a_n=1$.

Find

Convergence or divergence.

Solution
Write the running total
$$s_n=\underbrace{1+1+\cdots+1}_{n\ \text{terms}}=n$$

adding n copies of 1 gives n

Take the limit
$$\lim_{n\to\infty} s_n=\lim_{n\to\infty} n=\infty$$

the totals increase past every bound, so no finite limit exists

Answer $$\boxed{\text{Diverges to }\infty}$$
Check

Consistent with intuition: the terms never shrink, so the running total cannot possibly settle.

Checkpoint
§02.1 — recover a term and a sum from the totals●●○○○

A series has partial sums given in closed form. You can answer everything about it without ever seeing the terms one by one.

Given
  • The n-th partial sum is $s_n=\dfrac{3n}{n+1}$.

Find
  1. (a) What is the sum of the series?

  2. (b) What is the third term a_3?

Hint 1/4

The sum is the limit of the given totals; a single term is the jump from one total to the next.

Hint 2/4

Use $\displaystyle\sum a_n=\lim_{n\to\infty}s_n$ for part (a), and $a_n=s_n-s_{n-1}$ for part (b).

Hint 3/4

For (a) divide through by $n$; for (b) compute $a_3=s_3-s_2$ with $s_3=\tfrac{9}{4}$ and $s_2=2$.

Hint 4/4

The sum is $3$, and $a_3=\tfrac94-2=\tfrac14$.

Show solution
Part (a): limit of the totals
$$\lim_{n\to\infty}\frac{3n}{n+1}=\lim_{n\to\infty}\frac{3}{1+1/n}=3$$

equal degrees, so the limit is the ratio of leading coefficients

Part (b): term as a jump in the totals
$$s_3=\frac{9}{4},\quad s_2=\frac{6}{3}=2$$

evaluate the formula at n=3 and n=2

$$a_3=s_3-s_2=\frac94-2=\frac14$$

the third term is exactly what the total gained on the third step

Answer $$\boxed{\sum a_n=3,\quad a_3=\tfrac14}$$
Check

Check: a_3 should be positive and small since the totals are levelling off near 3, and 1/4 fits.

⚠ Adding the terms instead of taking the limit of the totals

the notation looks like an instruction to add, so the instinct is to start adding rather than to build s_n and take a limit

wrong$$\sum_{n=1}^{\infty}\frac1{2^{\,n}}=\frac12+\frac14+\cdots=\text{``keep going''}$$
right$$\sum_{n=1}^{\infty}\frac1{2^{\,n}}=\lim_{n\to\infty}\Big(1-\frac1{2^{\,n}}\Big)=1$$
⚠ Confusing the term a_n with the partial sum s_n

both are sequences indexed by n and both are written with subscripts, so they blur together

wrong$$a_3\stackrel{?}{=}s_3$$
right$$a_3=s_3-s_2$$

2.2The geometric series and the a over one minus r formula

Sums any series whose terms share a constant ratio $r$, giving $a/(1-r)$ the moment $|r|<1$.

The doubling series above blew up because its ratio was $2$; the halving series settled because its ratio was $\tfrac12$. That contrast is a theorem waiting to be written.

TheoremTheorem: the geometric series
Conditions
  • each term is a fixed multiple of the previous one: $a_{n+1}=r\,a_n$

  • $a$ is the first term actually added

$$\boxed{\ \sum_{n=1}^{\infty} ar^{\,n-1}=\frac{a}{1-r}\ \text{ if }|r|<1;\quad\text{diverges if }|r|\ge 1.\ }$$

If every term is r times the one before and r has size below one, the endless sum equals the first term divided by one minus the ratio. If r has size one or more, the terms do not shrink and the series has no sum.

Looks like this, but is not

The series $\displaystyle\sum_{n=1}^{\infty}\frac{n}{2^{\,n}}=\frac12+\frac24+\frac38+\cdots$ looks geometric because a power of two sits in every denominator.

The ratio of consecutive terms is $\dfrac{a_{n+1}}{a_n}=\dfrac{n+1}{2n}$, which changes with $n$ (it is $1$, then $\tfrac34$, then $\tfrac23,\dots$). A geometric series needs the same ratio every time, so $a/(1-r)$ does not apply here; the extra factor $n$ disqualifies it.

Summing 3 + 3/4 + 3/16 + ... with the geometric formula

Evaluate $\displaystyle\sum_{n=1}^{\infty} 3\left(\tfrac14\right)^{\,n-1}$.

Given
  • First term $a=3$ (put $n=1$)

  • ratio $r=\tfrac14$

Find

The sum.

Solution
Check the convergence condition
$$|r|=\tfrac14<1$$

so the formula applies and the series converges

Apply the formula with the correct first term
$$\frac{a}{1-r}=\frac{3}{1-\tfrac14}=\frac{3}{\tfrac34}=4$$

a is the first term you actually add, here 3

Answer $$\boxed{4}$$
Check

Check with a partial sum: 3 + 0.75 + 0.1875 = 3.9375 is already close to 4, and the remaining terms are tiny.

Turning the repeating decimal 0.272727... into a fraction

Write $0.\overline{27}=0.272727\ldots$ as a fraction in lowest terms.

Given
  • The block $27$ repeats forever.

Find

An exact fraction.

Solution
Write the decimal as a geometric series
$$0.\overline{27}=\frac{27}{100}+\frac{27}{100^2}+\frac{27}{100^3}+\cdots$$

each repeat shifts the two-digit block two places, i.e. multiplies by 1/100

Identify a and r, then sum
$$a=\frac{27}{100},\ r=\frac{1}{100},\quad \frac{a}{1-r}=\frac{27/100}{99/100}=\frac{27}{99}=\frac{3}{11}$$

the first term is 27/100, not 27; the ratio is 1/100 because the block has two digits

Answer $$\boxed{\dfrac{3}{11}}$$
Check

Check by division: 3 divided by 11 is 0.2727..., which is the decimal we started with.

Resolving the bouncing ball: total distance travelled

A ball is dropped from $2$ m and each bounce returns to $\tfrac34$ of the previous height. Find the total distance it travels.

Given
  • Initial drop $2$ m

  • each rebound height is $\tfrac34$ of the last

Find

Total up-and-down distance.

Solution
Separate the first drop from the bounces
$$D=2+\sum_{k=1}^{\infty}2\,h_k,\qquad h_k=2\left(\tfrac34\right)^{k}$$

after landing, bounce k rises h_k and falls h_k, contributing 2h_k; the very first drop of 2 has no matching rise

Sum the geometric part
$$\sum_{k=1}^{\infty}2\cdot 2\left(\tfrac34\right)^{k}=4\cdot\frac{3/4}{1-3/4}=4\cdot 3=12$$

first term of this geometric series is 4 times 3/4, ratio 3/4, so it sums to 12

Add the first drop
$$D=2+12=14$$

the lone initial drop plus the 12 metres of bouncing

Answer $$\boxed{14\ \text{m}}$$
Check

Check the size: the ball falls 2, then bounces roughly 3 + 2.25 + 1.7 + ... metres, and those clearly pile up to about 12, not to infinity.

The geometric formula was applied once; the only real work was reading off a and r correctly.

This closes the opening problem: infinitely many bounces, one finite total, because the ratio 3/4 is below 1.

Checkpoint
§02.2 — same terms, two starting indices●●○○○

The starting index decides what the first term a is. Read it off carefully in each case.

Given
  • $\displaystyle A=\sum_{n=0}^{\infty}\left(\tfrac23\right)^{\,n}$

  • $\displaystyle B=\sum_{n=1}^{\infty}\left(\tfrac23\right)^{\,n}$

Find
  1. (a) Find A.

  2. (b) Find B, and say how it relates to A.

Hint 1/4

For each sum, first write down the very first term you actually add.

Hint 2/4

Both are geometric with ratio $\tfrac23<1$, so each equals its own first term over $1-\tfrac23$.

Hint 3/4

For A the first term (at $n=0$) is $1$; for B the first term (at $n=1$) is $\tfrac23$.

Hint 4/4

So $A=\dfrac{1}{1/3}=3$ and $B=\dfrac{2/3}{1/3}=2$; they differ by the missing $n=0$ term, which is $1$.

Show solution
Sum A (starts at n=0, first term 1)
$$A=\frac{1}{1-\tfrac23}=\frac{1}{\tfrac13}=3$$

at n=0 the term (2/3)^0 = 1 is the first term

Sum B (starts at n=1, first term 2/3)
$$B=\frac{\tfrac23}{1-\tfrac23}=\frac{\tfrac23}{\tfrac13}=2$$

at n=1 the first term is 2/3

Relate them
$$A-B=3-2=1$$

the only difference is the n=0 term, which equals 1

Answer $$\boxed{A=3,\ B=2,\ A=B+1}$$
Check

Check: A should exceed B by exactly the dropped first term 1, and 3 = 2 + 1.

⚠ Using a/(1-r) without checking |r| < 1

the formula is memorable and the |r| condition is easy to skip when a problem is presented as routine

wrong$$\sum_{n=1}^{\infty}2\cdot 3^{\,n-1}\stackrel{?}{=}\frac{2}{1-3}=-1$$
right$$|r|=3\ge 1\ \Rightarrow\ \text{diverges (no sum)}$$
⚠ Reading the wrong first term a

the constant in front of the power is mistaken for a when the index starts at 0 or the exponent is n rather than n-1

wrong$$\sum_{n=1}^{\infty}5\left(\tfrac12\right)^{\,n}\stackrel{?}{=}\frac{5}{1-\tfrac12}=10$$
right$$a=5\cdot\tfrac12=\tfrac52\ \Rightarrow\ \frac{5/2}{1-1/2}=5$$

2.3Telescoping series: when the middle cancels

Sums a non-geometric series by splitting each term into a difference so consecutive pieces cancel and $s_n$ collapses.

Not every series is geometric, but some still have a partial sum you can write in closed form, because the terms are secretly differences that eat each other.

MethodMethod: telescoping a series
Conditions
  • the term can be written as a difference $a_n=b_n-b_{n+1}$

  • the sequence $b_n$ has a limit

$$\boxed{\ a_n=b_n-b_{n+1}\ \Rightarrow\ s_n=b_1-b_{n+1},\quad \sum_{n=1}^{\infty}a_n=b_1-\lim_{n\to\infty}b_{n+1}.\ }$$

If each term is one piece minus the next piece, then in the running total every interior piece is added once and subtracted once, so only the very first piece and the very last piece survive. The sum is the first piece minus the limit of the last.

Looks like this, but is not

The series $\displaystyle\sum_{n=1}^{\infty}\frac1{n^2}$ has a tidy term and it is tempting to hunt for a telescoping split $\frac1{n^2}=b_n-b_{n+1}$.

No elementary $b_n$ makes $b_n-b_{n+1}=\frac1{n^2}$, so this series does not telescope; its convergence needs a different tool (the tests in the next section). Telescoping is a lucky structure, not a universal method.

Summing 1/(n(n+1)) by splitting into a difference

Evaluate $\displaystyle\sum_{n=1}^{\infty}\frac{1}{n(n+1)}$.

Given
  • $a_n=\dfrac{1}{n(n+1)}$

Find

The sum.

Solution
Split the term by partial fractions
$$\frac{1}{n(n+1)}=\frac1n-\frac1{n+1}$$

solving 1 = A(n+1)+Bn gives A=1, B=-1

Collapse the partial sum
$$s_n=\Big(1-\tfrac12\Big)+\Big(\tfrac12-\tfrac13\Big)+\cdots+\Big(\tfrac1n-\tfrac1{n+1}\Big)=1-\frac1{n+1}$$

every interior fraction cancels with its neighbour; only 1 and -1/(n+1) survive

Take the limit
$$\lim_{n\to\infty}\Big(1-\frac1{n+1}\Big)=1-0=1$$

the surviving tail 1/(n+1) tends to 0

Answer $$\boxed{1}$$
Check

Check with the figure: the cancellation diagram leaves exactly 1 - 1/(n+1), whose limit is 1.

A telescoping series that diverges: sum of ln(n/(n+1))

Evaluate $\displaystyle\sum_{n=1}^{\infty}\ln\!\frac{n}{n+1}$, or show it diverges.

Given
  • $a_n=\ln\dfrac{n}{n+1}=\ln n-\ln(n+1)$

Find

The sum or a divergence statement.

Solution
Recognise the difference form
$$a_n=\ln n-\ln(n+1)$$

a logarithm of a quotient is a difference of logarithms, which is already b_n - b_{n+1} with b_n = ln n

Collapse and take the limit
$$s_n=\ln 1-\ln(n+1)=-\ln(n+1)\ \xrightarrow[n\to\infty]{}\ -\infty$$

the telescope leaves -ln(n+1), and ln(n+1) grows without bound

Answer $$\boxed{\text{Diverges to }-\infty}$$
Check

Note the warning it carries: here a_n = ln(n/(n+1)) tends to ln 1 = 0, yet the series diverges. Shrinking terms did not save it.

Telescoping tells you the fate directly through b_n: convergence needs lim b_{n+1} to exist, and here it does not.

Checkpoint
§02.3 — a telescoping sum with a spacing of two●●●○○

The split still works when the two pieces are two steps apart; you just carry an extra surviving term at each end.

Given
  • $\displaystyle\sum_{n=1}^{\infty}\frac{1}{(2n-1)(2n+1)}$

  • partial-fraction hint: $\dfrac{1}{(2n-1)(2n+1)}=\dfrac12\!\left(\dfrac{1}{2n-1}-\dfrac{1}{2n+1}\right)$

Find
  1. (a) Find the sum.

Hint 1/4

Use the given split, then write out a few partial-sum terms to see what cancels.

Hint 2/4

With $a_n=\tfrac12\!\left(\tfrac{1}{2n-1}-\tfrac{1}{2n+1}\right)$ the telescope leaves the first piece and the last piece.

Hint 3/4

$s_n=\tfrac12\!\left(1-\dfrac{1}{2n+1}\right)$, since the surviving front piece is $\tfrac{1}{1}$ and the back piece is $\tfrac{1}{2n+1}$.

Hint 4/4

Let $n\to\infty$: the sum is $\tfrac12(1-0)=\tfrac12$.

Show solution
Collapse the partial sum, keeping the 1/2
$$s_n=\tfrac12\!\left[\Big(1-\tfrac13\Big)+\Big(\tfrac13-\tfrac15\Big)+\cdots+\Big(\tfrac1{2n-1}-\tfrac1{2n+1}\Big)\right]=\tfrac12\!\left(1-\tfrac1{2n+1}\right)$$

the interior fifths, thirds and so on cancel; keep the 1/2 in front the whole time

Take the limit
$$\lim_{n\to\infty}\tfrac12\!\left(1-\tfrac1{2n+1}\right)=\tfrac12$$

the tail 1/(2n+1) vanishes

Answer $$\boxed{\tfrac12}$$
Check

Check the first term: 1/(1*3) = 1/3, and the total 1/2 is only a little more, which fits since later terms are small.

⚠ Forgetting the surviving last term in the limit

once the middle cancels, it is easy to declare the sum equals b_1 and stop, ignoring -b_{n+1}

wrong$$\sum_{n=1}^{\infty}\Big(\tfrac1n-\tfrac1{n+1}\Big)\stackrel{?}{=}b_1=1\ \text{(by luck right here)},\ \text{but}\ s_n=1-\tfrac1{n+1}$$
right$$\sum=b_1-\lim_{n\to\infty}b_{n+1}=1-0=1$$
⚠ Dropping the constant factor from the partial-fraction split

when the two linear factors differ by more than 1, the split carries a coefficient like 1/2 that is easy to lose

wrong$$\frac{1}{(2n-1)(2n+1)}\stackrel{?}{=}\frac{1}{2n-1}-\frac{1}{2n+1}$$
right$$\frac{1}{(2n-1)(2n+1)}=\frac12\!\left(\frac{1}{2n-1}-\frac{1}{2n+1}\right)$$

2.4The harmonic series diverges, even though its terms vanish

Shows by grouping that $\sum 1/n$ has no sum, the standard warning that $a_n\to 0$ is not enough for convergence.

We have summed several series exactly; here is a famous one whose terms shrink to zero and that still has no sum, and it is the reason the next test has to be worded so carefully.

TheoremTheorem: the harmonic series diverges
Conditions
  • the terms are $a_n=\dfrac1n$, which do tend to $0$

$$\boxed{\ \sum_{n=1}^{\infty}\frac1n=1+\frac12+\frac13+\frac14+\cdots=\infty.\ }$$

The sum of the reciprocals of the whole numbers grows without bound. The individual terms get as small as you like, but there are so many of them that the running total still climbs past every ceiling.

Looks like this, but is not

Because the terms $\tfrac1n$ march down to $0$, the running total looks like it should flatten out and settle, exactly as $\sum\tfrac1{2^n}$ did.

Shrinking terms only stop the total from jumping; they do not stop it from creeping upward forever. The grouping shows the creep never runs out: every doubling of the term count buys another $\tfrac12$, and you can double as often as you please.

Why halving every term does not rescue the harmonic series

Does $\displaystyle\sum_{n=1}^{\infty}\frac{1}{2n}=\frac12+\frac14+\frac16+\cdots$ converge?

Given
  • $a_n=\dfrac{1}{2n}$, which tends to $0$.

Find

Convergence or divergence.

Solution
Pull the constant out
$$\sum_{n=1}^{\infty}\frac{1}{2n}=\frac12\sum_{n=1}^{\infty}\frac1n$$

every term carries the same factor 1/2, so it factors out of the whole sum

Use the harmonic result
$$\frac12\cdot\infty=\infty$$

a nonzero constant times a divergent series still diverges; the 1/2 cannot tame an infinite total

Answer $$\boxed{\text{Diverges}}$$
Check

Check the intuition: the terms are just half the harmonic terms, and half of something unbounded is still unbounded.

A shifted harmonic tail still diverges

Does $\displaystyle\sum_{n=1}^{\infty}\frac{1}{n+5}=\frac16+\frac17+\frac18+\cdots$ converge?

Given
  • $a_n=\dfrac{1}{n+5}$

Find

Convergence or divergence.

Solution
Recognise it as the harmonic series with the first five terms removed
$$\sum_{n=1}^{\infty}\frac{1}{n+5}=\sum_{m=6}^{\infty}\frac1m$$

reindex with m = n+5; this is the harmonic series starting from the sixth term

Apply the finite-edit rule
$$\text{dropping } 1+\tfrac12+\tfrac13+\tfrac14+\tfrac15 \text{ (finite) cannot change divergence}$$

removing finitely many terms never converts a divergent series into a convergent one

Answer $$\boxed{\text{Diverges}}$$
Check

Consistent with the convention on finite edits: only the tail decides convergence, and this tail is the harmonic tail.

Neither a constant multiple nor a finite shift can rescue a divergent series; both are recurring exam traps.

Checkpoint
§02.4 — does a smaller-looking harmonic-type series converge?●●○○○

A classmate says: the terms of $\sum \tfrac{1}{3n}$ are tiny, only a third of the plain reciprocals, so surely this one adds up to a finite number.

Given
  • Claim: $\displaystyle\sum_{n=1}^{\infty}\frac{1}{3n}$ converges.

Find
  1. (a) True or false, with a one-line reason.

Hint 1/4

Ask whether a constant factor can change convergence at all.

Hint 2/4

Pull the constant out: $\sum\tfrac{1}{3n}=\tfrac13\sum\tfrac1n$, and reuse the harmonic result.

Hint 3/4

Since $\sum\tfrac1n$ diverges, one third of it is still an unbounded total.

Hint 4/4

So the claim is false: $\tfrac13\cdot\infty=\infty$.

Show solution
Factor and compare to harmonic
$$\sum_{n=1}^{\infty}\frac{1}{3n}=\frac13\sum_{n=1}^{\infty}\frac1n=\frac13\cdot\infty=\infty$$

the 1/3 factors out; the remaining series is the divergent harmonic series

Answer $$\boxed{\text{False (diverges)}}$$
Check

Check: the terms 1/3, 1/6, 1/9, ... are exactly one third of 1, 1/2, 1/3, ..., so their totals are one third of unbounded totals.

⚠ Believing a constant multiple can make a divergent series converge

shrinking each term by a fixed factor feels like it should help, but it only rescales an already-infinite total

wrong$$\sum\frac{1}{100\,n}\stackrel{?}{<}\infty$$
right$$\sum\frac{1}{100\,n}=\frac{1}{100}\sum\frac1n=\infty$$
⚠ Reading tiny terms as a guarantee of a finite sum

the terms of the harmonic series obviously go to 0, and the mind equates small terms with a controlled total

wrong$$a_n=\tfrac1n\to 0\ \stackrel{?}{\Rightarrow}\ \sum a_n<\infty$$
right$$a_n\to 0\ \text{is necessary, not sufficient: } \sum\tfrac1n=\infty$$

2.5The Test for Divergence

Rejects any series whose terms fail to reach $0$; silent, by design, whenever the terms do reach $0$.

The harmonic series showed that shrinking terms are not enough for a sum. Turn that around and you get the fastest test there is: if the terms do not even shrink to zero, the series is doomed.

TheoremTheorem: Test for Divergence
Conditions
  • $a_n$ is the general term of the series $\sum a_n$

$$\boxed{\ \text{If }\lim_{n\to\infty}a_n\neq 0\text{ or does not exist, then }\sum_{n=1}^{\infty} a_n\text{ diverges.}\ }$$

If the terms do not settle down to zero, the running total keeps getting kicked by pieces of a fixed size and cannot possibly converge. The test only ever proves divergence; it never proves convergence.

Looks like this, but is not

It is tempting to read the test as a two-way rule and conclude that $a_n\to 0$ proves $\sum a_n$ converges.

That direction is exactly what the harmonic series refutes: $\tfrac1n\to 0$ and yet $\sum\tfrac1n=\infty$. When $a_n\to 0$ the Test for Divergence returns no verdict at all, and you must reach for a different test.

Rejecting a series whose terms head to a nonzero limit

Determine whether $\displaystyle\sum_{n=1}^{\infty}\frac{n}{2n+1}$ converges.

Given
  • $a_n=\dfrac{n}{2n+1}$

Find

Convergence or divergence.

Solution
Take the limit of the terms
$$\lim_{n\to\infty}\frac{n}{2n+1}=\lim_{n\to\infty}\frac{1}{2+1/n}=\frac12$$

equal-degree quotient, so the limit is the ratio of leading coefficients

Apply the Test for Divergence
$$\tfrac12\neq 0\ \Rightarrow\ \text{diverges}$$

the terms settle near 1/2, so each new one pushes the total up by about a half

Answer $$\boxed{\text{Diverges}}$$
Check

Check: adding numbers that stay near 1/2 forever obviously sends the total to infinity.

Terms that oscillate: the limit does not exist

Determine whether $\displaystyle\sum_{n=1}^{\infty}(-1)^{\,n}$ converges.

Given
  • $a_n=(-1)^{\,n}$, i.e. $-1,+1,-1,+1,\dots$

Find

Convergence or divergence.

Solution
Examine the limit of the terms
$$\lim_{n\to\infty}(-1)^{\,n}\ \text{does not exist}$$

the terms flip between -1 and +1 and never settle

Apply the Test for Divergence
$$\text{limit not }0\ (\text{in fact DNE})\ \Rightarrow\ \text{diverges}$$

the partial sums bounce -1, 0, -1, 0 and never converge

Answer $$\boxed{\text{Diverges}}$$
Check

Check by partial sums: they read -1, 0, -1, 0, ..., which has no single limit, matching the verdict.

A limit of the terms that fails to exist counts as not zero for this test, so oscillating terms are an immediate rejection.

Checkpoint
§02.5 — can this test ever prove the harmonic series diverges?●●●○○

Careful reading of what the test can and cannot do. A classmate applies the Test for Divergence to $\sum\tfrac1n$ and reports that, because $\lim\tfrac1n=0$, the test returns no verdict here.

Given
  • Claim: applied to $\displaystyle\sum_{n=1}^{\infty}\frac1n$, the Test for Divergence is inconclusive, because $\lim_{n\to\infty}\tfrac1n=0$.

Find
  1. (a) True or false, and say what the test actually returns here.

Hint 1/4

Ask what the test needs to see in order to say anything.

Hint 2/4

The Test for Divergence fires only when $\lim a_n\neq 0$; if $\lim a_n=0$ it is silent.

Hint 3/4

Here $\lim_{n\to\infty}\tfrac1n=0$, so the test returns no verdict at all.

Hint 4/4

So the claim is true: the harmonic series does diverge, but by grouping, not by this test — here the test itself is silent.

Show solution
Check the trigger of the test
$$\lim_{n\to\infty}\frac1n=0$$

the test only concludes divergence when this limit is nonzero or missing

Read off the verdict
$$\text{limit}=0\ \Rightarrow\ \text{test inconclusive}$$

a zero limit tells the test nothing; divergence must come from the grouping argument

Answer $$\boxed{\text{True}}$$
Check

Check against the concept: this is the exact case the converse-is-false warning was written for.

⚠ Concluding convergence from a_n -> 0

the test is remembered as a rule about the terms going to zero, and the mind completes it into a two-way street

wrong$$a_n\to 0\ \stackrel{?}{\Rightarrow}\ \sum a_n\ \text{converges}$$
right$$a_n\to 0\ \Rightarrow\ \text{test says nothing; try another test}$$
⚠ Claiming the Test for Divergence proves convergence

having one fast test, students hope it is a full decision procedure rather than a one-sided rejection rule

wrong$$\lim a_n=0\ \stackrel{?}{\Rightarrow}\ \text{Test for Divergence proves it converges}$$
right$$\text{The test can only ever prove DIVERGENCE, never convergence.}$$

2.6Combining series: linearity and its one trap

Adds and scales convergent series term by term, and flags the case where a convergent plus a divergent series must diverge.

Once you can sum the standard types, real problems hand you a mix; the rules for adding and scaling series let you take them one recognizable piece at a time.

RuleRule: linearity of convergent series
Conditions
  • $\sum a_n$ and $\sum b_n$ both converge

  • $c$ is a constant

$$\boxed{\ \sum c\,a_n=c\sum a_n,\qquad \sum(a_n\pm b_n)=\sum a_n\pm\sum b_n.\ }$$

You may pull a constant out of a convergent series and you may add two convergent series term by term, in each case getting the sum you would expect. The catch is that both pieces must already converge for the term-by-term rule to be legal.

Looks like this, but is not

The rule $\sum(a_n+b_n)=\sum a_n+\sum b_n$ looks like it should hold no matter what, so one might split $\sum\big(\tfrac1n-\tfrac1n\big)$ into $\sum\tfrac1n-\sum\tfrac1n$.

The left side is $\sum 0=0$, but the right side is $\infty-\infty$, which is meaningless. The rule needs both series to converge; here neither does, so splitting is illegal even though the combined series is perfectly well behaved.

Summing a mixture of a geometric and a telescoping series

Evaluate $\displaystyle\sum_{n=1}^{\infty}\left(\frac{3}{4^{\,n}}+\frac{1}{n(n+1)}\right)$.

Given
  • geometric piece $\dfrac{3}{4^{\,n}}$

  • telescoping piece $\dfrac{1}{n(n+1)}$

Find

The total sum.

Solution
Confirm both pieces converge, then split
$$\sum\frac{3}{4^{\,n}}\ \text{and}\ \sum\frac{1}{n(n+1)}\ \text{both converge}$$

both are standard convergent types, so linearity is legal

Sum the geometric piece
$$\sum_{n=1}^{\infty}\frac{3}{4^{\,n}}=\frac{3/4}{1-1/4}=1$$

first term 3/4 at n=1, ratio 1/4

Sum the telescoping piece
$$\sum_{n=1}^{\infty}\frac{1}{n(n+1)}=1$$

computed earlier: it collapses to 1 - 1/(n+1) -> 1

Add the two sums
$$1+1=2$$

linearity lets the two convergent totals be added

Answer $$\boxed{2}$$
Check

Check the pieces separately: the geometric part is under 1 (0.75 + 0.19 + ...) and the telescoping part is under 1 too, and both edge up to 1.

A convergent piece plus a divergent piece

Determine whether $\displaystyle\sum_{n=1}^{\infty}\left(\frac{1}{2^{\,n}}+\frac1n\right)$ converges.

Given
  • $\sum\dfrac{1}{2^{\,n}}$ converges to 1

  • $\sum\dfrac1n$ diverges

Find

Convergence or divergence.

Solution
Classify each piece
$$\sum\frac{1}{2^{\,n}}\ \text{converges},\quad \sum\frac1n\ \text{diverges}$$

the first is geometric with r=1/2, the second is harmonic

Apply the convergent-plus-divergent rule
$$\text{convergent}+\text{divergent}=\text{divergent}$$

if the sum converged, subtracting the convergent geometric part would make the harmonic series converge, which it does not

Answer $$\boxed{\text{Diverges}}$$
Check

Check the logic by contradiction: a finite total minus a finite total is finite, but the harmonic total is not, so the combined total cannot be finite.

A single divergent ingredient poisons the whole sum; you cannot average it away with a convergent one.

Checkpoint
§02.6 — scale and add two convergent series●●●○○

Use linearity to evaluate a combination of two geometric series you can each sum by formula.

Given
  • $\displaystyle\sum_{n=1}^{\infty}\left(\frac{2}{3^{\,n}}+\frac{5}{6^{\,n}}\right)$

Find
  1. (a) Find the total, summing each piece and adding.

Hint 1/4

Split into two geometric series and sum each on its own before adding.

Hint 2/4

For $\sum \tfrac{2}{3^{n}}$ the first term is $\tfrac23$ and ratio $\tfrac13$; for $\sum\tfrac{5}{6^{n}}$ first term $\tfrac56$, ratio $\tfrac16$.

Hint 3/4

So the pieces are $\dfrac{2/3}{1-1/3}=1$ and $\dfrac{5/6}{1-1/6}=1$.

Hint 4/4

Add them: $1+1=2$.

Show solution
First geometric piece
$$\sum_{n=1}^{\infty}\frac{2}{3^{\,n}}=\frac{2/3}{1-1/3}=\frac{2/3}{2/3}=1$$

first term 2/3, ratio 1/3

Second geometric piece
$$\sum_{n=1}^{\infty}\frac{5}{6^{\,n}}=\frac{5/6}{1-1/6}=\frac{5/6}{5/6}=1$$

first term 5/6, ratio 1/6

Add
$$1+1=2$$

both pieces converge, so their totals add

Answer $$\boxed{2}$$
Check

Check each ratio is below 1 (1/3 and 1/6), so both formulas were legal to use.

⚠ Splitting a series when the pieces diverge

linearity is memorized without its hypothesis, so it gets applied to pieces that do not converge

wrong$$\sum\Big(\tfrac1n-\tfrac1n\Big)\stackrel{?}{=}\sum\tfrac1n-\sum\tfrac1n=\infty-\infty$$
right$$\sum\Big(\tfrac1n-\tfrac1n\Big)=\sum 0=0\ \text{(do not split divergent pieces)}$$
⚠ Hoping a convergent piece can rescue a divergent one

adding a well-behaved series feels like it should calm the sum down, but divergence dominates

wrong$$\sum\Big(\tfrac{1}{2^{n}}+\tfrac1n\Big)\stackrel{?}{<}\infty$$
right$$\text{convergent}+\text{divergent}=\text{divergent}$$
First moves on any series question

Every convergence or sum question in this section, and the opening line of most series questions later in the course.

  1. Terms to zero?

    Compute $\lim_{n\to\infty}a_n$. If it is not $0$ (or fails to exist), stop: the series diverges by the Test for Divergence.

  2. Geometric?

    Check whether $a_{n+1}/a_n$ is a constant $r$. If so, and $|r|<1$, the sum is $a/(1-r)$ with $a$ the first term; if $|r|\ge 1$ it diverges.

  3. Telescoping?

    Try to split $a_n=b_n-b_{n+1}$ by partial fractions or a log identity. If it splits, $s_n=b_1-b_{n+1}$ and you take the limit.

  4. None of these?

    If the terms go to $0$ but the series is neither geometric nor telescoping, this section cannot finish it; the tests of the next section take over.

Where it goes wrong
  • Skipping step 1 and grinding on a series that dies instantly.

  • Calling a series geometric when the ratio is not actually constant.

Turning a repeating decimal into a fraction

Any repeating decimal such as 0.overline(27) or 2.3overline(45).

  1. Find the block

    Identify the repeating block and its length $k$; the ratio will be $r=10^{-k}$.

  2. First term

    The first term $a$ is the block placed just after the decimal point, e.g. $\tfrac{27}{100}$ for a two-digit block.

  3. Sum

    Add any non-repeating front part separately, then use $a/(1-r)$ for the repeating tail.

Where it goes wrong
  • Using $r=1/10$ for a two-digit block instead of $r=1/100$.

  • Taking $a$ to be the block itself (27) rather than the block over its place value (27/100).

Geometric with |r| < 1: converges

$\displaystyle\sum_{n=1}^{\infty}\left(\tfrac12\right)^{\,n-1}$

Given
  • ratio $r=\tfrac12$

Find

The sum.

Solution
Ratio below 1, so sum by formula
$$\frac{a}{1-r}=\frac{1}{1-\tfrac12}=2$$

first term 1, ratio 1/2 < 1

Answer $$\boxed{2}$$
Check

Totals 1, 1.5, 1.75, ... close in on 2.

Geometric with |r| >= 1: diverges

$\displaystyle\sum_{n=1}^{\infty}\left(\tfrac32\right)^{\,n-1}$

Given
  • ratio $r=\tfrac32$

Find

Convergence or divergence.

Solution
Ratio at least 1, so terms grow
$$|r|=\tfrac32\ge 1\ \Rightarrow\ \text{diverges}$$

the terms 1, 1.5, 2.25, ... increase and never shrink

Answer $$\boxed{\text{Diverges}}$$
Check

Totals 1, 2.5, 4.75, ... run off with no ceiling.

Both are geometric with first term 1; only the size of $r$ differs, and that single number decides convergence.

How to tell them apart

Compute $|r|$ before anything else: below 1 means sum $=a/(1-r)$, at least 1 means diverges.

Terms to zero, converges: 1/(n(n+1))

$\displaystyle\sum_{n=1}^{\infty}\frac{1}{n(n+1)}$

Given
  • $a_n=\tfrac{1}{n(n+1)}\to 0$

Find

The sum.

Solution
Telescopes to a finite value
$$s_n=1-\tfrac1{n+1}\to 1$$

the split 1/n - 1/(n+1) collapses

Answer $$\boxed{1}$$
Check

A genuine finite sum.

Terms to zero, diverges: 1/n

$\displaystyle\sum_{n=1}^{\infty}\frac1n$

Given
  • $a_n=\tfrac1n\to 0$

Find

Convergence or divergence.

Solution
Grouping shows unbounded totals
$$s_{2^{k}}>1+\tfrac{k}{2}\to\infty$$

each doubling block adds more than 1/2

Answer $$\boxed{\text{Diverges}}$$
Check

Same shrinking terms, opposite fate.

Both have terms tending to $0$, yet one converges and one diverges, which is precisely why $a_n\to 0$ can never be the end of the story.

How to tell them apart

$a_n\to 0$ only clears the Test for Divergence; you still need a real method (telescoping here, grouping there) to decide.

Scaffolding comes off
The common skeleton
  1. Name the standard series type (geometric or telescoping).

  2. State and check that type's convergence condition.

  3. Get a closed form for the partial sum s_n by the type's method.

  4. Take the limit of s_n, then sanity-check the number.

1 · fully worked

Rung 1 worked in full: sum of 3*(1/4)^(n-1)

Compute $\displaystyle\sum_{n=1}^{\infty} 3\left(\tfrac14\right)^{\,n-1}$.

Given
  • first term $a=3$

  • ratio $r=\tfrac14$

Find

The sum.

Solution
Step 1 and 2 name the type and check the condition
$$\text{geometric},\ |r|=\tfrac14<1$$

constant ratio 1/4, size below 1, so it converges and the formula is legal

Step 3 get the closed form for s_n
$$s_n=\frac{a(1-r^{\,n})}{1-r}=\frac{3\big(1-(1/4)^{n}\big)}{3/4}=4\big(1-(1/4)^{n}\big)$$

the geometric partial-sum formula with a=3, r=1/4

Step 4 take the limit and check
$$\lim_{n\to\infty}4\big(1-(1/4)^{n}\big)=4$$

the moving piece (1/4)^n vanishes

Answer $$\boxed{4}$$
Check

Partial sums 3, 3.75, 3.9375 already sit just under 4.

2 · you write the reasoning

An easier one, and this time you write the reasons. Compute $$\sum_{n=1}^{\infty}\left(\tfrac12\right)^{\,n-1}$$ and, in the empty column, say why each line is allowed.

  1. reasoning

    The ratio between consecutive terms is the constant 1/2, and its size is below 1, so this is a convergent geometric series and the formula applies.

  2. reasoning

    Apply the geometric partial-sum formula with first term a = 1 and ratio r = 1/2.

  3. reasoning

    As n grows, (1/2)^n tends to 0, so the only moving part disappears and the total is 2.

3 · find the buried error

Harder, and the work is done for you, badly. Two of the four steps are wrong. Find both. The problem: sum the telescoping series $$\sum_{n=1}^{\infty}\frac{1}{n(n+2)}.$$

the two buried errors (2)
⚠ step 2

The partial-fraction split is missing its coefficient. Solving $1=A(n+2)+Bn$ gives $A=\tfrac12,\ B=-\tfrac12$, so $\dfrac{1}{n(n+2)}=\dfrac12\!\left(\dfrac1n-\dfrac1{n+2}\right)$, with a factor of $\tfrac12$.

when the two factors differ by more than 1, the numerators are not plus and minus 1, but the eye expects the clean 1/n - 1/(n+2) pattern from the differ-by-1 case.

right

Keep the $\tfrac12$: $\dfrac{1}{n(n+2)}=\dfrac12\!\left(\dfrac1n-\dfrac1{n+2}\right)$.

⚠ step 3

With a gap of 2, two terms survive at each end, not one. The collapse leaves $s_n=\tfrac12\!\left(1+\tfrac12-\tfrac1{n+1}-\tfrac1{n+2}\right)$, because 1 and 1/2 both fail to find a partner.

the differ-by-1 telescope leaves a single front survivor, and that habit gets copied to the gap-2 case where in fact two front terms and two back terms are left over.

right

$s_n=\tfrac12\!\left(1+\tfrac12-\tfrac1{n+1}-\tfrac1{n+2}\right)\to\tfrac12\cdot\tfrac32=\tfrac34.$

4 · the bare problem
§02 — bare geometric sum, no scaffolding●●●○○

Same skeleton, no help this time. Name the type, check its condition, get s_n, take the limit, and sanity-check before you look.

Given
  • $\displaystyle\sum_{n=1}^{\infty} 5\left(-\tfrac13\right)^{\,n-1}$

Find
  1. (a) Find the sum.

Hint 1/4

Two things to settle first: which type this is and whether it meets that type's condition.

Hint 2/4

It is geometric with first term $a=5$ and ratio $r=-\tfrac13$; since $|{-1/3}|<1$ it converges.

Hint 3/4

Apply $\dfrac{a}{1-r}=\dfrac{5}{1-(-1/3)}=\dfrac{5}{4/3}$; the sign of $r$ stays negative inside $1-r$.

Hint 4/4

So the sum is $\dfrac{5}{4/3}=\dfrac{15}{4}$.

Show solution
Check the condition
$$|r|=\tfrac13<1$$

converges, so the formula applies

Apply the formula, watching the sign
$$\frac{5}{1-(-\tfrac13)}=\frac{5}{1+\tfrac13}=\frac{5}{4/3}=\frac{15}{4}$$

1 - r with r = -1/3 becomes 1 + 1/3 = 4/3

Answer $$\boxed{\tfrac{15}{4}}$$
Check

Check with partial sums: 5, 5 - 1.667 = 3.333, 3.333 + 0.556 = 3.889, closing in on 3.75.

Full exam-style question

Exam-style: sum a mixture of a telescoping and a geometric seriesexam format

Evaluate $$\sum_{n=1}^{\infty}\left(\frac{2}{n(n+1)}+3\left(\tfrac15\right)^{\,n}\right).$$

Given
  • telescoping piece $\dfrac{2}{n(n+1)}$

  • geometric piece $3\left(\tfrac15\right)^{\,n}$

Find

The exact value of the sum.

Solution
Confirm both pieces converge so linearity is legal
$$\sum\frac{2}{n(n+1)}\ \text{telescopes};\quad \sum 3\left(\tfrac15\right)^{n}\ \text{geometric},\ |r|=\tfrac15<1$$

both are standard convergent series, so the sum may be split

Sum the telescoping piece
$$\sum_{n=1}^{\infty}\frac{2}{n(n+1)}=2\sum_{n=1}^{\infty}\Big(\frac1n-\frac1{n+1}\Big)=2\cdot 1=2$$

the inner sum collapses to 1, then the factor 2 comes along by linearity

Sum the geometric piece (mind the first term)
$$\sum_{n=1}^{\infty}3\left(\tfrac15\right)^{n}=\frac{3\cdot\tfrac15}{1-\tfrac15}=\frac{3/5}{4/5}=\frac34$$

at n=1 the first term is 3*(1/5)=3/5, not 3; ratio 1/5

Add the two sums
$$2+\frac34=\frac{11}{4}$$

linearity permits adding the two convergent totals

Answer $$\boxed{\dfrac{11}{4}}$$
Check

Independent check: the telescoping part is exactly 2, and the geometric part 0.6 + 0.12 + 0.024 + ... clearly totals about 0.75, so 2.75 = 11/4 is right.

Two standard sums plus one addition; the only trap is reading the geometric first term as 3 instead of 3/5.

Exam mixtures are almost always linearity over pieces you already know: split, sum each, add.

Practice

A · concept 4 questions
1§02.4 — shrinking terms and convergence●●○○○

A study partner writes down a rule of thumb and wants you to confirm it before the quiz.

Given
  • Rule of thumb: if $\lim_{n\to\infty}a_n=0$, then $\sum a_n$ converges.

Find
  1. (a) True or false? If false, name a counterexample.

Hint 1/4

Ask whether the terms going to zero has ever been shown to be enough on its own.

Hint 2/4

The condition $a_n\to 0$ is necessary for convergence but not sufficient.

Hint 3/4

The harmonic series $\sum\tfrac1n$ has $a_n\to 0$ yet diverges, which kills the rule.

Hint 4/4

So the statement is false, with $\sum\tfrac1n$ as the counterexample.

Show solution
Produce the standard counterexample
$$a_n=\tfrac1n\to 0,\ \text{but}\ \sum\tfrac1n=\infty$$

the harmonic series is the canonical witness that a zero limit does not force convergence

Answer $$\boxed{\text{False}}$$
Check

Check: this is exactly the converse the Test for Divergence warns against.

2§02.2 — is the geometric formula unconditional?●●○○○

Another rule of thumb from the same partner, this time about geometric series.

Given
  • Rule of thumb: $\displaystyle\sum_{n=1}^{\infty} ar^{\,n-1}=\frac{a}{1-r}$ holds only when $|r|<1$; for $|r|\ge 1$ the series diverges.

Find
  1. (a) True or false? If true, say what goes wrong once $|r|\ge 1$.

Hint 1/4

Ask what has to be true of $r$ for the running totals to settle.

Hint 2/4

The formula only holds when $|r|<1$; otherwise the series diverges.

Hint 3/4

For example $r=2$ gives terms $1,2,4,\dots$ whose totals blow up, yet the formula would print $\tfrac{1}{1-2}=-1$.

Hint 4/4

So the statement is true: $|r|<1$ is exactly the condition the theorem carries, and $r=2$ shows it is genuinely needed.

Show solution
Confirm the condition with a large ratio
$$r=2:\ \sum 2^{n-1}=1+2+4+\cdots=\infty,\ \text{yet}\ \tfrac{1}{1-2}=-1$$

a divergent series cannot equal a finite negative number, so the stated condition $|r|<1$ is genuinely required

Answer $$\boxed{\text{True}}$$
Check

Check: a sum of positive growing terms can never be negative, confirming the $|r|<1$ hypothesis is genuinely needed.

3§02.5 — which series the Test for Divergence rejects●●○○○

The Test for Divergence only rejects; pick the one series below that it actually rejects.

Given
  • Four series to weigh, each judged only by $\lim_{n\to\infty}a_n$.

Find
  1. (a) Which series diverges by the Test for Divergence?

Hint 1/4

Apply one move: take the limit of the terms of each series.

Hint 2/4

The test rejects a series exactly when $\lim a_n\neq 0$; a zero limit leaves the test silent.

Hint 3/4

Three of these have $a_n\to 0$ (so the test says nothing); one has $a_n\to 1$.

Hint 4/4

The one with $a_n=\tfrac{n}{n+4}\to 1\neq 0$ is rejected by the test.

Show solution
Take the limit of each general term
$$\tfrac{n}{n+4}\to 1,\quad \tfrac{1}{n+4}\to 0,\quad \tfrac{1}{2^{n}}\to 0,\quad \tfrac{1}{n(n+1)}\to 0$$

only the first fails to reach 0

Reject the nonzero-limit series
$$\lim\tfrac{n}{n+4}=1\neq 0\ \Rightarrow\ \text{diverges}$$

the other three are left undecided by this test

Answer $$\boxed{\sum \tfrac{n}{n+4}}$$
Check

Check: n/(n+4) is near 1 for large n, so adding it forever clearly runs off.

4§02.6 — can scaling fix a divergent series?●●○○○

Last rule of thumb before the computations.

Given
  • Rule of thumb: multiplying every term of a divergent series by a small constant like $\tfrac{1}{1000}$ can make it converge.

Find
  1. (a) True or false, with a one-line reason.

Hint 1/4

Ask whether a constant factor can change the fate of a series at all.

Hint 2/4

By linearity $\sum c\,a_n=c\sum a_n$, so a nonzero $c$ never changes convergence, only the value.

Hint 3/4

If $\sum a_n=\infty$ then $\tfrac{1}{1000}\sum a_n=\infty$ as well.

Hint 4/4

So the statement is false; scaling cannot rescue divergence.

Show solution
Factor the constant out
$$\sum \tfrac{1}{1000}a_n=\tfrac{1}{1000}\sum a_n=\tfrac{1}{1000}\cdot\infty=\infty$$

a nonzero multiple of an unbounded total is still unbounded

Answer $$\boxed{\text{False}}$$
Check

Check: the harmonic example one-third and one-hundredth both still diverge.

B · computation 5 questions
1§02.2 — a plain geometric sum●●○○○

A routine geometric sum to lock in the formula and the first-term reading.

Given
  • $\displaystyle\sum_{n=1}^{\infty} 5\left(\tfrac23\right)^{\,n-1}$

Find
  1. (a) Find the sum.

Hint 1/4

Read off the first term and the ratio before touching the formula.

Hint 2/4

Geometric with $|r|<1$ sums to $a/(1-r)$; here $a$ is the $n=1$ term.

Hint 3/4

At $n=1$ the term is $5\left(\tfrac23\right)^{0}=5$, and $r=\tfrac23$.

Hint 4/4

So the sum is $\dfrac{5}{1-2/3}=\dfrac{5}{1/3}=15$.

Show solution
Check and apply
$$|r|=\tfrac23<1,\quad \frac{5}{1-\tfrac23}=\frac{5}{\tfrac13}=15$$

condition holds, so the formula gives 15

Answer $$\boxed{15}$$
Check

Check: 5 + 3.33 + 2.22 + ... piles up toward 15.

2§02.3 — a telescoping sum with a half factor●●●○○

A telescoping series where the split carries a coefficient; keep it through the collapse.

Given
  • $\displaystyle\sum_{n=1}^{\infty}\frac{1}{(2n+1)(2n+3)}$

  • split: $\dfrac{1}{(2n+1)(2n+3)}=\dfrac12\!\left(\dfrac{1}{2n+1}-\dfrac{1}{2n+3}\right)$

Find
  1. (a) Find the sum.

Hint 1/4

Write out two or three partial-sum terms using the given split and see what cancels.

Hint 2/4

With the $\tfrac12$ out front, the telescope leaves the first surviving piece minus the last.

Hint 3/4

$s_n=\tfrac12\!\left(\tfrac13-\tfrac1{2n+3}\right)$, since the front survivor is $\tfrac13$ (from $n=1$).

Hint 4/4

Let $n\to\infty$: the sum is $\tfrac12\cdot\tfrac13=\tfrac16$.

Show solution
Collapse keeping the coefficient
$$s_n=\tfrac12\!\left[\Big(\tfrac13-\tfrac15\Big)+\Big(\tfrac15-\tfrac17\Big)+\cdots\right]=\tfrac12\!\left(\tfrac13-\tfrac1{2n+3}\right)$$

the first term at n=1 is 1/(3*5), so the front survivor is 1/3

Limit
$$\lim_{n\to\infty}\tfrac12\!\left(\tfrac13-\tfrac1{2n+3}\right)=\tfrac16$$

the tail vanishes

Answer $$\boxed{\tfrac16}$$
Check

Check first term 1/(3*5)=1/15; the total 1/6 is a bit larger, consistent with small positive tails.

3§02.2 — a repeating decimal as a fraction●●○○○

Convert a purely repeating decimal to an exact fraction using a geometric series.

Given
  • $0.\overline{72}=0.727272\ldots$

Find
  1. (a) Write it as a fraction in lowest terms.

Hint 1/4

Express the decimal as a sum of copies of the repeating block, each shifted two places.

Hint 2/4

It is geometric with first term $\tfrac{72}{100}$ and ratio $\tfrac{1}{100}$.

Hint 3/4

Sum: $\dfrac{72/100}{1-1/100}=\dfrac{72/100}{99/100}=\dfrac{72}{99}$.

Hint 4/4

Reduce: $\dfrac{72}{99}=\dfrac{8}{11}$.

Show solution
Series form and sum
$$\sum_{k=1}^{\infty}\frac{72}{100^{k}}=\frac{72/100}{99/100}=\frac{72}{99}$$

first term 72/100, ratio 1/100

Reduce
$$\frac{72}{99}=\frac{8}{11}$$

divide top and bottom by 9

Answer $$\boxed{\tfrac{8}{11}}$$
Check

Check: 8 divided by 11 is 0.7272..., matching the decimal.

4§02.1 — sum and a term from a partial-sum formula●●○○○

You are given the running totals directly; extract the sum and one term.

Given
  • $s_n=\dfrac{4n}{2n+1}$

Find
  1. (a) Find the sum of the series.

  2. (b) Find a_2.

Hint 1/4

The sum is the limit of the totals; a term is a difference of consecutive totals.

Hint 2/4

Use $\sum a_n=\lim s_n$ and $a_2=s_2-s_1$.

Hint 3/4

$\lim\dfrac{4n}{2n+1}=2$; and $s_2=\dfrac{8}{5},\ s_1=\dfrac{4}{3}$.

Hint 4/4

So the sum is $2$ and $a_2=\dfrac{8}{5}-\dfrac{4}{3}=\dfrac{4}{15}$.

Show solution
Sum as a limit
$$\lim_{n\to\infty}\frac{4n}{2n+1}=\frac{4}{2}=2$$

equal degrees, ratio of leading coefficients

Second term as a jump
$$a_2=s_2-s_1=\frac{8}{5}-\frac{4}{3}=\frac{24-20}{15}=\frac{4}{15}$$

subtract consecutive totals

Answer $$\boxed{\sum a_n=2,\ a_2=\tfrac{4}{15}}$$
Check

Check: a_2 is small and positive since the totals are already near 2 and still creeping up.

5§02.2 — rewrite before summing●●●○○

The series is geometric once you rewrite it; do the algebra first, then read off a and r.

Given
  • $\displaystyle\sum_{n=1}^{\infty}\frac{3^{\,n}}{4^{\,n+1}}$

Find
  1. (a) Find the sum.

Hint 1/4

Pull the constant part of the denominator out to expose a clean ratio.

Hint 2/4

Rewrite $\dfrac{3^{\,n}}{4^{\,n+1}}=\dfrac14\left(\dfrac34\right)^{\,n}$, then it is geometric.

Hint 3/4

First term (at $n=1$) is $\dfrac14\cdot\dfrac34=\dfrac{3}{16}$, ratio $\dfrac34$.

Hint 4/4

Sum: $\dfrac14\cdot\dfrac{3/4}{1-3/4}=\dfrac14\cdot 3=\dfrac34$.

Show solution
Rewrite and pull out the constant
$$\sum_{n=1}^{\infty}\frac{3^{n}}{4^{n+1}}=\frac14\sum_{n=1}^{\infty}\left(\frac34\right)^{n}$$

4^{n+1}=4*4^n, so 1/4 factors out

Sum the inner geometric series
$$\sum_{n=1}^{\infty}\left(\tfrac34\right)^{n}=\frac{3/4}{1-3/4}=3$$

first term 3/4, ratio 3/4

Multiply back
$$\frac14\cdot 3=\frac34$$

restore the pulled-out constant

Answer $$\boxed{\tfrac34}$$
Check

Check: first term 3/16 = 0.1875, then 0.1406, ..., piling toward 0.75.

C · exam level 3 questions
1§02.2 — bouncing ball total distance●●●○○

A worked-example type dressed as a quiz. A ball is dropped and rebounds to a fixed fraction of each height forever; the tape adds every rise and fall.

Given
  • Dropped from $4$ m, each rebound reaches $\tfrac12$ of the previous height.

Find
  1. (a) What is the total distance travelled?

Hint 1/4

Separate the single first drop from the up-and-down pairs of every bounce.

Hint 2/4

Total $=4+\sum_{k\ge 1}2\,h_k$ with $h_k=4\left(\tfrac12\right)^{k}$, since each bounce goes up $h_k$ and down $h_k$.

Hint 3/4

The bounce sum is $\sum_{k\ge1}2\cdot4\left(\tfrac12\right)^{k}=8\cdot\dfrac{1/2}{1-1/2}=8$.

Hint 4/4

Total $=4+8=12$ m.

Show solution
Set up drop plus bounce pairs
$$D=4+\sum_{k=1}^{\infty}2\cdot4\left(\tfrac12\right)^{k}$$

the first drop 4 is unmatched; each later bounce is a rise and a fall of h_k

Sum the geometric bounces
$$\sum_{k=1}^{\infty}8\left(\tfrac12\right)^{k}=8\cdot\frac{1/2}{1-1/2}=8$$

first term 8*(1/2)=4, ratio 1/2, sum 8

Add the drop
$$D=4+8=12$$

combine

Answer $$\boxed{12\ \text{m}}$$
Check

Check: 4 down, then 2+2, 1+1, 0.5+0.5, ... adds 8 more, giving 12.

2§02.6 — a mixture summed by linearity●●●○○

Evaluate a two-piece series by summing each recognizable part and adding.

Given
  • $\displaystyle\sum_{n=1}^{\infty}\left(\frac{1}{2^{\,n}}+\frac{1}{n(n+1)}\right)$

Find
  1. (a) What is the total?

Hint 1/4

Split into a geometric part and a telescoping part; each is a standard sum.

Hint 2/4

$\sum\tfrac{1}{2^{n}}$ is geometric and $\sum\tfrac{1}{n(n+1)}$ telescopes; both converge, so linearity applies.

Hint 3/4

The geometric part is $\dfrac{1/2}{1-1/2}=1$ and the telescoping part is $1$.

Hint 4/4

Add: $1+1=2$.

Show solution
Geometric piece
$$\sum_{n=1}^{\infty}\frac{1}{2^{n}}=\frac{1/2}{1-1/2}=1$$

first term 1/2, ratio 1/2

Telescoping piece
$$\sum_{n=1}^{\infty}\frac{1}{n(n+1)}=1$$

collapses to 1 - 1/(n+1) -> 1

Add
$$1+1=2$$

both converge, so add the totals

Answer $$\boxed{2}$$
Check

Check: each piece is a familiar sum equal to 1, so the total is 2.

3§02.5 — what a zero limit of terms tells you●●●○○

A conceptual multiple choice of the kind that separates students who memorized the test from those who understand it.

Given
  • A series $\sum a_n$ is known only to satisfy $\lim_{n\to\infty}a_n=0$.

Find
  1. (a) Which conclusion is correct?

Hint 1/4

Recall exactly which direction the Test for Divergence runs.

Hint 2/4

$a_n\to 0$ is necessary for convergence but does not decide it either way.

Hint 3/4

So from $a_n\to 0$ alone you cannot tell whether it converges (like $\sum 1/2^n$) or diverges (like $\sum 1/n$).

Hint 4/4

The correct conclusion is that no conclusion follows yet; another test is needed.

Show solution
Exhibit both fates under the same hypothesis
$$\sum\tfrac{1}{2^{n}}\ \text{converges},\quad \sum\tfrac1n\ \text{diverges},\ \text{both with}\ a_n\to 0$$

one hypothesis, two outcomes, so it cannot decide

State the verdict
$$a_n\to 0\ \Rightarrow\ \text{test inconclusive}$$

the divergence test only fires on a nonzero or missing limit

Answer $$\boxed{\text{No conclusion yet}}$$
Check

Check: the two witnesses share the hypothesis but disagree on the outcome.

D · interleaved 3 questions
1§02.1 — a sequence versus its series●●●○○

This mixes the previous section with this one. The same formula defines a sequence and the terms of a series, and they need not share a fate.

Given
  • $a_n=\dfrac{n}{n+1}$

Find
  1. (a) Does the sequence a_n converge, and to what?

  2. (b) Does the series sum(a_n) converge?

Hint 1/4

Answer two different questions: one about the terms settling, one about the running totals settling.

Hint 2/4

For (a) take $\lim a_n$ (a sequence limit); for (b) apply the Test for Divergence to those same terms.

Hint 3/4

$\lim_{n\to\infty}\dfrac{n}{n+1}=1$, so the sequence converges to $1$; but then $a_n\to 1\neq 0$.

Hint 4/4

So (a) converges to 1, and (b) the series diverges by the Test for Divergence.

Show solution
Part (a): the sequence limit
$$\lim_{n\to\infty}\frac{n}{n+1}=1$$

the terms settle on 1, so the sequence converges

Part (b): the series test
$$a_n\to 1\neq 0\ \Rightarrow\ \sum a_n\ \text{diverges}$$

terms that do not go to 0 force the running total to run off

Answer $$\boxed{\text{seq}\to 1,\ \text{series diverges}}$$
Check

Check the distinction: a convergent sequence of terms can still give a divergent series whenever the limit is not 0.

2§02.1 — recover terms from a running-total formula●●●○○

Blends the sequence limit skill of the previous section with the partial-sum definition of this one.

Given
  • The partial sums are $s_n=3-\dfrac{2}{n+1}$.

Find
  1. (a) Find the sum of the series.

  2. (b) Find a_3.

Hint 1/4

The sum is a sequence limit; a single term is a difference of consecutive totals.

Hint 2/4

Use $\sum a_n=\lim s_n$ and $a_3=s_3-s_2$.

Hint 3/4

$\lim\left(3-\dfrac{2}{n+1}\right)=3$; and $s_3=3-\dfrac24=\dfrac{5}{2}$, $s_2=3-\dfrac23=\dfrac{7}{3}$.

Hint 4/4

So the sum is $3$ and $a_3=\dfrac{5}{2}-\dfrac{7}{3}=\dfrac{1}{6}$.

Show solution
Sum as a limit
$$\lim_{n\to\infty}\Big(3-\frac{2}{n+1}\Big)=3-0=3$$

the correction 2/(n+1) vanishes

Third term as a jump
$$a_3=s_3-s_2=\frac{5}{2}-\frac{7}{3}=\frac{15-14}{6}=\frac{1}{6}$$

subtract consecutive totals

Answer $$\boxed{\sum a_n=3,\ a_3=\tfrac16}$$
Check

Check: a_3 is small and positive because the totals near 3 are barely still climbing.

3§02.2 — geometric series with a parameter●●●●○

Connects the sequence limit of a power to the geometric-series sum, and previews power series to come.

Given
  • $\displaystyle\sum_{n=1}^{\infty} x^{\,n}$, where $x$ is a real number.

Find
  1. (a) For which x does the series converge?

  2. (b) For those x, what is the sum in terms of x?

Hint 1/4

Treat this as geometric in the variable $x$; the ratio is $x$ itself.

Hint 2/4

A geometric series converges exactly when $|r|<1$, so here you need $|x|<1$.

Hint 3/4

The first term (at $n=1$) is $x$ and the ratio is $x$, so the sum is $\dfrac{x}{1-x}$.

Hint 4/4

Thus it converges for $|x|<1$, with sum $\dfrac{x}{1-x}$.

Show solution
Part (a): the ratio is x
$$|r|=|x|<1\ \Leftrightarrow\ -1<x<1$$

geometric convergence needs ratio-size below 1; for |x|>=1 the terms do not tend to 0

Part (b): sum with first term x
$$\sum_{n=1}^{\infty}x^{n}=\frac{x}{1-x}\quad(|x|<1)$$

first term at n=1 is x, ratio x, so a/(1-r)=x/(1-x)

Answer $$\boxed{|x|<1,\ \ \tfrac{x}{1-x}}$$
Check

Check at x=1/2: the sum should be (1/2)/(1/2)=1, and indeed 1/2+1/4+1/8+...=1.

Mistake ledger (12 entries)
⚠ Trying to add the terms instead of taking the limit of the totals

the notation looks like an instruction to add forever, but the sum is defined as a limit of partial sums

wrong$$\sum_{n=1}^{\infty}\frac1{2^{n}}=\frac12+\frac14+\cdots=\text{keep going}$$
right$$\sum_{n=1}^{\infty}\frac1{2^{n}}=\lim_{n\to\infty}\Big(1-\frac1{2^{n}}\Big)=1$$
⚠ Confusing a term a_n with a partial sum s_n

both are subscripted sequences, so they blur together; a term is the jump between consecutive totals

wrong$$a_3\stackrel{?}{=}s_3$$
right$$a_3=s_3-s_2$$
⚠ Using a/(1-r) without checking |r| < 1

a sum of positive growing terms cannot be a finite negative number; the |r|<1 hypothesis was skipped

wrong$$\sum_{n=1}^{\infty}2\cdot3^{n-1}\stackrel{?}{=}\frac{2}{1-3}=-1$$
right$$|r|=3\ge 1\ \Rightarrow\ \text{diverges}$$
⚠ Reading the wrong first term a in a geometric series

the constant in front is not a when the exponent is n or the index starts at 0; a is the first term actually added

wrong$$\sum_{n=1}^{\infty}5\left(\tfrac12\right)^{n}\stackrel{?}{=}\frac{5}{1-\tfrac12}=10$$
right$$a=5\cdot\tfrac12=\tfrac52\ \Rightarrow\ \frac{5/2}{1-1/2}=5$$
⚠ Dropping the constant factor from a partial-fraction split

when the factors differ by more than 1, the split carries a coefficient that is easy to lose

wrong$$\frac{1}{(2n-1)(2n+1)}\stackrel{?}{=}\frac{1}{2n-1}-\frac{1}{2n+1}$$
right$$\frac{1}{(2n-1)(2n+1)}=\frac12\!\left(\frac{1}{2n-1}-\frac{1}{2n+1}\right)$$
⚠ Forgetting the surviving last term when telescoping

after the middle cancels it is tempting to keep only the first survivor and forget the last

wrong$$s_n\stackrel{?}{=}b_1$$
right$$s_n=b_1-b_{n+1},\quad \sum a_n=b_1-\lim_{n\to\infty}b_{n+1}$$
⚠ Believing a constant multiple can make a divergent series converge

a nonzero factor rescales an infinite total but cannot make it finite

wrong$$\sum\frac{1}{100\,n}\stackrel{?}{<}\infty$$
right$$\sum\frac{1}{100\,n}=\frac{1}{100}\sum\frac1n=\infty$$
⚠ Reading tiny terms as a guarantee of a finite sum

small terms stop the total from jumping but not from creeping upward forever

wrong$$a_n=\tfrac1n\to 0\ \stackrel{?}{\Rightarrow}\ \sum a_n<\infty$$
right$$a_n\to 0\ \text{is necessary, not sufficient: }\sum\tfrac1n=\infty$$
⚠ Concluding convergence from a_n -> 0

the divergence test is one-directional; a zero limit leaves the question open

wrong$$a_n\to 0\ \stackrel{?}{\Rightarrow}\ \sum a_n\ \text{converges}$$
right$$a_n\to 0\ \Rightarrow\ \text{Test for Divergence says nothing}$$
⚠ Claiming the Test for Divergence proves convergence

having one quick test, students hope it decides both directions, but it only rejects

wrong$$\lim a_n=0\ \stackrel{?}{\Rightarrow}\ \text{converges by the test}$$
right$$\text{The test can only ever prove divergence, never convergence}$$
⚠ Splitting a series into divergent pieces

linearity needs both pieces to converge; splitting divergent pieces produces meaningless differences

wrong$$\sum\Big(\tfrac1n-\tfrac1n\Big)\stackrel{?}{=}\sum\tfrac1n-\sum\tfrac1n=\infty-\infty$$
right$$\sum\Big(\tfrac1n-\tfrac1n\Big)=\sum 0=0$$
⚠ Hoping a convergent piece rescues a divergent one

a single divergent ingredient dominates; adding a tame series cannot calm the total

wrong$$\sum\Big(\tfrac{1}{2^{n}}+\tfrac1n\Big)\stackrel{?}{<}\infty$$
right$$\text{convergent}+\text{divergent}=\text{divergent}$$
Formula card
Sum of a series as a limit of partial sums
$$\sum_{n=1}^{\infty}a_n=\lim_{n\to\infty}s_n,\quad s_n=a_1+\cdots+a_n$$

the limit must exist and be finite for convergence

Geometric series
$$\sum_{n=1}^{\infty}ar^{\,n-1}=\frac{a}{1-r}\quad(|r|<1)$$

$|r|<1$; $a$ is the first term actually added; diverges if $|r|\ge 1$

Geometric partial sum
$$s_n=\frac{a(1-r^{\,n})}{1-r}\quad(r\neq 1)$$

any $r\neq 1$; take $n\to\infty$ and use $r^{n}\to 0$ when $|r|<1$

Telescoping series
$$a_n=b_n-b_{n+1}\ \Rightarrow\ \sum_{n=1}^{\infty}a_n=b_1-\lim_{n\to\infty}b_{n+1}$$

the term splits into a difference and $\lim b_{n+1}$ exists

Harmonic series diverges
$$\sum_{n=1}^{\infty}\frac1n=\infty$$

none; the standard example that $a_n\to 0$ does not force convergence

Test for Divergence
$$\lim_{n\to\infty}a_n\neq 0\ \Rightarrow\ \sum a_n\ \text{diverges}$$

only concludes divergence; if $\lim a_n=0$ it is inconclusive

Linearity of convergent series
$$\sum c\,a_n=c\sum a_n,\qquad \sum(a_n\pm b_n)=\sum a_n\pm\sum b_n$$

both $\sum a_n$ and $\sum b_n$ must converge

Check yourself

Close the page and write out, from memory: the definition of the sum of a series in terms of partial sums; the geometric formula with its condition and how you read off the first term; how a telescoping series collapses; why the harmonic series diverges; and exactly what the Test for Divergence can and cannot conclude. Then reopen and compare; every gap is a block to reread, and none of this is scored.

  • State that the sum of a series is the limit of its partial sums, and recover a term as s_n minus s_(n-1)?

    c-series-as-partial-sums

  • Sum a geometric series after checking |r| < 1, reading off the correct first term even when the index starts at 0?

    c-geometric-series

  • Split a term into a difference and collapse the telescoping partial sum, keeping any coefficient and every surviving boundary term?

    c-telescoping

  • Explain, by grouping, why the harmonic series diverges despite its terms tending to zero?

    c-harmonic

  • Apply the Test for Divergence to reject a series, and say why it is silent when the terms tend to zero?

    c-divergence-test

  • Combine convergent series by linearity, and recognise that a convergent plus a divergent series diverges?

    c-algebra-of-series

Glossary (10 terms)
seriesseri

The sum of the terms of an infinite sequence, defined as the limit of its partial sums.

partial sumkısmi toplam

The finite total of the first n terms of a series, written s_n.

convergeyakınsamak

To have partial sums that close in on a single finite number, which is then the sum.

divergeıraksamak

To fail to converge, either by running off to plus or minus infinity or by oscillating with no limit.

geometric seriesgeometrik seri

A series in which each term is a fixed multiple r of the previous one.

common ratioortak oran

The fixed number r that multiplies one term of a geometric series to give the next.

telescoping seriesteleskopik seri

A series whose term is a difference b_n minus b_(n+1), so the partial sums collapse to first minus last.

harmonic seriesharmonik seri

The series of reciprocals of the whole numbers, the sum of 1 over n, which diverges.

Test for Divergenceıraksaklık testi

The rule that a series diverges whenever its terms do not tend to zero.

toplam indisi

The counting letter n in a sum, a dummy symbol that does not appear in the final value.

What comes next
§03 · Convergence tests: the Integral, Comparison and Alternating Series Tests

We can now sum a geometric or telescoping series exactly and reject a series whose terms miss zero, but the Test for Divergence goes silent the moment the terms do tend to zero. The next section fills that silence with real tests, the Integral Test, the Comparison Tests and the Alternating Series Test, which decide the many series this one leaves open.

Sources
  • James Stewart, Calculus, 9th edition, section 11.2 (Series) The definition, the geometric-series theorem, the telescoping and harmonic examples and the Test for Divergence follow this text, the required book for the course.
  • MATH 102 course plan: weekly schedule and assessment weights Week 2 covers section 11.2; the Quiz and Homework weights of 10 percent each come from the grading scheme.
  • Grouping (Oresme) argument for the divergence of the harmonic series The block-doubling lower bound s_(2^k) > 1 + k/2 is the standard elementary proof and is reproduced in the harmonic block.

Spotted something missing or wrong? tell us · share your own notes or an old exam.