Intuition The one core idea
A never-ending sum of shrinking positive numbers is really just a staircase of rectangles , and the smooth curve those rectangles hug is something we already know how to measure: area . If we can measure the area, we can decide whether the staircase reaches the sky or stops at a finite height.
Before you can read the Integral Test note , you must own every symbol it throws at you. This page builds them in order , each from the picture it stands for. Nothing here assumes you have seen calculus notation before line one.
Definition How to read the
::: self-test lines
Throughout this page you will see short lines shaped like Question ::: Answer . These are flip-card prompts : read the part before the three colons, answer it in your head, then check yourself against the part after the three colons. The three colons are just the divider between question and hidden answer.
Definition Sequence and the term
a n
A sequence is an ordered list of numbers, one for each counting number 1 , 2 , 3 , … . We write the number in slot n as a n , read "a-sub-n ". The little n underneath is just the slot number .
The picture: imagine numbered mailboxes in a row. Box 1 holds the number a 1 , box 2 holds a 2 , and so on forever.
Intuition Why the topic needs this
The whole subject is about adding these box-contents together. So we first need a clean name — a n — for "the number in box n " before we can talk about summing them.
a n exampleFor the list 1 , 2 1 , 3 1 , … the rule is a n = n 1 , so a 5 = 5 1 .
∑
The tall symbol ∑ (Greek capital "sigma", the S-sound) is just a shorthand for "add these up" . Underneath it we write where the slot number starts, above it where it stops.
∑ n = 1 4 a n = a 1 + a 2 + a 3 + a 4 .
WHAT this says: start at n = 1 , plug each slot number into a n , keep adding until n = 4 .
Definition An infinite series
When the top of the sigma is the sideways-eight ∞ (infinity — "keeps going, no last number"), we mean we never stop adding:
∑ n = 1 ∞ a n = a 1 + a 2 + a 3 + ⋯
This endless sum is called a series .
∞
We cannot literally add infinitely many numbers by hand. The entire point of the Integral Test is to decide the fate of such an endless sum without doing the impossible addition. So the symbol ∞ is the problem we are trying to tame, not a number we compute with.
∑ n = 1 3 n 1 equals1 + 2 1 + 3 1 = 6 11 .
S N
Since we can't add forever, we stop at slot N and look at that finite total:
S N = ∑ n = 1 N a n = a 1 + a 2 + ⋯ + a N .
This running total is the ==N -th partial sum==. Capital N is our chosen stopping point.
The picture: watch the total grow as you add one more box each time. S 1 , S 2 , S 3 , … is itself a new sequence — the sequence of running totals.
Definition Convergence of a sequence, written with
lim
We say a sequence of numbers S 1 , S 2 , S 3 , … ==converges to the limit L ==, written
lim N → ∞ S N = L ,
if the values S N get and stay as close as we like to the single finite number L once N is large enough. In plain words: pick any tiny tolerance; past some slot, every S N sits within that tolerance of L . If no such finite L exists (the values grow without bound or never settle), the sequence diverges .
Intuition Why partial sums are the real definition
"Does the infinite sum settle?" is literally the question "does lim N → ∞ S N exist and equal a finite L ?" Every convergence test — including the integral test — is really a statement about how S N behaves. That is why the parent note keeps writing S N in its inequalities.
n -th term precondition
Before any convergence test, glance at the terms themselves. If a series ∑ a n converges, then its terms must shrink to nothing :
lim n → ∞ a n = 0.
Turned around: if lim n → ∞ a n = 0 (or the limit doesn't exist), the series diverges immediately — no further work needed. This is the n-th term test for divergence .
Intuition Necessary, not sufficient
Terms going to zero is a doorway you must pass through, not a guarantee of arrival. Passing it (a n → 0 ) only means you are allowed to ask the deeper question the integral test answers. Failing it ends the story on the spot. The harmonic series has a n = n 1 → 0 yet still diverges — proof that the doorway alone decides nothing.
Why is a n → 0 only a precondition, not a verdict? It is necessary but not sufficient; ∑ n 1 has a n → 0 yet diverges, so a real test is still required.
Definition Increasing and bounded above
A sequence is increasing if each term is ≥ the one before: S 1 ≤ S 2 ≤ S 3 ≤ ⋯ .
It is bounded above if there is a ceiling number M that no term ever passes: S N ≤ M for all N .
The least upper bound (supremum) of the sequence is the smallest such ceiling — the lowest number L that still satisfies S N ≤ L for all N .
If the a n are all positive , then adding a positive amount each step means the partial sums only ever go up — they are automatically increasing.
Intuition The monotone ceiling picture
Picture the running totals climbing a ladder, never stepping down. A ceiling M blocks them from above, so they cannot climb forever — they must pile up just beneath the lowest height nothing ever exceeds (the least upper bound). That "must settle" fact is exactly what the integral test uses in its final step: the integral supplies the ceiling M .
Why "positive terms" matters With positive terms the totals only rise, so a ceiling forces convergence; with mixed signs totals can wobble and this reasoning breaks.
Definition Function and its curve
A function f is a machine: feed in a number x , out comes one number f ( x ) . Its graph is the smooth curve you get by plotting the point ( x , f ( x )) for every input — the height of the curve above position x is the output f ( x ) .
[ 1 , ∞ )
The test always works with f defined on the domain [ 1 , ∞ ) — every real input x from 1 onward, going right forever. The square bracket "[ " means 1 is included ; the "∞ ) " means there is no right-hand endpoint. (In some problems we may start at another whole number N instead, giving [ N , ∞ ) , but the default is [ 1 , ∞ ) .)
The bridge to sequences: if we set a n = f ( n ) , then our discrete boxes are just the curve's heights read off at the whole-number inputs x = 1 , 2 , 3 , … inside that domain. The rectangles of the series sit under (or over) the smooth ramp of the function.
Intuition Why swap a sum for a function
A jagged sum is hard; a smooth curve has a tool we know — area . By turning a n into f ( n ) we unlock that tool. This is the single move that makes the integral test possible.
Definition Positive, continuous, decreasing (on
[ 1 , ∞ ) )
These are the three quality checks on f the test demands, all across the domain [ 1 , ∞ ) :
Positive : the curve stays above the x -axis, f ( x ) > 0 — so all rectangle areas are genuine additions, none cancel.
Continuous : the curve has no breaks or jumps — so the area under it actually exists.
Decreasing : the curve falls as you move right, so a later input never gives a taller output. This means on each width-1 strip the left edge is the tallest point and the right edge is the shortest. Because every strip leans the same way, the rectangles can be compared with the area in one consistent direction — left-edge rectangles always sit above the curve, right-edge rectangles always below.
Definition The integral sign
∫
∫ a b f ( x ) d x is the exact area trapped between the curve y = f ( x ) and the x -axis, from x = a on the left to x = b on the right. The stretched-S ∫ is another "sum" symbol — a sum of infinitely many razor-thin strips. The d x means "of thin strips along the x -direction".
Intuition Why area answers our question
The rectangles of the series and the area under the curve are so close they trap each other. If the smooth area is finite, the boxed rectangles can't total more than that finite ceiling — the series converges. If the area is infinite, the rectangles are dragged up with it. Area decides the fate of the sum. This is the heart of the whole method.
Definition Improper integral and the limit
lim
To measure area out to infinity we can't just plug in ∞ . Instead we compute the area out to a movable wall b , then slide b to infinity and watch:
∫ 1 ∞ f ( x ) d x = lim b → ∞ ∫ 1 b f ( x ) d x .
The symbol lim b → ∞ reads "the value this approaches as b grows without end". If it lands on a finite number the integral converges ; if it grows forever it diverges . See Improper integrals .
Why we need a limit here We can never "reach" infinity, so we track the area to a wall b and see what number it approaches as the wall retreats forever.
p and negative powers
In n p 1 , the number p is the power the slot number is raised to. A negative exponent just means "one over": x − p = x p 1 . Bigger p makes the terms shrink faster , which is exactly the knob that decides convergence.
Definition The natural logarithm
ln
ln x answers the question "to what power must the special number e ≈ 2.718 be raised to get x ? " It appears because the area under the curve y = 1/ x turns out to equal ln x . It grows, but agonisingly slowly — slow growth that still reaches infinity is the punchline of the harmonic series.
Meaning of x − 2 x 2 1 — one over x squared.
Why ln shows up The area under y = 1/ x from 1 to b is exactly ln b , and ln b → ∞ , which is why ∑ n 1 diverges.
The diagram below (rendered where Mermaid is available) shows the build order; if it appears as plain text, read it top-to-bottom as the same dependency chain described in §1–§8.
Monotone convergence theorem
Function f x and its graph
Positive continuous decreasing
Improper integral with limit
Test yourself — reveal only after answering aloud.
What does a n mean in plain words? The number sitting in slot n of an ordered list.
What does ∑ n = 1 ∞ a n mean? Add every list-term forever, starting at slot 1.
What is the partial sum S N ? The finite running total a 1 + ⋯ + a N , stopping at slot N .
Write convergence of the series with limit notation. ∑ a n converges when lim N → ∞ S N = L for some finite L , and that L is the sum.
State the n -th term precondition every test assumes. If ∑ a n converges then lim n → ∞ a n = 0 ; if a n → 0 the series diverges outright.
State the Monotone Convergence Theorem. An increasing sequence bounded above converges, to its least upper bound (supremum).
Why does "increasing and bounded above" force convergence? Rising totals blocked by a ceiling must pile up under their least upper bound.
On what domain must f satisfy the three conditions? [ 1 , ∞ ) — every real input from 1 rightward, forever.
Name the three conditions on f and one word for why each is needed. Positive (no cancelling), continuous (area exists), decreasing (clean rectangle comparison).
What does ∫ 1 b f ( x ) d x measure? The exact area between the curve and the x -axis from 1 to b .
What does lim b → ∞ do to an integral? Slides the right wall to infinity and reports the number the area approaches.
What does x − p equal? 1/ x p .
What question does ln x answer? To what power must e ≈ 2.718 be raised to give x .
Integral test — proof, p-series (index 4.3.7) — the topic these foundations unlock.
Improper integrals — the limit-to-infinity area machine of §7.
Harmonic series — where the slow ln growth of §8 bites.
n-th term test for divergence — the necessary (not sufficient) precondition of §4.
Comparison test · Limit comparison test — sibling tests built on the same a n = f ( n ) picture.
Riemann zeta function — names the value of ∑ 1/ n p once it converges.