Visual walkthrough — Direct comparison test

Step 0 — What a "series" even is (drawing the pile)

Before any test, we must picture the thing we're testing.

WHAT. A sequence is just an endless list of numbers, one for each counting position $n=1,2,3,\dots$. We write the $n$-th number as $a_n$ (read "a-sub-n"). So $a_1$ is the first number, $a_2$ the second, and so on.

A series means we add them all up, forever:

WHY draw bars. We cannot literally add infinitely many numbers by hand. So we picture each $a_n$ as the area of a thin bar of height $a_n$ standing at position $n$. Adding the series = collecting total area under all the bars.

PICTURE. In the figure, each burnt-orange bar has height $a_n$. Notice the bars shrink — a hint (not a guarantee!) that the total area might stay finite.

Step 1 — The partial sum: adding up to a stopping point

We can't finish an infinite sum, so we watch it grow one bar at a time.

WHAT. The partial sum $S_n$ is the running total after $n$ bars:

Here $S_n$ is a single number: the accumulated area of the first $n$ bars. As $n$ climbs, $S_n$ climbs — we get a new sequence of totals $S_1,S_2,S_3,\dots$.

WHY. "Does the series add up to something finite?" becomes a cleaner question: does the sequence $S_n$ settle toward a fixed number as $n\to \infty$? ($n\to \infty$ means "as $n$ marches off forever.") We have swapped an impossible infinite addition for a limit of ordinary numbers.

PICTURE. The plum staircase: each step up from $S_n$ to $S_{n+1}$ is exactly one new bar of height $a_{n+1}$.

Step 2 — The staircase only goes up (monotone)

WHAT. Because every bar height $a_n\ge 0$, the gap $S_n-S_{n-1}=a_n\ge 0$. So:

This chain reads "each total is no smaller than the one before it." A sequence that never decreases is called monotone increasing.

WHY this is the payoff of non-negativity. If a bar could be negative, the staircase could step down, wobble, and never settle — the whole argument would collapse. This is exactly why DCT forbids negative terms. (For sign-changing sums you'd switch to Absolute convergence or the Alternating Series Test.)

PICTURE. Left panel: all bars $\ge 0$, staircase rises steadily. Right panel: a sneaky negative bar makes the staircase dip — the counter-example we are ruling out.

Step 3 — Domination: one series sits under another, term by term

Now bring in the comparison partner — this is where $b_n$ enters.

WHAT. Introduce a second non-negative sequence, written $b_n$ (the benchmark or comparison sequence), that dominates $a_n$:

"Dominates" just means: at every position, the $b$-bar is at least as tall as the $a$-bar. This is the first time $b_n$ appears — it is a series we already understand, chosen on purpose to sandwich the one we don't.

WHY. We usually don't understand our own series, but we do understand a benchmark like a p-series or a Geometric series. By pinning our bars underneath (or above) a benchmark, the benchmark's known behaviour leaks onto us.

PICTURE. At each $n$, a short orange bar $a_n$ nested inside a taller teal bar $b_n$. The orange area is trapped inside the teal area, column by column.

Step 4 — The ceiling: if the big series converges, a finite roof appears

WHAT. Suppose the taller series converges — its totals $T_n$ climb toward a finite number we'll name:

Chaining with Step 3:

WHY $T_n\le B$. The middle inequality needs a word. Since every $b_k\ge 0$, the teal staircase $T_n$ is itself increasing (same argument as Step 2, now applied to the $b$'s). An increasing sequence never overshoots the value it climbs toward, so $T_n\le {\lim }_{m\to \infty }{T}_m=B$ for every $n$. That closes the loop: the teal staircase stays under its own limit $B$, and the orange one stays under the teal one — so orange stays under $B$.

This installs a fixed horizontal ceiling at height $B$ above our orange staircase. Our staircase is bounded above — the phrase means "never rises past some fixed line."

PICTURE. Orange staircase rising, teal staircase rising above it, and a dashed plum line at height $B$ capping both. The orange steps can never poke through the plum line.

Step 5 — The engine: Monotone Convergence closes the case

WHAT. We now have exactly the two hypotheses of the theorem for the orange staircase $S_n$:

1. it is increasing (Step 2), and
2. it is bounded above by $B$ (Step 4).

WHY it's forced. An always-rising staircase that can never cross a ceiling has nowhere to run off to. It gets squeezed into a corner and must settle onto some limiting height $L\le B$. Hence $S_n\to L$, i.e. $\sum a_n$ converges. ∎

PICTURE. The orange steps crowd closer and closer to a limit line $L$ just under the ceiling $B$ — the gaps between steps vanish.

Step 6 — The mirror case: divergence flows UP (contrapositive)

WHAT. Now suppose instead the small series $\sum a_n$ diverges — its staircase $S_n\to \infty$ (grows without bound). Since Step 3 gave $T_n\ge S_n$, the taller totals are pushed even higher:

So $\sum b_n$ diverges too. ∎

WHY this is the only other useful direction. Knowing the small one blows up forces the big one (sitting above it) to blow up. This is logically the same statement as Step 5, read in reverse (its contrapositive).

PICTURE. An infinite floor: the orange staircase rockets upward past every horizontal line; the teal staircase, always above it, is dragged up too.

Step 7 — Degenerate & edge cases (the ones that trip people)

Case A — the inequality only holds "eventually," and shifting the index. DCT needs $a_n\le b_n$ only for $n\ge N$ (some starting index). Split the sum at $N$:

Convergence is decided entirely by the tail: define a shifted partial sum $S_n^{\prime }={\sum }_{k=N}^n{a}_k$ that starts counting from $N$. Adding a constant (the finite front chunk) to every term of a sequence never changes whether it settles to a finite limit — it just slides the whole staircase up by a fixed amount. So the front chunk is harmless, and on the tail the inequality $a_k\le b_k$ holds throughout. This is why "eventually" is enough.

Case B — a zero bar. If some $a_n=0$, the staircase simply doesn't rise at that step. Perfectly allowed: $0\le b_n$ still holds. Zeros are harmless.

Case C — the messy numerator trick. When a term wobbles, bound it by its maximum to get a clean partner. Example from the parent: since $-1\le \sin n\le 1$,

where $3$ is the largest the numerator $2+\sin n$ can be. Now recall the p-series rule: $\sum \frac{1}{n^p}$ converges exactly when $p>1$ and diverges when $p\le 1$. Here $p=2>1$, so $\sum \frac{1}{n^2}$ converges, hence $\sum \frac{3}{n^2}=3\sum \frac{1}{n^2}$ converges too, and our wobbly series converges.

Case D — negative terms present. Then Step 2 fails outright: the staircase can dip and Monotone Convergence has no grip. DCT is simply not applicable — reach for Absolute convergence or the Limit comparison test instead.

PICTURE. Four mini-panels: (A) ignored early bars shaded grey, (B) a flat zero-height step, (C) the wobbly $\sin$ bars capped by a flat $3/n^2$ lid, (D) a red-crossed dipping staircase.

The one-picture summary

Everything above compressed into a single frame: two staircases, one nested under the other, plus the two arrows.

• Orange = our unknown series $\sum a_n$. Teal = trusted benchmark $\sum b_n$.
• Bars satisfy $0\le a_n\le b_n$, so orange staircase $\le$ teal staircase at every step.
• Down-arrow (convergence): teal hits a finite ceiling $B$ ⇒ orange, trapped beneath, must also settle. ✅
• Up-arrow (divergence): orange rockets to $\infty$ ⇒ teal, forced above it, must also rocket. ✅
• Engine throughout: increasing + bounded ⇒ converges, which is why every bar must be $\ge 0$.

Recall checks