Convergence tests sit at the centre of the AP Calculus BC series-and-sequence unit, and comparison tests are the workhorse most students reach for when the standard toolkit runs dry. The comparison tests for convergence on the AP Calculus exam are the Direct Comparison Test and the Limit Comparison Test, and the difference between them is rarely a matter of taste. Pick the wrong one and you will spend three extra minutes on the free response, lose the thread of the justification, and ship an argument that the reader's rubric is built to dock. This article unpacks both tests, the geometric intuition behind them, the exact mechanical steps the rubric expects, and the prep strategy that turns a fragile 'I think it converges' into a defensible four-line justification.
What the two comparison tests actually say
The Direct Comparison Test (DCT) is a qualitative inequality. If you can show that every term of your unknown series is bounded above by a known convergent p-series (or below by a known divergent series) and that the terms are positive, the unknown series must inherit the convergence behaviour. There is no computation, only an inequality chain. The Limit Comparison Test (LCT) is a quantitative version. You build a companion series whose terms you already understand, take the limit of the ratio of the two general terms, and read the convergence answer from the limit's value.
Two facts decide which test you should reach for first. First, DCT demands a genuine inequality that holds for every term beyond some index N. LCT only needs the limit to exist and be a positive finite number. Second, LCT is essentially a sanity check on the inequality you are about to write: if the limit is 0 or infinity, DCT can still work, but you have to invert the chain. In practice, on the AP Calculus exam, the LCT is the safer default because the limit machinery is uniform and the rubric's language matches it almost word for word.
It helps to remember that both tests are conditional. They only work on series with non-negative terms. The moment negative terms enter, you must either take absolute values first and test for absolute convergence, or use the Alternating Series Test separately. The 2014 BC exam famously featured a series that mixed signs and asked whether it converged conditionally; candidates who tried to apply DCT without an absolute-value step were answering a different question.
The rubric language you should mirror
- State the test you are using by name.
- Define the companion series and show the limit calculation with a single line of algebra.
- Conclude with the p-series or geometric series whose behaviour is being inherited.
- Write one sentence in plain English that ties the limit value to the conclusion.
Direct comparison test: when the inequality is obvious
Some series come dressed in a comparison you can read off in a single glance. Consider the series sum from n equals 1 to infinity of 1 over (n squared plus 1). The numerator is 1, the denominator is n squared plus 1, and n squared plus 1 is bigger than n squared for every n greater than or equal to 1. That gives 1 over (n squared plus 1) less than 1 over n squared for every n. Since sum 1 over n squared is a convergent p-series with p equal to 2, the smaller series converges by the Direct Comparison Test. The full justification takes four lines and never uses a limit.
Now consider a divergent-looking sibling: sum 1 over (n minus 1 over 2). Each term is bigger than 1 over n for every n at least 2, because the denominator is smaller. The harmonic series diverges, so the larger series diverges by DCT. Notice the direction: for a divergent comparison, you want the unknown series to be larger than the divergent reference, not smaller. Students who invert this direction are usually the ones who lose the point for the conclusion.
Direct comparison has one structural weakness that the AP exam knows how to exploit. The inequality must hold for every term in the tail. If the comparison only works for the first few terms, you must say so and explicitly use the tail definition of convergence. The test is about the behaviour at infinity, and the rubric wants to see that you know the first finitely many terms do not matter. A safe form is: 'For all n greater than or equal to 2, a sub n is less than b sub n, and sum b sub n converges, so by the Direct Comparison Test the original series converges.'
Common pitfalls and how to avoid them
- Comparing to a series that does not actually converge or diverge in the direction you need.
- Forgetting the positivity condition, then writing the test on a series that has negative terms.
- Producing an inequality that flips at some specific n, such as n equals 1, and not handling the tail explicitly.
Limit comparison test: the default on a free response
The Limit Comparison Test is the engine most AP Calculus students should reach for first. The procedure has four mechanical steps, and once the steps are in muscle memory the test takes roughly 90 seconds of working time. Step one: identify the dominant term in the general term. Step two: write the companion series by stripping the dominant term. Step three: compute the limit of the ratio of the two general terms. Step four: read the result.
Take the series sum from n equals 1 to infinity of 1 over (n plus 1 over n). The dominant behaviour is 1 over n, so the companion is the harmonic series sum 1 over n. The ratio a sub n over b sub n equals n over (n plus 1 over n) equals n squared over n squared plus 1. The limit as n approaches infinity of n squared over n squared plus 1 is 1. Because the limit is a positive finite number, both series share the same convergence behaviour. The harmonic series diverges, so the original series diverges by the Limit Comparison Test.
Now consider sum from n equals 1 to infinity of 3 to the n over (4 to the n minus 1). The dominant base is 3 to the n over 4 to the n, which behaves like (3 over 4) to the n, a geometric series with ratio less than 1. The companion is sum (3 over 4) to the n. The ratio of the original general term to the companion is 4 to the n minus 1 in the denominator of the original multiplied by 4 to the n, which simplifies to 4 to the n over 4 to the n minus 1, tending to 1. The companion converges, so the original converges. This is the standard BC free-response shape, and the steps map onto the rubric's 1, 2, 3, 4 point structure almost line for line.
There is one numerical fact worth memorising. If the ratio limit is 0, the comparison series must converge. If the limit is infinity, the comparison series must diverge. If the limit is some positive constant, the two share behaviour. The BC exam occasionally constructs a problem where the limit is 0 to test whether the student remembers that the chain of reasoning is one-directional.
Choosing between direct and limit comparison in exam conditions
When a free-response problem hands you a series with a clean algebraic manipulation, the Direct Comparison Test will be the shorter argument. When the inequality requires careful bounding or a sign-controlled substitution, the Limit Comparison Test will save you the trouble. The way to choose quickly is to ask one question: can I write the unknown general term as a known convergent or divergent term plus a remainder that is provably smaller? If yes, write the inequality and finish with DCT. If no, set up the LCT and let the limit argument carry the conclusion.
Consider a contest-style problem: sum from n equals 2 to infinity of 1 over (n times the natural log of n). The direct inequality is not obvious because the function grows slowly. The LCT companion is sum 1 over n. The ratio is 1 over the natural log of n, which goes to 0, not a positive constant. This signals that DCT with the harmonic series will fail, because 1 over (n ln n) is smaller than 1 over n and the harmonic series diverges, so a smaller divergent series tells us nothing. To save DCT you would need a smaller convergent series such as 1 over n to the 1.1, but that requires knowing the exact p-test threshold. The cleaner move is to switch to the integral test or to use LCT against 1 over n to the 1.01 and conclude divergence that way.
Speed matters because the AP Calculus exam is paced. BC students have roughly 90 minutes for six free-response problems, which works out to about 15 minutes per problem, and the series question usually comes last. If you spend eight minutes on the comparison setup, you are eating into the next problem. The LCT costs about three minutes in practice; DCT can cost less, but only when the inequality is genuinely obvious. Train yourself to default to LCT and to switch down to DCT only when the bound jumps out at you.
| Situation | Best test | Reason |
|---|---|---|
| General term is a small perturbation of a known series, e.g. 1 over (n squared plus 1) | Direct Comparison | Inequality is one line and the bound is provable for all n |
| General term has competing growth, e.g. 3 to the n over 4 to the n plus 1 | Limit Comparison | Limit of ratio reduces to a positive constant cleanly |
| General term is dominated by a logarithm, e.g. 1 over n ln n | Limit Comparison with a p-series near the boundary | DCT against harmonic gives no information |
| Series alternates in sign | Take absolute values first, then DCT or LCT | Comparison tests require non-negative terms |
Worked example: a complete BC-style free response
Suppose the problem reads: 'Determine whether the series sum from n equals 1 to infinity of n plus 2 over 2 to the n converges or diverges. Justify your answer.' The first move is to recognise the denominator as exponential growth, which means the series almost certainly converges and the LCT against a geometric companion is the natural choice. The companion is sum 1 over 2 to the n, a geometric series with ratio 1 over 2 and first term 1 over 2. The ratio a sub n over b sub n equals (n plus 2) over 2 to the n times 2 to the n, which is just n plus 2. The limit of n plus 2 as n approaches infinity is infinity, not a positive finite number, so LCT in its standard form does not give a clean answer.