The ratio test is a method for determining whether an infinite series converges or diverges. It works by comparing consecutive terms: you take the limit of the absolute value of each term divided by the previous term, and the result tells you whether the series adds up to a finite number or grows without bound. Sometimes called the d’Alembert ratio test after the French mathematician who developed it, this is one of the most practical convergence tests in calculus, especially when your series involves factorials or exponential expressions.
How the Ratio Test Works
Given an infinite series with terms a_n, you compute the limit L as n approaches infinity of |a_(n+1) / a_n|. In plain terms, you’re looking at the ratio between each term and the one before it, then asking what happens to that ratio as you move further and further out in the series. The result L gives you three possible outcomes:
- L < 1: The series converges absolutely.
- L > 1 (or L = infinity): The series diverges.
- L = 1 (or the limit doesn’t exist): The test is inconclusive, and you need a different method.
The logic behind these cutoffs comes from a comparison to geometric series. A geometric series with common ratio r converges when |r| < 1 and diverges when |r| ≥ 1. The ratio test essentially asks whether your series behaves like a geometric series in the long run. If consecutive terms shrink by a factor consistently less than 1, the series must converge. If they grow or stay the same relative to each other, it diverges or the test can’t tell.
When L Equals 1: The Inconclusive Case
The trickiest part of the ratio test is understanding why L = 1 tells you nothing. When the limit lands exactly on 1, the series could converge absolutely, converge conditionally, or diverge. All three outcomes are possible, so you can’t draw any conclusion.
This happens consistently with certain types of series. The ratio test always returns L = 1 for p-series (series of the form 1/n^p) and for any series built from rational functions of n. For example, applying the ratio test to the harmonic series (1/n) gives a limit of 1, even though the harmonic series diverges. Apply it to 1/n^2 and you also get 1, even though that series converges. The test simply cannot distinguish between the two. If you recognize your series as a p-series or a ratio of polynomials in n, skip the ratio test entirely and use the p-series test or a comparison test instead.
Where the Ratio Test Excels
The ratio test is the go-to choice when your series contains factorials or terms raised to the nth power. Factorials are what make it shine, because dividing (n+1)! by n! simplifies beautifully: everything cancels except a single factor of (n+1). This clean cancellation turns what looks like a complicated limit into something straightforward.
Consider a series whose terms include n! in the denominator and 5^n in the numerator. When you form the ratio of consecutive terms, the factorials collapse and the exponential terms reduce to a single factor of 5. The limit becomes (n+1)/5, which grows to infinity. Since infinity is greater than 1, the series diverges. Without the ratio test, reaching that conclusion would require considerably more work.
Series involving expressions like 2^n, n^n, or products of factorials and exponentials are all natural candidates. If you see a factorial and no nth-power terms, reach for the ratio test first. If you see nth-power terms but no factorials, the root test is typically a better fit.
Finding the Radius of Convergence for Power Series
One of the most common applications of the ratio test is determining where a power series converges. A power series has the form c_n(x – a)^n, where c_n are the coefficients, a is the center, and x is the variable. The ratio test lets you find the radius of convergence, R, which tells you how far from the center the series produces finite values.
The process has a few steps. You form the ratio of consecutive terms, which simplifies to (c_(n+1) / c_n) times (x – a). Then you take the absolute value and compute the limit as n goes to infinity. Call the resulting coefficient N (the part that doesn’t depend on x). Setting N|x – a| < 1 gives you the radius of convergence R = 1/N. If the limit comes out to zero regardless of x, the series converges everywhere and R is infinite. If the limit is infinite for every x except the center, the series converges only at x = a and R is zero.
There’s one important detail: the ratio test determines the open interval of convergence but can’t settle what happens at the endpoints. Plugging each endpoint into the series produces a specific numerical series, and you need to test those separately with another method. Using the ratio test at the endpoints almost always gives L = 1, which is inconclusive. The integral test, comparison test, or alternating series test are better choices for checking endpoint behavior.
A Step-by-Step Example
Suppose you want to test the series whose nth term is n! / 5^n. You form the ratio of consecutive terms:
|a_(n+1) / a_n| = |(n+1)! / 5^(n+1)| divided by |n! / 5^n|
The factorials simplify: (n+1)! / n! = (n+1). The exponentials simplify: 5^n / 5^(n+1) = 1/5. So the ratio becomes (n+1)/5. As n approaches infinity, this limit is infinity, which is greater than 1. The series diverges.
Now consider n^2 / 3^n. The ratio of consecutive terms gives (n+1)^2 / (3 · n^2). As n grows large, (n+1)^2 / n^2 approaches 1, so the whole limit is 1/3. Since 1/3 < 1, the series converges absolutely. The exponential term 3^n in the denominator dominates the polynomial n^2 in the numerator, and the ratio test captures that cleanly.
Relationship to Absolute Convergence
When the ratio test tells you a series converges, it specifically tells you the series converges absolutely. Absolute convergence means that even if you replaced every term with its absolute value, the series would still converge. This is a stronger statement than plain convergence: every absolutely convergent series is convergent, but not every convergent series is absolutely convergent.
This distinction matters for series with mixed positive and negative terms. If you have an alternating series and the ratio test gives L < 1, you know the series converges absolutely, not just conditionally. If the ratio test gives L = 1, though, you’re back to needing another approach. The alternating series test, for instance, can establish conditional convergence in cases where the ratio test stays silent.
Choosing the Right Convergence Test
The ratio test is powerful but not universal. Knowing when to use it and when to pick something else saves time and frustration. Series with factorials, products of factorials, or expressions like n^n are ideal candidates. Series that are purely polynomial fractions (rational functions of n) will always give L = 1, so use the limit comparison test or p-series test instead.
When your series has terms raised to the nth power but no factorials, the root test often works more efficiently. It extracts the nth root instead of forming a ratio, which simplifies expressions like (2/3)^n more directly. The ratio and root tests agree whenever both give a definitive answer, so it comes down to which one simplifies the algebra faster for the particular series you’re working with.