Series convergence tests
A practical guide to choosing a convergence test, checking its hypotheses, and stating a valid conclusion.
Which convergence test should I try?
First check whether the terms approach zero. Next recognize geometric, p-series, or telescoping forms. For positive rational or algebraic terms, try comparison or the integral test. For alternating signs, check absolute convergence and the Alternating Series Test. Factorials and nth powers usually suggest the ratio or root test.
Always check the terms first
The nth-term test can prove divergence immediately, but it can never prove convergence.
Inconclusive case
A zero term limit only means another test is needed.
Benchmark series to recognize
Recognize geometric and p-series before trying a more complicated test.
Geometric series
p-series
Telescoping series
Write partial sums and check whether the uncanceled term has a finite limit.
Comparison tests
Use these for positive-term series that resemble a known benchmark such as a p-series.
Direct comparison: convergence
Direct comparison: divergence
Limit comparison
Assume aₙ and bₙ are positive eventually.
Integral test
Use when the terms come from a function that is positive, continuous, and decreasing eventually.
Integral test
When aₙ = f(n) and the hypotheses above are satisfied.
Integral-test remainder bound
Here Rₙ is the error after the Nth partial sum.
Alternating series
Check both that the magnitudes decrease eventually and that they approach zero.
Alternating Series Test
Assume bₙ > 0 eventually.
Alternating-series error
The truncation error is no larger than the first omitted magnitude.
Ratio and root tests
These are especially effective with factorials, exponentials, and quantities raised to the nth power.
Ratio test setup
Root test setup
Decision for either test
Absolute and conditional convergence
Testing absolute values is often the cleanest way to classify a series with mixed signs.
Absolute convergence
Conditional convergence
Power-series intervals
Use the ratio or root test to find the radius, then test both endpoints separately.
Power-series form
Radius pattern
Endpoint warning
A test name alone is not a justification.
State the limit or comparison you computed, verify the relevant hypotheses, and explain exactly what the test proves. Remember that several tests have inconclusive cases requiring a different method.