What Is a p-Series?
A p-series is the family Σ1/nᵖ = 1 + 1/2ᵖ + 1/3ᵖ + ⋯, indexed by a single exponent, and its convergence question has a complete, crisp answer:
Σ 1/nᵖ converges exactly when p > 1.
The family is the measuring stick of series theory. Because its verdict is known for every p, it anchors comparison arguments: a series whose terms are eventually smaller than a convergent p-series converges; larger than a divergent one, diverges. Most convergence questions in a first course reduce, sooner or later, to a comparison with some 1/nᵖ, which is why internalizing the p > 1 boundary matters more than memorizing almost any other single fact in the chapter.
How to Use This Calculator
Enter p. The calculator states the test, delivers the verdict, and computes partial sums (S₁₀, S₃₀, S₆₀) with the running totals plotted, settling visibly for p = 2, drifting upward without a ceiling for p = 1, and crawling ambiguously for the boundary-hugging p = 1.01, a deliberately instructive case where sixty terms cannot reveal the (true, but distant) convergence. The test's certainty where numerics are ambiguous is exactly the point.
Why p = 1 Is the Boundary
The integral test makes the threshold transparent: Σ1/nᵖ behaves like ∫₁^∞ dx/xᵖ, and that improper integral converges precisely for p > 1 (the antiderivative x^(1−p) dies at infinity) and diverges for p ≤ 1 (at p = 1 it's ln x, growing without bound). The series inherits the integral's fate, the same p-test that governs improper integrals, ported to sums. See the Improper Integral Calculator for the continuous twin of this page.
At the boundary itself sits the harmonic series Σ1/n: terms → 0, partial sums → ∞, growing like ln N. It needs 12,367 terms to pass 10 and over 10⁴³ terms to pass 100, divergence at a geological pace, and the permanent reminder that shrinking terms do not settle the question.
Worked Example
Test Σ1/n². Since p = 2 > 1, it converges, and this particular member has a celebrity value, Euler's π²/6 ≈ 1.6449 (the Basel problem). The calculator's S₆₀ ≈ 1.6284 sits below the limit with tail ≈ 1/60, matching the integral-test tail estimate. Contrast p = ½: terms shrink, but only like 1/√n, and the partial sums scale as 2√N, past every bound, as the plot's steady climb shows.
Common Mistakes to Avoid
- Placing the boundary at p = 0 or including p = 1. The harmonic series diverges: p must strictly exceed 1. "p ≥ 1" is the single most common mis-remembering.
- Judging by term size instead of the test. 1/n and 1/n² look similar termwise; their series live on opposite sides of the divide. The exponent, not the visual impression, decides.
- Expecting the test to give the sum. It certifies convergence only. Values like π²/6 (p = 2) and π⁴/90 (p = 4) come from deeper machinery; odd p sums (like p = 3, Apéry's constant) remain without closed forms.
- Trusting partial sums near the boundary. p = 1.01 converges to about 100.6, but its partial sums at n = 60 have barely reached 4.7, numerics near the threshold mislead in both directions, which is precisely what the test is for.
Real-World Applications
The p-series boundary governs real phenomena that accumulate diminishing contributions. In network science, degree distributions with tail exponent p straddle the same threshold: moments diverge or converge according to p-series logic, deciding whether average behavior even exists, the mathematical heart of "heavy-tailed" internet and social-network models. Zipf's law in linguistics (word frequencies ∝ 1/rank) is a harmonic-series statement, and the divergence of Σ1/n is why vocabulary coverage grows logarithmically with corpus size.
The p = 2 member is everywhere in physics, Σ1/n² structures appear in vibrating string overtones and quantum energy levels, while the Riemann zeta function ζ(p) = Σ1/nᵖ, this page's family viewed as a function of p, sits at the center of analytic number theory and the distribution of primes: Euler's product formula links ζ to primes directly, making the humble p-series the gateway to the deepest open problem in mathematics, the Riemann Hypothesis.
Frequently Asked Questions
What are the known exact values of p-series?
Even exponents have closed forms in powers of π: ζ(2) = π²/6, ζ(4) = π⁴/90, ζ(6) = π⁶/945, Euler found them all. Odd exponents are mysterious: ζ(3) ≈ 1.2021 (Apéry proved it irrational in 1978) has no known closed form, one of analysis' famous open questions.
How is the p-series used in comparison tests?
As the yardstick: to test Σ(n+3)/(n³+1), note the terms behave like 1/n² for large n, and limit-compare with the convergent p = 2 series. Identify the dominant powers, form the effective p, and the verdict follows, the standard workflow for rational-function terms.
Why does the integral test apply here?
Because 1/xᵖ is positive, continuous, and decreasing, the test's requirements, making the series and integral converge or diverge together, with the tail bounds ∫_N^∞ ≤ tail ≤ ∫_{N-1}^∞ thrown in. The p-series is the integral test's flagship application.
What happens for p ≤ 0?
The terms don't even shrink, 1/n⁰ = 1 forever, and negative p makes them grow, so the series diverges by the divergence test before subtler questions arise. The interesting regime is 0 < p ≤ 1: terms vanish, sums still explode.
Is there anything between the harmonic series and convergence?
Yes, a whole hierarchy: Σ1/(n ln n) diverges, Σ1/(n (ln n)²) converges, logarithmic corrections slice the boundary infinitely finely. Each layer falls to the integral test, and each shows the p-test is the first word on the boundary, not the last.
How fast do convergent p-series converge?
The tail after N terms is roughly N^(1−p)/(p−1), so p = 2 has tail ≈ 1/N (slow: three decimals needs a thousand terms), while p = 4's tail ≈ 1/(3N³) settles far faster. Bigger p buys speed; p near 1 converges in principle and crawls in practice, as the p = 1.01 chip dramatizes.
Why is the boundary at p equal to 1 so important?
The p-series converges for p greater than 1 and diverges for p at or below 1. The harmonic series at p equal to 1 is the razor's edge, diverging despite its terms shrinking to zero, which surprises nearly every student.