Bourbaki

Proposition 3

Tag 08JE, permanent. Cite it as [Bourbaki, Tag 08JE].

Functions of a Real Variable, chapter VII, § 2. THE GAMMA FUNCTION IN THE COMPLEX DOMAIN

For every real x>0x > 0 one has

∫1x+1log⁡Γ(t) dt=x(log⁡x−1)+12log⁡2π \int_1^{x+1} \log \Gamma(t)\ dt = x(\log x - 1) + \frac{1}{2} \log 2\pi
(Raabe’s integral).

First we establish formula (13) for x=0x = 0. Since log⁡Γ(x)∼log⁡1x\log \Gamma(x) \sim \log \frac{1}{x} as xx tends to 0, the integral ∫01log⁡Γ(x) dx\int_0^1 \log \Gamma(x)\ dx converges. Further, the function log⁡Γ(x)\log \Gamma(x) decreases on ]0,1]]0, 1] (VII, p. 310); for every α>0\alpha > 0 one thus has

1n∑k=1qlog⁡Γ(kn)⩽∫0αlog⁡Γ(x) dx, \frac{1}{n} \sum_{k=1}^q \log \Gamma \left( \frac{k}{n} \right) \leqslant \int_0^\alpha \log \Gamma(x)\ dx,

qq being the largest integer such that q/n⩽αq/n \leqslant \alpha. Since ∫0αlog⁡Γ(x)dx\int_0^\alpha \log \Gamma(x) dx tends to 0 with α\alpha and also 1n∑k=q+1nlog⁡Γ(kn)\frac{1}{n} \sum_{k=q+1}^n \log \Gamma \left( \frac{k}{n} \right) tends to ∫α1log⁡Γ(x)dx\int_\alpha^1 \log \Gamma(x) dx as nn tends to +∞+\infty (II, p. 57, prop. 5) one has

∫01log⁡Γ(x)dx=lim⁡n→∞1n∑k=1nlog⁡Γ(kn). \int_0^1 \log \Gamma(x) dx = \lim_{n \to \infty} \frac{1}{n} \sum_{k=1}^n \log \Gamma \left( \frac{k}{n} \right).

But, by (12), the right-hand side of this formula is the limit of

n−12nlog⁡2π−12log⁡nn, \frac{n-1}{2n} \log 2\pi - \frac{1}{2} \frac{\log n}{n},

whence

∫01log⁡Γ(x)dx=12log⁡2π. \int_0^1 \log \Gamma(x) dx = \frac{1}{2} \log 2\pi.
(14)

Next we remark that from the identity

log⁡Γ(x+1)=log⁡Γ(x)+log⁡x \log \Gamma(x+1) = \log \Gamma(x) + \log x

one deduces, on integrating, that for x>0x > 0

∫0xlog⁡Γ(t+1)dt=∫0xlog⁡Γ(t)dt+∫0xlog⁡tdt. \int_0^x \log \Gamma(t+1) dt = \int_0^x \log \Gamma(t) dt + \int_0^x \log t dt.

But the integral on the left-hand side is also equal to ∫1x+1log⁡Γ(t)dt\int_1^{x+1} \log \Gamma(t) dt. Thus, by (14),

∫1x+1log⁡Γ(t)dt=∫0xlog⁡tdt+12log⁡2π=x(log⁡x−1)+12log⁡2π. \int_1^{x+1} \log \Gamma(t) dt = \int_0^x \log t dt + \frac{1}{2} \log 2\pi = x (\log x - 1) + \frac{1}{2} \log 2\pi.

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Transcribed from the printed volume. Not translated.