The Ramanujan's master theorem states that: $$ \int_0^{\infty}x^{s-1}\sum_{n=0}^{\infty}\frac{(-1)^n}{n!}a_nx^ndx=\Gamma(s)a_{-s} $$

I found a really strange proof recently on a personal blog: Define $$ \tau *a_n=a_{n+1} $$ we have $$ {\int_0^{\infty}x^{s-1}\sum_{n=0}^{\infty}\frac{(-1)^n}{n!}a_nx^n\text{dx} \\=a_0\int_0^{\infty}x^{s-1}\sum_{n=0}^{\infty}\frac{(-1*\tau_a*x)^n}{n!}\text{dx} \\=a_0\int_0^{\infty}x^{s-1}\text{e}^{-\tau_ax}\text{dx} \\=\frac{a_0}{\tau_a^s}\int_0^{\infty}{(\tau_a x)}^{s-1}\text{e}^{-\tau_ax}\mathrm{d\tau_a x} =\frac{a_0}{\tau_a}\Gamma(s) \\=a_{-s}\Gamma(s) } $$

My question is, how can I understand the above proof? Can anyone give a rigorous explanation for the above proof? For example, I totally have no idea about the meaning of $\color{blue}{\mathrm{d\tau_ax}}$(in the sense of analysis).

Would it be related to Umbral calculus?

Also, I want some further readings if possible.

  • $\begingroup$ See if this link lends you some help:dauns01.math.tulane.edu/~vhm/papers_html/rmt-final.pdf $\endgroup$ Dec 21, 2020 at 15:46
  • $\begingroup$ May I suggest that you change the title to "Ramanujan's Master Formula: A proof and relation to umbral calculus" $\endgroup$ Dec 21, 2020 at 20:20
  • $\begingroup$ ok, thanks for the suggestion $\endgroup$
    – FFjet
    Dec 22, 2020 at 2:18
  • 2
    $\begingroup$ Formula, or recipe, or better yet heuristic, rather than theorem, is more in line with R's application of the algorithm (unmatched by any other math mage). For me, it's a quick way of guessing the Mellin transform (or related extensions) of functions from their Taylor series expansion (or other types of expansions) and a guide for constructing Newton interpolations (standard and extended). Don't get hung up on on general proofs and precisely delineating the borders of its applicability--that's not the intention or in the spirit of R's gift. $\endgroup$ Dec 26, 2020 at 14:19
  • 1
    $\begingroup$ $\tau$ gets defined, but not $\tau_a$, which is what actually shows up in the formulas. $\endgroup$ Feb 25, 2021 at 1:51

3 Answers 3


The RMF is definitely related to umbral calculus via the modified Mellin transform (MMT) pair and symbolic extension of the iconic Euler gamma function integral. The proof you copied? I don't know. The MMT pair allows for interpolation of the coefficients of generating functions, often directly connected to sinc and/or Newton interpolation.

First consider the MMT and its inverse

$$\tilde{f}(s) = MMT[f(x)] = \int_{0}^{\infty} f(x) \; \frac{x^{s-1}}{(s-1)!} \; dx$$

$$f(x) = MMT^{-1}[\tilde{f}(s)] = \frac{1}{2 \pi i} \int_{\sigma - i \infty}^{\sigma + i \infty} \frac{\pi}{\sin(\pi s)} \tilde{f}(s) \frac{x^{-s}}{(-s)!} \; ds .$$

Then the RMF holds for a class of functions such that

$$f(x) = e^{-a.x} = \sum_{n \geq 0} \frac{(-a.x)^n}{n!} = \sum_{n=0} a_n \frac{(-x)^n}{n!} = \sum_{n=0} \tilde{f}(-n) \frac{(-x)^n}{n!} \; ,$$

that is, such that we may close the complex contour to the left (e.g., in the sense of the limit of a semicircle with its radius expanding to infinity) for $0 < \sigma < 1$ and $0 < x < 1$ when $F(s)$ has no singularities/poles within the contour. This rep allows an extension of the RMT (and the Mellin transform) to cases in which poles are present in $F(s)$ and other ranges of $x$.

Also note (see, e.g., Gelfand and Shilov's "Generalized Functions") the relation

$$D_x^{m+n+1} \; H(x) \frac{x^m}{m!} = H(x) \frac{x^{-n-1}}{(-n-1)!} = \delta^{(n)}(x)$$

reflected in the two (of several) reps of the fractional differintegro op equivalent under analytic continuation

$$\frac{x^{\alpha-\beta}}{(\alpha-\beta)!} = \frac{d^{\beta}}{dx^\beta}\frac{x^{\alpha}}{\alpha!}=\int_{0}^{x}\frac{z^{\alpha}}{\alpha!}\frac{(x-z)^{-\beta-1}}{(-\beta-1)!} dz = \frac{1}{2\pi i} \oint_{|z-x|=|x|}\frac{z^{\alpha}}{\alpha!}\frac{\beta!}{(z-x)^{\beta+1}}dz ,$$

with $H(x)$ the Heaviside step function.

So, under the conditions above,

$$\tilde{f}(-n) = \int_{0}^{\infty} f(x) \; \frac{x^{-n-1}}{(-n-1)!} \; dx = \int_{0}^{\infty} e^{-a. x} \; \delta^{(n)}(x) \; dx = a_n,$$

and this suggests the analytic continuation and relation to umbral calculus

$$\tilde{f}(s) = \int_{0}^{\infty} f(x) \; \frac{x^{s-1}}{(s-1)!} \; dx = \int_{0}^{\infty} e^{-a.x} \; \frac{x^{s-1}}{(s-1)!} \; dx = (a.)^{-s} = a_{-s}.$$

The iconic guiding example is the Euler gamma function integral rep with $(a.)^n = a_n = c^n$

$$ (a.)^{-s} = a_{-s} = c^{-s} = F(s) = MT[f(x)= e^{-c\; x}] = \int_{0}^{\infty} e^{-c \; x} \; \frac{x^{s-1}}{(s-1)!} \; dx = \frac{1}{c^{s}}.$$

Another useful example, which vividly illustrates the relation to the Appell Sheffer sequences of umbral calculus (of which the $x^n$ with e.g.f. $e^{x}$ is the basic example), is the integral rep for (what I call) the Bernoulli function, simply related to the Hurwitz zeta function and generalizing the Bernoulli polynomials,

$$ B_{-s}(z) = (B.(z))^{-s} = \int_{0}^{\infty} e^{-B.(z)t} \; \frac{t^{s-1}}{(s-1)!} \; dt $$

$$ = \int_{0}^{\infty} \frac{-t}{e^{-t}-1} \; e^{-zt} \frac{t^{s-1}}{(s-1)!} \; dt = s \; \zeta(s,z)$$

where the e.g.f. for the Bernoulli polynomials with $(b.)^n = b_n$ the Bernoulli numbers is

$$e^{B.(x)t} = e^{(b.+x)t} = e^{b.t} e^{xt} = \frac{t}{e^t-1} \; e^{xt}.$$

Note that

$$B_n(z) = -n \; \zeta(1-n,z),$$

$$B_n(1) = -n \; \zeta(1-n,1) =-n \; \zeta(1-n) (Riemann) = (-1)^n B_n(0) = (-1)^n b_n.$$

Through this characterization, it is not too difficult to show that the Bernoulli function inherits all the elegant properties of a regular Appell sequence, such as $D_z \; B_{s}(z) = s \; B_{s-1}(z)$.

Riemann knew all this stuff. Ramanujan intuited it. Hardy formalized it. I stumbled across it on a journey starting from the ladder ops of QM and a brief comment by my old math prof Stallybrass about the sequence $D^{m+n} H(x) \frac{x^m}{m!}$ in his integral transforms class an eon ago.

For application to defining fractional powers of operators, see my answer and comments therein to the MO-Q "What does the inverse Mellin transform really mean?" and several of my blog posts, such as "The Creation / Raising Operators for Appell Sequences."

Added 2/5/21:

Other examples of interpolation of $a_n$ for the exponential generating funcrtion $g(t) = e^{a.t}$ from the MMT of $f(t) = g(-t) = e^{-a.t}$, or, conversely, surmising the MMT of $f(t)$ from the Taylor series coefficients of $g(t)$ via $a_n \; |_{n \rightarrow -s} =a_{-s} =\tilde{f}(s)$:

1) $\;g(t) = \cos(t) = \sum_{n \geq 0} \cos(\pi \frac{n}{2}) \; \frac{t^n}{n!}, $

$\; \; \; \; \;f(t) = g(-t) = \cos(t) = \sum_{n \geq 0} \cos(\pi \frac{n}{2}) \; \frac{t^n}{n!} ),$

$\; \; \; \; \;\tilde{f}(s) =\cos(\pi \frac{s}{2})$ for $0 < Re(s) < 1,$

2) $\;g(t) = \sin(t)= \sum_{n \geq 0} \sin(\pi \frac{n}{2}) \; \frac{t^n}{n!},$

$\; \; \; \; \;f(t) = g(-t) = \sin(-t) = \sum_{n \geq 0} \sin(-\pi \frac{n}{2}) \; \frac{t^n}{n!},$

$\; \; \; \; \;\tilde{f}(s) =-\sin(\pi \frac{s}{2})$ for $-1 < Re(s) < 1,$

3) $\;g(t) = \frac{1}{1-t} = \sum_{n \geq 0} \; n! \; \frac{t^n}{n!},$

$\; \; \; \; \;f(t) = g(-t) = \frac{1}{1+t} = \sum_{n \geq 0} \cos(\pi n) \; n! \; \frac{t^n}{n!},$

$\; \; \; \; \;\tilde{f}(s) =(-s)! $ for $0 < Re(s) < 1,$

4) $\;g(t) = \frac{1}{1+t} = \sum_{n \geq 0} \cos(\pi n) \; n! \; \frac{t^n}{n!} ,$

$\; \; \; \; \;f(t) = g(-t) = \frac{1}{1-t} = \sum_{n \geq 0} \; n! \; \frac{t^n}{n!},$

$\; \; \; \; \;\tilde{f}(s)=\cos(\pi s) (-s)!$ for $0 < Re(s) < 1,$

5) $\;g(t) = \ln(1-t) = \sum_{n \geq 0} \; -(n-1)! \; \frac{t^n}{n!} ,$

$\; \; \; \; \;f(t) = \ln(1+t) = -\sum_{n \geq 0} \cos(\pi n) \; (n-1)! \; \frac{t^n}{n!},$

$\; \; \; \; \;\tilde{f}(s) = -(-s-1)! $ for $-1 < Re(s) < 0,$

6) $\;g(t) =\sum_{n \ge 0} \frac{x^n}{n!} \frac{t^n}{n!}, $

$\; \; \; \; \;f(t) = J_0(2 \sqrt{xt}) =\sum_{n \ge 0} (-1)^n \frac{x^n}{n!} \frac{t^n}{n!},$

$\; \; \; \; \;\tilde{f}(s) = \frac{x^{-s}}{(-s)!}$ for $0 < Re(s) < \frac{3}{4},$

7) $\;g(t) = e^{-t^2} =\sum_{n \ge 0} \cos(\frac{\pi n}{2}) \; \frac{n!}{(\frac{n}{2})!} \; \frac{t^n}{n!}, $

$\; \; \; \; \;f(t) = g(-t) = e^{-t^2},$

$\; \; \; \; \;\tilde{f}(s) = \cos(\pi\frac{ s}{2}) \; \frac{(-s)!}{(-\frac{s}{2})!} = \frac{1}{2}\frac{(\frac{s}{2}-1)!}{(s-1)!} \;$ for $ Re(s) > 0.$

8) $\;g(t) =\sum_{n \ge 0} \; \frac{a_{\bar{n}}\; b_{\bar{n}}}{c_{\bar{n}}} \; \frac{t^n}{n!} = F(a,b;c;t)$, the hypergeometric function, where, e.g.,

$\; \; \; \; \;a_{\bar{n}} = \frac{(a+n-1)!}{(a-1)!}$, the rising factorial,

$\; \; \; \; \;f(t) = g(-t) = F(a,b;c;-t),$

$\; \; \; \; \;\tilde{f}(s) = \frac{a_{-\bar{s}} \; b_{-\bar{s}}}{c_{-\bar{s}}},$ see the Mellin-Barnes contour integral.

  • $\begingroup$ Thanks for your answer and suggestions. I'll pay attention to that next time. $\endgroup$
    – FFjet
    Dec 22, 2020 at 2:28
  • $\begingroup$ As usual, one has to be careful about signs, so I put a little more detail in an answer to the the related MO-Q mathoverflow.net/questions/353282/… $\endgroup$ Dec 22, 2020 at 2:35
  • $\begingroup$ Thx, I'll read it. $\endgroup$
    – FFjet
    Dec 22, 2020 at 2:40
  • $\begingroup$ Another simple application of the RMF is illustrated in mathoverflow.net/questions/381935/… $\endgroup$ Jan 23, 2021 at 19:52
  • $\begingroup$ A relation to Newton interpolation, which often has a larger region of convergence than the MMT, is given in mathoverflow.net/questions/192146/… $\endgroup$ Feb 5, 2021 at 21:28

This line of argument goes back to 1874 papers by J. W. L. Glaisher and J. O’Kinealy. A discussion of this early work and a critical examination (conditions on $a_n$ for which the theorem holds) is given in Ramanujan's Master Theorem, by Amdeberhan et al.

See also this related MSE discussion.

  • 2
    $\begingroup$ A lot of nonsense around Ramanujan's master theorem. In your linked article they only define $\Phi(x)$ by its power series for $|x|$ small but they don't mention we are looking at its analytic continuation (has to be proven it exists) in $\int_0^\infty \Phi(x)x^{s-1}dx$. $\endgroup$
    – reuns
    Dec 21, 2020 at 15:53
  • $\begingroup$ Carlo, you seem to always have an encyclopedia of references at your fingertip. Do you know of any ref on the operator $(x \; D_x \; x )^n = x^n D_x^n x^n$ related to the Laguerre polynomials? In the day, Stokes and others published short notes on this op, and someone refed Glaisher, I believe, but I could never find the ref. $\endgroup$ Dec 21, 2020 at 19:51
  • $\begingroup$ @TomCopeland -- are you referring to $L_n(x)=(1/n!)\exp(-D_x x D_x)(-x)^n$ ? (I have not seen the $x D_x x$ variety) $\endgroup$ Dec 21, 2020 at 20:40
  • $\begingroup$ @CarloBeenakker, see oeis.org/A132440 for $(xDx)^n$ and tcjpn.wordpress.com/2012/11/29/… for the relation to $(DxD)^n$. Recall also $n! L_n(-:xD:) = (:Dx:)^n$ where $(:AB:) ^n= A^nB^n$. $\endgroup$ Dec 21, 2020 at 21:03

The rigorous proof of Ramanujan master theorem is the easiest part, the hard one is to find the precise conditions we need, ie. in which cases it starts to fail. I'll rename $\Gamma(s)\phi(-s)$ into $F(s)$.

If $F(s)$ is analytic and has exponential decay on a strip $\Re(s)\in (a,b),b >a>0$ then (Cauchy integral formula) so does $F''(s)$ on $\Re(s)\in (a+\epsilon,b-\epsilon)$, the inverse Mellin transform $f(x)=M^{-1}[F(s)](x)=\frac1{2i\pi}\int_{c-i\infty}^{c+i\infty} F(s)x^{-s}ds, c\in (a,b)$ converges and is analytic for $arg(x)\in (-r,r)$ and $x^{v-1}f(x)$ is integrable on $(0,\infty)$ for $v\in (a,b)$ so we have the convergent Mellin transform $F(s)=M[f(x)](s)=\int_0^\infty x^{s-1}f(x)dx$ for $\Re(s)\in (a,b)$. If also $f$ is equal to a power series on $(0,\delta)$ then $F(s)$ extends meromorphically to $\Re(s)<b$ with simple poles at $-n$ of residue $f^{(n)}(0)/n!$.

Forgetting about $F$, if we ask that $f(x)=\sum_{n\ge 0} \frac{a_n}{n!} x^n$ converges for $x\in(0,\delta)$ and extends analytically to $(0,\infty)$ such that $x^{b-1}f(x)$ is integrable for some $b>0$ then $F(s)=\int_0^\infty x^{s-1}f(x)dx$ is analytic for $\Re(s)\in (0,b)$ and it extends meromorphically to $\Re(s)<b$ with simple poles at negative integers of residue $a_n$.

On the space of such sequences $(a_n)$, with $(T^k a)_n= a_{k+n}$, the linear map $U(x)(a) = (e^{xT} a)(0)=\sum_{n\ge 0} \frac{a_n}{n!} x^n$ for $x\in (0,\delta)$ extends analytically to $(0,\infty)$ and the theorem is that the obtained linear map $V(s)(a)=\frac1{\Gamma(s)}\int_0^\infty x^{s-1}U(x)(a)dx,\Re(s)\in (0,b)$ extends analytically to $\Re(s)< b$ and $V(-k)=T^k$ for all integer $k\ge 0$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.