BOOK I REVIEW

Ramanujan: Essays and This volume has some ratherunusual contents

too which we shall describe before moving on
Surveys
to its mathematical contents. First, the only
four extant photographs of Ramanujan are
BSury reproduced and their backgrounds discussed.
Then, there is an interesting account by a
Ramanuian: Essays and Surveys British physician DBA Young on Ramanu-
Bruce C Berndt and Robert A Rankin jan's illness and its (faulty) diagnosis. It has
Hindustan Book Agency been widely believed that tuberculosis was
New Delhi 110 016, 2003.
the cause of Ramanujan's death but the doc-
Price: Rs.4S0/-
www.hindbook.com
tor explains in detail how and why the
diagnosis was faulty and opines that hepatitic
The fact that Ramanujan made path-breaking amoebiasis was the most likely cause of his
discoveries without much formal training illness. Apparently, inadequately treated
and was a misfit under the regular curriculum amoebiasis can be a permanent infection even
(he failed in the FA examination), has been a though the patient could spend long periods
source of mystery to many. This intriguing without obvious symptoms. A briefbiography
aspect has fired the imaginations of several of Janaki Ammal (Ramanujan's widow) by
authors who have penned his biography. Bruce Berndt appears in this volume and is
Almost all of them have dealt more with his followed by an in terview wi th her by P Nan dy.
personal life than with his mathematical work. There is also a biography of Narayana Aiyar,
This volume can be said with some justifi- the Chief Accountant of Madras Port Trust
cation to be an attempt at a more mathematical and Ramanujan's close friend, who supported
biography. A facsimile editionofRamanujan's and encouraged him while in Madras. In fact,
famous notebooks had been published in 1957 Ramanujan's first letters to Hardy were
by the Tata Institute of Fundamental written with the help of Narayana Aiyar.
Research, Bombay and, as such, they were
unedited. A highly interesting discussion is that of the
books studied by Ramanujan in India. It is
One of the two editors, Robert Rankin, passed well known that Ramanujan studied G S Carr's
away just before this volume was published. A synopsis of elementary results in pure
The other editor, Bruce Berndt, has devoted, mathematics and adapted its style in recording
since May 1977, all of his research efforts to his results. However, one of the veritably
editing Ramanujan's notebooks. No amount intriguing questions that arose was how
of praise is adequate acknowledgement of his Ramanujan was able to prove hundreds of
contribution to the mathematical community identities in the theory of modular elliptic
in this regard. functions without learning complex analysis

-R-ES-O-N-A-N--C-E-I-M--ay--2-00-4---------------~~-----------------------------8-7
 BOOK I REVIEW

ina conventional manner. The editors unearth Ramanujan's 'lost' notebook, had written a
the fact that in 1914, the library of the Uni- beautiful account of this discovery as well as
versity of Madras contained a copy of the of the contents of the 'lose notebook in the
book, The applications of elliptic functions by American Mathematical Monthly, 1978. This is
A G Greenhill. This book also does not use reproduced in this volume. In 1976, while
any complex analysis. They go on to point out sorting through some materials from the estate
other features which seem to indicate that ofG N Watson, Andrews found a manuscript
Ramanujan must have read this book and was of about 100 pages in Ramanujan's hand-
influenced by it in his work on modular writing. Andrews contends that these were
equations. written during the last year of Ramanujan's
life, after returning to India, when he was in
An extremely fascinating overview of Rama- 'severe pain and only skin and bones.' Of the
nujan's three notebooks is given by Berndt. contents, only some results on 'mock theta
He counts some 3254 results recorded in functions' came to be known earlier through
them. But, as is well known, as much as half the last letter that Ramanujan wrote to Hardy.
of these may have been rediscovery. As a Perhaps Watson filed this lost notebook away
possible reason for Ramanujan recording only and forgot about it. Watson gave a lecture in
the final result, and not the proofs, Berndt 1935 on some of Ramanujan's results on the
says that like most Indian students of those mock theta functions quoted in the last letter
times, Ramanujan wrote on a slate and rubbed mentioned above. The title of the lecture
out proofs; paper was expensive. Berndt was, 'The final problem: an account of the
emphasizes that Ramanujan thought like mock theta functions.' He says, "I doubt
other mathematicians and not through divine whether a more suitable title could be found
intervention or anything of that sort. He just for it than the title used by John H Watson,
thought with more insight than most people. MD, for what he imagined to be the final
Berndt arranges the contents of this chapter memoir on Sherlock Holmes." Concluding
into 12 basic topics and discusses each in his address, Watson says, "Ramanujan's
some detail. For instance, people who think discovery of the mock theta functions makes
ofRamanujan primarily as a number-theorist it obvious that his skill and ingenuity did not
may be surprised to learn that several hundred desert him at the oncoming of his untimely
theorems of his on theta functions and end. As much as any of his earlier work, the
modular functions, are at the interface of mock theta functions are an achievement
analysis and number theory. Berndt says that sufficient to cause his name to be held in
infinite series were Ramanujan's first love,
lasting remembrance."
and his talents in working with them were
possessed perhaps only by the great Euler.
Another captivating article in the volume
George Andrews who discovered the so-called describes 58 problems that Ramanujan had

-------------------------------~---------------R-E-S-O-N-A-N-C-E-I--M-a-Y-2-0-0-4
88
 BOOK I REVIEW

posed in theJournal o/the Indian Mathematical The Borweins assert that the partial sums in
Society. Although these appear also in Rama- the above infinite series converge to the true
nujan's collected works, here they are not value more rapidly than any other calculation
merely reproduced but are analysed in some of IT until the 1970's. Each successive term
detail. Following this, Freeman Dyson's adds roughly eight more correct digits. The
article, 'A walk through Ramanujan's garden', Borweins bettered Ramanujan's result in 1987.
published immediately after Ramanujan's In an article in Scientific American of 1988,
birth centenary, is reproduced. The following they say, "Iterative algorithms (where the
samplings from his article bear witness to his output of one cycle is taken as the input for
feelings on Ramanujan's mathematics: the next) which rapidly converge to pi were,
in many respects, anticipated by Ramanujan,
"My love affair with Ramanujan's mathe- although he knew nothing of computer
matics began 48 years ago. The wonderful programming. Indeed, computers not only
thing about Ramanujan was thathe discovered have made it possible to apply Ramanujan's
so much, and yet left so much for other people work but have also helped to unravel it. Sophis-
to discover. I have intermittentlybeencoming ticated algebraic manipulations software has
back to Ramanujan's garden and, every time, allowed further exploration of the road
I find fresh flowers blooming." Ramanujan travelled alone and unaided 75
"In the cold dark evenings, while I was years ago."
scribbling these beautiful identities amid the A sense of incredulity prevails on reading
death and destruction of 1944, I felt close to these words when one pictures Ramanujan
Ramanujan. He had been scribbling even sitting and writing on a slate and erasing with
more beautiful identities amid the death and his sleeve!
destruction of 1917."
Finally, there is the extremely educative and
Ramanujan's vast work on hypergeometric fascinating article by Atle Selberg on
series was surveyed by Richard Askey in an Ramanujan's work and how his own work
article published in 1988. This is also repro- was influenced by it. Selberg says, "Ramanu-
duced here. Finally, I come to my two favou- jan's particular talent will seem to be primarily
rite articles in this volume. One is by J M of an algebraic and combinatorial nature. He
Borwein and P B Borwein on 'Ramanujan developed it, for a long time in complete
and pi.' Ramanujan wrote a paper, 'Modular isolation, really without any contact with
equations and approximations to ITwhere one other mathematicians. He had, on his own,
of his formulae reads acquired an extraordinary skill of manipula-
tion of algorithms, series, continued fractions
~ _ y8 ~ (4n)!(1103 + 26390n).
and so forth, which certainly is completely
'if - 9801 f=o(n!)4396 4n
unequalled in modern times."

--------~~------
RESONANCE I May 2004 89
 BOOK I REVIEW

"He might very well have become a theory- In December 2003, there was a meeting of the
builder if he had had a different and more American Mathematical Society in Bangalore
conventional start and training as a mathe- and, George Andrews gave a talk on 'Rama-
matician. Even then, in what has been left in nujan and partial fractions.' He said that after
his work, there seems quite clear evidence several years trying to recognise and under-
that he had developed, on his own, a theory of stand the specific insights that Ramanujan
modular forms and equations." had, he feels now that a lot of Ramanujan's
identities are based on his highly developed
At. another place, Selberg recounts some
intuition to recognise partial fraction expan-
intriguing facts on Ramanujan's work on
sions in diverse situations.
partitions. For instance, he points out that in
his very first letter, Ramanujan has written an Ramanujan's work on cusp forms and their
exact formula for an infinite product analo- generalisations playa central role in present-
gous to that defining the function p(n) of par- day mathematics and other deep discoveries
titions of n; he had suggested also the exis- like the mock theta functions are still in the
tence of such a formula for pen) itself. Selberg early stages of being understood. Thus,
opines that perhaps Hardy did not fully trust Ramanujan's stature in mathematics has only
Ramanujan's insight and chose an expression been growing over the years and this volume
for pen) which gives only an asymptotic will be a valuable source of initiation into
formula. Perhaps, Ramanujan did not press Ramanujan's mathematics for some time to
this point out of respect for Hardy. Later, come. To sum up:
Rademacher came out with an exact formula
Ramanujan did mathematics somehow;
for p(n), whose similarity with Ramanujan's
we still can't figure out even now.
exact formula mentioned above is so striking
He left his mark on 'p of n',
that itis surprising that Hardy and Ramanujan
wrote Jr in series quite often.
did not end up with it themselves.
The theta functions he called 'mock'
Selberg speculates about what would have are subject-matter of many a talk.
happened had Ramanujan come into contact He died very young - yes, he too!
with Hecke, a great mathematician of talents He was only thirty-two!
more similar to Ramanujan's. he says that His name prefixes the function T.
that might perhaps have brought out new Truly, that was his last bow!
things in Ramanujan which did not come
into fruition by his contact with Hardy. But,
he adds that Hardy deserves the greatest credit B Sury, Indian Statistical Institute, Bangalore 560 050,
for recognizing Ramanujan's originality and India. Email: sury@isibang.ac.in

assisting him and his work in the best way he

could.

-90--------------------------------~----------------R-ES-O--N-A-N-C-E-I--M-a-Y-2-0-0-4