MARIN MERSENNE, 15881648

SHAILESH A SHIRALI

Is it possible for any one person to the serve as a human equivalent of the World Wide Web? And could this have been possible in the post-Renaissance world four centuries back, when letters could take weeks to travel from one point to another? In this article we document the story of one individual who appears to have done just this: Father Marin Mersenne (15881648), who lived in France during an era of intellectual giants, from his country and elsewhere: Ren Descartes (15961650), tienne Pascal (15881651) and his son Blaise Pascal (16231662), Pierre de Fermat (16011665), Gilles Roberval (16021675), Girard Desargues (15911661), Galileo Galilei (15641642), Johannes Kepler (15711630), Evangelista Torricelli (16081647) and others, with whose names are linked so many great discoveries and inventions in mathematics and mechanics: the marriage of algebra and geometry (Descartes), the birth of probability theory (Fermat and Pascal), results in number theory (Fermat), results in geometry and hydrostatics (Pascal), advances in projective geometry (Pascal and Desargues). Against this backdrop, the work of Mersenne in sound and number theory may seem modest. But we shall show that Mersennes real contribution lay in quite another area. Living in an era which was enormously productive for mathematics and mechanics the decades preceding Newton and Leibnitz, and the decades during which the seeds of the calculus were being sown his role was absolutely central.

F IGURE 1. Father Marin Mersenne in his cell (source: [2])

SHAILESH A SHIRALI

Early years. Marin Mersenne was born in September 1588, in a town called Oiz. His parents were of modest means, but they took pains to attend to his education, sending him to a Jesuit School indeed, the same school attended by Descartes some years earlier. He was expected to join the Church (we nd this theme occurring time and again in such biographies!), but his eagerness to study further had a decisive role in shaping his future. On his way to Paris for further studies in philosophy and theology, he stayed at a convent of the Order of the Minims, and this brief contact touched him deeply. He completed his studies in the Sorbonne in 1611, and promptly joined the Order of the Minims, in Paris.

The Minims. The Order had been set up two centuries earlier. They regarded themselves as the least of all the religions; hence their name, minims. They gave a high place to study and scholarship. It must have been these aspects of the Order that so attracted Mersenne. It was their practice to be clothed in a habit made of coarse wool, and this is what we see in all the surviving portraits of Mersenne. By 1614 Mersenne was teaching philosophy and theology in the monasteries of the Order. Around this time he discovered and explored the curve known as the cycloid. (We shall say more about this later in the article.) It was during this period that his characteristic style of work began to take shape: the way he maintained links with scholars and exchanged ideas with them. Mersenne stayed at the Place du Royale monastery in Paris from the second half of this decade to the end of his life a stay of over thirty years, supported for the most part by the Church. As his biography states, . . . The Minims realised that the biggest service he could give was through his books and they never asked any more of him (see [5]).

Transformation. A remarkable transformation happened in Mersennes life over the 1620s. His initial writings were mostly on religious topics, and he would have been regarded as pro Aristotle and anti Galileo. But in the late 1620s he studied the criticisms made against Galileo very closely, and by the early 1630s he was convinced that Galileos ideas needed to be propagated. By this time he had also decided for himself that alongside religious studies he would devote his time to science and mathematics. This interest in combination with his meetings with scholars soon gave rise to an extraordinary and unprecedented tradition, in which he began to keep contacts with a number of scientists and mathematicians (including some whom he never met): Descartes, Fermat, tienne Pascal and Blaise Pascal, Roberval, and many others. He set up meetings with them in which they would discuss their work. This informal academy began to be known as Acadmie Mersenne. At one such meeting Mersenne persuaded Roberval to work on the cycloid, and this brought forth rich dividends.

Mersennes work in music and sound. Meanwhile Mersenne continued his own research. Here is an example of one of his discoveries, in sound (he had always had a deep interest in acoustics and music). The Pythagoreans had discovered, early on, that as one varies the

MARIN MERSENNE, 15881648

F IGURE 2. Cover page of Mersennes Harmonie Universelle (source: [4]) length of a vibrating string, the pitch of the resulting note varies in inverse proportion to the length. When one halves the length of the string, the pitch doubles; we ascend through one octave. This simple arithmetical relationship along with the fact that pleasing musical harmonies correspond to simple ratios like 2 : 1, 3 : 2, 5 : 3 and so on, had deep implications for the way the Pythagoreans viewed the world, for it suggested that the world had to be constructed on mathematical principles. But pitch must depend on other factors as well; one only has to look at the different strings of a guitar or sitar to see this. Mersenne was the rst person to investigate this matter in a systematic and scientic manner, and he discovered the following beautiful law relating pitch to the tension, the length and the thickness of the vibrating string: tension . pitch of sound length (diameter)2 Mersenne was not content with the purely empirical discovery and strove hard to establish it using the known principles of mechanics. All his thoughts and discoveries on this topic are contained in a book he published in 1630, Harmonie Universelle. His interest in music and the theory of musical harmony also led him to explore a topic which in those days would not have seemed to have anything to do with mathematics: permutations and combinations!

The human web. Through the 1630s he continued his work in encouraging others and making their results known to others. It was said of him: To inform Mersenne of a discovery meant to publish it through the whole of Europe. For example, following a visit to Italy in

SHAILESH A SHIRALI

the early 1640s he learnt of Torricellis work on the barometer and immediately set about publicizing it, while at the same time conducting many experiments of his own in the same eld. Even Galileo beneted from his zeal. Characteristically, Mersenne set about verifying all of Galileos ndings when he heard about them; for example, the law governing the speed 1 2 gt . In his usual of a falling body the formula every school student remembers as s = 2 way he then made Galileos work known to others. Another scientist who beneted greatly from the kindness shown by him was Christiaan Huygens (16291695); he must have been just a teenager at the time. Mersenne posed a challenge to him: to determine the equation of the catenary. But this challenge proved too difcult for the young Huygens. A major puzzle posed to historians by Mersennes life and work is this: How did he manage to pursue his free exchange of scientic ideas (in particular, those concerning the ideas of Copernicus) during an era when many others could not (for example, Galileo)? The answer is not clear but may have to do something with Mersennes style of work and his way of relating to people, and also his close relationship with the Church. (Many centuries after Mersenne, the American essayist and poet Ralph Waldo Emerson (18031882) wrote: The music that can deepest reach, and cure all ills, is cordial speech. Mersenne would surely have agreed. Probably he could have taught us a thing or two about people skills and E.Q.!) Mersennes letters run into thousands of pages. After his death in 1648, letters were found in his cell from seventy eight different correspondents. Read chronologically, they offer a very insightful glimpse at how mathematics and mechanics were evolving during this period of ferment. It is no exaggeration to say that he was the creator of a scientic academy that stretched across the length and breadth of Europe.

T HE CYCLOID As a case study to illustrate Mersennes work, we describe his work related to the cycloid. This is the curve traced by a point on a circle which rolls without slipping upon a straight line. It consists of an endless series of arches, called cycloidal arches (see Figure 3). In the rst half of the seventeenth century, the cycloid and hyperbola occupied a pride of place in mathematics, for a simple reason: they provided a wonderful laboratory for the testing of new techniques. The circle had been known from ancient times; formulas were known for the area enclosed by a circle and by a segment of a circle. Similarly, results were known (thanks to Archimedes) for the volume and surface area of a sphere; there was little left to test. Results were also known about the class of objects next higher in complexity: the conic sections. Archimedes had squared the parabola: he knew how to

MARIN MERSENNE, 15881648

F IGURE 3. One arch of a cycloid, generated by a circle rolling on a line

Cycloid

Semi-ellipse

F IGURE 4. Cycloidal arch and semi-ellipse, superposed

nd the area of a segment of a parabola. Investigations into the hyperbola had already commenced, and new results were around the corner. So the cycloid and hyperbola were made to order for the times; newly discovered techniques could be tested on them: for computing area, using Cavalieris new theory of indivisibles (which would give way to the innitesimal of later generations), and for drawing tangents, using the methods due to Fermat and Descartes. Both curves were simple to describe, yet sufciently complicated that the methods of Euclidean geometry could not unravel all their secrets. Galileo was one of the rst to begin work on the cycloid, and Mersenne probably heard about the curve from him. He wondered whether the cycloid is merely half of an ellipse, cut along its major axis. Galileo determined the area by the simple technique of cutting the cycloidal arch and the generating circle from the same material and comparing their weights (trust the practical minded Galileo to come up with this method, which a pure mathematician would never think of!); he found that the area under the cycloid is roughly three times the area of the circle. That is, if the generating circle has radius r, then the area of the arch is (roughly) 3 r2 . It was Mersenne who introduced the curve to Roberval and urged him to determine its area. Roberval succeeded in doing so, and showed that Galileo was right: the area is indeed three times that of the generating circle. (Compare this with the estimate for area obtained by supposing the gure to be a semi-ellipse. Since the semi-axes of the ellipse are r and 2 2 2r, the area is 1 2 r 2r = r . This is about 4.7% larger than Galileos gure. Figure 4 shows the two curves, superposed.)

SHAILESH A SHIRALI

2
B

1
I A

II

F IGURE 5. Equation of a cycloid generated by a circle of radius 1 unit

Quadrature of the cycloid. The way in which Roberval worked out the area is typical of the times: by using simple, down to earth pre-calculus reasoning, coupled with Cavalieris method of indivisibles (as distinct from the Greek method of exhaustion). We shall explain his reasoning using modern algebraic language which makes it easier to understand. First, let us determine the equation of a cycloid (see Figure 5). Consider two positions (I and II) of the rolling circle in which the point A has moved to B. The movement can be regarded (in the physics sense) as the resultant of two movements: a movement along the arc of the circle in position I, from A to B , followed by a horizontal movement from B to B. (Note the parallelogram AB BC.) If we denote by t the angle subtended by arc AB at the centre of circle I, then we easily get the coordinates of B; they are: (t sin t , 1 cos t ). This yields a parametrization of the cycloid (the rst arch is generated by 0 t 2 ). Arguing thus, Roberval focuses his attention on half the arch (and half the circle) and constructs an auxiliary curve (see Figure 6) essentially by subtracting the sideways movement produced as a result of rolling. To do this he draws a horizontal line through the gure (), giving rise to the segment EF and the point of intersection E (on the cycloid). Then he measures off a length E F equal to EF . The locus of the F is the auxiliary curve. Now by invoking Cavalieris theory of indivisibles he infers that the area of region II enclosed by the auxiliary curve and the cycloid (i.e., by arcs PF R and RE P) is the same as the area of 1 the semicircle III, i.e., 2 r2 . It remains only to nd the area of region I. The curve PF R is a sinusoidal curve, and it is easy to argue in a number of different ways, by physics reasoning or by invoking the fact that cos( /2 t ) = cos( /2 + t ), that the area of region I is half the area of rectangle PSRQ; indeed, that the curve PF R divides the 1 rectangle PSRQ into two congruent halves. Hence the area of region I is 2 r 2r = r2 . 3 It follows that the area of half the cycloidal arch is 2 r2 , and hence that the area of the 2 complete arch is 3 r .

MARIN MERSENNE, 15881648

III

II

F IGURE 6. Robervals construction of the auxiliary curve

Cavalieris theory of indivisibles is not known by that name now; it is called Cavalieris Principle, and it states that if a family of parallel lines intersects two planar regions in pairs of segments with equal length, then the areas of the two regions are equal. (It has a three dimensional version as well: if a family of parallel planes intersects two solids in pairs of planar regions with equal cross-sectional areas then the volumes of the two solids are equal.) Roberval did not reveal the method by which he deduced his result. (In fact it became known only after his death.) While this too was typical of the times, in his case it had to do with the fact that he had to justify holding his ofcial position by composing problems for a scholarship competition, and it was therefore in his self-interest to keep his methods secret. Roberval also found a way of nding the slope at any point of a cycloidal arch. (He kept this method a secret too! However in this case it was Descartes who came out with the better solution. But we shall not discuss this problem here.) It is interesting to report on the tone of the correspondence between Mersenne and Descartes, Fermat and Roberval. Characteristically, Descartes has nothing favourable to say about almost anyone (other than himself, of course). For example he dismisses Robervals accomplishment with the caustic statement: Roberval has laboured overmuch to produce so small a result. (Robervals response: Prior knowledge of the answer to be found has no doubt been of assistance; see [6].) Even Fermat was moved to write the following words in a letter dated 1638 to Mersenne: I will not send you anything else for M. Descartes, since he imposes such harsh regulations on innocent discussions, and it makes me happy to tell you that I have yet to nd anyone here who is not of my opinion. Descartes cannot have been a particularly pleasant person to know!

SHAILESH A SHIRALI

The cycloid continued to feature in front line investigation even much later. For example, the solution to the brachistochrone problem (to nd the curve of steepest descent; problem posed by Johann Bernoulli in 1696, and solved independently by himself and by Jacob Bernoulli, Newton, Leibnitz and LHpital) turned out to be the cycloid, as did the solution to the tautochrone problem (to nd the curve for which the time taken by a particle sliding down the curve to its lowest point is independent of its starting point). This is a remarkable economy of nature, and Johann Bernoulli was moved to exclaim, Nature always tends to act in the simplest way, and so it here lets one curve serve two different functions . . . .

M ERSENNE

PRIMES

We close with a brief discussion of the topic for which Mersenne is best remembered in mathematics: Mersenne primes. Interest in such primes dates to Greek times. In Euclids great text The Elements we nd a perfect number dened as one for which the sum of the proper divisors equals the number itself. Example: 6; for its proper divisors are 1, 2, 3, and 1 + 2 + 3 = 6. Euclid had the following simple and pleasing theorem about such numbers. Theorem 1 (Euclid). Let n > 1 be a positive integer such that 2n 1 is a prime number. Then the number 2n1 (2n 1) is perfect. Since 23 1 = 7 and 25 1 = 31 are primes, Theorem 1 leads us to believe that the numbers 22 7 = 28 and 24 31 = 496 are perfect; and indeed they are. 2n1 P The proof of Theorem 1 is easy. Let P = 2n 1; then P > 1, and the proper divisors of are the numbers 1, 2, 22 , . . . , 2n1 together with P, 2P, 22 P, . . . , 2n2 P. It is easy to check, using the identity 1 + 2 + 22 + + 2n1 = P, that the sum of the proper divisors is 2n1 P.

In the middle of the eighteenth century Euler proved that every even perfect number arises from a prime of the form 2n 1. The two results in combination show that the even perfect numbers may be placed in one-to-one correspondence with primes of the form 2n 1. (But this proof is not so simple.) Mersenne spent a lot of time researching primes of this kind, and in his honour primes of the form 2n 1 are now called Mersenne primes. For short let us write Mn = 2n 1; then Mn is called the nth Mersenne number. It is easy to prove the following: If n is composite, then so is Mn . Indeed, if a is a proper factor of n then Ma is a proper factor of Mn . (More generally, if x, a, b are positive integers, with x > 1 and b > 1, then xa 1 is a proper divisor of xab 1.) Expressing this in contrapositive form we get the following simple result: If Mn is prime, then n is prime.

MARIN MERSENNE, 15881648

Unfortunately the converse of this statement is false; it is not true that M p is prime for every prime p. Here are the rst ten primes p for which M p is prime: (1) 2, 3, 5, 7, 13, 17, 19, 31, 61, 89.

The corresponding Mersenne primes are: M2 = 3, M3 = 7, M5 = 31, M7 = 127, M13 = 8191, M17 = 131071

M19 = 524287,

M31 = 2147483647,

M61 = 2305843009213693951,

M89 = 618970019642690137449562111. Note the absence of some small primes in (1): 11, 23, 29 and 37. Here is a beautiful result concerning the factorization of composite Mersenne numbers; it follows from the little theorem of Fermat: Theorem 2 (Euler). If p is prime and M p is composite, then the prime factors of M p leave remainder 1 when divided by 2 p.

For example, M11 = 2047 = 23 89; observe that 23 and 89 leave remainder 1 when divided by 22. Similarly, M23 = 8388607 = 47 178481, and both 47 and 178481 leave remainder 1 when divided by 46. Theorem 2 may have been known to Fermat, but Euler was the rst to publish a proof. (In passing we note that Theorem 2 provides yet another proof of the innitude of the primes. For any prime number p we consider the number M p = 2 p 1. Either this number is prime, and it clearly exceeds p; or the number is composite, in which case all its prime divisors exceed 2 p. Both possibilities lead to the existence of a prime number exceeding p. It follows that there is no last prime.) One of the tasks that Mersenne set for himself was to nd a way to identify the primes p for which M p is prime. It is not clear whether he knew Theorem 2 (he probably did know the little theorem of Fermat, as he was in close touch with Fermat), but in 1644 he wrote that 2 p 1 is prime for p = 2, 3, 5, 7, 13, 17, 19, 31, 67, 127, 257 but composite for the other 44 primes p < 257. It is remarkable that he could have made such a statement, for it is surprisingly accurate: there are just three primes which ought to be on his list but are not (61, 89 and 107), and just two which are on his list but ought not to be (67 and 257). It is unclear what the basis of his statement was; there is certainly no direct way he could have checked the primality of M p for, say, p > 25, as the numbers involved are too large. It has been suggested that the few errors in his list may be printers errors, and that Mersenne may well have been following this rule: Select only those primes p which differ by 1 from a power of 2, or by 3 from a power of 4. While this accounts for his list (except for the non-inclusion of 61), it is only a conjecture.

10

SHAILESH A SHIRALI

Mersenne primes continue to be a source of investigation; see [7]. It is not known whether or not there are innitely many such primes. Currently just 48 Mersenne primes are known, the largest being M57885161 ; it was found in February 2013. Recently a conjecture was put forward to which at present no counterexamples have yet been found: Conjecture. Let p > 2 be a positive integer. If any two of the following statements are true, then so is the third one: p differs by 1 from a power of 2, or by 3 from a power of 4; 2 p 1 is prime; (2 p + 1)/3 is prime. Closing remarks. Mersennes story is an extraordinary one. It is humbling to realize the strength of community feeling that lies behind his work.

R EFERENCES
[1] http://en.wikipedia.org/wiki/Marin_Mersenne [2] http://en.wikipedia.org/wiki/File:MarinMersenne.jpg [3] http://www-history.mcs.st-andrews.ac.uk/PictDisplay/Mersenne.html [4] http://gallica.bnf.fr/ark:/12148/bpt6k5471093v [5] http://www-history.mcs.st-andrews.ac.uk/Biographies/Mersenne.html [6] J. Martin, The Helen of Geometry, College Mathematics Journal, Vol. 41, No. 1, 1727, Jan 2010 [7] http://en.wikipedia.org/wiki/Mersenne_prime

S AHYADRI S CHOOL , T IWAI H ILL , R AJGURUNAGAR , P UNE 410 513 E-mail address: shailesh.shirali@gmail.com