MathDL - The MAA Mathematical Sciences Digital Library
Search

Search Loci: Convergence:

Keyword

  Advanced Search
Random Quotation

Veblen, Thorstein (1857-1929)

The outcome of any serious research can only be to make two questions grow where only one grew before.

The Place of Science in Modern Civilization and Other Essays.

See more quotations

The Mathematical Association of America
The National Science Digital Library Project
The National Science Foundation
Register Sign In

Loci: Convergence

Sums of Powers of Positive Integers

by Janet Beery (University of Redlands)

Conclusion

Sums of powers of positive integers have been of interest to mathematicians since antiquity.  Over the years, mathematicians in various places have given verbal formulas for the sum of the first n positive integers, the sum of the squares of the first n positive integers, the sum of the cubes of the first n positive integers, and so on.  Beginning as early as the tenth or eleventh century, general methods existed.  However, since each sum depended on the sums of the lower powers and required extensive new calculation, often done entirely verbally, in practice, these general methods did not result in calculation of formulas for sums of very high powers.  Thomas Harriot seems to have been the first to derive and write formulas for sums of powers using symbolic notation, but even he calculated only up to the sum of the fourth powers.  Johann Faulhaber used some symbolic notation and must have spent months, if not years, calculating formulas for sums of powers up to the 17th power for his 1631 Academia Algebrae.  Jakob Bernoulli also may have spent months or years calculating formulas for sums of powers up to the tenth power, but at some point he hit upon the pattern needed to compute relatively quickly and easily the coefficients of the formula for the sum of the cth powers for any positive integer c.  Although his method required one to have computed sums of lower powers, or at least to have recorded the Bernoulli numbers, it was efficient enough that Bernoulli himself accomplished the following amazing feat.

With the help of this table, it took me less than half of a quarter of an hour to find that the tenth powers of the first 1000 numbers being added together will yield the sum 91,409,924,241,424,243,424,241,924,242,500.  (Smith, p. 90)

Today we might write Bernoulli’s formula for the sum of the cth powers as $$\sum_{k = 1}^n {k^c }  = {1 \over {c + 1}}\sum_{m = 0}^c {c+1\choose m}B_m n^{c + 1 - m} $$

with the Bernoulli numbers Bm defined recursively by $$B_0 = 1;\, B_1 = {1 \over 2};\, {1 \over {m + 1}}\sum_{i = 0}^m {m+1\choose i}B_i  = 1$$

for m even and at least 2; and Bm = 0 for m odd and at least 3. Actually, $${1 \over {m + 1}}\sum_{i = 0}^m {m+1\choose i}B_i  = 1$$

defines the Bernoulli numbers recursively for every nonnegative integer m.  The first several Bernoulli numbers defined by this formula are B0 = 1, B1 = 1/2, B2 = 1/6, B3 = 0, B4 = –1/30, B5 = 0, B6 = 1/42, B7 = 0, B8 = –1/30, B9 = 0, B10 = 5/66, B11 = 0, ….  For more modern formulas for sums of powers and Bernoulli numbers, see the article by Apostol, especially pp. 178-179. 

 

Acknowledgments:  We are grateful to Patricia Cornez (instructor) and Michael Camp (student), of the University of Redlands Department of Computer Science, for preparing the animations for this article; to John Navarrette, student of mathematics and of German at the University of Redlands, for his assistance with translation; and to Sandra Richey, of the University of Redlands Library, for obtaining books and journals from libraries near and far.  We thank the British Library and Huntington Library for the use of manuscripts and rare books, the University of Dresden Digital Library for the use of a digital copy of Academia Algebrae, and Columbia University for permission to use a digital image from Academia Algebrae.  Finally, we thank the editors of Convergence, Victor Katz and Frank Swetz, for their encouragement and assistance.

Pages: | 1 |  2 |  3 |  4 |  5 |  6 |  7 |  8 |  9 |  10 |  11 |  12 |  13 |  14 |  15 |  16 |  17 |  18 |  19 |  20 |  21 | 

Beery, Janet, "Sums of Powers of Positive Integers," Loci (February 2009), DOI: 10.4169/loci003284



Discuss this article

start a new discussion thread  | show all 3 threads about this article

thread #1:

Karaji's solution

by Hasan Unal (posted: 01/09/2010 )

I just wonder if the Karaji Generalized his method of solutions? Is there any information about that. For instance did he found sum of fifth powers?

Reply:

Al-Karaji and higher powers by Janet Beery (posted: 01/20/2010 )
My understanding is that there is no evidence that al-Karaji derived a "formula" for the sum of the fourth, fifth, or higher powers. His justification for his formula for the sum of the cubes was ingenious but al-Haytham's idea seems more readily generalizable.

add your reply

thread #2:

On the symmetry of the resulting polynomials

by Christopher Yeleighton (posted: 05/01/2011 )

== Conjecture == It would be interesting to know that * sums of odd powers can be factored as a polynomial on (n ⋅ (n + 1) / 2) and is therefore symmetric with the center at (−½); and that * sums of even powers can be factored to a function like above multiplied by (2 ⋅ n + 1) and therefore antisymmetric with the center at (−½). It would be interesting to know whether any rule concerning the polynomials thus obtained.

Replies:

Re: On the symmetry of the resulting polynomials by Rick Mabry (posted: 12/30/2011 )
Christopher, see the following article, especially section 3 (Faulhaber Polynomials). Although my browsers don't let me view the code you pasted, I think the article confirms some of what you're getting at (some of which might be gleaned from page 8 of the article we're viewing). A. F. Beardon, Sums of Powers of Integers, American Mathematical Monthly, Vol. 103, No. 3 (Mar., 1996), pp. 201-213.

More about symmetry: by Janet Beery (posted: 01/03/2013 )
The most succinct, beautiful, and perhaps even historically plausible explanation I've seen for the symmetry of the Faulhaber polynomials about -1/2 was in a recent article by Reuben Hersh in the College Math Journal 43:4 (Sept. 2012), pp. 322-4. His approach? Extend the polynomials to negative values of n.

add your reply

MathDL Homepage MathDL Homepage National Science Digital Library The Mathematical Association of America