# Euler's Constant Solution

#### Tyger

In Mathematics a little substitution can work wonders. I'm going to show you how to solve a 150 year old riddle by using a little substitution.

Euler's Constant (Euler-Macheroni Constant), represented by &gamma; value .5772156649 has left mathematicians with two perplexing questions. Can it be represented as the sum of a series, and is it an irrational or transcendental number?

Normally &gamma; is represented as the limit of this expression:

(a) 1 + 1/2 + 1/3 ... +1/n &minus; ln(n)

however note that if we replace ln(n) with ln(n + 1/2)

the limiting value is approaced much more quickly. In fact any other substitution e.g. 2/5 or 3/5 doesn't approach the limit as rapidly. This turns out to have crucial significance for finding the answer for it allows us to change (a) so that each term in the harmonic series corresponds to a term in another series. This will result in &gamma; being represented as the difference between an infinite series and an infinite array, which answers our first question. Here is how each term appears:

(b) 1/n &minus; ln(n + 1/2) + ln(n &minus;1/2) which equals

(c) 1/n &minus; ln{(n &minus;1/2)/(n + 1/2)}

using the standard series for ln(x/y) which is

(d) 2&Sigma; {1/(2m + 1)}[(x &minus; y)/(x + y)]^2m + 1, for m = 0 and up.

Substituting for x and y and including the harmonic series term

(e) 1/n &minus; 2&Sigma;{1/(2m + 1)}(1/2n)^2m + 1

but note that the harmonic series term and the first term of the logarithmic series cancel. Also for n = 0 we have to add ln(2) so our final expression for the value of &gamma; is
&Sigma;
(f) ln(2) &minus; 2&Sigma;&Sigma; {1/(2m + 1)}(1/2n)^2m + 1 for m =1 and up and n = 1 and up.

Despite the fact that the array is two dimensional it converges rapidly. I haven't bothered to try to prove if it is irrational or transcendental, the former can be seen to be true almost by inspection, the latter should not be difficult to prove. I'm sure some young able-brained math whiz will have no trouble with it.

Last edited:

#### Hurkyl

Staff Emeritus
Gold Member
the former can be seen to be true almost by inspection
If you can see things like this by inspection, then you stand to become a very rich mathematician.

Staff Emeritus
Gold Member
Dearly Missed
What I want to know is, does that final formula converge to .5772156649 (at ten place accuracy). Anybody got some math software that will do the sums?

#### Tyger

As I said

the array converges very rapidly so that a few minutes with a pocket calculator will get good accuracy. Also each row of the array is a logarithmic expression minus the harmonic term, so it can be summed row by row. It's also possible to appproximate the remainder of the array. These are all methods that were used before they had math programs.

In any case barring some clerical error on my part, it does give the correct value for Euler's Constant.

#### Tyger

Brad, that is a very good link

but no I didn't see either formula there, the assymtotic form with n+1/2 or the series and array derived from it. I am convinced that this is new. Still amazing how many ways of deriving it without quite hitting the jackpot, so to speak. I actually found it quite a few years ago and am surprized that no one else seems to have. But it isn't very obvious, it required that I recognize that there was special significance to the rapidity with which assymtote converged.

That's cool stuff. Sort of like my new 'e' equation, but I won't put it here. I'll publish it in about 2-4 years once I get a degree and I can get a formal proof.

### The Physics Forums Way

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving