IL , " Third Supplementary Paper on the Calculationi of the Numerical Value of Euler 's Constant .
" By WILLIAM SHANKS .
Communicated by the Rev. B. PRICE .
Received February 29 , 1868 .
When n=5000 , we have 23v* E=-57721 56649 01532 86060 65120 90082 40243 10421 5933.5 93995 35988 05773 14949 71379 78029 07030 ( last term is +Bii 22 .
5002 Comparing the values of E obtained from taking n-500 , 1000 , 2000 ( given in former papers ) , and 5000 ( giveil in this ) , and assuming that the increase in the several values of E obtainable from taking n higher nuLmbers will be nearly constant , we may conjecture that the valtue of the 60th decimal last found in E will be increased I by taking n-5000 .
4 ; the 59th place will be increased 1 by taking n=5000 .
4 " ; in like manner the 58th decimal will be increased 1 by taking n=5000.4100 , and the 57th also I when n=5000 .
41000 .
It is certain , however , that when n is very large we may , numerically speaking , express E pretty nearly by Sn-loge n ; and indeed when n becomes infinite , the formula E= Snlogen- , + 2_n 2 .
n2 becomes E Snloge n , as giveln by Professor Price in his ' Infinitesimal Calculus .
' In the value of E last found , then , we deem it probable that at least 56 decimals will remain unchanged , whatever high values be given to n.

