Romyn Hitchcock, Popular Science Monthly (1881)
“ It may be thought that so small an error as five one-thousandths of an inch, in a bar more than three feet long, is not worthy of serious consideration; but the fact is, that any error that can be detected by the most refined instrumental means is of great consequence, especially for scientific work. The error mentioned above will appear as a relatively large one, when we state with what accuracy minute measurements may be conducted. Professor Rogers considers that the error in comparing the length of two metre bars need not exceed the one-millionth part of a metre. ”
