Ss AS Sethe her bbl a eee party | St = an = 
em a 
of bars at the temperature of melting ice. 457 
Differences in Length of Steel and Copper Meters. 
Date, 1883. §,—C, Mean. Remarks. 
March 4 +43°7 
4 +50°5 fe 
- 5 +452) +4471 Before heating C, 
45 5 +475 
bs 6 +48°8 
4 7 +45°1 | { 
9 +474 ; 
= 9 +51°4 Ze C, to near 212° F. 
- 10 +47°7 
- 10 +45°6 
el SOTy $4r9) 4462 ppt 
“% 12 +45°6 
Differences in Length of Steel and Brass Meters. 
Date, 1883. S,—B, Mean. Remarks. 
ft 
March 10 +40°1) a 
ng ll + 36°T é * B 
ua 12 4.36'1 +372 Before heating B, 
ee 13 +36°0 
3 13 aay After heating 
" 14 +36°7 +35°8 9° F. 
uc 15 +36°8 B, to near 21 
" 20 +38°4 i 
og +417) 439°6 a es berate) a 
e 23 +38°6 
The probable error of a single difference of length in the 
above tables may be derived preferably from the formula 
oe7454/ L el, 
in which m is the whole number of Seoul n the number of 
groups of comparisons and [vv] the sum of the squares of the 
diedieratiges between the mean and individual results for the 
several groups. Including the results for the glass bars given 
below, this probable error is found to be oh 8. ae 
The i th etive groups of comparisons of the 
eee el ee a ‘th a standard, do not differ 
steel, copper and brass meters wit 
temperature of melting ice) was produce 
cooling anes although the brass net was aprersne length- 
ened by heating and aay iby cooling ifference 
between the mean results ft e brass Fee after heating and 
after cooling it, viz: 38, fam * a measurable quantity is 
quite within the range which cult be inferred from the above 
Pp | 
