far the most uncertain. 
make At=:000 at 100°, the result is 
RESEARCHES IN THERMOMETRY. 
585 
I have therefore made in it the alteration necessary to 
At='012901¢— 000249372? + (00000120362. 
The real probable error of a single comparison of calculation with experiment 
is evidently ‘036 x °84=-:030, deduced from the arbitrary condition above referred 
to; the probable error as calculated from all the comparisons below (Table X.) 
is ‘085, a value exceeded by 18 of the actual errors, and not attained by 15 of 
them. 
Series. 
it 
(Air therm, 
TE, 
(Air therm. 
III. 
(Air therm. 
IV. 
(Air therm. 
Mi, 
(Air therm. 
VI. 
(Air therm. 
VII. 
(Air therm. 
WAEET. 
(Air therm. 
IX. 
(Air therm. 
X. 
(Air therm, 
2), 
t (Mere. ) 
28°79 
50°54 
76°90 
112°87 
25°12 
48:17 
74°78 
16°73 
30°29 
50°70 
66°80 
| 29:06 
54:50 
67°70 
17°45 
27°10 
/ 55°53 
67:00 
19-25 
28°86 
49°76 
65°61 
17-76 
28°54 
49°95 
65°77 
113°64 
15°88 
2717 
54:19 
74:14 
97-23 
11557 
At Found. At Cale. At poe ~ At 
ale. 
+'267 OS 074 
— ‘036 “1:70 — ‘206 
+:005 064 —'059 
+°226 010 ‘216 
+048 186 —'138 
—'008 ‘176 —'184 
—'120 ‘074 —'194 
+:'156 BIE, 004 
+:297 195 =O 
+103 ‘170 —'067 
+:103 108 —°005 
+:246 "193 053 
+:215 Mae 058 
+°170 103 ‘067 
+°404 155 "249 
+:460 “191 269 
+:141 153 —012 
+104 ‘107 —°003 
+'368 165 203 
+182 193 —011 
+:206 Nfs) 033 
+:123 113 ‘010 
+199 si "042 
+°343 235 ‘108 
+131 rie 2, —'041 
+160 tal 049 
—°086 013 —'099 
+041 147 —'106 
+°283 190 093 
+:342 158 184 
+:053 075 — ‘022 
—'034 003 —'037 
+°191 018 “173 
Actual errors of the above magnitude occur with about equal frequency in 
the work of preceding experimenters, and seem to be naturally incidental 
They depend partly upon the fact (first observed by 
to such operations. 
