189 PROCEEDINGS OF SECTION A. 
rods, which were three-eighths of an inch thick, were insulated 
thermally by wrapping in blankets, the whole being covered by a 
box with glass windows for reading thermometers placed within, 
and almost in contact with the bar. Under these circumstances, 
the steel was found to be affected only very slowly by changes in 
the outside temperature. By systematically arranged comparisons 
with the standard under as wide a range of temperature as was 
practicable, the temperature at which the rods were of standard 
length and the value of the co-efficient of expansion for each were 
determined. The bars gave most. satisfactory results, but as an 
additional precaution the comparisons with the standard were so 
arranged as to be made under conditions as nearly as possible 
those existing during the measurement. During the course of 
the experiments some interesting information was obtained as to 
what is technically known as “lag,” or the time which elapses 
after a change has taken place in the temperature of the body 
under investigation, before the full amount of this change can 
produce its effect on the thermometer, the converse of the 
problem usually met. 
The following table shows the length of the line and the actuai 
difference between the two measurements :— 
Measurement and Re-measurement—Base of Verification, 
Richmond. 
Sochon: aaa ath Meee ee wa Pine Steel 
Feet. Feet. Feet. 
If 9696-95860 9696°98541 — ‘02681 
IBLE. 6348 75810 6348 °73853 + *01957 
Ge 5317:22788 5317-24112 == Olson 
IAW, 5317°07039 5317:06660 + :00379 
WA 5106°99705 5107:010S0 — ‘01375 
Wak 5202:32449 520234920 — ‘02471 
36989'33651 36989°39166 == Obol5 
Mean length of base-line... .., 96989-36408 feet. 
Difference between measurement 
and re-measurement... ity 0-662 inch. 
