INSTRUMENTS USED IN ENGINEERING LABORATORIES. 544 
Test sheet forms were drawn up by the Committee, to be issued to 
different observers with the sets of test bars, on which the observations 
were to be recorded in a uniform manner. A sample of the form of ob- 
servation sheet prepared by the Committee is printed on the opposite page- 
In January 1895 two sets of test bars were sent out, to be circulated 
amongst those observers who had consented to make measurements for 
the Committee. A very large number of measurements have been made 
with great care, and the record forms have been returned to the 
Committee. 
It is proposed in the present report to give an analysis of these 
results. The following short statement gives particulars, so far as they 
are stated on the forms received, of the extensometers used. 
Extensometers used in Measurements made for the British Association 
Calibration Committee. 
Instrument used 
| 
H Observer 
Kennedy’s Extensometer. Lever or mechanical multiply- 
ing arrangement. 
Ewing’s Extensometer. Measurements taken by micro- 
scope with micrometer in eyepiece. 
Ewing’s Extensometer. : 
Goodman’s Extensometer. A mechanical multiplying in- 
strument of Kennedy type. 
Unwin’s Mirror Extensometer. The extension of the bar 
deflects a mirror. A scale is placed at a distance, and 
by a fixed telescope the reading of the scale reflected in 
the mirror is taken. Also a lever extensometer with 
triple compound levers. 
Mechanical multiplying instrument of Kennedy type. 
Bronze mechanical extensometer, Kennedy pattern. 
| A. B. W. Kennedy . 
J.A.Ewing . ‘ 
H.S.Hele Shaw . 
J.Goodman . : 
W.C. Unwin . . 
T. Hudson Beare . 
D.S. Capper . ‘ 
_1. Variations in Measurement of the Area of the Test Bars by different 
Observers. 
The following table gives the measurements of the test bars by 
different observers. They are satisfactory as showing that the error of 
measurement of area of a test bar, by different observers, seldom exceeds. 
about one-fifth of 1 per cent. That is the error reckoned from the mean 
value of the area given by several observers. The difference of measure- 
ment for any two observers may amount to 0:5 of one per cent. 
Taste I.—Comparison of Measuremenis of Test Bars. 
ry F ¢ | Distin- Observed | Mean value of pret ead Deviaticn in 
Meee? guishing} Observer area in | observed atea) po, nen per cent. 
Letter sq. in. in sq- in. | valuein sq. in. of mean value 
Flat A W.C.U. . | 1:0175 — ‘0002 — "02 
A. B.W. K. | 1:0170 — 0007 — 07 
T.H. B. . {| 1:0190 10177 +0013 +13 
J.G.. - | 1:0170 —°0007 — 07 
| | D.S.C. . | 1:0180 +°0003 +08 
