180 
ers, were used with a battery of ten Daniell 
cells in good condition. ‘These comparisons 
proved that the indications of the potential gal- 
vanometer could be relied upon as trustworthy 
within practical limits, and in the actual tests 
it alone was used on account of the greater 
convenience and rapidity with which observa- 
tions could be made. Further tests of its 
accuracy were made, however, which will be re- 
ferred to later. It is sufficient to say, that this 
instrument, in form and construction, is quite 
similar to the current galvanometer already 
described, except that the coil has a resistance 
of nearly seven thousand ohms. A key is 
placed in the circuit, so that the current passes 
through the coil only during the few moments 
necessary to secure a reading, thus preventing 
the heating of the coil. ‘The difference of po- 
tential in volts, between the two points to which 
the leading wires are connected, is found by the 
same process as is used for reducing the read- 
ings of the current galvanometer to amperes. 
In measuring the efficiency of the dynamos, 
wires were brought from their binding-posts to 
the galvanometer. In the arc-light machines 
the electromotive force was high, amounting to 
more than twelve hundred volts in the ‘Thom- 
son-Houston dynamo; and it was therefore 
desirable to introduce extra resistance in the 
galvanometer circuit. From resistance-boxes 
made by Elliot Brothers, an amount equal to 
seventeen times the resistance of the galva- 
nometer was thrown in, thus bringing the fall 
in potential in the galvanometer within easy 
range. Great care was taken to see that the 
coils were not heated during these measure- 
ments; and for this purpose the boxes were 
opened, and the coils exposed to the air, fre- 
quent examination being made to see that no 
rise in temperature took place. Precisely the 
same arrangement existed throughout the tests 
of both are systems. During the photometric 
tests the wires of the potential galvanometer 
were attached directly to the lamp under test, 
so that the fall in potential through the lamp 
only was measured. 
Photometric measurements. 
Unquestionably, the most difficult question 
to deal with, in work of this kind, is the ques- 
tion of photometry. The expression of illumi- 
nating-power in ‘ candles’ is a matter of great 
uncertainty, arising from the uncertain charac- 
ter of the standard, and also from the great 
inequality existing in the intensity and com- 
position of the lights which are brought into 
comparison. As the test was intended to be 
SCIENCE. 
[Vou. II., No. 54, 
purely competitive, the jury decided to ignore 
the question of ‘ candle-power’ entirely, and 
confine itself to a comparison of the lights 
under consideration. It is believed that the — 
adoption of this plan rendered the results 
free from many errors to which they would 
otherwise have been liable. 
The photometric comparisons were made by 
means of the ordinary Bunsen disk photometer, 
as modified by Letheby. Some preliminary ex- 
periments were made with one of Glan’s spec- 
trum photometers, for the use of which the jury 
was again indebted to the kindness of Professor 
Brackett. The adjustments of this instrument 
are delicate, and observations cannot be made 
so rapidly with it as with the ordinary disk 
photometer; so that, in consideration of the 
limited time at the disposal of the jury, it was 
decided not to attempt its general use through- 
out the tests. It was hoped and intended, in 
the beginning, to make a thorough examina- 
tion of the composition of the different lights ; 
but unforeseen delays in the preparation of 
other portions of the machinery of the test for- 
bade this. As the candle was not made use 
of, all the lights which were compared were 
more nearly of the same composition, and thus 
much of the difficulty in the use of the disk 
photometer did not appear. 
It was found most convenient to make the 
comparison of the arc-lights through one of the 
incandescent lamps, as the steadiness and con- 
stancy of these could be depended upon during 
the time necessary for a comparison. In these 
measurements, a long gallery in the basement 
of the main building, and adjoining the testing- 
room, made it possible to place the two lights 
which were being compared at a distance of fifty 
feet from each other. The line extended into 
the testing-room, where the photometer-bar, 
ten feet in length, was placed. An Edison in- 
candescent lamp, nominally of sixteen candle- 
power, was used as a standard. In the first — 
series of experiments, comparisons were made — 
with the arc-lamps in three different positions ; 
five readings of the position of the photometer- 
box and of the galvanometers being made at 
each position. The lamp was first suspended 
in its normal, vertical position ; then afterwards 
it was inclined at an angle of forty-five degrees, 
first with its base away from the photometer- — 
box, and afterwards with its base towards the 
same. After such a series had been completed — 
with one of the two lamps in competition, it was 
at once removed, and its place was supplied by — 
the other. On the following night the com- 
parison was continued, other lamps having been 
selected ; but the lamps were tested in only two 
. 
-». =| 
