352 
Proceedings of the lloyal Society 
was replaced by an ordinary metronome, adjusted so as to make a 
vibration every half second. This instrument, from the distinct- 
ness of its beats, was found much better adapted to our purpose. 
The time, by this means, could be registered with perfect accuracy 
to one-fourth of a second of time. 
The mean result of three observations was always taken with 
each different weight; and when there appeared to be any dis- 
crepancy, additional observations were made, in order to get a more 
exact average for the result. But except when the smaller weights 
were used this was quite unnecessary, as the results of the different 
observations agreed very well. 
The mean result with each weight was registered, forming 
Table I., where, in a line with each weight, is to be found the time 
in seconds required for the vanes to make forty-seven revolutions. 
Table I. 
Weight 
on cord 
in lbs. 
54 - 85 rr 
Square. 
Ri 
41-15 O' 
Round. 
S 2 _ 
1G6--3Q" 
Square. 
r 2 
125-8 □" 
Round. 
S 3 
345-2 □" 
Square. 
r 3 
264-8 O' 
Round. 
1 
82 
73 
145 
140 
61 
55 
110 
94 
2 
50 
46 
89 
77i 
138 
114 
3 
40 
361 
70 
59i 
106 
91 
4 
34 
31 
58 
51 
92 
76 
6 
27 
25 
47 
41J 
74 
61 
8 
23| 
211 
401 
36 
64 
53 
10 
21 
19£ 
37 
32 
56 
47^ 
1 2 
19 
17* 
34 
29 
52 
44 
14 
18 
161 
311 
27 
481 
41 
16 
16f 
151 
29ir 
25h 
45 
381 
18 
16 
14j 
28 
24“ 
43 
36^ 
20 
15 
13f 
261 
22 
41 
341 
To ascertain the absolute resistance of the surfaces, curves 
similar to those on Plate I. were laid down, where the ordinates 
represented pressure in pounds, and the abscissas velocities; the 
vanes now moving perpendicularly to their planes. In the same 
manner curves were laid down from Table II., which shows the 
weights required to give different velocities when the vanes moved 
in their own planes. The difference between the ordinates of a 
pair of curves belonging to the same vanes, the velocity being the 
