THE CONSTANTS OF THE CUP ANEMOMETER. 
1057 
—1*914 ; fourteen are =0, and nine are negative. But if we divide them into four 
consecutive and nearly isochronal groups the discordance is much less. 
s. 
'ime =161 "4 
v=6'219 
+=5*203 
v— +=1*016 
159-8 
5-508 
4-375 
1-133 
160-6 
6-241 
5-926 
0-3L5 
164-9 
5-369 
5-067 
0-302 
The extreme range here is 0"831 instead of 5'002, grouping them in pairs 
T=321‘2 -y=5 , 864 +=4791 v—v'=V07B 
325-5 5-799 5*489 0-310 
There can be little doubt that the total means are nearly correct, and these values of 
v—v differ from the mean one by +0-475 and —0”289. In general, v — v' will be less 
than this ; and if it be observed by inspecting the chronograph while an observation 
is proceeding that the ratio of A to A' varies notably, a longer time should be taken. 
(60.) As to the second point it is easily shown that no great errors can arise from 
assuming that V is truly given by the mean v. The mean V of a series is, taking the 
time into account, 
_SVT_ xSrT SyTx \/ z +t*\ 
~ ST ST~ ST 
Now the first of these =xX mean v. In instruments like K where </> is small, if we 
develop the radical in powers of + the second term becomes 
Si>T x 
and as the (/> term may be neglected the mean of radical becomes v / s i X mean velocity. 
When cf> is large the simplest course is (calling the radical B) to compute \ or 
what conies to the same thing C being the time-space, and compare this with 
the B computed with the mean v. Taking at random No. X. of Table XXI. whose 
(f)=: 343"28, we have for the separate groups whose A and A' are nearly uniform 
No. 
c. 
I. 
31-8 
6-081 
II. 
95-2 
7-411 
III. 
88-3 
7-350 
IV. 
49-35 
4-886 
V. 
32-5 
7-364 
VI. 
18‘68 
6-042 
315-83 
6-835 
sum rCE 
sum C 
20-633. 
R for mean ff=20 - 689. 
Bequiring the correction = 0‘056. 
Here the +’s do not range very widely, and 1 take a more aberrant set observed 
September 16, 1878, +=173*80. 
6 u 
MDCCCLXXX. 
