812 
DR. T. R. ROBINSON ON THE DETERMINATION OF 
The end of each series of a set is marked by a single line, and that of each of the 
two sets by a double one. It must be remembered that each set begins with a low 
value of v which increases in each term ; in each also there is a different brake friction. 
The values of AV' are not very close ; but this was to be expected from the uncertainty 
of W and F ; and as the aberrant ones are not regularly distributed they are evidently 
due to casual errors. Thus in the first series of the first set, six have large positive 
values ; but this is not the case in the corresponding terms of the other series, in 
which they are small and mostly negative, except the last, in which are four large 
negative and two positive. 
In the second set the three first series are of the same character as the previous 
one ; but there was evidently some disturbing influence at work in the last four. I 
have already mentioned what I believe this to have been. 
The brake friction of the fourth occurs in the last but one of the first set, and the 
last of No. III., without producing such anomaly, so it is not due to error in the 
measure of the friction. The results for No. III. are given along with those obtained 
by supposing y— 0 hi 
Table XVI. 
No. 
AV'. 
Do. y = 0. 
No. 
AV'. 
Do. 7 = 0. 
159 
-0-43 
-0-43 
181 
-0-64 
-0-68 
160 
— 0T7 
— 0T1 
161 
-0-27 
— 0T8 
182 
+ 2-68 
+ 2-68 
162 
+ 0-08 
+ 0-20 
183 
+ 1-49 
+ 1-20 
163 
-0-08 
+ 0-09 
184 
-0-09 
-o-oo 
164 
+ 0-33 
. . 
185 
-0-32 
-0-20 
165 
-0-06 
+ 0-28 
186 
-0-61 
-0-45 
166 
— 0T1 
+ 0T4 
187 
-1-08 
-0-69 
167 
+ 0-27 
+ 0-58 
188 
+ 0-67 
+ 0-90 
168 
-0-20 
-0-29 
189 
-0-33 
-0-07 
169 
+ 0-23 
-0-25 
190 
+ 0-33 
4~ 0*63 
170 
-0-25 
-0-30 
191 
-0-27 
-0-02 
171 
-0-25 
-0-22 
192 
-0'27 
+ 0-06 
172 
-0-75 
+ 0-20 
173 
-0-72 
-0-56 
193 
+ 078 
+ 070 
174 
-0-59 
-0-03 
194 
-0-35 
-0-25 
195 
— 0-40 
-0-31 
175 
-0-31 
—0-30 
196 
+ 0-39 
-0-36 
1 76 
a-qq 
177 
+ 1-30 
-0-99 
197 
-0'49 
-0-45 
178 
-0-77 
-0-66 
198 
-0-49 
-0-41 
179 
-1-03 
-0-63 
199 
— 0'46 
-0-35 
180 
— IT 7 
-0-98 
Here also V' is sufficiently well represented ; but it is to be noted that y= 0 gives 
rather better results than the other. The probable error = fi^OYSI, the maximum 
= + 2‘68, the minimum = — 0*98. For the other the probable error the 
maximum also — fi-2’68, the minimum = — Iff 7. 1 did not think it necessary to 
