THE CONSTANTS OF THE CUP ANEMOMETER. 
815 
inferior to the preceding. I have little doubt that with the true frictions they would 
give correct results. Aberrant as both are, I give the results in — 
Table XYIII. 
No. 
AV' 
Do. I. and III. 
No. 
AV'. 
Do. I. and III. 
200 
+ 0-22 
-0-02 
217 
+ 0-92 
+ 0-47 
201 
+ 2-46 
+ 1T5 
218 
+ 0-95 
+ 0-55 
202 
+ 2-38 
+ 1-72 
— 
- 
203 
+ 324 
+ 2-46 
219 
-2-83 
-2-99 
204 
+ 3-20 
+ 2-28 
220 
-2-21 
-2-66 
205 
+ 3-29 
+ 2-29 
221 
-1-96 
-2-58 
206 
+ 3-56 
+ 2-49 
222 
-1-37 
-2-00 
207 
+ 4-01 
+ 2-76 
223 
-1-38 
-2-35 
208 
+ 4T6 
+ 2-67 
224 
-0-44 
-1-64 
209 
+ 4-40 
+ 3 ’04 
- 
210 
+ 4-50 
+ 2-98 
225 
-2-57 
-3-08 
— 
226 
-2-47 
— 2T8 
211 
+ 0-47 
-0-23 
227 
-1-01 
-2-80 
212 
-0-70 
-1-39 
228 
-2-08 
-2-95 
213 
+ 0-09 
-0-80 
214 
+ 0T2 
-0-79 
229 
-3-02 
-3-17 
215 
+ 070 
-0-51 
230 
-2-81 
-2-96 
216 
+ 1-00 
-0-27 
231 
-2-42 
-3-01 
If, as I suppose, the viscid oil increased f and f" and lessened the brake frictions, 
there must be a certain part of the set where these opposite effects will balance each 
other ; and 'this seems to occur in the second series. The increase of F which would 
nullify AY' for 210 in the second system =139; that for 231 = — 225; both quite 
possible. 
(44.) No. V. cylinder cups is in marked contrast to No. IY. Its plotting (Plate 70) 
is more regular than any of the others. Minimum squares give a=9T52; /3 = 13*416; 
y=— 3 ‘481. The a is a little less than what I got by two doubtful measures, so I 
retain it, and have x— I '4659 ; z— 1 '7 69. The results of these are given in — 
Table XIX. 
No. 
AV'. 
No. 
AV'. 
No. 
AV'. 
No. 
AV'. 
232 
-0-81 
240 
-0-24 
247 
-0-49 
254 
-0-57 
233 
+ 0-39 
241 
-0-27 
248 
-0-56 
234 
— 0T9 
242 
-0-60 
249 
-0-51 
235 
-0-03 
243 
-0-42 
250 
-0-33 
255 
— 0-63 
236 
+ 0-12 
244 
-0-62 
256 
-1-06 
237 
+ 0-26 
245 
-0-59 
257 
-1-03 
238 
+ 0T8 
251 
+ 0-70 
239 
+ 0-14 
252 
— 0-36 
246 
-1-57 
253 
-1*1.6 
These might be considerably improved by a small increase of a ; for instance, il I 
had taken it =0 - 9 X 10'4, the error for 257 would have become — 0 '00. 
MDCCCLXXVIII. 5 M 
