XV .
" An Account of certain Experiments , on Aneroid Barometers , made at Kew Observatory , at the expense of the Meteorological Committee .
" By B. STEWART .
Communicated by the Meteorological Committee .
Received June 3 , 1868 .
In judging of the value of an instrument , such as an aneroid , it is not the mere extent of difference between its indications and those of a standard barometer that ought to guide us ; but it is rather the constancy of its indications under the various circumstances to which it may be subjected , that determines its value .
An aneroid may differ from a standard barometer at the ordinary pressure , and to a greater extent at other pressures ; but provided these differences can be well ascertained and remain constant , such an instrument ought to be regarded as valuable , just as much as a chronometer of known constancy , but of which the rate is wrong .
The circumstances which may be supposed to affect the indications of an aneroid may be classed under three heads , namely:(1 ) Time .
( 2 ) Temperature .
( 3 ) Sudden variations of pressure .
( 1 ) Time.-Of the influence of time , I am not able to say much ; Captain Henry Toynbee has allowed me to examine the various readings of an aneroid , which he carried about with him for many years in his voyages , and constantly compared with a standard barometer .
This aneroid ( which I shall call No. 1 ) was between 4 and 5 inches in diameter , and was compensated for temperature .
In July 1860 , as compared with a standard barometer , it read 0'025 in .
too low .
In September 1862 it read ( at the same temperature ) about 0'012 in .
too low ; while in March 1864 ( still at the same temperature ) it read about 0'020 in .
too low .
This instrument , which was well cared for , and which , being used chiefly on the surface of the ocean , was subjected neither to a very great nor to a very sudden change of pressure , must be allowed to have retained its character with great constancy .
This is the only definite information regarding the effect of time on these instruments which I have received .
( 2 ) Temperature.-A good aneroid is generally compensated byits maker for the effects of temperature , and the question to be investigated is , totwhat extent such compensations are trustworthy .
I record the results ( obtained at the Kew Observatory ) of subjecting six aneroids , each 42 inches in diameter , made by two different makers , to a very considerable range of temperature .
~No .
~of Correction at No. of instrument , F men 55 ?
F. 7 8 ? .
780 .
880 F. oo00 ?
F. 2 .
-I 135 -'I35 40 --I45 I45 3 .
-'o55 -'090 -o095 -'095 --'oo 4 .
-'o95 o5 -o095 -'o8o -o6o 5 , -'xo6 -'I06 -'III --'II I -'III 6 .
-'ioi -'III i-'oI -'Io6 i-'o6 7 .
--'o6 --o6 -'06 o6 -o6 i -'06 03 These results are , on the whole , very satisfactory , and appear to show that a well-made compensated instrument has its indications comparatively little affected by a very considerable temperature change .
It ought always to be borne in mind that an aneroid is not capable of being read to the same accuracy as a standard barometer , and that the ~of an inch is a very small quantity .
These temperature experiments were made at the ordinary atmospheric pressure .
I am unable to say what effect a change of temperature would have at a diminished pressure .
( 3 ) Sudden changes ofpressure.-A preliminary investigation , made at the request of Mr. De La Rue , into the behaviour of an aneroid belonging to the Italian Government , seemed to show considerable error at low pressures .
For the purpose of investigating the influence of sudden changes of pressure upon the indications of aneroids , I then applied to some of the best known makers of these instruments , for the loan of several , and through their courtesy in lending me a sufficient number , and for a sufficiently long times I have been enabled to investigate this influence at some length .
In the following experiments , the instruments were , to begin with , suspended vertically , at the usual atmiospheric pressure .
They were tapped before being read .
The pressure was then lowered an inch , and the instrument allowed to remain ten minutes at this pressure before being read , after having again been well tapped .
The pressure was thus reduced an inch every time , being allowed to remain ten minutes at each stage ; the instrument was always well tapped before being read , by means of an arrangement contrived for this purpose by Mr. R. Beckley .
The exhaustion was carried downwards to 19 inches in the case of those instruments in which the scale was sufficiently great , and the instrument was allowed to remain an hour and a half at its lowest pressure ; the air was then admitted an inch at a time , the previous arrangement as to time and tapping being followed .
TABLE I. Correction at 30 inches .
29 inches .
28 inches .
27 inches .
26 inches .
25 inches .
Up .
Down .
Up .
Down .
Up .
Down .
Up .
I_ cCt CS c^ C ?
s I ' 0 Down .
Up .
Down .
Up .
Down .
8 .
4 incles .
June 1867 .
*0 oo -o +'07 -'I2 +'io --'1 +'14 -'IO +I'5 -'o6 +'19 9 .
41inches .
June I867 .
+'03 -*7 +'03 -'I7 +'o6 -*I7 +'Io --I6 +'O -I'8 +'II 8 .
2nd trial .
July I867 .
+-o6 -'i6 +09 -'15 +'0 -*'2 +'14 -'I2 +'19 -'07 -+22 9 .
2nd trial .
July I867 .
+'707 +'7 -'O +'Io -0 -'07 +'3 -'07 +'I7 -'07 +'15 o. 2z-inches .
' June 867 .
+'o6 -I5 +09 -'S + '12 -'12 +'17 -'08 +'19 o6 +2zz I. 2 inches .
July 1867 .
-'03 -'20 --05 -21 -'02 -'22 -'03 -'22 oo --9 +'03 12 .
2 inches .
July 1867 .
'oo -*I9 +02 -'I9 +'o6 --i8 +'05 -'I3 +'12 --Io +'14 13.2 inches .
July 1867 .
--19 -44 -i6 -'41 -'o6 -'36 0oo --42 'oo -35 *co 14 .
4 inches .
June I867 .
+ 06 -'o +'07 -.o8 +'II -03 +'I5 -'oi +'i6 --oi +'16 15 .
4 inches .
August i867 .
+'OI -11 +'03 -'I5 00 -I5 +'01 -'I4 +'03 -'I2 +'03 I6 .
I2 ?
inches .
June 1867 .
+'04 -'2 +'03 --o1 +'o6 -'O +'07 -'o6 +'1o -'02 +'13 17 .
2 inches .
June 1867 .
+'15 -'Io +20 + -'05 +'29 +'02 +'37 +'08 +'44 +'I3 +'46 24 inches .
23 inches .
22 inches .
21 inches .
20 inches .
19 inches .
8 .
4 inches .
June 1867 .
-'07 +'20 o00 +'23 +'04 +'23 +'07 +'2I +'12 +'2I +-'8 +'18 9 .
42inches .
June 867 --'I7 +'II -'I4 +'09 -'10 +-II --'4 +-II +-'o +'IO +'05 +'05 8 .
2nd trial .
July 1867 .
-`07 +'24 '04 +'27 +'oI +'25 +'03 +-26 +'09 +'27 + 'i6 +'25 9 .
znd trial .
July 1867 .
-'o8 +'15 -'o8 +'4 -'o6 +'12 -'03 +-i6 *oo +16 +'5 +'I3 IO .
2 inches .
June I867 .
-'I3 +'24 -'3 +'20 'oo +'21 +-o6 +'20 +'12 +-17 +'I5 +'I7 11 .
2zinches .
July I867 .
-'i6 +'o6 -I5 +'o8 -'15 +'I -'12 +'09 07 0 -o6 +'"I 12 .
2 inches .
July 1867 .
--2 +'I4 -'II +'15 -05 +'24 +'02 +'24 O+'8 , +'27 +'07 +'26 I3 .
2 inches .
July 1867 .
-'34 'oo -33 '00 --'29 +0'o -'26 oo00 -'8 oo00 --3 -'05 14 .
4 inches .
June 1867 .
+'oI +'I3 '02 +'04 15 .
4 inches .
August 867 .
-'05 +'08 +'oI +09 16 .
z2inches .
June 1867 .
+oi +1'2 +'04 +'io 17 .
2 inches .
June 1867 .
+'18 +'49 +28 +'48 ' TABLE II .
Correction for the Down readings of the Aneroids of Table I. , supposed right at 29 inches .
noi .
o Size .
Date of trial .
29 inches .
8 inches .
27 inches .
z6 inches .
5 in ches .
2i nches .
inhezz inches .
zI inches .
z2 inches .
19 inches .
anerooid .
I 8 .
4 inches. .
June I867 ... .
oo -'0z 00 oo +004 +03 +'Io +'14 + ' 7 +'22 +z28 9 '41 inches. .
Jne 867 ... .
oo oo oo -01o oo '03 +-07 j +'3 +'I8 +'22 8 .
znd trial..July I867 ... .
oo +oI +04 +04 +09 +09 +'I2 +17 +'I9 '+ 25 +'32 9 .
znd trial. .
July 867 ... .
' oo + +05 +'05 +-05 + '0o+ f+o4 +'o6 +'9o +-Iz +'17 Io .
2inches. .
June I867 ... .
'o 'oo +'03 +'07 +'09 +o 2 ?
+'12 +-i5 +'z2 +'27 +'3 I. '2z inches. .
July I867 ... .
'oo --* O - ' -o 02 +'oI 04 +'o 05 S +c8 +'II +'4 12 .
2 inches..July I867 ... .
'o00 oo +'OI +o6 +'og +'07 +'c8 +'14 + +2i '27 +'26 13 .
inches ... July 867 ... .
-oo +'03 +'o8 +02z 09 +Io +'0 + +'I5 +'i8 + ' 26 +'3I I4 .
41 inches. .
June I867 ... .
oo +'02 +'7 +'09 +'09 +'i +'c 15 .
4 inches..August I867. .
oo -'04 -'04 -'03 -'oI +'o6 + ' i I6 .
2 inches. .
June 1867 ... 'oo0 +'o02 + ' +0o6 +'I +'r3 +'I6 17 .
2 inches..June 1867 ... .
00 oo '05 +'12 +-'8 +'-3 +28 +'38 r-1 O~ Go oF &amp; 3'1 ^S to EL cl r5 'i !
CD Cl y"d r ; tr , CP c L " Cl Cl~3 0 S " rj ff .
U cr r= ; ?3 tG , * &lt; - " .
* ?
( ._ rO oS i2 .
oC r " d^ CC t " ' XF 0 ?
i,5 If now we separate the results of Table II .
into two sets , one comprising large ( 4 to 4 ' inch ) aneroids and the other small instruments , we shall find the mean down correction for large aneroids to be as follows .
29in .
28in .
27in .
26 in .
25in .
24in .
23in .
22in .
z21in .
20in .
19 in .
Mean correction for large aneroids , graduated to i9 in ... 'oo 'oo +02 +'03 +'04 +'04 +'07 +'11 +'14 +'19 +'25 23in ... . .
-oo -o02 +'o0 03 1+'04 +'o8 -+'o In like manner we shall find for small aneroids , supposed right at 29 inches , the following mean correction:29 in .
28 in .
27in .
6 in .
258 24iin n. 23 in .
2 .
znin .
2o in .
19in .
Mean correction for smallaneroids , graduatedtoxgin ... 0oo +'oI +'02 +'03 +'07 +'o7 +'09 +'I2 !
+'I7 +'23 +'25 23in ... 0 ... oo +'03 +'07 +'I2 +'I6 +'20 + ?
27 It will be seen that there are two instances in which the same instrument has been twice experimented on .
Assuming that the mean of the two experiments represents the true correction for each of these instruments , we find:9 in .
28 in .
z7 in .
z6 i. 25 in .
411 in. .
22in .
19in. .
No. 8 .
Mean correction , deduced from two experiments ... .
'oo --oI -'2 +oz +'o6 +o6+'I+'I5 +8 +'23 4-'30 Mean minus first determination m. .
oo +'oi +'02 +oz '02 +'+03 -'O I +'o +c +i0 i'o0 2 In like manner:2o in .
28 in .
27 in .
26]in .
2z in .
z4.in .
2i l. zz in 2 .
21in .
2oin.l lin. _ _-_ _i _ii_o .
9 .
Mean of two experiments ... oo +-OI +o +03 +oz oz +'03 +'0 6 +'II +'I5 +'20 Mean minus first determination. .
*00 -+o +'0 z i'02 -+'-203 -- ' oz We see from these results that if aneroids , right to begin with , be subjected to a decrease of pressure similar to that to which they were subjected in these experiments(1 ) That a vell-constructed large aneroid will not go far wrong down to 24 inches , but after that pressure its reading will be considerably lower than that of a standard barometer , so that a large positive correction will have to be applied .
( 2 ) That small aneroids are less trustworthy than large ones , and probably cannot be trusted below 26 inches .
( 3 ) That if previous experiments are made upon an aneroid , we are enabled by this means to obtain a table of corrections which , when applied to future observations with the same instrument , will most probably present us with a much better result than had we not verified our instrument at all , and that by this means we may use our instrument down to 19 inches with very good results .
Let us now consider the up readings of these instruments , and let us suppose that each instrument is right to begin with , that is to say , while remaining an hour and a half at its lowest reading .
These corrections and up readings are exhibited in the following Table : TABLEr III .
f Correction for the up readings of the aneroids of Table I. supposed right ~N~O~ .
of with standard at lowest reading .
alleroid .
19 .
20 .
21 .
22 24 .
2 5 .
z6 .
27 .
28 .
29 .
30 .
8 .
*oo +'o3 +'3 j+'05 -+'o 5 -+'oz +'o -'03 --'04 -'08 --11 -*18 9 .
'oo +'05 +-o6-'o6 +04o 6 +'04 +'06 o6 +05 +5 +-oI -o02 --02 8 .
'oo +'oz +'oI '00 +o02 --oi -'03 -'o6 -'i --'I5 -I6 --I9 9 .
oo +'03 +'03 --'oI +'o 0 -+'02 +'02 +'04 oo -'03 --'o6 -'o6 o. *oo 'oo +'03 +'04 +'03 +'07 +'05 +'02 '00 -'05 -'08 --'1 II .
'oo --'04 --0 00oo -'03 -'05 -'o8 --'1 -'14 -'13 --'6 -'14 12 .
'oo -'o 0 -2 02 --'II -'12 -'12 -'I4 -'21 -'20 -'24 -'26 13 .
'oo +'05 +'5 +'o6-k'o5 +'05 +'05 +'05 +'05 -'oI --II --14 14 ... ... ... ... '00 +'09 +2 1 +12 +'11 +'07 +'03 +'02 15 ... ... ... 'oo -'OI -'o6 --o6 -*o8 --og -'o6 -'o8 i6 ... ... ... . .
'oo +'0 '3 0oo -'03 -'04 -'07 -'o6 17 ... ... ... .
In like manner for small aneroids we have the following result:g9 in .
z z. 2 .
i n. 52in zin2 i. 22 in .
25in .
in .
28in.29 in.30in .
Rightat 9 in ... oo 'oo +-'o +'02 -'oI -'oI -'02 -'04 -'07 --'io -'I5 -'I6 Right at 23 in ... ... ..oo + ' '00 -'02 -'07 -'II --'2 -'19 As before , there are two instances in which the same instrument was twice tried ; assuming the mean of the two trials to represent the truth , we find9Igi 2 .
i 2 .
3i .
4 25i 6 .
in .
in .
inin .
2in .
24in.in .
26in .
in .
in .
30in .
No. 8 .
Mean correction. .
'oo +'02 +0oz +'02 +'03 '00 --'o -'04 -'07 --1 -'I--I8 Mean minus first determination. .
'oo -'o0 / -- ' -'-o3 -'02 -'o0--2 --'OI -'03 -'03 -'o 0oo In like manner9 in .
20 in .
21 in .
22 in .
2i .
236 iinn in .
i27 in .
28 in .
29 in .
30 in .
No. 9 .
Mean correction. .
'oo -'04 +'0o4 + '03 +'03 04 +'04 + ' 04 + '0 -0oI --04 -- ' 04 Mean zinus first determination. .
'oo -'0o -'02 -'03 -'o --'02 -o02 --0 -'03 -'02 -'02 -'02 We may learn from these results that if aneroids which have been subjected for at least one hour and a half to the lowest pressures which they register , have the pressure increased by means of the gradual introduction of air into the receiver , after the manner already described .
( 1 ) That a well-constructed large aneroid will not go far wrong for about 8 inches above the lowest pressure .
( 2 ) That in this respect small aneroids are somewhat less trustworthy than large ones .
( 3 ) That if the instrument read be previously tested and its corrections ascertained , we may consider it trustworthy ( making use of these corrections ) for up readings throughout a greater range than if it had not been so tested .
I come now to consider whether a rapid change of pressure affects an aneroid after the experiment has been completed .
The following Table will exibit the results obtained in this direction .
TABLE IV .
ist Ex. ist Ex. 2ndEx .
2ndEx .
8 .
9 .
8 .
9 .
Io .
II .
I2 .
I3 .
I4 .
I6 .
Correction before experiment ... .
--io -'12 -'13 -'09 -'I2 -*II -'13 -'47 -'04 -'05 Immediately after experiment ... .
'oo +'03 +'06 +'07 +-o6 -'03 0oo -I9 +*o6 +'04 I8 hours after experiment ... ... -'07 -'03 +'04 'o02. .
-'Io -'07 -'34 +'OI +'oi 48 hours after experiment ... ... -8 ... ... -'03 ... .
-'37 3 days after experiment ... ... -'o8 -'05 ... ... ... --'oi --o* 3 weeks after experiment ... ... -'3 -'o ... .
--II ... ... -'07 --o6 It thus appears that if an instrument reads correctly before it is put into the receiver it will read too low immediately afterwards , and that it may be some considerable time before it recovers its previous reading .
The instrument cannot , therefore , be safely trusted for absolute determinations if it has been recently exposed to rapid changes of pressure .
The experiments hitherto recorded , in which an inch of pressure has been taken away or added every ten minutes , are perhaps analogous to ascents in a balloon , or descents from a mountain ; they are not , however , precisely analogous to mountain ascents , since a longer time than 10 minutes is usually taken to produce a change of pressure equal to 1 inch .
At the suggestion of Mr. Charles Brooke , a couple of aneroids were tested in April 1868 , with the view of rendering the experiment more analogous to a mountain ascent .
The pressure was reduced by half an inch at a time and at intervals of 30 minutes , the aneroids being well tapped .
The following corrections were obtained for down readings ( instruments supposed right at 30 inches ) .
TABLE V. At No. 8 .
No. 9 .
At No. 8 .
No. 9 .
inches .
inches .
30 'co 'oo 23'5 +'o8 -'02 29'5 'oo -03 23 +II -'03 29 '00 -'04 22-5 +'12 -'01 28-5 'oo -'03 22 +-'4 'oo 28 'oo00 -'03 21'5 +-'6 +'02 27'5 'CO -'02 21 +'I 7 -'04 27 ... ... .
20'5 -'20 +-o6 26'5 '00 -'02 20 +-'22 +'07 26 +'oI --'02 9'5 +'25 +o09 25'5 +'04 -'02 I9 +'27 +'1I 25 +'05 -'04 24'5 +'o6 -o02 24 +'05 'o General Sabine on Terrestrial Magnetism .
These results , when compared with the previous determinations for these same instruments , would seem to show that a somewhat better result is obtained whei the exhaustion is carried on more slowly , and hence that the corrections depend , to a considerable extent , on the nature of the treatment received .
No. 8 seems to be more constant under different treatment than No. 9 .
From all these experiments we may perhaps conclude as follows : ( 1 ) A good aneroid of large size may be corrected for temperature by an optician , so that the residual correction shall be very small .
( 2 t ) If an aneroid correct , to commence with , be used for a balloon or mountain ascent , it will be tolerably correct for a decrease of about 6 inches of pressure .
( 2 / 3 ) A large aneroid is more likely to be correct than a small one .
( 2 y ) The range of correctness of an instrument used for mountain ascents may be increased by a previous verification , a table of corrections being thus obtained .
( 3 a ) If an aneroid have remained some time at the top of a mountain , and be supposed correct to start with , then it will give good results for about 8 inches of increase of pressure .
( 3 3 ) A large aneroid is more likely to be correct than a small one .
( 3 y ) If the aneroid has been previously verified , it is likely to give a better result .
( 4 ) After being subjected to sudden changes of pressure the zero of an aneroid gradually changes , so that under such circumstances it ought only to be used as a differential and not as an absolute instrument , that is to say , used to determine the distance ascended , making it correct to begin with , or to ascertain the distance descended , making it correct to begin with , it being understood that the instrument ought to be quiescent for some time before the change of pressure is made .
Before concluding I ought to mention that most of the experiments herein described were undertaken and executed in a very careful manner by Mr. T. W. Baker .

