On the Expansion by Heat of Metals and Alloys .
By A. MATTHIESSEN , P.B.S. Received June 20 , -Read June 21 , 1866 .
IN a memoir " On the Expansion by Heat of Water and Mercury"* , I described a method of determining the expansion of bodies by weighing them-in water at different temperatures .
This method was chosen on account Qf its yielding accurate results with comparatively small quantities , for to purify large quantities of metal wouldentail immense labour and expense ; so much so , in fact,.that to purify sufficient quantities to .
make bars for the determination of the linear expansion would be practically impossible .
On determining the expansion of the metals by this n method , I fo und that they did not expand regularly between 0 ?
and 100 ? .
The difference , howev r , between therates of expansion between 0 ?
and 500 and 50 ?
and 1000 was found to be so great that part of it might be possibly attributed to errors in the determination of the coefficients of expansion of water , as the 'coefficients of expansion of the metals are comparatively small when compared with those of water .
Now , although the coefficient of expansion of mercury when determined by this method agrees with REGNAULT'S value , yet on account of this value being large , small errors in the water coefficients will not materially influence it ; for the volume of water at 40=1L0000 , at 100 ?
=1-04316 , that of mercury at 0=_1 0000 , at 100 ?
=1*0I815 , and that of copper , for instance , at 0 ?
=1 00000 , , at -1Q4 ?
=10049.98 , showing that the expansion of copper is very small compared with that of water or mercury .
This fact led me to make another series of check experiments by determining the linear expansion of a certain copper bar , as described in the first part of my paper above quoted , and then weighing a piece of it , turned to the shape of a double cone , in water at different temperatures .
The following are the results obtained with ' the copper bar ; and it may here be mentioned that copper does not behave in one respect like glass .
The glass rods , as there shown , do not return directly to their original length after being heated to 1000 and cooled rapidly ; copper , however , does so ; for no differences in the coefficients were observed after heatingthe rod to 100,.determining its expansion , allowing it to stand over night , and redetermining the coefficients .
In Table I. , T , T1 , TI , T3 , T4 indicate the temperatures in the order in which the observations were made ; a , b , c the increment in length in millims. of the rod between T , and T2 , T , and T3 , T and T , respectively ; the values in the three last columns are the coefficients of expansion of the rod between the observed temperatures .
The length of the rod was 1804 millims. and its diameter about 15 millims. TABLE I. T1 .
TA T3V T4 b. c T2-Tj T3-1 ' T3 T4 7.4 99.9 7.~~~~~~ 2~~807 2~~~503 0'03034 -00301,7 7-4 ... ... .
99,09 7:0 ... ... 2. .
28i7 28 3 ... ... ... .0003 0Q1 70 50 1 100 1 10A4 1E263 2 808 2 718 0 02590 0a03016 0-003030 8*4 51-2 100 370 1*252 2507 2-830 0*02925 0*03054 0 O3033 7.5 5Z19 003 78 1P295 2 805 2*803 0'02917 0*03023 0 03030 8s6 50 9 100 6 7'2 1238 23775 2*812 0*02927 0'03016 0*03011 Mean. .
002915 0-03026 Or the mean coefficient between 00and 50 ?
maybe taken=-0'029156 , and that between 00 and 100 ?
0 03(026 .
And taking the length of the rod at 0 ?
-1804 , it will be at 500=1805 4575 , at 1000=1808.0260 .
From these values the linear expansion of the copper rod can be expressed by the formula Lt-1804(1 +0 00001555t+0 0000000122t2 ) , or for any length of this sort of copper the formula for the correction of the linear expansion for temperature will be L , =LO(l +0 000001555t+0'00000-00122t2)I and that for the correction of the cubical expansion Vt=V0(1 + 000004665t + 0-0000000366t2 ) .
Two series of weighings in water were .
made with the piece cut from the end of the copper rod ; the results are given in Table II .
TABLE I. No. 1 .
Loss of weight in water =W .
( 1+at ) .
1 , d14 3'9057-40 39 5828X 50'7 3'91825 3 &amp; 96640 97*1 3 &amp; 81:940 3-97590 No. 2 .
1 046 3-s957 " '395850 54.2 3-912240 -39*6 97 3 39 1865 3 97570 The copper was slightly gilded to prevent the action of the water on it ; the water being reboiled before each weighing to drive out any absorbed air .
The values given in the third column express the volume of the copper in cubic centimetres , and are deduced from the observed loss of weight in water W , as described ( p. 245 ) in thepaper already quoted .
Calculatingformulae toexpress the volumes of copper at different tempexatures , we find for No. 1 .
Yt=3 95680(1 + 0O0004645t+O 0OOO33 6t2 ) , or if V=1 , then V10 =1'004981 ; No. 2 .
Yt= 3 95655(1+0 00004681t+00000000300t2 ) , or if VO=1 , then V,1= 1'004981 , formulae agreeing closely with that deduced from the determination of the linear expansion of the copper rod , namely , Vt=V0(1 + 00004665t+00000000366t2 ) , or if V0X1 , then Vr1 _005031 .
This memoir may be divided into two parts : I. On the Expansion by Heat of the Metals .
: II .
On the Expansion by Heat of Alloys .
I. On the Expansion by Heat of the Metals .
The metals employed for these experiments were purified in the manner described in a former paper * , and cast in a. well-smoked mould , which gave the casting the shape of a double wedge , as shown in fig. 7 t.# Owing to the action of the water on some of the metals and alloys , the castings had in some cases to be varnished or gilded , the latter method being more generally used .
To prove that the gilding or varnishing had no influence on the results , some of those metals on which water has no action were varnished or gilded in the one series and not in the other ( Series Nos. 3 and 4 , 14 and 15 , 11 and 12 ) .
The disposition of the apparatus and the method of observation was the same as described in the paper " On the Expansion of Water and Mercury .
" Sometimes observations were made commencing at the highest temperature , and cooling down without boiling out the water between the determinations made at the different temperatures .
( The series made in this manner will be headed once boiled .
) At others the observations were first taken at the lowest temperature , the water reboiled , then those at the highest temperature , and afterwards on cooling those at the intermediate one ( twice boiled ) .
And again , at others the water was reboiled between each .
set of observations ( thrice boiled ) .
It may be as well to point out some of the causes of failure in this method of deter* P]hilosophicad Transactions , 1860 , ,p .
177 .
t Philosophical Transaction.s , 1866 , :Plate XX .
mining the coefficients of expansion ; for although it appears a very simple one , yet it requires , I may say , very great care to ensure reliable results .
I. Very often a series was rendered valueless by small hairs or particles of dust falling into the water and attaching themselves to the fine platinum wire , as 0-1 or 02 milligramme difference in the weight makes a considerable difference in the expansion of those metals or alloys which have a low coefficient of expansion ; this source of error , was carefully guarded against ; in fact if on taking the cylinder out to reboil the water any particle of dust floated near the wires , the observations were considered worthless , and fresh ones taken after reboiling the water .
II .
When , for instance , the observations were being made at the highest temperature , if by chance the temperature sunk three or four degrees and gradually rose again , the air absorbed at the lower temperature would be expelled , and an air-bubble would sometimes attach itself to the metal and cause a false weighing .
II .
If the casting were not perfect , or were crystalline , and the air could not be completely expelled from the small cavities by boiling , results were obtained which did not agree together , owing to its expansion at the high and partial absorption by the , water at the lower temperature .
For this reason two series were always made with each metal and alloy , and where possible they were recast .
The values obtained when weighing the purified metals in water were as follow : TABLE III.-Cadmium .
Owing to this metal becoming crystalline at about 800 , four observations were made between 00 and 100 ? .
It will be seen that this change of molecular condition has no effect on this physical property .
No. 3.-Cadmium four times boiled .
Loss of weight T. in water=W .
W(1+at ) .
Calculated .
Difference .
8'0 6-29175 6*29250 6'29241 +0'00009 45*4 6*25080 6-31280 6-31309 -000029 77-4 6*16450 6*33255 6'33249 +0O00015 95*2 6*10230 6-34380 6-34377 +0-00003 Vt=6-28826(1 + 10-407991t+ 1 O- .
x0 163t2)* , or if Y0= 1 , then V,10= 100961 1 .
No. 4.-Cadmium varnished four times boiled .
8*1 5691135 5'91210 5 91182 +00002O8 47*9 5'86800 5-93265 5.93274 000009 75'2 5 79895 5'94885 5-94885 O00OOOO 93-8 5*73915 5-96040 596065 -0O00025 Vt=5.90793(l + 10-40'7991t+ 10-6O X 163t ) , or if Vo-1 , then Vlo=1O009611 .
( continued .
) No. 5.-Zinc varnished , thrice boiled .
Loss of weight T. in water=W .
W(l+at ) .
10O 7*4N245 7-42445 50-6 7-35985 7 45000 951 7 19460 7*48200 Vt=7-41863(1+10-4x 07719t+ 10-6 x 0.126te ) , or if V0=1 , then V10=1'008979 .
No. 6.-Zinc varnished , thrice boiled .
90 66975 6-69830 50o6 6*64010 6'72285 95.0 6-49345 6-74945 Vt=6 69304(1 + 10-4 x 018726t+ 1056 x0 0153t ) , or if V0=1 , then 0 V= 1008879 .
No. 7.-Lead varnished , once boiled .
14'0 4'55540 4-55880 5142 451635 4-57295 94'1 4-41840 4-58960 Vt= 4.55355(1 + 10x 0'8215t+ 106 x 040210t ' ) , or if V0=1 , then V10=-1008425 .
No. 8.-Lead varnished , once boiled .
1M4 4'40320 4*40695 529 4036270 4'42090 90'7 4-27960 4'43500 Vt'=440166(1 + x10-x08140t+ 10 x 0-0234t2 ) or if V0=J , then V10= 1'008374 .
No. 9.-Tin , once boiled .
10'2 366860 3-6696a 55.0 962850 368060 93'2 3'55520 369070 Vt=3-66730(1 + I04 x 0-6237t+ 10-6 x0*0656t2 ) , or if V0=1 , then V =1006893 .
No. 10.-Tin , once boiled .
8'2 4028625 4.28680 48'5 4'25005 4029800 94'8 4'14945 4'31250 Vt 4-28468(1 +1 0-4X 05964t+ 106 X 00922t ' ) , or if Vo 1 , then V1@= 1 006886No .
11.-Silver gilded , thrice boiled .
10'7 533570 ; 5'33750 52.7 5'28040 5$35035 94-3 516 280 5-36370 V-=1533433(1+ 10x0 5523t 4106x 0'0335t2 ) , or if VJ.=1 , then V1.=,1005858 .
( continued .
) No. 12.-Silver , thrice boiled .
Loss of weight T. , in water=W .
W(1+ae ) .
7*5 5549925 5.49975 50 &amp; 5 5 44655 5-51300 97.1 5-37080 5-52845 Vt ' -549754( ?
104X 0*5330t+ 106 X 0.0475t ' ) , or if V0=1 , then V1oO= 1-005805 .
No. 13.-Copper gilded , twice boiled .
10X2 5-82730 5 82900 59.8 5-74595 5*84280 95.1 5 63155 .
5-85400 V~t=5 82644(1 + 10-4 XO4223t+ 10-6 x 0.0791te ) , or if V0= 1 , then Vi0= 1P005014 .
No. 14.-Gold varnished , thrice boiled .
9.1 2-44605 2-44655 45.4 2-42625 2 45030 95 1 2-36250 2-45585 V-t=244565(1 + ?
0-4 x 0-4008t+ 10-6 x0 0397ti ) , or if V0=1 , then Vlw= 1 004405 .
No. 15.-Gold , once boiled .
10-7 2-44415 2-44490 47-8 2-42225 2-44880 93-0 2-36405 2-45380 Vt= 2.44381(1 + 10-4 x 04142t+ IO-( x 0'0276t2 ) , or if V0= 1 , then VW= 1 004418 .
No. 16.-Bismuth varnished , once boiled .
89 4-29915 4-29995 51 2 4*25350 4-30680 96.0 4-14795 4*31455 V 4-29857(1 + 10-4x 0 3591t+ 10-6 x 0.02941 ' ) , or if V0=1 , then V100= 1'003885 .
No. 17.-Bismuth varnished , twice boiled .
i i16 5 16400 s*16625 50.0 5.11255 5*17375 93.5 4-99200 5-18335 .
Vt=516416(1 + 10-4 X 003413t+106 X0 0599t2 ) , or if V0= 1 , then V , , , = 1 004012 .
No. 18.-Palladium ( purified metal , lent by Messrs. JOHNSON and MATTHEY ) , thrice boiled .
7.5 7 48500 7-48700 55.5 7'38975 7.49785 97*5 7-21130 7'50875 Vt , 7*48545(1 + 1.0-4 X 0*2708t+ 10- ' X0 0497t2 ) , or if Vr= 1 , then Vl= 1-003205 .
( continued .
, ) No. 19.-Palladium , same piece as used for the last series , repolished , once boiled .
Loss of Weight T. in water=W .
W(l+at ) 98 7-48420 7 48610 560 7 &gt; 738795 7.49785 97q4 7,2114 7 50855 Vt=748363(1+10O-4x OX357t+10-6x.00629qt ) , or if V1=1 , then Vloo-1O003420 .
No. 20.-Antimony gilded , once boiled .
11*7 6.26320 6*06590 57-2 5'98155 6 07420 97'1 5-84300 6-08245 Vt=6.06396(1 +10-4 X O2686t+10-6 X 00469 , t ) , or if VJ=1 , then V100=1-003155 .
No. , 21.-Antimony gilded , .
once b-oiled.\ 10-9 6*20525 6020745 61i0 61o1010 6.21705 97.8 5 97670 6-2247.5 Vt_= 620550(1 + 10x-4X 02854t+ 10-6 x 0.0325t2 ) , or if V0= 1 , then V10=1003179 .
No. 22.-Plati-num ( purified metal , lent by Messrs. JounsoN and MATTHEY ) , twice boiled8.5 2022465 2.22500 56.0 2119510 2-22775 97-3 2-14215 2 23025 Vt= 2 22452(1 +10-4 X 0.2516t+ 10-6 X 0.01 34t2 ) , or if V70= 1 , then V1jc-11002650 .
No. 23.-Platinum , same piece as used for the last series , repolished , twice boiled .
66 222402 2244O 51*8t 2I19890 2-22705 9617 2*14260 Vt=222402(1 +104x02591t+ 10-6+ 00074t2 ) , or if V0= 1 , then V1 , = -00e66s5 The means of the foregoing formula are put together in Table IV .
TABLE IV .
Formulm for correcting the cubical expansion .
Cadmium. .
a*** Vt=VO(1+10-4X 08078t 10-6 X 0-140t2 ) Zinc *e. .
**** Vt=V0(1 ?
+ 104 X O8222t+ 10 6X 0.0706t2 ) Lead ... ... .
Vt=VO(I + 10-4 X 0-8177t+ 10-6 X 00222t2 ) Tin. .
V=Vo(1 +10-4X 0-6100t+10-6X 00789t2 ) Silver..Vt VO(l + 10-4 X 2-5426t+ 1 o-6 X0 0405t2 ) Coppers ... ... Vt =Vo(I + 10-4 X 0-4463t ? ?
10-6 X 0.0555t2 ) Gold..Vt=VO( + 10-4 X 0-4075t+ 10-6 X0 0336t2 ) Bismuth ... Vt=VO(I + 10-4 X 0-35021+10-6 X 0-0446t ' ) Palladium * .
* METALS{ AND : ALLOYS .
869 On the Expansaon by Hat of Alloys .
The alloys were made in the manner described in a former paper* , and the resultts obtained are contained in Table VII .
I have grouped them together in the same way as I did when speaking of their electric conducting-power .
TABLIE VII .
No. 24.--n4 Pb gilded , thrice boiled .
Loss of weight T. in water= W W(1 +at ) .
12Q0 7'04010 7#04345 51.5 6-97425 7.06265 95.0 6.81730 708610 Vt= 7*03801(1 + 10i4 &gt; 06331t+10-6 X 0x0907t0 ) , or if V0= 1 , thei V1 = 1O007238 .
No. 25.-Sn4Pb .
gilded , once boiled .
15.0 7*03845 7*04475 5964 6.95o5o 7-06620 95'1 6-s1635 7,08560 VY=7.O3 17(1 + 104X 06068t+106 XO.107t2 ) , or if VO= , lthen Yo=I 1O007138 .
No. 26.-Pb4 Sn gilded , twice boiled .
11 0 5v26995 -527190 54 8 5'21670 5229115 95-8 510535 5-30965 Vt=5 26715(1 + 104x 058171t+ 10-6x 0 0263t2 ) , or if V0=1 , then V1-=1O008434 .
No. 27.-Pb4 Sn gilded , once boiled .
11 0 5'27395.5'27590 52t8 5 22465 529410 95 6 5-10975 5-31350 Vt=5-27124(1 + 10-4 x 0'8004t+ 10-1 x 0Q0401t2 ) , or if V0= 1 , then V10C= 1 008405 No. 28.-Cd Pb gilded , twice boiled .
8*6 ~ 536850 5-36935 56*4 5-31245 5-39260 9851 5'19620 5 41305 Vt=5-36519(1 .
+i 6O X'9012t ?
10-6 O0834t2 ) , or if Vo=1 , then Vloo= 1'909095 .
No. 29 .
Cd Pb gilded , once boiled .
912 -5*60060 5.60175 604 5.53285 5462790 91*7 5'44265 5 64415 Vt=5~5971 ( 1 + 10-4 X0 8998t+ 106 x 0601 St ' ) , or if Vo= 1 , then V_ 1'009180 .
( continued .
) No. 30.'-Sn4 Zn gilded , once boiled .
Loss of weight T. in water=W .
W(1 +at ) .
9'3 6g58530 6.58670 54-7 6*51495 6*60765 95 9 6.37195 6.62740 Vt=6058252(1 + 10-4 X 0,6S09t+ 10-6 X 0.0314t2 ) , or if V06=1 , then V1oo=1-007123 .
No. 31.-Sn4 Zn glded , once boiled .
10*7 6-58830 6-59055 60*9 6*49990 6.61330 94 5 6*38165 6-63095 Vt=6-5862o6(1+ 10-4x 05945t+0I 0.130t2 ) , or ifVO=1 , then V1oo=1-007245 .
No. 32.-Sn6 Zn gilded , once boiled .
11^01 6685455 6*85715 57.9 6-77180 6-87915 96.8 6'62940 6s89960 vt=6'85235(1 + 10-4 X 0'6205t+ 10-6 X 0.0948t1 ) , or if V0=1 , then Vi , ==1 007153 .
No. 33.-Sn6 Zn gilded , thrice boiled .
14-3 7-23135 7-23705 5207 7-16090 7-25575 91 4 7-01775 7-27610 Vt=7-23047(1 + 10 -4X 0,6268t+ 1 0x 0.0697t2 ) , or if V0= 1 , then Vin= 1'006965 .
No. 34.-Bi4 Sn gilded , thrice boiled .
16*2 5 49395 5'49995 53-2 5 43465 5 50795 91*1 5 32160 .
5*51635 Vt= 549651(1 +10-4 X 0-3842t+10-6x 0-013t ' ) , or if V0=1 , then V100= 1003974 .
No. 35.-Bi44 Sn gilded , once boiled .
10*8 6*10665 6ios885 57-8 6-02515 6'12040 94 0 5-90165 6 13005 Vt=6.10635(1 + 10-4X 0-3743t+ 106 X 0.0410t2 ) , or if V0= 1 , then V10o=1004153 .
No. 36.-Bi Sn , gilded , once boiled .
10.2 6 04220 6 04395 58.6 5.96205 6 &amp; 05875 96.8 5-83300 6-07065 VT= 6A04087(1 + 104 X 0A4987t+1 o-6 X , 010tI ) , or if V , 1 , then -T10-1A005097 .
ANgD ALOYS .
871:TABLE VII .
( continued .
) No. 3 !
.-Bi Sn2 gilded , once , biled .
Loss of weight T. in water=W .
W(1 +at ) , , 10 9 640,5175 6*05390- , 57.5 J5'97470 6-06820,0 97-6 5*83,920 6o08074o ; 1 V-= 6O05O59(1 + 1014X 0c5008t+ 10X00 , 925 ) , or ifV l , ten V i00510 .
No. 38--Bi24 Pb gilded , once boiled .
10-2 *V 672-710 6172905-1 56.6 6064040 6*74125 94*4:6'49825 6'75160 , Vt=6-72644(1 + 107 x0 3781t+ 10= X0 019 t2 ) , or if Up , then , O0 1003973 No. 39.-Bi24 Pb gilded , once boiled .
13*2 6011480 6*11865 50 4 6f05440 6 12800 90 9:5*92270 6.13865^ ; V 611543(i+ 10-4+O39^5t+ 1060x O 2456 ) , or if : V= I , then Vlo1-0004200* No. 40.-Bi Pb2 gilded , once boiled .
111 597110 5-97335 56'5 5 90675 5S99615 ; 96.8 5`78095 6*01655 _=59678O(l + 10-4 x0 '8362t+ 16x 0 O6782t2 ) , or if VJ=l , then V10=1,008440 .
No. 41.-Bi Pb2 gilded , once boiled .
10*6 6'17070 G17275 53-1 ff11340 6.19555 950 5 98 260 6*21850 Vt=6.16714(1 + 10x 0*8561t+ 10-6 x 0 &amp; 0216t2 ) , or it tV 1 , then Vimo 1I008777 .
No. 42.-Bi Pb2 , gilded , once boiled .
12-4 4517 255 5-17520 581 -5 11390 5*1955Q 96,8 5,000855 5 21300 .
Vt=Y5016976x(1 + iO x0 8463 + 10V X 0-0183t2 ) , ortif Ve 1 , then V100-1 008646 .
No. 43.-Alloy of copper and zinc , containing 71-0 per cent. by weight copper , determuined by --analysis,.gilded , thrice boiled .
98 761]20 7 6131m , 57'4 7'-51595 7 633115 97*1 7 3B020 7*65145 V6=7.60n30-( L+,4xv XO51O9t+ 10-i( 0.0614t ) , or if Ve-1 , then V1co=1l0.723 .
$72 DDR .
A. MATTiIESSEN ON THE XEPANSION TABLE VII ( continued .
) No. 44.-No .
43 repeated ... Loss of weight T. , in water=W .
W(I+a ) .
Calculated .
, 11,27*61090 7t61:385 58-9 7-51100 7-63405 96.5 7-35315 7'65120 V.= 7.60936(1 + 10-1 x 0f5213t+ 10-6 X 040503t ' ) , or if VoY 1 , then V10= 1i005716 .
No. 45.-Au Sn2 , thrice boiled .
9.8 5-24965 5'25100 SF6G 5419420 5-26025 97'4 54o6210 5 &amp; 27070\T~=5.2487(1 + 10-4 X 0'4124t ? ?
106 X 0'0149t2 ) , or .
i =1 , then V[1= 1'004273 .
No. 46.-Au Si2 , twice boiled .
10 55 44040 5-44215 55-7 5 &amp; 37295 5 45210 97*1 5'24700 -5146205 -V=5 43997(1 + 10 x03763t+ 106 X 040429 ? ? ? ?
) , or if Vo= Ithen VN= 1'00419L No. 47.-Au2 Sn , twice boiled .
12.0 5 &amp; 57075 5 57340 55.1 5'50440 5r58380 96'8 5'37535 5 59445 Vt=5.5-7061(1+i0-4X0 4132t+ 10-6X 00298e ) , or if X%=1 , then V10 1'004430 .
No. 48.-Au2 Sn7 , gilded , twice boiled .
1241 5*58500 5*58775 5-6-3 5-51560 5'59850 953 5'39450 5.60840 V1=5 58489(1 +10-4 x 0'4198t+ 10-6 X 0.0229 ? ? ? ?
) , or if V0=1 , then V % &amp; = 1-'004427 No. 49.-Ag4 Au , thrice boiled .
9-7 3'64395 3464485 3 &amp; 64485 51'4 3-60725 3-6S270 3-65270 9641 3*51950 3-66110 3*66110 , Vt=3'64303(1+10-4x05S163t ) , or if V0=1 , then V1 , c=lo005163 .
No. 5Q.O-Ag4Au , thrice boiled .
108 3-76475 3 76605 3 46694 58'5 3171525 3*77530 3'77532 92'S 3'6441.5 3'78200 3178200 V=-3'7G394(1+ 10-4 X 05170t ) or if V0=1 , then VM=1 005170 .
( continued .
) No. 51i-Ag Au , thrice boiled .
Loss of weight T. in water=W .
W(l+at ) .
Calculated , 8.6 3 99445 3X99510 3X99510 57X3 3X94340 4X00470 4X00465 98X4 3X85115 401270 4-01270 V=3 ?
99342(1 +10-4x 04906t ) , or if Vo ?
=1 , then V1oo= 1=004906 .
No. 52.-Ag Au , thrice boiled .
9.0 3*99005 3.99080 3-99080 58-3 3*93730 4*00055 400055 98.0 3-84800 400830 400831 V -=3-98904(1 + 10-4 X 04927t ) , or if Vo01 , then V01*004927* No. 53.-AgAu4 , thrice boiled .
9-6 0-90470 2-90540 606 2-86165 '2e91110 94-1 2-80700 2'91585 Vt=29O453(1 + 10X O2989t+ 1OX O2123t ' ) , or if Vol. then V1o=1loO4 1 19 .
No. 54.-Ag Au4 , thrice boiled .
8-4 2.98805 2-98850 56-7 2-94930 2-99425 93-5 2.88895 2*99970 Vt=2.98766(1 10-4x03241t+ 106 x 0114t2 ) , or if 0= 1 , then V1OO 4381 .
No. 55.-Alloy of silver and platinum , containing 666 per cent. by weight silver ( lent by Messrs. JOHNSON and MATTHEY ) , thrice boiled .
7*0 4-67260 4'67290 58-4 4-60940 4-68370 96.3 4*50995 4-69200 Vt =467147(1l+ 1Ox 0-4359t+ 1 0-6X o0213i2 ) , or if Vo=1 , then V10 = ] J004572 .
Not 56.-No .
55 repeated , thrice boiled .
94 4 67330 4 67435 59'4 -4-60800 4'68470 96'8 4 50930 4-69310 vt4= 4.67 !
252(1 + 10-4 X 04134t+ 10-6x 00431te ) , or if V'o=1 , then V1%0m 1004565 .
No. 57.-Alloy of gold and copper , containing 66'6 per cent. gold , gilded , once boiled .
11P0 4-52820 4*52985 59 4 4-46525 4'53965 97.7 4'36740 4'54835 Vt= 4'52781(1+10-4XO0*4O29+10-6x O.0629t ' ) , or if ' ' ?
41 , then V100= l*OO4658 .
( continued .
) No. 58.-No .
57 repeated , twice boiled .
Ioss of weight To in water=W .
W(l+a ' ) 100P 4P52845 4i52965 ; 54*2 : 4'417595 4-53850 96.9 -4,3i970 4*54815 Vt 4.54781 ( 1 ?
1IQ0-4 X O4OO2t ?
+1 O-X '0655t ) , or if V6= 1 , then V-1 1 V004657 .
No. 59.-Alloy of silver and copper , containinog 36 1 per !
cent. by weight , silver determined by analysis , thrice boiled .
8.5 3 19535 3-19585 52 5 3316165 3'20325 95.1 3 08880 3-2.10$10 Vt= 3.19447(1 + 1i0x 05065t+ 1O ` x 0327t2 ) , or if V0 = I , then V100 =1 005392 .
No. 60.-Not .
59 , .
: repeated , thrice boiled .
9'6 319525 3*19600 56-4 3n156o2 3-20380 96 0 3-08725 3.21125 Vt= 3.19454(1 + 10-4 x 04704t+ 1-6 x 0778tP ) , or if Vs= 1 , then Vet= 1005482 .
No. 61.-Alloy of silver and copper ; containing 716 per cent. by weight , silver determined by analysis , gilded , thrice boiled .
7io 5 85330 5-85370 -53 1I 5-78990 5i96770 97 4 5665130 5-88415 Vt= 5 851 55(1+ 1x 0-4421 + 1I x,0q128t ) , or if V= 1 , then V1=1'00570 .
No. 62.-No .
61 repeated , regilded , once boiled .
101 5*85335 5'85500 52*4 5 79205 5.86795 9:881 5'64930 5a8850Vt 5-85232(1 + 10-4 X &lt; 04406t+ 10x 0 .
l32t2 ) , or if V= 1 , then brigT 1-005726 .
The alloys No. 24 to 33 belong to the .
first group , namely , those made of the metals which , when alloyed with one another , conduct electricity in the ratio of their relative volumes ; from 34 to 48 to the third group , namely , those made of the metals which when alloyed with one anot-her , or with one of those belonging to the first group of metals and alloys ( these form the second group No. 49 to 62 ) , conduct electricity in a lower degree than that calculated from the mean of their volumes .
In Table VIII .
the mean formula are given for the correction of expansion by heat for the foregoing alloys ; in Table IX .
those for the correction of the linear expansion by heat .
Alloy .
Mean correction formulae for cubical expansion by heat .
Sn4 Pb ... ... .
V. VtVO(1 + 10-4x 0-6200t+ 10-6 x 0.0988t2 ) Pb4 Sn. .
Y= Vt-Vo(1 + 10-4 X 0087t+ 10 XO332e2 ) Cd Pb ... . .
Vt=Vo(I + 1o-4 X 0O90Q5I+ 10-6 x 0-0133t2 ) Sn4 Zn ... . .
Vt= Vo ( I + 10-4 x 0-6377t , + I0-5 X 0807t2 ) 5 ; a Zn.* *** * .
* Observed vol. at 00=1 , Calculated vol. at 0= 1 , 'Vol .
per cent. then vol. at 100 ?
= then vol. at 1000= Sn4 Pb.,. .
X 1W 22-28 of Pb 1-007188 1'007225 Pb4 Sn ... : 82 09 of Pb 1-008419 1.008128 , Cd Pb. .
58A49 of Pb 1P009138 1P008847 Sn4 Zn *..87-46 of Sn 1-007184 1P007144 SnD Zn * 91P28 of Sn 1-007058 1P007066 Bi44 Sn * .
085 of Sn 1-004064 1'003972 Bi Sn2..42-81 of Sn 1P005098 1P005207 Bi24 Pb. .
176 of Pb 1-004086 1-004026 Bi Pb2..46 26 of Pb 1P008621 1006007 Cu Zn ( 71 per cent. CU ) .
3385 of Zn 1P005719 1P006328 forming them is not surprising , as in all probability there exist chemical combinations between the two metals , -just as it may be said that the speck/ ic gravity of an alloy is approxirmately eqal to the mean specific gravities of the volumes of the component metals , so also from the foregoing we may deduce that the volume which an alloy will occupy at any temperature between O and 1000 is approximately equal to the mean of the component volumes of the metals at the same temperature , or , in other words , the cubical or linear coefficients of expansion by heat of an alloy between 0 ?
and 100 are approximately equal to the mean cubical or linear coefficients of expansion by heat of the component metals .
In Table XII .
I have given the values from Table X. , together with the observed and calculated specific gravities and conducting-powers of the alloys experimented with .
TABLE XII .
Vol. 0=1 , then vol. 100= Specific gravity .
Conducting-power .
Observed .
Calculated .
Observed .
Calculated .
Observed .
ralculated .
Sn4 Pb ew'w'1 007188 1O007225 8818 8203 1057 1063 Pb Sn 1 .
008419 1-008129 10.590 10i645 ' 8.28 8-43 Cd Pb e e1009138 1'008847 10-246 10246 12.61 132 Sn4 Zn ... 1007184 1]007144 ... ... 13,22 13-45 Sn6 Zn ... 1-007058 1-007066 ... ... ... . .
i2-66 12-84 Bid Sn ... 1004064 1-003972 9-803 9801 0f245 1P28 Bi Sn2 .
** 1-005098 1P005207 8772 8738 3.96 5.59 Bi24Pb ... 1-004086 1-004026 9-845 9.850 0-257 1-30 , Bi Pb2. .
1-008621 1i006007 1 0,956 10-541 209 4-23 Cu+Zn .
*1-005719 1-006328 ... ... 21170 70-20 ( 33.85 vol. Zn ) Au Sn2 1* P004233 1'005919 1833 11.978 14-27 35,51 Au2 Sn7 1 internal cavities owing to the -surface being rapidly cooled , and for this reasonI have not.given the absolute weights of the metals or the alloys .
It is worth remarking that the .
observed specific gravity of the alloy Bi Pb2 differs considerably-from the calculated one .
From the conclusions just drawn ' it would appear that the determinations of the expansion by heat of alloys give in general no indication as to their chemical nature , and ' , that this property belongs to that class of physical properties which does not indicate .
their chemical nature .
In the Report on the Chemical Nature of Alloys* , I have shown that , from the determinations of the electric conducting-power of alloys , we may gain an insight into their chemical nature ; and basing my calculations on the hypothesis there propounded , I am at present able to deduce the conducting-power of any alloy which may be considered as a solidified solution of the one metal in the other , .
although it may differ widely from that calculated from the mean conducting-power of the component metals , as shown in Table XII .
It is proper to point out that the coefficients given in Tables IV .
, V. , VIII .
, and IX .
, are those calculated from readings of the ordinary mercury-thermometer , the mercury being contained in a tube of glass .
Therefore , except at the two values 00 and 1000 , the temperatures indicated by the degrees of the thermometerI used will differI .
From the temperature corresponding to the same number of degrees of an airthermometer .
II .
From the'temperature corresponding to the same number of degrees of a mercurial thermometer , the mercury being contained in a tube of some substance absolutely unaffected as to its volume by heat , could such be found .
The proper corrections can be deduced from the following Table : Temperatures indicated by Temperatures indicated by ordinary air-thermometer .
thermometer at same time .
00 10.0 1008 20.0 20-14 30-0 30-18 40-0 40-20 50.0 5020 60-0 60-18 70-0 70-14 80-0 80-10 90-0 90-05 100-0 100-00 This Table is copied from RECKNAGEL .
REGNAULT found 50 ?
air-thermometer to correspond to 500.20 by mercurial thermometer .
Professor MILLER has sent me , the following investigation of the correction to be applied to reduce the temperatures indicated by the common mercurial thermometer to those indicated by a thermometer of the second kind mentioned above .
I have since found ' that IRECKNAGEL and POGGENDORFFt have given a similar investigation .
A'_____________.3 Let MB denote a mercurial thermometer in which A is the freezing- , B the boilingpoint , the capacity of the tube between A and B beingdivided into 100-parts of equal capacity .
Let M ' B ' denote a thermometer of material absolutely uninfinenced by heat , having exactly the dimensions of AB when at jO a ; At , B ' its reeingand boiling-points and : the capacity of A ' , B ' divided-into.100 parts of equal volume .
Let V be the capacity of MA at 00 , 100 K the capacity of AB at,0 .
Sincethe thermometers are exactly alike atO0 , thecapacity of M ' A ' will be V. Let the capacity of A ' B'= 1000 my , and , therefore , m the cubic expansion of mercury for 1 ?
in terms of the indications of thermometer MA B ' ; g the cubic expansion of glass or 1 ? .
When the thermometers are at the same temperature , let the ends of the columns of mercury stand at T T ' , t , t ' being the number of degrees between A 'and T , and A ' and T ' respectively .
The capacity of MA will now be V(i+it ' ) and that of AT will -be Kt ( 1+gtt ) and the volume of the mercury in MT will be V(1 +Mt ' ) .
Hence TV(1 +mt9=V(1+gt ' ) +Kt(1 +g ' ) Therefore Vo(n -g)t =Kt(1 gt ' ) .
But t and t ' are both 100 at the same time .
Hence M t7-)100=K1OO(l Og100 ) .
Therefore Kg t'=t J t0O =t-gt(100-t ' ) very nearly .
But g is very small and t ' very nearly equal to t. Hence t'=t-gt(100-t ) , very nearly .
g varies from 0 000023 to 0 000028 .
Taking the higher value , at 500 C. , t-t'= 0 000028 X50 X 50007 C. , at 250 0 .
, tt---=0000028 x 25 x 75 0 ?
052OC .
Hence the expansion of mercury is notproportional to its ascent in the tubeof a thermometer , and the difference of rates is a perfectly sensible quantity too large to neglect in many 'researches .
Part of the increased rate of expansion of various bodies ' at high temperatures is due to this error of the mercurial thermometer , whichindicates temperature between O ?
and 1000 too high'as measured by the true expansion of mercury. .
* PoGGENDo1Fr 's ' Annalen , ' vol. exxiii .
p. 115 .
t Ibid. vol. xli .
p. 372 .
6c 2 frequently used a thermometre a poids .
Let P be the weight of mercury contained in it at 00 C. , p the weight of mercury expelled when exposed to the temperature t by ordinary thermometer .
At temperature t by ordinary thermometer the volume of mercury below 0 is V(1 +gt ' ) , and the volume of the mercury above 0 ?
is Kt(1 +gt ' ) .
Therefore the volume of the whole of the mercury is V(1+gt')+Kt(1+gt ' ) .
Hence p Kt( + ?
gt ' ) Kt P-V(1 +gt ' ) + Kt(1 +gt ' ) V+ Kt Therefore p Kt -=pv .
When t=100 let p=b .
Therefore b KlOO .
P-b v ' therefore t P-b P 1O0P-P bP -b pt REGNAULT uses ppE as the measureof temperature , which is the same as l on the indication of a common mercurial thermometer uncorrected for the effect produced by the expansion of glass .
A small error is generally made in taking the boiling-point of water , 100 ?
, viz. the influence of latitude is not taken"into consideration .
Professor MILLER writes to me on this point the following:- " LAPLACE , in the fifth edition ' of his 'Systeme du Monde , ' states that he regards 1000 C. to mean the temperature of water boiling under a pressure equivalent to that of a column of mercury at 0 ?
and 760 millims. in latitude 450 .
" DULONG , and afterwards REGNAULT , assumed 1000 to be that of boiling water under normal pressure at Paris .
Now the pressure at1 760 millims. of mercury at the level of the sea in latitude 450 is equivalent to a column in latitude X at z metres above the level of the sea , the height of which is 760 ( -11322 ) ( 1-0-0025659 cos. 2X ) , where r= radius of the earth =6366198 metres .
" The latitude of Paris ( 480 50 ' 14 " ) is not very different from 450 .
REGNAULT'S place of observation was about 60 metres above the sea .
Hence a column of 760 millims. mercury at the level of the sea exerts the same pressure as a column of 759-75 millims. in REGNAULT'S laboratory .
" This will only make a difference in the boiling-point of 0'01 , a difference which may be neglected in ordinary work~ , but in normal researches ought to be brought into calculation ; as , for example , in MILLER'S normal research " 'On the Construction of the New Standard Pound"* .
At Abo this correction would amount to upwards of 0 ?
'25 .
In the text-books on ' chemistry and natural philosophy we find the coefficients of expansion of air given 0 003665 for each degree of the mercury-thermometer , which is not correct ; for as air expands regularly between O0 and 1000 according to the air-thermometer , a unit of volume of air measured at 00 will not occupy a ' volume at 20-14 mercury-thermometer equal to 1+0-003665 x 20K14=1-07380 but 1+0 i003665 x 20=1-07330 , showing a difference of 0 05 per cent. MILLER , in his paper already mentioned , p. 714 , says , " REGNAULT found the expansion of air from 00 to 1000 under constant pressure equal to 0-36706 of its volume at 00 ... .The ' difference between the mercurial and air-thermometers amounts to about 0 2 .
Hence the expansion of air between 00 and 500.2 is 0-18353 of its volume at 00 ; or between 00 and 50 ?
the ratio of the density of air at 0 ' to its density at t is 1 +0 003656 X t. " In the -present state of science it seems quite wrong that such a want of accordance should exist in our normal instruments , considering that otherwise the errors that exist between the instruments do not amount to 00.1 when properly made .
I have , had several normal thermometers made for my experiments by Messrs. NEGRETTI 'and ZAMBRA , and in every case , when compared with my Kew Standard , the agreement between them has been almost perfect .
It would be decidedly a step in the right direction if in future the fact that mercury does not expand regularly between 0 ?
and 1000 were taken into account in the construction of all normal thermometers ; and if some comparisons between the Kew .
Standard and the air-thermometer were made , then the readings of thermometers which have been compared with the Kew Standard might be easily corrected .
JOLLY* , in his research into the expansion of water , corrected his thermometers by comparing them with an air-thermometer ; unfortunately , however , he does not give a table containing these comparisons ; ; so that his results are obviously not comparable with those obtained with the mercury-thermometer , and therefore at present I am unable to compare his results with those obtained on the same subject by myself .
In conclusion , my thanks are due to Dr. M. BEHREND for carrying out the determinations for the Series 1 , 2 , 13 , 20 , 21 , 24 to 27 and 34 , to Mr. BASSETT , who aided me in carrying out the others , and to Mr. R. P. WRIGHT for having undertaken the greater part of the computations .

