XVII .
" On the Change in the Elastic Force of a Constant Volume of dry Atmospheric Air , between 32 ?
F. and 212 ?
F. , and on the Temperature of the Freezing-point of Mercury .
" By BALFOUR STEWART , M.A. , F.R.S. Received June 18 , 1863 .
( Abstract .
) The author gave a detailed description of his apparatus , and of the method employed in drying the air .
The result of his experiments gave as the coefficient of increase of elasticity of air of constant volume for 1 ?
F. 0'002040 , this being slightly different from that given by Regnault , which is 0'002036 .
He also finds that the temperature of the freezing-point of mercury is constant , and that its value on Fahrenheit 's scale , as given by an air-thermometer , is -37§ 93 , while as recorded by a standard mercurial thermometer it is -38 ?
'00 .
It is probable that this difference is owing to an anomalous contraction of mercury before it freezes , similar to the corresponding expansion of water ; but this effect in the case of mercury seems to be very small , and it may be said that a mercurial thermometer properly graduated , will denote the true temperature , even down to the freezing-point of the mercury itself .
XVIII .
'On the Degree and Weight of the Resultant of a Multipartite System of Equations .
" By Professor J. J SYLVESTER , F.R.S. Received May 25 , 1863 .
Let there be ( l n ) equations each homogeneous in any number of sets of variables , and suppose that the degrees of the several equations in respect to these sets are respectively a , b , c,. .
, , a l , , c l ) , ? . .
, le a Le t*no e 1 +o , 1+f3 , 1+y,. .
1+ X , then in order that the system may have a resultant , since the number of ratios to be eliminated is +/ 3 +y+ ... +X , this sum must be equal to n. Let ai +bi + cir+. .
+c:w=L , and let LL , , L2 , ... L= P , then 1st , the degree of the resultant in question in regard to the coefficients of the rth equation will be the coefficient of r. 7 .
ry ... in P Lr 2nd .
As regards weight .
By the weight of any letter in respect to any given variable is to be understood the exponent of that variable in the term affected with the coefficient ; and by the weight of any term of the resultant in respect to such variable , the sum of the weights of its several simple factors ; each term in the resultant in respect to any given variable has the same weight ; and this weight may also be proved to be alike for each variable in the same set , and may be taken as the weight of the resultant in respect to such set .
This being premised , we have the following theorem : The value of the weight of the resultant in respect to any particular set of the variables , ex. gr. the ( 1 + a ) set , will be the coefficient of pf+ .
a. ry. .
X. in P. In the particular case where a=f3=y ... =X , the above expressions for the degree and weight evidently become polynomial coefficients .
Thus , ex. gr. , if we suppose each equation linear in respect to the variables of each set , the degree of the resultant in respect to the coefficients of any equation will be rr(a+3+y ... +X ) 7r .
7r .
Try..7rX and its weight in respect to the ( 1 +a ) set will be r(l+a+/ 3+ ... + x ) 7r(l -a)7r,8 .
ry * .
*r(X ) In particular if each set is binary , so that a=-3=y ... X* 1 , the degree becomes 7r(n ) , and the weight r( + .
The above eores ae , I believe , altogether new .
The above theorems are , I believe , altogether new .
It may just be noticed ( as a passing remark ) that the total degree in the general case is the coefficient of..rY ... , xinP { + +'*'+ and the total weight the coefficient of the same argument in P +-+ Ra +a R

