Third Memoir upon Quantics .
By ARTHUR CAYLEY , Esq. Received March 13 , -Read April 10 , 1856 .
MY object in the present memoir is chiefly to collect together and put upon record various results useful in the theories of the particular quantics to which they relate .
The tables at the comImencement relate to binary quantics , and are a direct sequel to the tables in my Second Memoir upon Quantics , vol. cxlvi .
( 1856 ) p. 101 .
The definitions and explanations in the next part of the present memoir are given here for the sake of convenience , the further development of the subjects to which they relate being reserved for another occasion .
The remainder of the memoir consists of tables and explanations relating to ternary quadrics and cubics .
Covariant and other Tables , Nos. 27 to 50 ( Nos. 1 to 50 binary quantics ) .
Nos. 27 to 29 are a continuation of the tables relating to the quintic ( a , b , c , d , e , fx , y)5 .
No. 27 gives the values of the different determinants of the matrix ( a , 4b , 6c , 4d , e ) a , 4b , 6c , 4d , e b , 4c , 6d , 4e , f b , 4c , 6d , 4e , f determinants which are Iepresented by 1234 , 1235 , &amp;c. , where the numbers refer to the different columns of the matrix .
No. 28 gives the values of certain linear functions of these determinants , viz. L= 1256+ 2345-2.1346 L ' =3.12561346 8M =--1345+ 2.1246 8M'=-2346+2.1356 SN =1245+ 3.1236 8N ' =-2356+ 3.1456 80P = L'--3L 5.1346-3.2345 16P'=--5L'L=6 .
1256-3.1346-2356 .
At the end of the two tables there are given certain relations which exist between the terms of Tables 14 , 16 , 25 , 26 , 27 and 28 .
* The Tables 49 and 50 were inserted October 6 , 1856.--A .
C. Resuming now the general subject,54 .
The simplest covariant of a system of quantics of the form ( *:(X , §§ .
" ( where the number of quantics is equal to the number of the facients of each quantic ) is the functioinal determinant or Jacobian , viz. the determinant formed with the differential coefficients or derived functions of the quantics with respect to the several facients .
55 .
In the particular case in which the quantics are the differential coefficients or derived functions of a single quantic , we have a corresponding covariant of the single quantic , which covariant is termed the Hessian ; in other words , the Hessian is the determinant folrmed with the second differential coefficients or derived functions of the quantic with respect to the several facients .
56 .
The expression , an adjoint linear form , is used to denote a linear function Sx+ry+. .
, or in the notation of quantics ( M , 4 ... 3x , y,. .
) , having the same facients as the quantic or quantics to which it belongs , and with indeterminate coefficients ( , n. .
) .
The invariants of a quantic or quantics , and of an adjoint linear form , may be considered as quCantics having ( E , n ... ) for facients , and of which the coefficients are of course functions of the coefficients of the given quantic or quantics .
An invariant of the class in question is term ( d a contravariant of the quantic or quantics .
The idea of a contravariant is due to Mr. SYLVESTER .
In the theory of binary quantics , it is hardly necessary to consider the contravariants ; for any contiavariant is at once turned into an invariant by writing ( y , -a , ) for ( 0 , v ) .
57 .
If we imagine , as before , a system of quantics of the form ( *3Y *y , *. .
) , where the number of quantics is equal to the number of the facients in each quantic , the function of the coefficients , which , equalled to zero , expresses the result of the elimination of the facients from the equations obtained by putting each of the quantics equal to zero , is said to be the Resultant of the system of quantics .
The resultant is an invariant of the system of quantics .
And in the particular case in which the quantics are the differential coefficients , or derived functions of a single quantic with respect to the several facients , the resultant in question is termed the Discriminant of the single quantic ; the discriminant is of course an invariant of the single quantic .
58 .
Imagine two quantics , and for0m the equations which express that the differential coefficients , or derived functions of the one quantic with respect to the several facients , are proportional to those of the other quantic .
Join to these the equations obtained by equating each of the quantics to zero ; we have a system of equations , one of which is contained in the others , and from which therefore the facients may be eliminated .
The function which , equated to zero , expresses the result of the elimination is an invariant which ( from its geometrical signification ) might be 636 termed the Tactinvariant of the two quantics , but I do not at present propose to consider this invariant except in the particular case where the system consists of a given quantic and of an adjoint linear form .
In this case the tactinvariant is a contravariant of the given quantic , viz. the contravariant termed the 1Reciprocant .
59 .
Consider now a quantic ( *x , y , , )m , and let the facients x , y,. .
be replaced by x+pX , Xy+^Y,. .
the resulting function may , it is clear , be considered as a quantic with the facients ( x , p ) and of the form ( *X , Y,..)m The coefficients of this quantic are termed Emanants , viz. excluding the first coefficient , which is the quantic itself ( but which might be termed the 0-th emanant ) ; the other coefficients are the first , second , and last or ultimate emanants .
The ultimate emanant is , it is clear , nothing else than the quantic itself , with ( X , Y,. .
) instead of ( x , y , ... ) for facients : the penultimate emanant is , in like manner , obtained from the first emanant by interchanging ( x , y,. .
) with ( X , Y , ... ) , and similarly for the other emanants .
The facients ( X , Y,. .
) may be termed thefacients of emanation , or simply the new facients .
The theory of emanation might be presented in a more general form by employing two or more sets of emanating facients ; we might , for example , write Xx++pX+vX ' , ?
Xy+ pie +vYI , ... for x , y , ... , but it is not necessary to dwell upon this at present .
The invariants , in respect to the new facients of any emnanant or emanants of a quantic ( i. e. the invariants of the emanant or emanants , considered as a function or functions of the new facients ) , are , it is easy to see , covariants of the original 'quantic , and it is in many cases convenient to define a covariant in this manner ; thus the Hessian is the discriminant of the second or quadric emanant of the quantic .
60 .
If we consider a quantic ( a , b,..Xx , y,. .
) and an adjoint linear form , the operative quantic ( a , ab,..I , ,. .
) ( which is , so to speak , a contravariant operator ) is termed the Evector .
The properties of the evector have been considered in the introductory memoir , and it htas been in effect shown that the evector operating upon an invariant , or more generally upon a contravariant , gives rise to a contravariant .
Any such contravariant , or rather such contravariant considered as so generated , is termed an Evectant .
In the case of a binary quantic , 2(a.x , , 4 P2 637 the covariant operator ( , a,. .
; $ , -x ) may , if not with perfect accuracy , yet without risk of ambiguity , be termed the Evector , and a covariant obtained by operating with it upon an invariant or covariant , or rather such covariant considered as so generated , may in like manner be termed an Evectant .
61 .
Imagine two or more quantics of the same order , ( a , ,..Ix , y ) we may have covariants such that for the coefficients of each pair of quantics the covariant is reduced to zero by the operators aB , +B , +. .
B+Sb +. .
Such covariants are called Combinants , and they possess the property of being invariantive , quoad the system , i. e. the covariant remains unaltered to a factor prs , when each quantic is replaced by a linear function of all the quantics .
This extremely important theory is due to Mr. SYLVESTER .
Proceeding now to the theory of ternary quadrics and cubics , First for a ternary quadric , we have the following tables : Covariant and other Tables , Nos. 51 to 56 ( a ternary quadric ) .
No. 51 .
The quadric is represented by ( a , b , c , / , g , h 3x , ) , z ) ' , which meansaz +4by+c z+2fyz+2gzx+2h , y. No. 52 .
The first derived functions ( omitting the factor 2 ) are(a , h , g x , y , z ) ( h , b , f Xf y , z ) ( g , , c X : y , ) .
6,38 No. 53 .
The operators which reduce a covariant to zero are( h , b , 2f 3ag , af , a6)ay , ( 2g , f , c a , ba , a , g ) -x , z ( a , 2h , gX(a , ab , af)-ya ( g , 2f , c ?
a6^ , a^ f a)-yaZ ( a , h , 2g(a , af , aj )-z-x ( 2h , bfa h , a , , a , )-ay .
No. 54 .
The evector is ( a. aO , a , af , a , , a , h , , ) No. 55 .
The discriminant is a , h , g h , b , f g , f , c which is equal to abcaf2bg2 ch+2fgh .
No. 56 .
The reciprocant is a , h , g 2 , h , b , f g , j , c which is equal to ( bc-f2 , ca--g , ab-h2 , gh-af , hf-bg , fg--ch( 21 , , )2 .
The discriminant is , it will be noticed , the same function as the Hessian .
The reciprocant is the evectant of the discriminant .
The covariants are the quadric itself and the discriminant ; the reciprocant is the only contravariant .
Next , for a ternary cubic , we have the following Tables : Covariant and other Tables , Nos. 57 to 70 ( a ternary cubic ) .
No. 57 .
The cubic is U= ( a , b , c , f , g , h , i , j , k , lIx , y , z)3 , which meansax + by3+ cz3 +3fz + 3g2 ; x + 3hTxy +3iyz2+ 3jzx'+ 3kxy2 + 61yz .
639 The preceding Tables contain the complete system of the covariants and contravariants of the ternary cubic , i. e. the covariants are the cubic itself U , the quartinvariant S , the sextinvariant T , the Hessian HUl , and an octicovariant , say OU ; the contravariants are the cubicontravariant PU , the quinticontravariant QU , and the reciprocant FU .
The contravariants are all of them evectants , viz. PU is the evectant of S , QU is the evectant of T , and the reciprocant FU is the evectant of QU , Qr what is the same thing , the second evectant of T. The discriminant is a rational and integral function of the two invariants ; representing it by R , we have R=64 S3-T2 .
If we combine U and HUE by arbitrary multipliers , say a and 6j3 , so as to form the sum oaU+63H-IU , this is a cubic , and the question arises , to find the covariants and contravariants of this cubic : the results are given in the following Table:aU+6pHU =aU+63HU .
H( ?
U+6/ 3HU)= ( 0 , 2S , No. 68 .
T , 882 , P3)'U +(1l , 0 , -12S , --2T , P)3 ) HUE .
P(aU+6P3HU)= ( 1 , 0 , 12S , 4T , , 3)3PU +(0 , 1 , o , -4ScT , P)3QU .
Q(cU+6/ 3HU ) = ( 0 , 60S , 30T , 0 , -120TS , -24T2+576S3',3 P)PU +(1 , 0 , 0 , 10T , 240S2 , 24TS a , P)5QU .
S(aU+6S/ HUE ) = ( S , T , 24S2 , 4TS , T2-48S3 C , 3)4 .
T(caU+6S3HU ) =(T , 96S2 , 60TS , 20T2 , 240TS2 , -48T2S+4608S4 , -8T3+576TS31a , 3)6 .
R(aU+63HU ) = [ ( 1 , 0 , -24S , -8T , -48S2 ao , 3)4 ] 3R .
F(aU+6pIHU ) =(1 , 0 , -24S , -8T , --48S2 , j)4FU +(0 , 24 , 0 0 , --48Ta , 3)4(PU)2 +(0 , 0 , 24 , 0 , 96S 6 , 3)4PU .
QU +(0 , , , o , p ) , 8 , ( QT)2 .
We have , in like [ manner , for the covariants and contravariants of the cubic 6aPU+-3QU , the following Table-No .
69 .
6zPU+PQU =6aPU+PQU .
H(6aPU-i3 , QU)=-(-2T , 48S2 , 18TS , Tr2+ 16S3c , 3P)PU + ( 8S , T , -8S2 , -TS , P)3'QU .
P(6aPU+PQU ) = ( 32S2 , 12TS , T2+32S3 , 4TS2 jX , P)3U +(4T , 96S2 , 12TS , T2-32S3a5 , P)3HU .
4Q2 645

