]\gt ; II .
A Ninth Memoir on Quantics .
By Professor CAYLEY , F.fi .
S. Received April 7 , \mdash ; Read May 19 , IT was shown not long ago by Professor GORDAN that the number of the irreducible covariants of a binary quantic of any order is finite ( see his memoir " " Beweis dass jede Covariante und Invariante einer binaren Form eine ganze Function mit numerischen Coefficienten einer endlichen Anzahl solcher Formen ist Crelle , t. 69 ( 1869 ) , Memoir dated 8 June1868 ) , and in particular that for a binary quintic the number of irreducible covariants ( including the quintic and the invariants ) is , and that a binary sextic the number is .
From the theory given in my " " Second Memoir on Quantics Phil. Trans. 1856 , I derived the conclusion , which , as it now appeal .
S , was erroneous , that for a binary quintic the number of irreducible covariants was infinite .
The theory requires , in fact , a modification , by reason that certain linear lelations , which I had assumed to be independent , re really not independent , but , on the , linearly connected together : the interconnexion in question does not occur in to the quadric , cubic , or quartic ; and these cases respectively the theory is true as it stands ; for the quintic the interconnexion first presents itself in legard to the degree 8 in the coefficients and order 14 in the variables , viz. the theory gives colrectly the numbel of covariants of any degree not exceeding 7 , and also those of the degree 8 and older less than 14 ; but for the order 14 the theory as it stands gives a non-existent irreducible covariant , viz. we have , according to the theory , , that is , of the form in question there are 10 composite ts connected by 6 , and therefore equivalent to 10\mdash ; 6 , asyzygetic covariants ; but the number of asyzygetic covariants being , there is left , according to the theory , 1 reducible covariant of the form in question .
The fact is that the 6 syzygies being interconnected and equivalent to 5 independent syzygies only , the composite covariants are equivalent to 10\mdash ; 5 , the full number of the asyzygetic covariants .
And similarly the theory as it stands $oives a non-existent irreducible covariant .
The theory being thus in error , by reason that it omits to take account of the interconnexion of the syzygies , there is no difficulty in conceiving that the effect is the introduction of an infinite series of nonexistent irreducible covariants , which , when the error is corrected , will disappear , and there will be left only a finite selies of irreducible covariants .
Although I am not able to make this correction in a general manner so as to from the theory that the number of the irreducible covariants is finite , and so to present the theory in a complete form , it nevertheless appears that the theory can be made to accord with the facts ; and I reproduce the theory , as well to show that this is so as to exhibit certain new formulae which appear to me to place the theory in its true light .
I remark that although I have in my Second Memoir considered the question of finding the number of irreducible covariants of a given degree in the coefficients but of any order whatever in the variables , the better course is to separate these according to their order in the variables , and so consider the question of finding the numb of the irreducible covariants of a given degree in the coefficients , and of a given order in the val.iables .
( This is , of coul .
, what has to be done for the enumelation of the irreducible coYariants of a given quantic ; and what is done completely for the quadric , the cubic , and the quartic , and for the quintic up to the degree 6 in my Eighth Memoir , Phil. Trams .
1867 .
) The new formulae exhibit this separation ; thus ( Second Memoir , No. 49 ) , writing instead of , we have for the quadric the expression , showing that we have irreducible covariants of the degrees 1 and 2 respectively , viz. the quadric itself and the discriminant : the new expression is , showing that the covariants in question are of the actual forms and respectively .
Similarly for the oubic , instead of the expression No. 65 , , we have , exhibiting the irreducible covaliants of the forms , and , connected by a syzygy of the form ; and the like for quantics of a higher order .
In the present Ninth Memoir I the last-mentioned formulae ; I carry on the theory of the quintic , extending the Table No. 82 of the Eighth Memoir up to the degree 8 , calculating all the syzygies , and thus establishing the interconnexions in virtue of which it appears that there are really no irreducible covaliants of the forms , and .
I reproduce in part GORDAN'S theory so far as it applies the quintic , and I give the expressions of such of the 23 covariants as are not given in my former memoirs ; these last were calculated for me by Mr. W. BARRETT DAVIS , by the aid of a grant from the Donation Fund at the disposal of the Royal Society .
The paragraphs- of the present memoir are numbered consecutively with those of the former memoirs on Quantics .
Article Nos. 328 to \mdash ; Reproduction of my original Theory as to the Number of the Irreducible Covariants .
328 .
I reproduce to some extent the considerations by which , in my Second Memoir on Quantics , I endeavoured to obtain the number of the irreducible ianls of a given binary quantic Considering in the first instance the covariants as functions of the coefficients without regarding the variables , and attending only to the following , a covariant is a rational and integral homogeneous function of the coefficients ; PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
, if , . . .
are covariants , any rational and integral function homogeneous in regard to the coefficients , is also a covariant , \mdash ; we say that the covariants X , , of the same degree in regard to the coefficients , and not connected by any identical equation ( where are quantities independent of coefficients . .
are asyzygetic covariants , and that a covariant not expressible as a rational and integral function of covariants oflower degrees is an irreducible covariant ; and it is assumed that we know the number of the asyzygetic covariants of the degreesl , 2 , 3 .
; say , these are , . . .
, or , what is the same thing , that the number of the asyzygetic covariants of the degree , or form , is equal to the coefficient of in a given function where have purposely written , as a representative of the coefficients . .
in place of the of my Second Memoir .
329 .
The theory was , that determining , . . .
by the conditions : that is , throwing into the form the index would express the number of irreducible covariants of the degree less the number of the ( irreducible ) linear relations , or syzygies , between the composite or non irreducible covariants of the same degree .
Thus , there would be covariants of the degree 1* ; these give rise to composite covariants of the degree 2 ; or , assuming that these are connected by syzygies , the number of } geticcom .
posite covariants of the degree 2 would be ; and thence there would be , that is , irreducible covariants of the same deglee ; so that ( irreducible invariants less syzygies ) is 330 .
The syzygies are here irreducible syzygies ; for , calling , . . .
the covariants of the degree 1 , there is no identical relations between the terms , : imagine for a moment that we could hay l such identical relations ( viz. this very well be the case if instead of the functions , . . .
, we were dealing with the same number of other quadric functions of these quantities ) , that is , relations not establishing any relation between , . . .
, and besides these non-identical relations as above ; then the number of irreducible invariants would be , and the number of irreducible syzygies being as before , the difference would be not * For the case of covariants , is of course ; but in the investigation the term covariant properlystands for any function satisfying the conditions and but .
The identical relations are here relations between composite covariants , and the effect ( if any such relation could subsist ) would , it appears , be to increase ; between syzygies such identical relations do actually exist , and the effect is contrariwise to diminish the ; we may , for instance , for the degree have irreducible covariants tess irreducible syzygies 331 .
Assume for a moment that , for a given value of is positive ; but for the term it would of course follow that there was the deglee in question a certain number of irreducible covariants ; and it was in this manner that Iwas led to infer that the number of the covariants of a quintic was infinite\mdash ; viz. the transformed expression for the number of asyzygetic covariants is coeff. in a product which does not terminate , and as to which it is also assumed that the series of negative indices does not terminate .
332 .
The principle is the same , but the discussion as to the number of the irreducible covariants becomes more precise , if we attend to the covariants as involvin not only the coefficients but also the variables ; we have then to consider the covariants of the , or , say , of the form ( and order ) , and the number of the asyzygetic covariants of this form is given as the coefficient of in a given function of , ( I write instead of the of my Second Memoir in the formulae which contain and ) : by taking account of the composite covariants and syzygies , we successively determine , from the given number of asyzygetic covariants for each value of and , the number of the irreducible covariants for the same values of and .
This is , in fact , done for the quintic in my Eighth Memoir up to the covariants and syzygies of the degree 6 .
But before resuming the discussion for the quintic , I will consider the preceding cases of the quadric , the cubic , and the quartic .
Article Nos. 333 to 336.\mdash ; New formulae for the number of Asyzygetic Covariants .
333 .
For the quadric , the number of asyzygetic covariants coeff. in ' ( see Second Memoir , No. 35 , obsel.ving that is there , and that the subtraction of successive coefficients is effected by means of the factor in the numerator .
See also Eighth Memoir , No. 251 , where a like form is used for the quintic ) .
Writing -for , and for , this is coeff. in .
PBOFESSOR CAYLEY'S NINTH BMOIR ON QUANTIC3 .
The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is The development is which is , where ; and , since contains only negative powers , the required number is coeff. in indicating that the covariants are powers and products of ( ax2 and ) , the quadric itself , and the discriminant .
Compare Second Memoir , No. 49 , according to which , writing therein for , the number of asyzygetic covariants is coeff. in 334 .
For the cubic the number of asyzygetic covariants is coeff. in ; or transforming as before , this is coeff. in the function is here , where ( that this is so may be easily verified ) ; and since the second term contains only negative powers , the required number is coeff. in .
The formula , in fact , indicates .
that the covariants are made up of , the cubic itself , the Hessian , the cubicovariant , and the discriminant , these being connected by a syzygy of the degree 6 and order 6 .
Compare Second Memoir , No. 50 , according to which the number of covariants of degree is coeff a in .
335 .
For the quartic the number of asyzygetic covariants is coeff. in ; or tl.ansforming as , this is coeff. in : the function is here , where ; and the second term containing only negative powers , the required numbel is coeff. in .
The formula indicates that the covariants are made up of , the quartic itself , the Hessian , the quadrinvariant , the cubinvariant , and the cubicovariant , these being connected by a syzygy of the degree 6 and order 12 .
Compare Second Memoir , No. 51 , according to which the number of covariants of degree is coeff. in 336 .
For the quintic the number of asyzygetic covariants is coeff. in ; or transforming as before , this is coeff. in The developed expression is PROFE@SOR CAyLEy'S NINTH MEMOIR ON but.here there is not any finite function such that this development is .
The numerical coefficients are of course the same as those in the development of the untransformed function ; viz. they are the numbers given in the third column of Table No. 82 ( Eighth Memoir ) , and also ( carried further ) in the third column of the following Table , No. 87 .
And we can , from the discussion of these coefficients , deduce the form of , viz. this is where , for shortness , I have written stand for , and so in other cases : moreover in the third column of the numerator the ( 9 ) shows that the factor is , and so in other cases : this will be further explained presently .
Compare herewith the form , Second Memoir , No. 52 , viz. the number of asyzygetic covariants of the degree is coeff. in each index being , it will be observed , equal to the number of factors in the numerator , less the number of factors in the denominator , in the corresponding column of the new formula .
Article Nos. 337 to 346.\mdash ; The 23 Fundamental Covariants .
337 .
GORDAN'S result is that the entire number of the irreducible covariants of the binary quintic is .
I represent these by the letters , . . .
, , identifying such of them as were given in my former Memoirs on Quantics with the Tables of these Memoirs , and the new ones , with the Tables Nos. 90 , 91 , 92 , 93 , 94 , 95 { of the present Memoir .
PROFESSOB CAYLEy'S NINPH MEMOIR ON QUANmCS .
Table No. 87.\mdash ; Identification of the 23 irreducible covariants of the binary quintic .
Table No. I ( GQ I ( GQ I ( GQ I ( GQ I ( GQ 14 15 16 17 18 19 20 21 22 23 24 83 84 ' 25 ( jr ) 29 338 .
Table exhibits the generation of the several covariants ; viz. denotes denotes \amp ; c. ( see post , No. 348 ) .
The column , \amp ; c. shows GORDAN'S notation , and the generation of his 23 forms ( written as with him for \amp ; c. ) : it will be observed that the forms are not identical ; if the calculations had been made de novo , I should have adopted his values , simply omitting numerical factors of the several forms ( thus every term of contains the factor 2 .
) : of course the presence of these numerical factors renders the , / , , \amp ; c. as they stand inconvenient for the expression of lesults ; and the numerical fixation of the values was no part of GORDAN'S object .
But by reason of the existing Tables the change of notation is in fact more PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
than this ; thus instead of being a submultiple of , that is , of , is in fact ; and so in other cases .
If the occasion for it arises , there is no difficulty in expressin any one of the forms / , , \amp ; c. in terms of the ; thus in the instance just referred to , , we have and whence ; also ; and therefore , finally , 339 .
I remark upon the value , that is the complete value of a covariant , the leading coefficient of which is given in Table No. 86 of my Eighth Memoir ; the form , omittin a numerical factor ( if any ) , would have had smaller numerical coefficients , but there is in the form actually adopted the advantage that it vanishes for , that is , when the ntic has two equal roots .
340 .
I now form the following Table No. 88 , viz. this is the Table No. 82 of my Eighth Memoir , carried as far as , but with the composite covariants expressed by means of the foregoing letters , . . .
, ; instead of giving the syzygies as in Table No. 82 , I transfer them to a separate Table , No. 89 .
In all other respects the arrangement is as explained , Eighth Memoir , No. 253 ; but in place of I have written* , to denote new covariant , new syzygy , derived syzygy , respectively ; and I have , as to the terms respectively , introduced the new symbol to denote an interconnexion of syzygies , as appearing by the Table No. 89 , and as will be further explained .
Table No. 88 .
Table No. 88 ( continued ) .
PROFESSOR CAYLEy'S NINTH MEMOIR ON QUANTICS .
Table No. 88 ( concluded ) .
CABK , HIABJ , \ldquo ; ABCD , ABCE , ABCF , 341 .
The syzygies and interconnexions of syzygies are given in Table No. 89 .
Table No. 89 ( continued ) .
342 .
In illustration take any one of the lines of Table No. 88 , for instance the line there are here 6 composite covariants , but the number of asyzygetic covariants is ; there must therefore be syzygies ; we have however ( see Table No. 89 ) two derived syzygies of the right form , viz. these are which are designated as , and there is consequently no new syzygy Z. But in the line , ACH , , FI there are 7 composite covariants , but the number of asyzygetic covariants is ; there must therefore be syzygies .
One of these is the derived syzygy which is designated by ; the other is a new syzygy ( see Table No. 89 ) , designated by Z. 343 .
Take now the line , CEF ; there are here lO composite covariants , but the number of irreducible covariants is ; there should therefore be 10\mdash ; 6 , syzygies .
There are , however , the 5 derived syzygies \amp ; c. ( see Table No. 89 ) designated by ; since these are equivalent to 4 syzygies only there must be lidentical relation between them ( designated by ) , viz. this is the equation obtained by adding the several syzygies , multiplied each by the proper numerical factor as shown Table No. 89 .
PROFESSOR CAYLEY'S NINTJI MEMOIR ON QUANTICS .
344 .
Again , for the line , BCI , BDF , CDE there are here 10 composite covariants , but only 5 irreducible covariants ; there should therefore be 10\mdash ; 5 , syzygies ; we have in fact the 6 derived syzygies \amp ; c. ( see Table No. 89 ) designated by 62 ' ; these must therefore be connected by 1 identical relation ( designated by ) , viz. this is the equation obtained by adding the several syzygies , each multiplied by the proper numerical factor as shown Table No. 89 .
345 .
These two cases are in fact the instances which present themselves where a correction is required to my original theory .
The two identical relations in question were disregarded in my original theory , and this accordingly gave the two non-existent irreducible covariants and .
And reverting to No. 336 , these give in the denominator of the factols .
In virtue hereof , writing , we have in the factor , agreeing with the function .
And we thus see that the denominator factors of do not all of them refer to irreducible covariants ; viz. we have each referring to an irreducible covariant , but and each referring to an identical relation or interconnexion of syzygies .
And we thus understand how , consistently with the number of the irreducible covariants being finite , the expression for may be as above the quotient of two infinite products ; viz. there will be in the denominator a finite number of factors each referring to an irreducible covariant , but the remaining infinite series of denominator factors will refer each factor to an identical relation or interconnexion of syzygies .
But I do not see how we can by the theory distinguish between the two classes of factors , so as to determine the number of the irreducible covariants , or even to make out affirmatively that the number of them is finite .
346 .
The new covariants are as follows:\mdash ; ( Remarks added March , It will be observed that the Tables are printed in a slightly different form from the preceding ones ; this has been done in order to show at a glance in each column the set of terms which contain a given power of , and in each such set the terms which contain a given power of The numerical verifications are also given , not only for the entire column , but for each set of terms containing the same power of ( viz. the equal sums of the positive and negative coefficients are shown by a number with the prefixed sign ) ; and in Table 95 the verification is given in regard to the subsets containing the same powers of and ; as to these subsets , the sums of the positive and negative coefficients are not in all cases equal , but a singular law manifests itself ( see p. PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
Table No. 90 ( Covariant O ) .
4 3 7 6 ce \mdash ; 8 6 3 4 4 1 18 cd cde3 \mdash ; 32 6 9 cd cd 4 1 NTH IIEMOIR ON QUANTICS .
( Covariant P ) .
6 6 166 134 PROFESSOR CAYLEY'S NINTH IIEMOIR ON QUANTICS .
Table No. 92 ( Covariant R ) .
IIDCCCLXXI .
cdef2 ] 8 18 cd 32 33 S2 24 .
Coefficient of -\mdash ; PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
Table No. 93 ( continued ) .
Table No. 94 Covariant ) .
Coefficient of Coefficient of PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
Table No. 94 ( continued ) .
Coefficient of Coefficient of Table No. 95 Covariant ) .
coefficient .
coefficient .
PROFESSOR CAYLEY'S NINIH MEMOIR ON QUANTICS .
Table No. 95 ( continued ) .
coefficient .
coefficient .
Table No. 95 ( continued ) .
coefficient .
coefficient .
56 360 1996 : 420 900 \mdash ; 1280 576 576 162 432 de 124 \mdash ; 8020 CAYLEY'S NINTH MEMOIR ON QUANTICS .
Table No. 95 ( continued ) .
coefficient .
coefficient .
4 : , \mdash ; 688 \mdash ; 2160 er \mdash ; 3420 \mdash ; 3510 lS40 \mdash ; 1 \mdash ; 2360 \mdash ; 1680 \mdash ; 1296 \mdash ; 1312 \mdash ; 2060 \mdash ; 8020 S540 : 4 23924\mdash ; 140 6 78 72 44 332 496 216 304 \mdash ; 312 \mdash ; 320 860 960 d \mdash ; 2160 420 540 700 lS40 \mdash ; 1530 96 80 \mdash ; 3240 \mdash ; 1120 3360 4800 8160 \mdash ; 6000 \mdash ; 1312 9360 72 3408 2156 \mdash ; 2360 \mdash ; 5808 90 78 Table No. 95 ( continued ) .
coefficient .
coefficient .
a 576 : S24 \mdash ; 5808 3240 6240 2608 \mdash ; 8160 1620 162 702 90 3640 796 a \mdash ; 600 18 36 \mdash ; 1 SO 184 \mdash ; 108 18 264 756 368 732 540 2520 1350 144 376 1440 \mdash ; 1530 6360 \mdash ; 6000 1350 2344 \mdash ; 9260 1720 : : \mdash ; 168 648 a 4920 918 2608 4260 38 832 1112 \mdash ; 2060 2026 \mdash ; 56 PROFESSOR CAYLEY'S NINTH MEMOIR ON ANTICS .
Table No. 95 coefficient .
20624\mdash ; 116 rical coefficients that we have as follows:\mdash ; coefficient .
viz. in the coefficient , the coefficients of are , that is the sum of the positive coefficients is , and the sum of the negative coefficients is .
But in the coefficients are , that is , the sum of the positive coefficients is , and the sum of the negative coefficients is ; and so in other cases .
The total sum , viz. the sum of the positive coefficients and that of the negative coefficients ( taken as a positive number ) are each , and so in the coefficient there is the same total sum ; which is as it should be , since there are in a different order the same numerical coefficients .
PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
Article Nos. 347 to 365.\mdash ; Sketch of Professor GORDAN'S prooffor the finite Number , , of the Covariants of a Binary Quintic .
347 .
I propose to reproduce the leading points of Professor GORDAN'S proof that the binary quintic has a finite system of 23 covariants , viz. a system such that every other covariant whatever is a rational and integral function of these 23 covariants .
348 .
Derivation.\mdash ; Consider for a moment any two binary quantics of the same or different orders , and which may be either independent quantics , or they may be both or one of them covariants , or a covariant , of binary quantic .
We may the series of derivatives where , however , there is no occasion to use the notation ( as this is simply the product ) , and the succeeding derivatives may ( when there is no risk of ambiguity ) be written more shortly , \amp ; c. ; in all that follows the word ' ' derivative\ldquo ; ( GOBDAN'S Uebereinanderschiebung ) is to be understood in this special sense .
349 .
The degree of the derivative is the sum of the degrees of the constituents ; the order of the derivative is the sum of the orders less ; it being understood throughout that the word degree refers to the coefficients , and the word order to the variables .
In speaking generally of the covariants or of all the covariants of a quantic , or of the covariants or all the covariants of a given degree or order , we of course exclude from consideration covariants linearly connected with other covariants ( for otherwise the number of terms would be infinite ) ; but unless it is expressly so stated , we do not carry this out rigorously so as to make the system to consist of asyzygetic covariants ; .
it is assumed that the system is complete , but not that it is divested of superfluous terms .
350 .
Theorem A.\mdash ; The covariants of a quantic of a given degree can be all of them obtained by derivation from and the covariants of the next inferior degree .
In particular for the degree 1 the only is the quantic itself ; for the degree 2 the covariants are .
: using for a moment to denote each of these in succession , the covariants of the third degree are , . . .
; and so on .
351 .
Suppose that the covariants of the second deglee , ( ff)2 , are in this order represented by then the covariants of the third degree written in the order may be represented by , the covariants of the fourth degree written in the order , , . .
, may be represented by , and so on : we thus obtain in a definite order the covariants of a given degree ; say , these are , . . . .
: any term is said to be a later term than the preceding terms , and an earlier term than the following ones , \amp ; c. Observe that each term is a derivative , the derivatives of an earlier are earlier than those of a later ; and as regards the derivatives of the same , the derivative with a less index of derivation is earlier than that with a greater index of derivation , or , what is the same thing , those are earlier which are of the higher order .
352 .
The series .
is not asyzygetic ; we make it so , by considering in succession whether the several terms , . . .
respectively are expressible .
as linear functions of the earlier terms , and by omitting every term which is so expressible .
The reduced series thus obtained is called , . . .
Observe that not every is a but that every is a ; every therefore arises from a derivation upon and a certain term ; which term supposing series reduced like manner is a linear function of celtain of the .
Each later is derived \amp ; om later , or it may be from the same 's as an earlier ; viz. if the later is derived from , then the earlier is derived , it may be , from , or from , but so that there is not in the series any term later than And if , considering any as thus derived from certain of the , and in like manner .
each of these 's as derived from certain of the , and so on , we descend to any preceding series , it will appear that the is derived from a certain number of the terms of this series .
353 .
The quadricovariants , . . .
are of different orders , and consequently asyzygetic .
They form therefore a series such as the -series , and they may be represented by , . . . . .
Supposing to be of the order is of the order of the order of the order , and so on .
Those terms which are of an order greater than , are said to be of the form ( agreeing with a subsequent more general definition of W ) ; those which are of an order equal to or less than , are said to be of the form ; so that the earlier terms of the series are , and the later terms are ; viz. the terms taken in order , beginning with the earliest , are , . . .
354 .
By what precedes any particular is derived from certain terms , of the series .
This series , , , may stop short of the terms , or it may include a certain number of them , say , . . . .
The terms derived from the 's are in the sequel denoted by PROFESSOR CAYLEY'S NINTH MEMOIR ON QUANTICS .
355 .
Every covariant wlmtever is a form or sum of forms such as , . . .
; wnting in regard to any such expression , Zind .
, . . .
( viz. is the sum of all those indices , \amp ; c. which belong to a term containing the symbolic number 1 , the sum of all the indices , \amp ; c. which belong to a term containing the symbolic number 2 , and so on ) then each of the numbers , . . .
is at most , that is , . . .
may be any of them , but they cannot be any of them negative ; the degree of the function is , and its order is .
It is to be further observed that the form is a function of the differential coefficients of of the orders \amp ; c. respectively .
It follows that if , . . .
are none of them , the form in question may be obtained from a like form belonging to a quantic of the next inferior order by replacing therein the coefficients , . . .
by , \amp ; c. respectively : for example , if denote the cubic function , then the Hessian hereof is ; the like form in regard to the quadric is , which is ; and substituting herein for respectively , we have the Hessian of the cubic .
A covariant of derivable in this manner from a covariant of the next inferior quantic is said to be a special covariant .
356 .
Revertin to the form ; if , as before , \amp ; c. are each of them ; if thele is at least one index which is or that which , and if the order be , then the form , or any sum of such forms , is said to be a form or covariant W. Every covariant is thus a special covariant , but not conversely .
In the particular case , the form is which will be a form if , or , what is the same thing , , that is if the order be .
Hence , as already mentioned , the covariants of the degree 2 are , or else , according as the order is greater than , or as it is equal to or less than 357 .
Theorem B.\mdash ; If any covariant be expressible as the sum of a form and of earlier 's than itself , then forming the derivative , either this is not a form , or being a form , it is expressible as the sum of a form and of earlier 's than itself ; or , what is the same thing , , if it be a form , is ( like the original T ) the sum of a form and of earlier 's than itself .
Hence also every form is the sum of a form , and of forms derived from the functions , . . .
, say or , what is the same thing , every covariant whatever is of the form 358 .
The proof that for a form of the order the number of covariants is finite , depends on the assumption that the number is finite for a form of the next inferior order : this being so , the number of the special covariants of will be finite ; .
say these are ' is itself one of the series , but we may separate it , and speak of the form and its special covariants ) : the forms are functions of the special covariants , and hence every covariant whatever of is of the form ; but it requires still a long investigation to pass from this to the theorem of the existence of a finite number of forms V such that every covariant whatever is .
I pass this over , and reproduce only the investigation for the case of the quintic .
359 .
Starting from the assumed system of forms , , , , , ( jr ) , ( fi ) , , ( ji ) , ( pi ) , , ( ioe ) , ( 17 ) , , , say , the 23 forms , it is to be shown that every other covariant whatever of the quintic is of the form .
The special covariants are , , which are forms ; the only form is , so that instead of writing , every covariant whatever of is ; so that it remains to show that every form is ; or , what is the same thing , that if be any form whatever , then that ( Hi ) and ( Hi)2 are each of them .
360 .
In order to show that every covariant of a degree not exceeding is , it will be sufficient to show that the several forms ( Hi ) and ( Hi)2 of a degree not exceeding are each of them ; and if for this purpose we assume that it is shown that every covariant of a degree not exceeding is , then in regard to the forms and of the degree , it will be sufficient to show that any such form is a function of covariants of a degree inferior to 361 .
First for the form : we have ; and hence we see that ( Hi ) will be if only ( Ui ) is always .
In forming the derivative of with the several covariants , we may omit itself , and also the four invariants ( ii)2 , ( ir)2 , , , since in each of these cases the derivative is .
We have therefore to consider the derivative of with , , , , ( fi ) , , , ( pi ) , , ( ioe ) , , respectively : the first seven of these are each of them ; the remaining eleven are each of them of the form Now is a linear function of , and , that is is a function of covariants of a lower degree than itself .
362 .
Next for the form , we have , a linear function of ; and we hence see that be if only is always .
PROFESSOR CAYLEY'S NINTH IOIR ON QUANTICS .
In forming the second derivative of with the several covariants , we may omit as before the four invariants , and also omit the four linear covariants , 7 , ; we have therefore to consider the second derivatives of with , , , ( jr ) , , , ( ji ) , ( pi ) , lespectively : the first six of these are each of them ; the lemaining nine are each of the form .
Now is a linear function of , , and .
The first two of these are terms of the same form ; , as a covariant of a lower degree than , is , and hence will be if only is being here any one of the functions , and being any one of the functions , ( fp ) , , , ( fi ) ; ( pi ) ( ri ) ( ioe ) ( 17 ) .
363 .
For equal to any one of the last eleven values , the form is , which is , and is thus a function of covariants of a lower ; there remains only the derivatives formed with two of the functions , or of one of these with or 7 .
But these are all other than the derivatives , , , , , , ( joe ) , , , , and since , the derivatives containing will depend upon covariants of a lower degree ; there remain therefore only ( fj ) , , , , ( pr ) ; ( foe ) , , : each of these can be actually calculated in the form .
Hence finally , assuming that every covariant of a degree inferior to is , it follows that every covariant of the degree is ; whence every covariant whatever is , viz. it is a rational and integral function of the covariants U. 364 .
It will be observed that , for , the proof depends on the theorems , a linear function of , , , , do .
do .
do .
which are theorems relating to any three functions whatever .
365 .
I remark upon the proof that the really fundamental theorem seems to be tlJat which I have called theorem A. As to the forms it is difficult to see a such forms are to be considered , or what the essential property involved in their definition is ; and in fact in a more recent paper , ' Die Simultanen Systeme binaren Formen\ldquo ; ( Clebsch and Neumann , t. 2 ( 1869 ) , see p. 256 ) , Professor GORDAN has modified the definition of the forms by omitting the condition that the older of the function shall exceed ; if it were possible further to omit the condition of at least one index being , and so only retain the conditions , \amp ; c. , each of them , then the essential property of the forms would be that any such form was a rational and integral function of the special covariants formed , as above , by means of the quantic of the next inferior order .
And moleover , as regards the theolem , there seems sometbing NDCCCLXXI .
rect and artificial in the employment of such a property ; one sees no reason why , when a system of irreducible covariants is once written down , it should not be possible to show that the derivatives of with the original quantic are each of them , instead of having to show this in regard to the derivatives of with the several covariants : as legards the quintic , where there is a single covariant , the quadric function , there is obviously a great abbreviation in this employment of in place of ; but for the higher olders , assuming that the proof could be conducted by means of the quantic itself , it does not appear that there would be even an abbreviation in the employment in its stead of the several covariants .
The like remarks apply to the proof in the lastmentioned paper .
I cannot but hope that a more simple proof of Professor GORDAN'S theorem will be obtained\mdash ; a theorem the importance of which , in reference to the whole theoly of forms , it is impossible to estimate too highly .

