First Memoir on the Theory of Analytical Operations .
By the Rev. R. MURPHY , M.A. Fellow of Caius College , Honorary Member of various Philosophical Societies .
Communicated by J. W. LUBBOCK , Esq. F.R.S. Received December 8 , -Read December 22 , 1836 .
§ 1 .
1 .
THE elements of which every distinct analytical process is composed are three , namely , first the Subject , that is , the symbol on which a certain notified operation is to be performed ; secondly , the Operation itself , represented by its own symbol ; thirdly , the Result , which may be connected with the former two by the algebraic sign of equality .
rrhus let a be the subject representing , we may suppose , some quantity , b the symbol for multiplication by b , and c the result or product ; for greater distinctness let the subject be enclosed in square brackets , the analytical process in this case is [ a ] b= c. Again , let x ' be the subject , 4 .
a symbol of operation denoting that x must be changed into x+ h , and ( x + h ) ' will evidently be the result , or [ xn ] A ( x + h)8 .
Again , let ax be the subject , Aa symbol of operation , which indicates that we are to subtract the subject itself from that which it becomes when x is changed into x+ h , which is usually called taking the finite difference , then the result is +h Axa a , or [ ax ] A [ ax](ah 1 ) Lastly , let d , denote the operation of taking the finite difference , and after dividing it by h , then putting h= 0 , which is the same as finding the differential coefficient of the subject , which we may suppose represented by u , then [ u ] XX d2B The operations written as above are monomial , consisting of only one term and polynonmial operations give the sums or the differences of the results of the respective ruonomials of which they are formed , according as these monomials are affected by the signs4 + or- , Thus if 1 as an operation ibe understood as the multiplying of the subject by unity which leaves it unaltered , and the symbols g , A have the same signification as in art .
I1 then [ u ] ( 4i-1)= [ u]A [ u ] ( A+ Q)= [ u &gt; A , where the subject t is any quantity whatever .
When general relations , such as these , between different symbols exist independ &amp; ently of the particular value of the subject , we may abstract the consideration of the latter , and the sign between symbols of operation being understood to indicate that they are universally equivalent , the symbols used in art .
1 would have the following relations independent of the subject .
A , =T , A=A+1 A 4v-1I when h is put = 0 .
3 .
A comnpound operation consists of a series of simple defined operations , monom"ial or polynomial , the subject of each individual in this series after the first being the result of all the preceding operations .
Thus [ x2 ] a 4 , A d , denotes that first 2 must be mntvltiplied by a , which gives a -2 ; then the operation 4 , v , which denotes the p uttitig xIh for a , gives a ( x + h)2 ; next the operation A will give a ( x +2 h)2_ a ( x + / )2 , or a ( 2 hx+3 h2 ) ; and lastly , the symbol of taking the differential coefficient relative to x gives 2h a : the final or complete result is therefore X2I a4A dw =2h a. When all the symbols in a compound operation are exactly the same , then for abridgment the whole operation is represented by writing an index to the right of the symbol for the simple operation over it , this index expressing the number of times the simple operation is repeated .
Thus [ xg ] 4,3 [ x 4 , .
, = [ ( X + h)2 ] ( +2 h)2 ] 4 ( x +3 h)2 .
But when the simple operations are different , they must be written consecutively in the order in which they are to be performed , unless that order of arrangement by the mutual relations of the operations should be indifferent .
Thus [ Xi]:1 r4 = [ x + ` ] 4 , ( x + h ) n+I [ X " ] 4 , i= [ ( ' + It ) ] Xv _i ( iv + p4 ) " But M1as lax x =a(x +I)t E'.4 a [ ( x + h ) ] a=a ( x + h)n .
in the latter case the order of the operations is indifferent , because the operation 4 does not act on the multiplier a , and for the contrary reason the order of x4 isfixed in the first case .
Operations are therefore relatively fixed or free in the first case a change in the order in which they are to be performed would affect the result , in the second case it would not do so .
In a compound operation any part of the symbols may be taken conjointly with the subject in the square brackets , their result being the subject for the compound operation of the remaining symbols .
Thus [ E2 ] a 4 , A d , = [ azl2 ] %4 A d ?
2 .
4 .
Linear operations in analysis are those of which the action on any subject is made up by the several actions on the parts , connected by the sign + or - , of which the subject is composed .
Let p denote the operation of multiplying by a quantity p , then [ a + bJp= [ al p+ [ b ] P ; this operation is therefore linear .
Let 4 , denote the operation of changing x into x+ h , then if f ( x ) , P ( x ) be any functions of x , we have [ f ( x)+ p ( ) ] --=f ( x +h ) + p(x h ) [ f ( x ) ] + [ p(x ) ] kp which shows that 4 , is also a linear operation .
Let X+t represent the subject acted on , and 0 , 0 ' any linear operations , then [ X+$1 ( 0+d ) = [ X+ ]d+ [ X+]0 = [ X ] + [ R + [ X ] 0 ' + RI 0 ' = XI ( 0+ )+X ( 0+ ) ; hence polynomial operations of which the parts are linear possess themselves the same character .
Thus A the operation of Finite Differences is linear , because A= 4-1 , the operation 4 , of changing x into x+h and the multiplying by unity being both linear .
Also [ X + e ] 0 f= [ X 0+t 0 ] 0 ' [ X ] 00+ [ X 0 0 , which shows that the compounds of linear operations are also linear .
The operation of taking the differential coefficient is therefore linear , for the opera 2nB2 tion A of finite differences and that of dividing by h being both of that character , the compound A. h is generally linear , and must remain so in the limiting state when h vanishes .
Hence every function of a linear operation is itself of the same class of operations .
5 .
The composition of polynomial linear operations is effected in the same manner as algebraic multiplication ; with , however , this peculiarity the order of the compound operations , when not relatively free , must be strictly preserved .
Thus let 0 , 0 ' , ' represent linear operations : then lu ] ( 0 + 0 ' ( )= Lul ( 0 I+0'+ 0 : ' + 0 ' ' ) ; for i ' being linear will act on each of the parts [ u ] 0 , [ uJ 0 , which form their subject .
Again , when 0 , 0 ' are relatively fixed , [ u ] ( O + 0')2 will not , as in algebraic involution , be identical with [ u ] ( 02 +20 0 ' + 012 ) , its correct value being [ uJ ( 0 +00+00+ 012 ) ; which , however , is the same as the former expression , when 0 , 0 ' are relatively free ; for then 0 ' =0 0 ' .
Similarly ( 0 + 01)3 = ( 03 + 0 ' 02 +0 0 ' 0+ 012 0 ) + ( 02 0+ o00+0 012 + 013 ) ; or putting 0(2 ) 0 ' for the sum of the compound operations in which 0 twice enters , and 01 and 0 , ( 2 ) for the sum of those containing 0 once and 01 twice , this may also be written ( 0 + 03)= 03 + 0(2 ) 01 +0 01(2 ) + 0'3 and employing a similar notation , we shall have , when n is a positive integer , ( 0 + 0 ' ) " = 0( ) + O(n-1 ) 0 ' + 0( ?
-2 ) o(2 ) + 0(2 ) 01(n-2 ) +0 1(4-1 ) + 01(n ) .
The term 0( " I ) 0 ' in this formula is the sum of n terms , formed by placing 0 ' at the beginning , at the end , and in all the n2 intermediate positions of the expression o0 0 .
( n -1 ) times .
Similarly , by the known theory of permutations , 0(n-2 ) 01(2 ) is the sum of terms &amp; c Hence when 0 , 01 are relatively free , we have ( O + 0 ' ) " = On -J n ( in-1 fi +n ( n on-'201 +0 et-1 + Awn , 6 .
The following theorems are immediately deducible from the general expansion of the preceding article : Since A= 4-1 , therefore ^nn n-1 ~~~~n(n -1 ) n2 n ( n-l)nt n-3 tA =()- ) = 4H nt2 ++ 20"-I2 ; +*** 1 ; or , if we introduce the subject f ( 4 ) , and observe that jlf ( x ) ] An =f ( ?
+n I ) , then Anf(a ) =nf( +nh)--nf{a ( n-l).h } r Again , 4'A + 1 , therefore =( &gt; + I)n= I +nA+ j IA2+ ( n -7a n or , introducing the subjectf ( x ) , f(x+nh)=f(iv ) +n Af(x ) + 1 ) l)(N- ) + ... ( .
) Again , I= 4-A , therefore A~n = ' , ; n-I A nn-1 ) n-2A,2_ n(n1 )(n ) 2 ) .
A3 ... ... or f(iV)=f(r+nh)-nAffx+(n--1 ) I.l } ( AII A hAn(nf-1 ) ( n + 2 ) 3 { X+ ( n3 ) h } , &amp; c ( k In the expansion ( II .
) put nh= k , or n= : therefore f(x +k ) =f(X ) +kf ( IV ) + ic ( k-(X ) +k k. ( Ajf(i ) , &amp; Now suppose n to increase infinitely , k remaining constant , the quantity h , which is the increment of xA , diminishes infinitely , and the operation y in the limiting state when h vanishes , becomes dQ .
Hence f ( x + k ) =f ( x ) +k dxf ( x ) + df ( x)+ 1 d3f( ) + , &amp; c ( IV* ) The expansions ( II .
) ( IV .
) are TAYLOR'S theorems for the development of functions by means of their finite differences and their differential coefficients respectively .
Again , if h be written for k in the expansion ( IV .
) , and the subject be omitted , it becomes hk dx 1t3d3 + hdx+ +j d+ , &amp; C. ( V. ) and A hd +-l +p1 , &amp;c. § 3 .
7 .
The expansion given for the operation 4 ' , of changing x into x+ hI , possesses remarkable properties , which we propose to develop in the present section , from the importance of the theorem of TAYLOR , which it expresses .
Representing , as usual , by 4 ' the operation of changing x into x+ hI , and by 4 " that of changing v into x+h ' , the quantities hI , h ' being independent of x , and , lastly , denoting by By the operation which changes x into xv +h+ h ' , we have obviously the identity and putting for these symbols their expansions found in the preceding article , we get { 1+hd + C++ _+ &amp; c } { l1+h d 2+ h+ dj3i+ , &amp; c } l hd+ 1.4 1. .
+ &amp; C.x1 ... S ) ' -1 +++ ( h + hk)d +(Ajd +hkij j ' + &amp; C. a relation which may be verified by actually compounding the two polynomials of the first member .
Now in this act of verification the operations hd hI ' d have only such properties as are common to any two linear operations which are relatively free : hence if 0 , 01 represent any such operations , we have generally { 1 +0+ A.2+ y.2i , + , &amp; c}* { 1 + '+ i+ i+ + , &amp; C } _I + t ?
0 + i ) + i(6+ 612 +( + ' ) I.Q-2 and it is easy to extend an identity exactly similar to any number of operations which 6712 ?
, t3 are all relatively free ; for in introducing a new polynomial , I+ 01 +i+ 13 2 + , &amp;c. , we have only to regard 0+ 0 ' as itself a free linear operation , and therefore the result would be I+ ( 0 + o+ or ) + ( d + 0t + ill )g , +(0 + $t + &amp; 1)3 +C 8 .
If the subject be a function of two variables , x and y , then using i to denote that x must be changed into x+ A , and i ' .
, that y must become y+ k , these opera .
tions are relatively free , it being of no consequence which operation is first performed ; therefore the operations AZ , Ay of taking the corresponding finite differences are also free ; from whence , lastly , the differentiations relative to x and y , represented by di , die , must be of the same character .
Now since A1+h d+ h+ gi9 + had3 + &amp; Aix =I + hdx+-ix 1 + , &amp; C. and y=I+ kdy + k2 dy2 ++ &amp; c therefore , by the general identity of ( 7 .
) , we have ( h d4 +k die )2 ( k d+k die )3 54- , x lp + ( hdx+ k die )+ + &amp; C -- ) .
And now introducing the subject f ( xy ) , and expanding the terms in the right member of this identity by the formula given in the preceding section , it will become , in the common notation , + h , y +k ) =f( , y ) + -f( x2/ h ?
d )j~ + dJ &amp; Cy ) 1 dx ~~ dx~.-+ d x3 1.2.3 d/ f q+ d* h+ d( , Y ) hk +df ( xy ) h1i &amp; + *f ) .
k+ d+ --dx2 die *fc2 d2 f ( x y ) kg d3 fX(Zy ) h Vf + ~~+ '--1 &amp; +d3 , f(x , y ) V If a third variable z becomes z+I by a third operation q.i " , then the actual conposition of the equivalent polynomial 1+I dz +d d3 + &amp;c. gives + ' + ' +=1+ ( h dx+k die +Idz ) + ( A dx+k die +ld92+1 4 hdw+ kd , + die + , &amp;c. ; and so this method may be extended , whatever be the number of variables .
9 .
Let 0 denote any linear operation , and make 0= I +0 ?
+ ++ then ) is itself a linear operation ; and if in ( 7 .
) we put 0= 0= _ = , &amp;c. , and suppose the number of these operations to be n , we have by that article =1 +rIn 0 + , 1+ rI 3 z3 Again , suppose +sn+.2m2+ +.2m3+ &amp; C m being any positive integer , we have by the same principles 02 I0 .
]+n 3+ it2 02 n3 03 Hence if p denote an operation which , repeated m times , gives Q and in which wre 1n shall employ the notation em , then fp denotes the same as O'9 , the latter being the operation which , repeated m times , is the same as O ' .
With this meaning understood , it follows that Om= I+ n10 0+ no0 !
2 7030 n3 &amp; 3 Lastly , if we put I ?
02 30Q= l-n 0+ 2 .
3+ ' + &amp; c*5 compound this by the formula of ( 7 .
) with the operation @ -l n n=1l 23 03 we have Q on l+ ( n-n ) .
0+ ( nln)20 + &amp;c. =l or if we introduce any subjectf ( x ) ; eQ , A being necessarily free , [ f ( X ) ] fl e2 [ ( X ) ] e } fl =f ( V ) .
Hence fA , ( p ' are mutually inverse operations , the action of the one on the result of the other restoring the original subject .
f2 may be represented by p-n .
attaching to this symbol the meaning here assigned .
If n represent any quantity , positive negative , integer , or fractional , understanding the conventional notation by the definitions laid down , it follows , that if 01+n 4+y { ; + [ 3+ , S &amp; c ) , then shall oX-z +a+f+ l2 &amp; C. 1 0 .
Suppose d to be simply the operation of multiplying by unity , then '3= 1+1 +j+y-g+ , &amp;c. ; and putting as usual s for the sum of this series , ' represents the operation of multfi plying by s ; put therefore in the formulk of art .
9 .
1 for 4 , and X for By and then +n+ -+ l1 , + , &amp; C. The properties of this series when any way involved are common , as has been seen , to those in a series where 0 , any linear operation , is put for n , and therefore we may write the purely symbolical identity 6 01+0+1 _____3_ where 0 may be an imiaginary multiplier , or any species of linear operation Thus if 0I d= and As denote the operation of changing x into a+ Ia , we have A= A~~~~h dx Similarly A(_x+k dy all of which are proved by the formula of art .
8 .
Ha Hd ving seen in the course of the investigations of this section the signification of the indices of operations when fractional negative or even purely symbolical of linear operations , it is easy to prove by similar steps that in all cases where 0 , 0 are relatively free , ( 4 +t d ) =.+ n 41+L ( 71 p d2 10 2_ &amp; for since ( 6 + 4ry ( 0 + f)'= ( 0 + O)n + n , it follows that the composition of the polynomials t0 +n ra+1s0 2+ &amp; c }* { m !
Amv-1A , ,m(fll ) Am-2At2 _ &amp; -(n-1).dn-2(n ?
m ) ( n r Mr-ld+ 1 ) - ) M2I25 = S4tn+ ( n + m ) O +m-1 + 1n + *_ Q+ M.+ 2 &amp;c. ; and since nothing in the actual verification of this identity depends on their being integers , for which case the expansion has been proved , the identity holds generally , and therefore if m= n , and we take p , such polynomials , we have { 6+n~1OI+1n(n1 ).dn-2012+ ... P~0np+ p10 ( n)(Onp ')0n { f +n0+l(- ) f2f2J = nPrnpff~-lD + ( P 2_ 02 { f + nJ 0 + .
}_ 4 ?
+nqrq-1 { ) + &amp;c. Put n=qPqqP { p P Y } + di +* * and in general if [ U ] -=y then [ y ] O=zt and therefore [ y ] O ' 0= y , the compound operation Ol 0 or 00 being equivalent to no operation .
To invert a compound operation we must invert the order as well as the nature of the component operations , which rule must be strictly observed when the latter are relatively fixed .
For let [ X ] =y [ y ] 0t =z and therefore [ x ] v D ' z. Then [ z ] 0I ' =y [ y ] = =x and. .
[ Z ] 0-1 D-1 = a , which proof is applicable , whatever be the number of the component operations .
If all the component operations be alike , we have then but to change the sign of the index to obtain the inverse , thus if [ U ] t ' y , then [ y ] -n = as in the last section .
13 .
The consideration of inverse operations leads to the introduction of the appendage , which when the operations are linear must be annexed to the result to give it the most general value of which it is susceptible , for the inverses of such operations are themselves linear ; thus if d be a linear operation , [ X + ]= X]0 +[ d. Put [ X ] 0= X1 , [ a ] 0 =1 hence rX1 + a1 ] 0=X+f [ X1 ] 01 + [ NJ ] d- ) , which shows the linearity of 0 ' .
Noxv suppose the nature of 0 to be such that [ P ] 0 0 , the subject P being thus in some way connected with the nature of the operation 0 , then if we , suppose [ X ] 0= y , we have also [ X + P ] d= y , hence [ y ] 0 ' X+ P , this being the same as writing y+0 for y , since [ 0 ] d1 P. The appendage therefore in a linear operation is the result of its action on zero ; P will express aformn , but its magnitude must be susceptible of an infinity of values , that is , it contains arbitrary constants which enter as multipliers , for if a be such a constant , we have in general [ X ] a0= [ X ] 0 a ; and supposing X= 0 , we have [ O. a ] 0= [ 0 ] 0 a : therefore whatever particular value may be assigned to [ 0 ] 0 , a more comprehensive value is attained by its arbitrary multiplication by a. A multiplier is the most general form in which the operation represented by a can enter when X is a function of but one variable ; but it admits of other forms more extended in cases of several variables , as may easily be perceived : thus [ fy ] representing a function independent of x , then [ fy ] By f/ ( y ) , 4 ' , denoting the operation of changing x into a+ h : hence [ f ( y ) ] A. = 0 .
Then if X be any function of a , and d be any particular value of [ X ] A1 we shall have more generally [ X ] -A 1= which includes the former ; since , the form off ( y ) being arbitrary , we have v alone , amongst the infinite number of values of +f ( y ) .
In compound operations , the appendage obtained by the first simple operation becomes a new subject for the succeeding operations , each of which may in like manner introduce a new appendage .
14 .
The operation 4iS taken directly or inversely , is incapable of introducing any appendage : for suppose [ 0 ] ?
' =9 ( x ) , then [ p ( v ) ] 4 0 , or P ( xv + I ) = 0 , which identity being general , we may put x instead of x+ h , which gives 9 ( x ) = 0 ; from which it also follows that [ 0 ] 4 7n = 0 .
To find the appendage introduced by d7 ' , suppose in the same way [ 0 ] dijl = c(x ) , hence [ I ( x ) ] dz 0 , therefore [ 9 ( x ) ] dZ2 =0 [ 9 ( x ) ] dx3 = 0 , &amp;c. ; and since 9(x+ h ) ( x ) +hdx+ 1 .
d2 + hence we have p ( x + I ) = ( x ) , and h being arbitrary , we may put it = x , therefore p ( x ) =o ( 0 ) , that is , i ( x ) is constant relative to x ; if therefore C be any arbitrary quantity independent of x , we have [ 0 ] di , ' = C. Again , [ 0 ] dS2 = [ 0 ] dx- ' dz,1 = [ C ] dj- ' + [ 0 ] dx- ' , as above stated ; but since [ C]I dZ = C , and [ 0 ] dj ' = C ' , any constant therefore [ 0 ] dj2 =Cx+ C ' , and in general [ 0 ] d -n =A Xn-1 + As n-2 + A3XAWNS ?
.,. .
+ An , where Al A2 ... .
An denote constant multipliers .
Lastly , let [ 0 ] Ax1 = ( x ) , or [ 9 ( x ) ] Ax _ 0 , therefore 9 ( x + h)=9 ( x ) , and by TAYLOR'S theorem ( dividing by 9 ( x ) , 0= p:(X ) ) A+ 1(P ( x ) h12 + &amp;c. where If ( x ) 9 " ( x ) &amp;c. are the differential coefficients of 9 ( x ) ; this identity being independent of x , the latter quantity must disappear from the series : put therefore 9 ( ) ) , n being independent of x ; hence e~~~~~~I / ( ) l(x ) =-he * ( x ) ( X ' ) = ( x ) , &amp; C therefore h ' 2C2a also since ( x ) ( x + h ) change x into x+h ( x + h ) ( x +2 h ) and ?
n n n/ generally @ ( x ) @ ( x +m h ) , m being an integer ; and since q ( X ) =Ca ms satisfies this equation , it follows that +n+2n+3 n , &amp;c. satisfy the equation ( a. ) , the complete appendage will therefore be nxV-i 2nxVi -n xV-I 2nx/ Al A +A2AS + ... . ?
B1 Ah + .
the number of constants being infinite .
15 .
When the simple operations which compose a compound one are relatively firee their places are transmutable , but when fixed a mutation of places will require an alteration in the operations themselves .
Let As denote the changing of x into x+h as before , and let if be any operation affecting x , or , which is the same , fixed relative to Ax , then considering dX as a subject , put [ Of ] xd = d ' , and if the compound operation [ u ] Ox Asx be proposed , its value by transmutation is [ u ] Ax d ' , for in the first compound operation Ax affects all the preceding symbols as forming its subject .
Again , let [ u ] dx A= y be proposed for transmutation , we have y= [ u]dx(j41 ) [ u ] ( .4 .
' d0 ) and putting Ax +1 for 4 's and 6x Ax for the finite difference of d considered as one operation , we have y =[u ] ( A d ' + ox Ax ) = [ u ] dx Ax .
Lastly , divide this identity by h , and then put h= 0 .
When x becomes dx , and O ' becomes &amp; x , we get for the transmutation of O , dxn [ u ] 6x dx = [ u ] ( d Ox + dx ) .
These formulae of transmutation separated from the subject are respectively Air ?
= Ax dx AS dx x=x Ax + ox x Oxdx= dmO , + do When Ox is constant or not dependent on x , then d ' a =_ xx=0 Ax Wx = O. and these formulae will then express merely that Ox , xn &amp;c. are relatively free .
For example , let 0 , simply denote a multiplier , then O. , 4 will be Od + also a multiplier , dx Ax= + , f , which may be represented by 'd and dxa = the limit of '+ A -when h vanishes , or the differential coefficient of 0 , which may be denoted by 0 ' , C then the general formula becomes ; U ] Ox +-kp= U ] 'kx Ox +h [ t ] OX AX = [ U ] ( Ax ax h +%ox ) [ u ] Ox d= [ u ] ( dx Ox + '4)e Again , let the subject bef ( xy ) , and suppose now 4 to be the operation of changing intoy +P ( ) , then or x+ his the operation of changing y into y+p(+ h ) , x AX or '4 is the operation which changesf ( x , y ) into f { x , y +P ( x + h)}-f { Ixy +p ( W ) } =f { x , y+r ( x ) } +r dPI ( x ) h+ &amp;c. -f { xy +q ( ) } and therefore 4 d , will give f'+ ' and is equivalent to 4 die @ ' ( x ) ; in this die n case we should have f ( x , y ) ] dd= [ f ( xi y ) ] ( d. Ox + Ox die 9 ' ( x ) ) , which result may be also deduced by putting for -its equivalent symbol s ( T ) which gives ( 6id-dy ( x ) die p " ( x ) Ox dy4 ( x ) .
, where p ' ( x ) is written for ( This example shows how operations may themselves be the subjects of other operations .
16 .
We now proceed to consider the transformed values of 0+r ' n4 ) dx x when n is any positive integer .
First suppose therefore dX -+X2 = PX -i _~ ( PX 4-w Now PX regards px solely as the subject of the operation 49 and by the first formula ; therefore p ' ?
.42 4.4 ~2 ASX Awe =X w4s and in general if we suppose then but ___ t~~~t ' . ?
= ?
- ' by the first formula ; therefore n4 , n4 1 ?
n44= 4$ This general formula may be more readily deduced by considering that '4 &lt; is the operation of changing x into x+n h , and consequently 4n may at once be substituted in the first formula for b4 .
Secondly , Px Ax= AOx AsX +O Ax ) therefore dX AX Ax dx As tZ + d7z .
Ax ) but d ' ?
Ax = Ax .
0 4 , A2 + Ox 'S A , writing ex'4 for O ; and similarly O p\x AX = x ' * 4i A+xA X2 whence Kx apt A X2 .
fx 2+2 Ax 0 gv i~ Ax + tx / \20 generally suppose ox AX nil = AX n-1 0 .
4 &lt; 'x1 + ( nI ) Axge2 .
d 4 , n-2 AX ( n 1 ) ?
xn-3 axn-3 42 Now if we write 0 4 , x for dO in the fundamental formula , we have OX ASm AXS A. O4 1AS + 4 ; 1X AX each term when we put for m , n 1 , n -2 , &amp;c. successively , will thus be divided into two , which being placed in two distinct lines will give egO a~=AiS .
Oyted+ ( nl)Ax'~'* OR ~ Axn-l +(n)(n-Q)A n-2 X AX nA2 2 &amp; C +a ni f } mn-1 'd o( oA -2e Ox q -i2__ &amp; +Ax OX Yx AX + i(n AX 0 n-2A2 + &amp;c. =.A,70 , ,4 n+ nAx O~ 4-x A+ ( n-.-1 Axn-2.0x n-2A,2 + &amp; C. which is the general formula sought for .
Divide now by Pn and observe that becomes d. and i becomes unity as a multiplier when h= 0 , hence the third general formula 0 , , d7= d 0+ +nd1.O d+( ).dn2 , T2 ?
&amp; c which when 0 , , represents quantity is the theorem commonly called LEIBNITZ'S .
17 .
We next proceed to investigate the formulae for negative indices .
First since '4 , ' denotes simply the changing x into x -h , we may write + : , for 4ix in the first formula .
Therefore 0x ~.47 ='x10 , Yx more generally ox AIPx =4'xOx 4pxn Secondly , since OAX = AOXkp + OX A therefore A Ox AxOx+A -oi A Put now 0$ .
1 fr Ox ) and therefore Ox for Oxx hence Aj ox 4 , &gt; -1 Ax = ox + AX- , tO4 A0 therefore 0 AX- , -A X OX b '-A'O4 &lt; A OX ' .
Put Ox 4Si ' for Ox , thence we have = -A .
Ox +A~ -2 xx A -2 A2zA 1 or we continue this process indefinitely OX AA X-1 -A .
ox ig X0 AXitZ + AXin px A ft ; -X 2 &amp; C. which is the same as the general formula for 0x Ax ' when n=_ Again OxAx S= A1 0x1x-1 A -1 _A* o , &lt; 4i-1 A -2 but Awe ow-1 Ol awe _A -1 A-2 00 A-2 _A2 $OX a-2A Ax I A7 ~ and X___ _X ________ A -.0 Yo Ax A _1 =A2O x-2 A_A2 Ox -2 A2A - ' Hence XX_ Xt z\ , , = .
aX .
n , , + ?
_2 2 A , *2 , b , -2 As AD _ -1 w , ^ ox AX-12A1X ox / x 2XOX X XAxA7+ x2 , x &lt; AXA and in a similar manner it is easy to prove generally in a terminating series Ot SE -n =A -n o , -nn -fn -1 Af .
X AAI n =nn 2 .
A2 - &amp; C. XXXXX or in an Infinite series , xZ -A ( .
ln-(n+l ) .
-+1)+ A(+2)o4 , Q-(n+2)A 2- &amp;c. Thirdly , divide Ax by hk and then put h 0 , whence 0 I ' d'Q-d'0 dd ' xx x1 .
x -d2 Oxdxx+d OXC Ox x 1_dx _n - , X dx O , dO dx2nO &amp; C Ot d-zd~ , O Before proceeding further in the search of the fundamental formulae for the transformation of operations , we shall exemplify the theory which precedes by inverting binomial operations and applying the results to some simple cases .
Let 0 , 0 ' denote two linear operations relatively fixed or free , and let us seek the value of ( 0-0 ' ) Put ( 0 0 ' ) 0 ' + t , ; the latter being the difference of two linear operations must itself be linear .
Hence 1 ( 0-1 + 41 ) ( 0-0 ' ) =1 0f +1 ( 0 -0 ) therefore n1 ( 0 0 ' ) 0 0 ' .
Similarly put -t = 0-1 o ' 0-1 + 42 which gives 41 ( 0-0 ' ) = 0 ' 0 ' ( 0 ' 01)2 +n ( 0 0 ' ) whence 42 ( 00 ' ) = ( 001)2 so again put ( 0-1 01)2 01 + 13 -3 ( 0-1 0')3 0-1 + 14 &amp;c. = &amp;c. We thus obtain ( 0 0)= 01 + ( 0-1 0 ' ) 0-1 + ( 0 ' 01)2 0-1 ++ ( 0-1 o0)'-1 0- " + E , where Y7 represents the compound operation ( 0 ' O')n ( 0 _ 0 ' ) ' .
The same formula continued to infinity would be obtained by first putting 0- ' ( 100 ' 0')- ' for ( 0 0 ' ) ; and since the operations represented respectively by I and 0-1 0 ' are relatively free , we should have by art .
I 1 .
( i 0-1 0 ' ) ' =1 + 0-1 o ' + ( 0-1 01)2 + &amp;c. ad infin .
When 0 .
, 0 ' are relatively free the theorem becomes ( 001)-1 = 0-1 + 0-201 + 0012 + 0-40'3 + ... .
0-n0'n+ 0-0'i ( 0 0')-i .
19 .
For a first example suppose AZ to denote the finite difference , on the supposition that by the operation 4y the quantity x is changed into x+ h. Then Aod is the finite integral , and in the usual notation of analysts is denoted by do we have therefore [ f ( x ) ] N= [ f( ) ] 1 ) = [ f( ] 1 ) + =f(x h ) +f(x-2 h ) +f ( xr-3 h ) +*.X +JfS(ul)h } + Ifs(x-nh ) ; where it may be remarked that if Ox be an integer , the final terms of the series would be vf ( 2 h ) +f ( h ) +f ( 0 ) , at any of which , if we suppose the series to stop , its finite difference would be obviouslyf ( x ) .
For the next example suppose the subject to be f ( x 4 y ) , and that by the operation 4h , x receives an increment h , and y the same increment by the operation 4 .
, then it is obvious that [ f ( x + y ) ] ( 4i- , 0 , therefore [ f ( x + y ) ] ( Ay - .
)= 0 ; hencef ( r + y ) must be included in the general value of [ 0 ] ( As , 4 ) Now [ 0]1[0 ] { A + A , -2 A , + iAYS A2+a A , , + &amp;c. } and also [ 0 ] A=p ( x ) an arbitrary function of a. [ 0 ] Ay2= ( &lt; ).Y , omitting the appendage , which being a function of x would not vanish with y , and which in the succeeding terms , if included , would only gene .
rate another series similar to that now formed from p ( x ) , and consequently in the present case not add to its generality .
Again , [ 0 ] Ay75 = Afor.y * ( y ) ] ( y + h ) , y y ( y ) _2 Similarly Hence [ 0 ] ( A Ax)-1 =P ( x ) +Y At p ( x ) +.A2 ( ) +y Ah ) ( _vh ) .
A3 p ( X ) + &amp;c. ; and sincef ( x + y ) is included in this general expression , the particular form to be assigned to the arbitrary function p ( x ) is known by making y0 which gives p ( x ) =f ( x ) in this formula the ordinary notation has been employed .
Suppose , for instancef ( x ) a= and h= 1 , then Anf ( x ) -a ( a -I ) ' ; therefore M+Y-atlya2.(a~l ) Y1la~-)2+y(Y9)(17 ).ax~a 1)3 , &amp; C. or putting a1+ b , and expunging aX from both sides , ( I +b)-vI+ , y. b +.v ( .v2-')(Y-2 ) .
P , &amp; C. which is the binomial theorem , whatever may be the value of.f For the next example let us take the same subject , f ( x + y ) , and let dx , die denote the differential coefficients relative to x and y ; then since [ f ( x + y)]gy-A iMrcccxvii .
2 D put I= 0and AX becomes d being similarly die ; therefore [ if ' ( + y ) ] ( die -dx ) 0 hencef ( x + y ) must be included in the general value of [ 0 ] ( die -d ) -I But [ 0 ] ( dy-do ) [ 0 ] ( d -1 + d-2 d+ de d2+ d-4 d 3+ &amp; C. ) Now [ 0 ] d-=(x ) an arbitrary function of x ; therefore [ 0 ] d -2 = ( x ) y omitting the appendage for the reason abovementioned also [ 0 ] die _s ( x ) .
since [ y2 ] d 2y ; and [ 0 ] d7y =P ( x)* .
e substituting we have LO ] ( die dx ) _ ( x ) +yd Lx + , Y d9 v y3 ( ) +f df ) &amp; C. employing the common notation the form of p ( x ) is determined by making y0 which gives p ( x ) =f ( x ) therefore PX + , Y ) +f W + .
if W+ fg W/ ( ) , where f ' ( x ) / ( x ) , &amp;c. are the successive differential coefficients of f ( x ) this 's TAYLOR'S expansion .
If we putf ( x ) =a and for the limit of a1 write log ( a ) , we get from this ay= +y .
( a ) +Y ( .a)2 + , &amp; C These examples suffice to show the mode and use of the inversion of binomial ope-r rations .
§ 7 20 .
To return to the general theory , suppose 0 , , x to represent three operations connected by the equation 0 : =i a , where the subject is omitted , the identity being supposed general ; the symbol s represents an operation which may be said to be ntermediate to those designed by 0 , z. If either of the extreme operations 0 , z be given , and the intermediate i be also given , the other extreme may be readily found for 0 = : ; X so and z= C'1 0:.s A remarkable property of intermediate operations is that they are also intermediate between any operations which are the same functions of the extremes .
For let 01X then performing the operation X 0oZ -_ X2 put now for iz its equivalent operation 0 1 , and we have 02t I *21 Similarly if we suppose 0n-1 = ?
then on-1 X= Ie but 1z 0 : ; therefore I , = xn .
Again , suppose the subject in the last equation to be one on which the operation 0 ' has been performed , then that equation becomes X -fn , Xn or con l &gt; S-An .
Again , suppose K an operation satisfying the equation n=i K. We have by the parts of the proof already given in this articles or 6= , K =K hence nn From these premised equations it follows , that iff ( 0 ) !
( z ) represent the aggregates of any similar powers of the operations 0 , z , with the same coefficients , we must have generally f ( 0 ) .
I if ( ) .
By this theorem , iff ( 0 ) be known , f ( z ) can be found , supposing that we know i the operation intermediate to 0 , z. 21 .
We shall now apply this theorem to cases where 0 , i are given , and therefore n known , as above shown .
Let one extreme 0 represent the operation of differentiating relatively to x , and the intermediate i that of multiplying by 8a " , then we have d ~a X.a X U~~~~ therefore ax but by § 5 .
0x dx dx 0 , + Ox dxg therefore ?
-as d =dw --aSa ( dx a ) Ca x Hence da =x the extreme required .
Now if O si i , it has been shown that f ( 0 ) .
s= sf ( 4 ) .
In this case therefore f(dx ) .
sax = axf ( d-a ) To find the intermediate operation between dx +b and dx + C ; b and c not containing x. Putf ( dx ) = dx +b in the last identity , and ba=c we then have ( dx + b ) .
6(b-c)x = 6(b-c)x ( dx + c ) f ( dx + b ) .
f(b-c).x s(b--)xf ( dx + c ) .
22 .
Suppose the intermediate operation * to denote s ' P considered as a multiply , P being a function of x , of which the differential coefficient is P , and 0 to represent dx as before , it is required to find z. Since d$ =P* therefore dx But -P 'dx xs-%p+ 6P dx dz , -'P p ?
P s= ( d -P)-P Hence z~~d P. Corollary ; f ( d * aP = Pf ( dx -P ) .
And if 'Q be a function , of which Q is the differential coefficient , we have in like manne f ( dx ) *Q s'Qf ( dx -Q ) hence Pf(dx P ) .
=5 if ( dx Q ) s 'O or ( dx P ) * QP _ 60Pf ( di ) that is , s'(0P ) is the intermediate to the operations dxP dx-Q .
23 .
Let Corollary ; f ( d , ) Is=Of ( dx ( x ) d~y ) ; and as in the last article , if / ' be the operation of changingy into y +f { p ( x ) -F ( x ) } .
Then -f(d-F ( x ) die ) s 'f ( d'-p ( x ) di ) The example last given is capable of being extended to any number of variables in the subject of the operations ; if that subject contain x , y , z , u , &amp;c. , and ?
denote the operation of changing y into y+p ( x ) , z into z+f ( x ) , u into u+Q ( x ) , &amp;c. , where 'p ( x ) ( x ) fZ ( x ) , &amp;c. represent any functions of x , then will s be the intermediate operation , of which d. and do p ' ( x ) .
die ' ( x ) dc -2 ' ( x ) .
d.- , &amp;c. form the extremes , the accents being used to represent the differential coefficients of the functions over which they are placed .
24 .
Let one of the extremes , 0 , be now supposed to represent the operation A. of finite differences , on the supposition that h is the increment of x. Let s represent a multiplier P. , constant or variable with x. Then X M= pIAX P , But by § 5 , PX1A =$ A P-'h + PX AxP thence P+ and then f ( AX ) .
Pxf : , PX+ P Ax .
LX Thus , if pa h.i Pxt += a hep1=(a -)o~ P~.P X ah4a.a we obtain f ( AX ) .
ah =a if(a + a-l ) .
Let i denote the operation of changing x into x+ h , then As = Ax + 1 , and the general formula of this article becomes f ( AX ) P= Pxf ( 4p -X Again , if the subject be a function of x and y , let s represent the operation of changing y into y+p ( x ) and t that which would change y into yA { ip ( x ) } , where Ax p ( x ) according to the common notation stands for p ( x + I ) -p ( x ) , then by a similar procedure we shall obtain the identity f ( AX ) if ( = -1 ) 25 .
In the identity f ( Ait 7JA_ 1 hence f(Ax ) ( I + b ) = ( 1 + b)hf ( Ax b ) .
Now by the nature of an identity both sides in this expression would be alike , if expanded according to the powers of b , the coefficients of like powers must therefore be equal , and in consequence the identity must remain if b instead of being a multiplier represent any operation which is free , relative to AX , thus if b= I , I+ b-= and therefore wX f(AX ) .
( A= f( AY ) As a particular example of this , suppose the subject to be @ ( y ) independent of x , and f ( Ax ) merely to stand for A. , theii [ @ ( my0 ] ( AX ) = O. , , the general identity therefore becomes L( , As = -0y ) , o 01 [ ( + k ) ] ( Ax -AY ) 0 ?
1 k being the increment of y ; now [ @ ( Y +W=(+ ki.j ) h=(+k$ +V ) =P ( y + ) which verifies the result deduced from the general identity .
26 .
Thus when the intermediate operation and one extreme are given the other extreme may generally be found , but it seems more difficult to discover the interme-l diate operation when both extremes are known ; here follow some examples of the latter process .
Let 0 represent any linear operation , dx that of taking the differential coefficient of the subject relative to x , and * the required intermediate .
Then 0 : =d= dx + perform on both -sides the inverse operation sol .
Hence 0d+I dxi or Suppose I= $f(Xd ?
) 1 +f(0d , ) + { f ( 0 , } &amp;c. then by § 3 , _ ?
-f(0-dx ) = f(O ( d ) +f(O-dx ) &amp; C. and id= -f(d-).f( xd ) d , also ( 0 -dl ) i ( 0 d ) of ( -dx ) Now 0 -d is free relative to its own functions , hence 0d=f ( O d , ) dx ; therefore f ( 0- .
dx ) = ( O-dx ) .
d , -1 .
from whence i is known .
For a second example , suppose i the operation of changing x into x+ h , A that of taking the finite difference on the same hypothesis , 0 any linear operation , then i is required to be such that ( '4 0 ) = A:0 .
By art .
15 , 0= LAs ' , + 1A Substituting we obtain To satisfy this identity , suppose ?
, v,1 =.0 ; now , + is free relative to * , hence prefix to each side the operation i4k , which only alters the subject , which is perfectly general ; therefore ~= T+ .
0 ; hence the preceding supposition fully satisfies , therefore we have this theorems , if I be determined such that 0 , then shall 0)= sA 0 , and f(0)I if ( A O ) e As a particular case , suppose 0 to be a multiplier P. , then swill be another multiplier v , , such that or ( log vX ) AX = log P~ , from whence v or : is determinable .
§ 8 .
27 .
In this section we shall give some examples of the use of the formulae investigated in the last section .
In general dxSax = Sam ( d , -a ) , therefore d-a = ax d Sax .
Put for a in this formula the terms of the series 0 , h , 2 h , ... .
( n 1 ) .
h , and comrn pound all the binomials which result from these substitutions , hence dh(d,.-h)(dx.-2(I1h ) ... . .
( d.-(n--l)h)-d ?
( x hxdX 62h-1 ) ( .
No wSdx SdS r ... 0 dx &amp; d ( n1)hx Nw dx ch -h d. , putting y=S -hx therefore dx ( dx h ) ( dx 2 h ) ... .
( dx , -(n 1 ) h ) -dfy " .
hn eC Example .
The expansions of ( a + b ) ' , viz. a ' + nan-1 b +.n-2 )2 +b &amp;c. G11W ~~~~~~~~ ~ .
and being identical when n is a quantity , ought to remain so when n is a linear operation , to verify which , suppose n= now it has been shown that [ f ( W ) ] JdX -f ( x + h ) ; hence [ f ( fx ) ( a + b)dx = Z{x + 1. .
( a + b ) } .
But ( a + b)dxad+d adab +( ad b2 + 1 .
-I .
43+ , &amp; C. = ad { l +d +d 2 % + 23(tQ ) But by this article dx(d 1 ) ( d 2 ) ... ( dx n Jr 1 ) =dy22*n .
if v Ex therefore ( a + b)d = adx { 1+ T. ( by)2 +a ( ) + &amp; C ) and now introducing the subjectf ( 4 ) we get ( L , ' d+f( + 1.a ) -Z b2dy2 d2f(x +1 ( a ) ) which series , if we substitute for x its value 1. .
y and putf ( 1. .
y)(y ) , becomes p ( ay + by ) p ( ay ) + by .
d ; f(ay ) + b2y sedo ~ ( ay ) + ' &amp; C. a die + 1.QJ a ' d/ which being also deducible from TAYLOR'S theorem gives the required verification .
28 .
Let vl v2 v3 .
v. represent functions of x as multipliers , or other operations fixed relative to d , , , it is required to find the polynomial arranged according to the powers ( usually called orders ) of di , , which shall be equivalent to the compound operation represented by v1 d , V2 dv3 d , ... .v .
da .
It is easy to see that this polynomial cannot contain the powers of the differential coefficients of any of the multipliers , that is such as , &amp;c. ; moreover , if we suppose v , = saJ , then de thus the order of the differential coefficients of v , in the required result will be the same as the power of the multiplier a when we substitute ga " ' for v , ; and the same method will apply to discover how V2 v3 &amp;c. are involved ; and when a1 a2 &amp;c. do not enter as multipliers we thence know that V1 v2 &amp;c. are themselves multipliers , and not their differential coefficients .
This being considered , the question is reduced to find the value of the compound operation S ?
l XdZ &amp; 2 Xd.a3 .
d arranged according to the decreasing indices of dT , and then where we have aim &amp; ^ if we put dxtV ' and when we have a2spa2 put dx v % &amp;c. , we shall obtain more readily the result of the question proposed .
Now by the last section d= ( dx + c ) * therefore X1 x 2xdx = ( dK + al ) .
( al+a2)x do = ( d + l ) ( dK + Ml + a2 ) 6(vla2)x Similarly Sell D dakxd2 daxd E.a dx = ( dt + a , ) ( d , , + al + a2 ) ( ds + -a1 + a2 + a3 ) 6( ++X and generally , * * &amp; d = ( d + A , ) ( d + a , + a2 ) ... ( d + al +2 **+ a ) '+a)X If therefore we expand according to the decreasing indices of da , the compound operation , ( dx + a,1 ) ( di + all TV a2 ) * Represent this product by xa + A1 xe + A2 X2 ++ Amxn +.-* + An-lx + A , the general coefficient Am being the sums of products , each of which contain m factors .
For Al it is easily seen that its value is n a , + ( n 1 ) a2 + ( n -2 ) 3 +.2 a0n- , + An Again , A2 consists of products such as a , a25 Ml,35 c &amp; 2 &amp; 3 &amp;c. , and pure powers , as s12a 2 &amp;c. 2 ' the general form of the first class of terms is As a. , and we now proceed to find its coefficient , or the number of times this combination occurs , which number may be denoted by ( as a , ) , and supposing p less than q , no factor preceding x+ a , + a2 + .
, , will be concerned in forming the combination in question , and in the factor itself and the succeeding ones the terms preceding a , , may also be neglected .
The factors commencing from the above , arranged horizontally , will form this diagram .
X+ a , + M* + a , X+ a , + ***+ MP +:a , )+ ' X ?
a , +.M+.p + ?
Ip+1 + ap+2 0 em 0e00 0. .
9 ' .
* .
0* &lt; &gt; X+ al +. .
alp + ap+1 + " p+2 ++*q*X+ aC1 * *1 + + they are therefore the same as in ( 1 + cP)n -1+that is((n +-which is half of the formula obtained by putting q= p. Hence A n.(n-1),2+(n-1)(n-f2 ) 2+( -12)(n-3 ) 21 1.Q2 +2 1.2 '3 + ( n-1 )(n-2 ) ?
s , 3 +(n-1 ) ( n-3 ) ?
6j ?
4 +.* *+ ( n 2 ) ( n 2 ) a a3 + ( n 2 ) ( n 3 ) a2'Z4 Lastly ( a.3 ) is the same as if all the terms M1 M2 &amp;c. were zero , except MP and is therefore ( n -p ?
+ l)(n -p ) ( n -p -i 1.2.3 More generally P ' ) = ( n-p§ 1)(n-p).(Pttimes ) We have thus investigated the coefficients of every combination which enter the whole product , and then if only Ve substituted for any general symbol ( P ) , the d x ' besbtttdfraygnrlsmo c ) required development is completely obtained .
It may be remarked that the coefficients of the combinations of consecutive terms are pure powers , thus ( a MD ) = ( n1)2 M23 ( n2)2 &amp;c. 1 M2 M3 = ( n -2)3 .
29 .
By the preceding investigation we have obtained the following general formula , in which the subject is any function of x : vdxvdxv3d.V dn ( n-1 ) + ( n-2)-d + d4v .
V1 , V2.Vi3.*. .
un td ~dd *r + ( n-1 ) ( n-2)jv d vd-3r ( n -3 ) VI~~~~~~ ( rz-2 Nnv ' d2 .
+v V , dr ( n2 ( n-3)d vdx. .
*v~dxZ+o + ( n- ) 4+ ( n ( Zvi3 ) vd + , &amp; c*.* ********* v* * and similarly if we put in the general formula vl = 1 , and write v , for v2 , v2 for V3 , &amp;c. and then multiply by v. , finally making all equal to v , we obtain ( d v~n~dnvnr(n1).n dn-1 n(1 d ) + n2 ( n )(n In(3n 5 ) n dt2 ( n -2 ) .
( n n pzld2vl + d , , , &lt; ,..s.4~~ v dnz2 + -v ~ 1+ &amp; C. Put now v for v in this formula , whence v ( dZ v 1)=dx " ~ nlrt.v ~~~ ~xd x 4 .
n -2 ( n ( n ) n(n + n)--2 des ( n -2)(n 1).v nI d2 % } 30 .
Change of the independent variable .
When u is a function of v. , and y of x , then it is easily shown that du6 = du .
( ?
die ) , or d. ( dy= omitting the subject u ; hence by substitudy dx dx ~ .
~ dxJ9 tion in the preceding general formula we have d ( dd - ?
( n -1)._ n 1ifdyV-ntdly ( dx *dnJ2 * Xdx * xd2J + n~-2f(n-Q2)(n-1).n(n+1 ) die -n-2 dy2 ( n-2).(n-1 ) , n die \ -nd3y +X d. 4 '\dxJ *Vd z ) 2 .
3 *kdxJ 'dz3 k+ &amp;c. Thus , for example , if n= 3 , d3u dou ( die )S d3 u ( d f4 d'y+ duf3 dy5 ( d2y\2 ( die \4 d3y ) d-dxi\dX !
*3'dxw2\dxl *dx2 'wd xdd x\dxl d x''x and d3 _ fdy\ ( dly\2 din -4 dly d2~\X d x/ Vd 'J *d3* As I am not aware that any formula has heretofore been given for the general change of the independent variable , I shall here add a reinvestigation of the same subject on simple principles .
When x is changed into x+ h , suppose y to become y+ k , and u to be changed into u+ 1 .
Now u+I may be expressed by TAYLOR'S theorem in two ways , du d2u k2 d3u h3 du dP u k2 d3 u k3 U+ k+ d + and since it is visible that h may be expanded in the form Ak+B ki2 + , &amp;c. , it will d7'u kn be unnecessary to consider in I any term after C 1 .
7 .
Hence is the coefficient of kP , when for h we substitute its value in terms of k in the polynomial .
dnu ?
n-1 _d It -2 du H= d 2 : .
h , ` h+n ( z-l 1 ) dh + ... _{ 2 .
3 4. .
d It .
da~ , dxn-l ' dxn-2 ' Again , k is given in terms of h by TAYLOR'S theorem .
k -~- .
+ ~yA"+ d3y ?
, &amp; C. d_de0 dv .
]2 +da-3 1.2 S3c Put for abridgement dgy ( d yXy -I d3y ( d y ' h2 d4 y Id h3 Y d9*dX *dx d +d3* kdx 2.3+dd4 !
* 7 mdxj.4 + ) &amp; c The equation for determining h in terms of k becomes then h{k ( dY ) -hY } =0 .
In a memoir on the Resolution of Algebraic Equations , published in the fourth volume of the Transactions of the Cambridge Philosophical Society , I have proved the following rule .
If p ( )0 be an equation which contains only entire positive powers of x , and f ( x ) any other function of the same kind , of which the differential coefficient or derived function isf ' ( x ) , then the value off ( x ) will be found by taking the coefficient of in the expression-If ( x ) 1. .
Applying this rule to the case before us we find that H is the coefficient of in the formula and consequently die is the coefficient of in the same formula .
The first ( n 1 ) terms in the expansion of the logarithm do not contain kV and may therefore be neglected , instead then of I.{ 1-C ( t. -Y ) } we may use the series I ( k [ V"Y}J+4 { .
( I )Jn &amp; c= -S and the value of dH .
dnu h-1 a n2 Jrn(n1)(n-2).d nhn 3 +..1.2 ... n duz d____ dx kfn in the product of both which series the coefficient of h being sought will give the required value of Tand this is manifestly of the form A du dnu A dn-2 ++A du Ai d~+ A2de ' 3 +~. .
A dx A1 is the coefficient of -h in n hnS , or of h in nS A2.off inn ( n 1 ) h2 S. orof 'D .
n ( n 1).S &amp; c Now if we observe that Y contains h as a factor it follows that the coefficient of kn in the series S is of the form an +2 1+ N-2+ ( : and consequently A , =noal A =n(n-1)oc2 A3 =n(n-l)(n2 ) a3 ... .A -n ( n1)(n-2 ) ... .
Lae Therefore d~nU d'u dn-1U dn-2u d =n ad~ +n ( n )2 .
djn 1+ n ( n -1 ) ( n 2 ) v3 n2 &amp; C. jy ~~~~~ xdxnBy taking in fact the coefficient of kn in S and multiplying it by hn we find that the product , viz. a+ a2 h+ OC3 h2 + ... .
cn hnl + , &amp;c. is equivalent to 1d in n { d ynn+ 1dy _n 2 ( n + )(n+2 ) die \ n *dxJ ddx I*+ 2 xJ 2.3 * dx .
Y*. .
I die ?
J n { d y-n-1 d9 , y d3y h &amp; ~-~n * xdc\dac ( n § 1)(n+Q ) ( die )-n-3 ( d2y )3 4- !
2.5 *Vdz 2d W2 n+ ( dy'yv2 d y\ ( d3y\ ( die &lt; ~z1 d4y +Q i !
d 2* V2dx2J\2.32 3d 3/ d 2dx2.3S.4 ?
In general let ( d2 , y 1d3y hJ d4y h , &amp; .
ofio m a &amp; ( ?
+ dx.3 +d~X4Q~3.4 + , -- &amp; C.)m +r , O+m , I , h+ ym , 2It2 &amp; C Then 1 -1 _ ( n +r 1 ) ( n +r 2)..0 .
( n J+ m 2 ) { dy8-f-f~ ( _ I am3 .
mI ) Txj gnXs 0 ( n +I ) ( n + ( n + m-9 d W\ -n -mW 2r T b.3 e(mg &gt; 1x + k~~~~~dx ( n+)(n+Q1)(n+ ... ( 2 )( d ) 2-n+2 Q.3 ... ... ... d ... -x,1 _ din Cd wnbeing : &gt; le Corollary .
Put u x. them d nX= ( a-1 ( n-2,.]0 were ( _L)- .
( n + n ) ( n + 2 ) ... ( 2 n-2 ) tdy2-+ ( n + 1 ) ( n + !
2 ) ... ( 2n-$ ) ( dy2n+2.3..d .
( n-d2 ) * ( d ) n-2 , V* *(-1 * in { n(n Jr1 ) * *(2n-2).*dYLy 1 -(n -1).n ( n + 1 ) * .
**(2n 3 ) *x 6 , ) *Y 2,1 ( n 2)d(n1)((2n-4 ) die \ *3d2 Thus 2 dxYy2 , o 02..\j]3iiff ) d-y3= d3)4* yo

