I. " The Higher Theory of Elliptic Integrals , treated from Jacobi 's Functions as its basis .
" By F. W. NEWMAN , Esq. , M.A. , Professor of Latin in University College , London .
Comnmunicated by the Rev. Dr. BOOTH .
Received March 3 , 1859 .
( Abstract .
) The peculiarly beautiful properties of these integrals , as treated by Jacobi and ( in his two sutpplements ) by Legendre , are obtained through so very elaborate and difficult a process , that few students can afford the time to study them .
Professor De Morgan , in his 'Integral Calculus , ' declines to enter even the Lower Theory , on the ground that the subject requires a detailed treatise .
That in some sense it is analogous to trigonometry , which no one would desire to be treated fully in the di erential and integral calculus , has been recognized by several writers .
Legendre , in his second supplement , sixth section , took the first steps toward treating Jacobi 's fiunctions ( A and 0 ) on a wholly independent basis , by investigating their properties from the series which they represent : but after only two pages of this sort he aids his research by assuming their relations to elliptic integrals as already established , which shows that he was not seeking for a new basis of argument , but only for new properties .
The author of the present paper proposes ( for didactic purposes ) to commence the higher theory from these functions .
The first division of his essay is purely algebraic and trigonometrical , not introducing the idea of elliptic integrals at all .
Adopting as the definition of the functions A and 0 the two equations A ( q , x ) = 2ql(sin x ql2 sin 3+ q23 sin 5x &amp;c. ) 1 O(q , x ) -= 1 -2qu cos 2x + 2q22 cos 4x-2q3 3 cos 6x + &amp;c. it demonstrates by direct algebraic methods many properties of great generality , of which we shall here specify1 .
If Vb stands for ( q , 0 ) and Vc for ( q , which iS 0(q , ?
r ) ( q , p7r)'wihs shown to yield 62+ c2 1 ; and if , further , A ' 00 stand for A(q , x+ ` w ) , O(q , x+ 21r ) ; we get the four equations ( equivalent to two only ) 2+ 6A02cO2 ; A02 + 6A2 c002 ; 02-6002=cA ; 002_602=cA02 ; from which it directly follows , that if t is an arc defined by the AA equation / 6 tan w= ^ , we shall have simultaneously &gt; / c sin w =- ; .A/ C C0 o(l-c'sin ' )= Vb e The symbol A(c , w ) , A(w ) or A represents / ( l -c2 sin2 ( w ) in this theory .
2 .
It is further shown that AOd -A dA __A2(k7w ) .
000 ; dx dx whence is easily obtained 0CCA(c , w ) .
Also 0 , do _e doo=A2(Z7 ) .
A AO .
dxdx dx 3 .
By direct multiplication of two trigonometrical series , it is found that A(q , x)A0(q , y)= A(q2 , x-y).0(q2 , x +y ) +0(q2 , x-y ) .
A(q2 , x ?
+y ) ; 0(q , x)O0(q , y)=A(q2 , x-y ) .
A(q2 , -+y ) +0(q2I x-y).0(q2 , x+y ) .
From the property marked ( 2 ) we obtain the connexion of the fLnctions A , 0 with elliptic integrals .
For , if F(c , c ) , as uisual , stands for Idw V V( 1-C2 Sinl2 w ) it yields F(c , )OCx ; or F( , , This introduces the seconid and principal part of the essay .
An easy inference from ( 3 ) is , that A ( q2 , X+ y ) Ay+ AyA ?
x 0(q2 , x+y)-OxoOy+ OyOX ' and consequently that if q is related to q2 and to xy by the same law as o is to q and to x , while c , is to q2 what c is to q , we obtain 4~ Vc sin c sin ( o + 0 ) nfc}~~~ Ys-ln ?
CW ) + A(C , - ) , wheu F(c , , x7 ) F(c , -w)+F(c , 0 ) This formula bas the peculiarity of comprising Euler 's integrals , with the integrations of Lagrange and of Gauss ; namely , if W =0 , we get the scale of Lagrarnge .
If 0=0 , the scale of Gauss is obtained .
But if we introduce a new variable X , such that F(c , t)=F(c , w)+F(c , 0 ) , we eliminate i1 by aid of the last result ( which is Vc sin=A( ) and obtain ____ ; csinQo ?
)0 ) which is equivalent to Euler 's integration .
The author believes this generalization to be new .
5H Ile proceeds ( assuming now the theory of Lagrange 's scale ) to prove the higher theorems by much simpler processes .
E being the , second elliptic integral , he writes G for OF , and V for g'GdF , and ouLt of the integration I log A= 'VY-V ( where V , is to q2 and 2z , what V is to q and x ) , he deduces V-log .
0(q , x ) 0(q , 0 ) by a process fundamentally that of Legendre , Second Supplement , ? ?
196 .
This is the equation by which E , and indirectly the third integral IT , is linked to the functions AG .
6 .
We may further point out , as perhaps new , the developments of AO in the case when q is very near to 1 .
Let r be related to 6 as 11 1.I oq to c ; then log- .
log =2 .
If log -=ra , and =ru , qrg A/ a O ( q , 7rU + _r ) rU2 + r(1 ?
t)2 + r(2 ibt)2+ r(3 ?
l)2+ &amp; c Va , A|t(q , u , + _7r)=rU2_r(1 ?
u)2 ?
r-(2 ?
U)2_ r(3jt6)2+ &amp; ... f in which the double sign denotes two terms , which must both be included .
But besides , if the syrnbol +(r ) stand for [ ( lIr')(1 r')(I-t r)..v .
&amp; .
] -1 , Va. ( r).O(q , 7r+ 7 ) =ru2 .
( 1 +rl ?
2u)(l +1a3 ?
2u)(l +r5 ?
2u ) 1 Va. ( r ) .
A(q , ru+ 2w)r2 ( 1 r1 +2u)(l 2r302u)(l ?
r5L2u). .
From these formulae not only all of Gudermanii 's developments for calculating elliptic integrals in every case are deducible , but others also , it seems , of a remarkable aspect , in the difficult case of q and c being extremely near to 1 .
We produce the two which seem to be simplest .
Let B be to 6 Sin. .
what C is to e , and Tan x represent --- , where 2 Sin x stands for -e_T and 2 Cos for e + .
Then when c is very near to 1 , we compute G and thereby E from the series , B. G(C , wo ) ( I -Tan )+ ( 1-Tan WX ) ( 1-Tan + ) + ) -Tan2w ) ( -Tan27r+x\ +1 -Tan3\x The third elliptic integral is in the same case deduced from a series of the form tnz.yj{-tan-(tanj .
Tanx ) + { j-taii-(tanlj .
Tan 7r } { j-tan-(tan).Tan 7}+ -{jtan1(tapj .
Tann2w ) } - &amp; C ... .
NTO L. I X. 3 B } Finally , the essay developes above thirty series which rise out of this theory , nearly all of which are believed to be new .
The most elegant of them may find a place here .
Writing , for coniciseness , C so related to c that F(c , kw)=7 C , and. .
F(c w)=C , we have ta ) .
+ sin 2 +1 sil 4x sin 6x ( , =+COSw7a + Cos 2qra 3Cos 37-a+ C b ) CA(e +2 cos 2x +2 cos 4 ?
+2 cos6x &amp; Coswa + Cos 27ra Cos 3a(I CC2 sin sin 2x 2 sin 4x 3 sin 6x ( c).~C22si2w= + -+ + &amp; .
Coswa Cos 2w:a Cos3a I1-cos 2x 1-cos4 IX 1cos 6x ( d ) V(c , @)= .
-+ 2.-.+..+ &amp; C ... e sin wa sin 2wa a sin 3wca ( e ) kC .
G(c , w)= sin 2x sin 4x sin 6x ( e ) .
'C-G(c , w)= + +_ .
+ &amp; C ... .
sin ?
ra + sin 27ra sin 37ra This is virttually eq. 49 of Legendre 's Second Supplement , ? ?
7 .
In eq. 53 of the same , he has a developmeent of sin2 a , which is given by Mr. Newman in a notation similar to eq. ( c ) above .
I -os x l-cos 6x +1l-coslOx -2)2gA( 5 )= S'in7rt 3* Sii37 , Sin5wa + &amp; cew sin x sin 3x sin 5x '- ?
Cc SIRl_(O = .
+ .
&amp; C ... .
2~ ) Csn= +7+ Sin 2ra Sln s-ra Sin 2-ra cos + cos 3x cos 5x ( h ) .
Cc cos +++ 2COSra Cos 3ra COS Aa -1C / 1 -A _sin x sin 3x sin 5x T4C +A Sin 7ain 5wa Moreover , Jacobi 's two celebrated theorems follow as a corollary from the general propositions here established .

