]\gt ; 217 IX .
The Third Elliptic Integral and the Ellipsotomic Problem .
A. G. GREENHILL , F.R. Received Deoember 22 , 1903 , \mdash ; Read January 21 , 1904 .
THE ABEL Centennial Ceremony , held in Christiania , September , 1902 , has directed the attention of mathematicians to the great influence of ABEL on modern analysis , and to the history of elliptic functions , and of the foundation by CRELLE of the " " Journal fur die reine und ewanlte Mathematik ABEL'S article in the first volume of ' CRELLE'S Journal , ' 1826 , ' ' Ueber die Integration der Differential-Formel ( A ) , wenn und ganze Functionen sind is of great importance as indicating the existence of what is now called the pseudo-elliptic integral ; the present memoir is intended to show the utility of this integral in its application to mechanical theory .
Provided with a list of these integrals , proceeding from the simplest case and extending as far as possible , the student of applied mathematics will be able to effect the complete solution of many interesting mechanical problems now abandoned in an unfinished state ; at the same time , the exploration along the simplest lines of progress is effected of the general analytical field , and mathematical research is guided in a path likely to arrive at useful development in the theory of elliptic functions .
The problem of the Division , and thence also of the Transformation , of elliptic functions is solved incidentally ; as also the difficulties raised by LEGFNDRE in his letter to ABEL , of January 16 , 1829:\mdash ; " " Mais de la nait un difficulte ' sir la division generale des fonctions elliptiques .
" " Admettrons-nous cette difference enorme entre les fonctions de espece parametre logarithmique et les fonctions a parametre circulaire , savoir que les premieres peuvent s'exprimer par un fonction de deux variables , facilement reductible en table , et que les autres no le peuvent pas ?
Il aurait donc par le fait quatre especes de fonctions elliptiques au lieu de trois , et la quatrieme serait bien plus composee que la troisieme .
C'est un point qui merit d'etre examine et mis au clair .
Js le recommande a votre investigation et a celle de M. JACOBI VOL CCIIL .
2 '3.8.04 But a mere change of sign in ABEL'S results is sufficient to pass from the logarithmic to the circular form of the third elliptic integral , and as it is the circular form which is of almost invariable occurrence in dynamical problems , we shall adopt it as our standard form .
Applications of this third elliptic integral are introduced in the course of the memoir to these problems , such as POINSOT'S herpolhode , and the spinning top or gyroscope , the spherical catenary , the velarium , and the elastica under uniform normal pressure .
References to former articles on the same subject in the 'Proceedings of the London Mathematical Society , ' vols .
25 and 27 , are given in the sequel in the abbreviated form , L.M.S. , 25 or 27 ; frequent reference is made also to KIEPERT'S articles on the theory of elliptic functions in the 'Mathematische Annalen ' ( ' Math. Ann vols .
26 , S2 , 37 ) .
1 .
Working then with the third elliptic integral in the circular form , when the elliptic parameter is a fraction of the imaginary period , we change the variable in the standard form of WEIERSTRASS , ( 1 ) , by putting ( 2 ) , , where is a real fraction , so that the integral changes to ( 4 ) ( where the elliptic arguments and may be supposed for a moment to be normalised to the Jacobian form ) , and is an elliptic function of which we may denote by , differing from WEIERSTRASS'S by a constant , so that .
, while Considering that ( HALPHEN , 'Fonctions Elliptiques , ' 1 , p. 222 ) ( 5 ) is an algebraical function of when is an aliquot part of a period , we take ( 6 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
where ( 7 ) .
and denote this integral by I ( v ) , and work with it as our standard form of the elliptic integral of the III .
kind .
The function employed by HALPHEN is now ( s ) , ( 9 ) , and is a Lame function of the first order , satisfying his differential equation ( 10 ) ; .
thence the Lame functions of higher order may be derived by differentiation , as shown by HERMIT , 'Comptes Rendus , ' 1877 , and these can be employed in the problems considered by Professor G. H. DARWIN , in the 'Phil .
Trans 197 , 1901 , " " Ellipsoidal Harmonic Analysis In the Hermite-Jacobi notation we may take ( uzsv ) , or ( uznv ) ( 11 ) .
2 .
Next introduce the and employed by HALPHEN ( ' F.E. , ' 1 , p. 103 ) , which may be connected with the , and of ABEL'S notation ( ' ( Euvres , ' II .
, p. 155 ) by ( 1 ) and put ( 2 ) , ( 3 ) ; then if we put ( 5 ) , ( 6 ) .
The multiplicative values , ( 7 ) , is ' , is ' , . .
is ' ( nv ) ( 8 ) , can now all be expressed rationally in terms of and , with the functions of HALPHHiN ( ' F.E. , ' 1 , p. 102 ; also ABEL , '(Euvres , ' II .
, p. 159 ; MONTARD PONCELOT , ' Applications d'analyse et de geometry , ' .
I. , Paris , 1862 ) , by means of the recurring formulas , ' ( 9 ) .
Thus ( 10 ) , ( 11 ) , ( 12 ) , .
( 13 ) , .
( 14 ) , ' ( 16 ) .
3 .
For the determination of ( nv ) we have the formula HALPHEN , ( 1 ) ; whence , by logarithmic differentiation , ( 2 ) .
Now for every homogeneity factor , and as elliptic functions of degree zero , ( 3 ) , ( 4 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
and ( 5 ) , ( 6 ) , so that ( 7 ) .
Thus , putting , 3 , 4 , .
, ( 8 ) , ( 9 ) , ( 10 ) , ( 11 ) , ( 12 ) , ( 13 ) , ( 14 ) , , ( Ov ) ( 16 ) , ( 17 ) , .
4 .
When is an aliquot part of the imaginary period , so that ( 1 ) , and is congruent to a period , then ( 2 ) , so that ) is given by equation ( 7 ) , S 3 , by putting , with the additional condition that ( 3 ) , or ( 4 ) .
Thus , , ' ' ' , so that adding ( 11 ) and ( 12 ) , 11 .
' and , nilarly , THIRD ELLIPTIC INTEGRAL AND THE LIPSOTOMIC PROBLEM .
ellipsotomic problem of the determination of the division-values of the elliptic functions resolves itself thus into a consideration of the curve in and given by this rational }ation ( , which may thus be called the equation , by analogy with the clotomic equation for the circular functions ; and the problem may be considered solved when .
and can be expressed , rationally or irratio1lally , in terms of a parameter ; according to POINCARE ( : Bulletin de la Societ Mathe'matique de France , ' 1883 , 11 , I ) .
112 ) this can always be effected by uniform functions of an independent variable .
The details have been carried out in the ' Proceedings of the London Mathematical Society ' ( L.M.S. ) , 25 , for values of up to 22 inclusive , omitting 19 , which still remains awaiting solution .
When is an even number , or is a root of so that the cubic can be resolved into factors , and we can employ the functions of the Second Stage of LEGENDRE and JACOBI , as required in [ most dynamical applications .
But when is odd , this resolution cannot be effected algebraically , and STRASS'S functions of the First must be employed .
The ellipsotomic relation and ( 5 ) are equivalent , so that odd , , , . .
( 6 ) , even , ) ) , . .
( 7 ) .
Writing for , and for ( 8 ) ( HALPHEN , 'F.E .
, ' 1 , pp. 102 , 198 ) ; and changing into , and dividing , ( 9 ) .
But ( 10 ) , so that ( 11 ) , and generally and this is KIEPERT'S function , defined in ' Math. Ann. , ' 32 , p. 6 .
According to HALPHEN ( ' Math. Ann 15 , p. 359 ; 'Comptes Rendus , ' March , 1879 ) ( 13 ) .
5 .
The integral employed by ABEL , changed to the circular form , can be written ( 1 ) , ( 2 ) , and , as pointed out in the Archiv der Mathematik und Physik , ' 3 .
Reihe , I. , p. 72 , this is exactly in the form required in the problem of LEvy 's ' Elastica , ' with reduce it to our standard form in ( 4 ) , S1 , by putting ( 3 ) , ( 4 ) , so that ( 5 ) , ( 6 ) , and therefore ( 7 ) , ( 8 ) , is a factor of corresponding to More generally , with ( 9 ) , so that is the value of corresponding to ; then in ABEL'S notation , changing the sign of and to obtain the circular form , ( 10 ) , ( 11 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
on putting ( 12 ) , and ( ) ( ) ( 13 ) , which is a perfect square in the form ' ( 14 ) , implying that ( 15 ) , ( nv ) ' ( ) ( 16 ) .
( 17 ) , which also satisfies ( 16 ) .
Then ( 18 ) , , VOL. CCIII.\mdash ; A. requiring the relations ( nv ) ( 20 ) , ( 2 ] .
) , , ( 22 ) .
Now ( 23 ) , ( 24 ) , so that ( 25 ) , and taking ( 26 ) , so that ( nv ) ( 27 ) , ( 28 ) , we obtain from ( 17 ) , ( 21 ) , ( 22 ) , is ' ( 29 ) , ( 30 ) , ( 31 ) ; and integral ( 32 ) , on taking ( 33 ) .
Putting ( 34 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
so that ABEL'S recurring formula for ( ' ( Euvres , ' 2 , p. 157 ) , ( 35 ) , becomes ( 36 ) , and on comparison with the Weierstrassian formula , ( 37 ) , we infer that ( 38 ) .
From another formula , ( 39 ) , ABEL'S relation ( 40 ) , and ( 41 ) ; so that ( 42 ) .
Writing for ABEL'S , his is given by ( 2 , p. 161 ) ( 43 ) , as before in ( 33 ) , since ( 44 ) .
With ( 45 ) , ( 46 ) .
But with ( 47 ) , ( 48 ) .
6 .
In the simplest case of ( 1 ) , the form ( 6 ) , S1 , is illusory ; but we take ( L.M.S. , 25 , p. 210 ) I ( 2 ) ; and putting ( 3 ) .
Writing for ( 4 ) .
In ABEL'S form ( ' ( Euvres , ' 2 , p. 163 ) , ( 5 ) ; and with ( 6 ) .
Or , in the circular form , ( 7 ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
In the next case ( L.M.S. , 25 , p. 213 ) , , I ( 2 ) , ( 3 ) , ( 4 ) , ( 5 ) , and is the icosahedron , and KIEPERT'S In ABEL'S form , with has the factor ( 6 ) , and putting , ABEL'S integral , with ( 7 ) .
Putting I ( 8 ) , and the degree of the expressions is h. allied , with great gain of symmetry .
The degree is halved with greater ease by putting in I , and now I ( 9 ) , derivable from the preceding I ( v ) by writing for , and for PROFESSOR A. G. GREtNHILL ON THE 8 .
This suggests that in the general case of it is simpler to work with ( 1 ) , and to put ; and then with ( 2 ) , ( 3 ) , we shall have I ( 4 ) , according as is even or odd ; and the results are of one-quarter the deglee that would be given by ABEL'S method of the periodic continued fraction ; and since ( 5 ) , the determination of , . .
can be carried out by a consideration of the re'duites ( HALPHEN , ' F.E. , ' 2 , p. 576 ) in preference to continued fractions , once the coefficients of in and have been assigned .
9 .
Thus ( L.M.S. , 25 , p. 222 ) for , is a unicursal , in which ( 2 ) , and ( 3 ) ( 4 ) , I ( 5 ) , .
( 6 ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
Introducing a normalising homogeneity factor , so that the substitution should correspond to ( 7 ) , derived from , ( 8 ) , ( 9 ) , , , ( 12 ) .
In the notation of KLEIN-FRICKE , ' Modulfunctionen , ' 2 , p. 399 , ( 13 ) , ( 14 ) .
10 .
With ( L.M.S. , 25 , p. 232 ) ( 1 ) , is a unicursal , in which , ( ) , ( 3 ) , and from the relation ( 4 ) , ( 5 ) , - ( 6 ) , I ( 7 ) , , ( 8 ) .
The substitution will correspond to with a normalising factor ( 9 ) , and ( 10 ) , ( 11 ) , ; while ( 13 ) , unchanged by the substitution .
Also , a quantity required in the transformation of the mnth order ; and ( 15 ) , ( 16 ) , ( 17 ) , ( KIEPERT , 'Math .
Ann 32 , p. 66 ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
Math. , leads by the substitutions ( 1 ) to the bicursal ( 2 ) , ( 3 ) ; ( 4 ) , \mdash ; , ( 5 ) .
We now find ( 6 ) , I ( 7 ) , , ( 9 ) , , ( 11 ) , using detached coefficients .
We find also ( 12 ) , where \mdash ; \mdash ; ( 13 ) and the values of and are given in L.M.S. , 27 , p. 455 .
VOL. CCIII.\mdash ; A. 2 The normalising factor is ( 14 ) , but so far this has not made evident any symmetry of results ; it will be noticed that our parameter here is an elliptic function of ( L.M.S. , 27 , p. 429 ) ( 15 ) , which is of a period out of phase with that required to lead to KLEIN'S results .
12 .
' Math. by the substitutions , ( 1 ) , leads to a with class , in which ( 2 ) , ( 3 ) , and we find , using detached coefficients of ascending powers of ( 4 ) , I ( 5 ) , ( 6 ) , ( 7 ) , , ( 8 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
( 9 ) , , ( 10 ) .
By means of an appropriate homogeneity factor , we can express ( 11 ) in such a manner that the substitutions correspond to and to , and or ( 12 ) .
13 .
, ( L.M.S. , 25 , p. 258 ; 'Math .
Ann 52 , p. 485 ) , is reducible to a bicursal Changing in L.M.S. , 25 , p. 259 , into , and normalising by a homogeneity factor ( 1 ) , I ( 2 ) , ( 3 ) ( 4 ) , ( 5 ) , ( 6 ) , ( 7 ) , Differentiating ( 2 ) and equatin coefficients of we find ' ( 10 ) .
\mdash ; ( 11 ) ; ' ( 12 ) ; THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
( 13 ) ; ' ( 14 ) .
These calculations , as well as for and 13 , and their verification , were carried ont for me by Mr. J. W. HICKS , of Greenwich Observatory .
Putting ( 15 ) , then , since ( 16 ) , we find ( 17 ) , .
The substitution changes into into 80 , . .
, so that and are obtained from and by writing the coefficients in reverse order .
( 19 ) , ( 20 ) .
Again , since ( 21 ) , we find ( 22 ) ( 23 ) , and in the coefficients run in reverse order , 2.38 PROFESSOR A. G. REENHILL ON THE So also ( 24 ) , and , are obtained from , by writing the coefficients in reverse We find also ( 25 ) ; ( 26 ) ; ( 27 ) , ( 28 ) , and these are unchanged by the substitution .
Also ( 29 ) and ( 30 ) .
With ( 31 ) , ( S2 ) , the expressions for \mdash ; are lowered in degree ; for , ( 34 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
\mdash ; ( 36 ) , ( 37 ) , , ( 40 ) ; , , so that ( 43 ) ; ( 44 ) , 14 .
The case of can also be derived by a trisection of , and so generally when is any multiple of 3 .
For when ( 1 ) is expressed in the form ( 2 ) , this implies that ( 3 ) ; and now ( 4 ) , ( 5 ) , ( 6 ) , so that ( 7 ) , reducing to ( 8 ) , a iacobian quartic for , which can be resolved .
For and putting .
To identify with the preceding results in S 13 , put ( 11 ) , and then the associated octahedron irrationality 15 .
For , the quartic for in terms of is given in L.M.S. , 25 , p. 264 , obtained by the substitution in of This irreducible quartic is made reducible by putting , and with becomes ( 2 ) , ( 3 ) .
The alternating function ( 4 ) , and the division values are associated with elliptic functions of an argument ( 5 ) , ( 6 ) , .
By the transformation , 4 ( ( 8 ) , ( 9 ) , and a comparison can be made with the equations of KIEPERT ( ' Math. Anm 37 , p. 386 ) 16 .
The next case of presents difTiculties not yet surmounted , although it was ] ) oped that the analogy with would the clue required .
In the case of , the substitution of ( 1 ) THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
in ( 2 ) , S 11 , makes the ellipsotomic relation equivalent to ( 2 ) , an addition equation of the elliptic functions and of the argument ( 3 ) , and differing in phase by one-fifth of a period ; and the five division values of the arguments ) , are derivable from each other when considered as elliptic functions of , 2 , 3 , 4 .
The connexion with KLEIN'S parameter is made through ( 4 ) , and the quintic transformation , if ( 5 ) , and then \mdash ; IIOHK ( 6 ) , an elliptic-function addition equation ( L.M.S. , 25 , p. 244 ; 27 , p. 428 ) .
If analogy is to help us in passing from to 19 , the ellipsotomic equation should be reducible to the relation ( 7 ) , where , in terms of FRICK'S and ( ' Math. Ann 40 ) , ( 8 ) , with the addition of the cubic transformations ( 9 ) , and the same functions of This combination of ( 7 ) and ( 9 ) leads to an equation of the degree between and , functions of the elliptic argument ( 10 ) .
OL .
CCIII . .
2 I The nine division values of the argument should now be functions of an argument , and thence derivable from each other , being grouped in sets of three In passing horizontally in these sets , the substitution connecting and is ( 12 ) .
The additional cubic transformation ( 13 ) leads to a multiplication relation of the order , connecting and ( 14 ) .
The relation of the order between and is of the order in and , representing a in the coordinates and But so far the various transformations of , as given in L.M.S. , 25 , lead to a of the fifth degree in each variable , and the reason of this is a mystery still .
17 .
For , applying the trisection equation ( 8 ) S 14 , with the relations of ( 1 ) , ( 2 ) .
Put ( 3 ) , ( 4 ) , ( 5 ) , ( 6 ) , .
THIRD ELLIPTJC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
The substitution gives , and ( 8 ) , ( 9 ) , ( 10 ) , , ( 11 ) , ( 12 ) , ( 13 ) .
18 .
For the class of the modular equation is 2 , so that simple relations cannot be anticipated ; but as the class is zero for , it is possible that the ellipsotomic equation may be susceptible of reduction ( L.M.S. , 25 , p. 275 ) .
19 .
As mechanical applications of the preceding integrals of the First Stage we may cite the case of LEvy 's " " Elastica mentioned in S 5 and discussed in the ' Math. Ann 52 , the Spherical Catenary ( L.M.S. , 27 ) , and the Velarium surfaces considered in F. KoTTER'S ' Inaugural-Dissertation ' ( Halle , 1883 ) .
Take an umbrella with straight ribs , and hold its axis vertical , as an illustration of a velarium .
If the gore laid out flat forms a sector of a circle , then it is obvious that for any other angle between the radii formed by the ribs , the edge and its concentric lines form spherical catenary curves , as shown by equations .
But with triangular gores ( KoTTER , ' Diss p. 38 ) the projection of the edge on a horizontal plane is given by ( 1 ) , with ; and this is reducible immediately to our standard form ( 1 ) , S 5 , by putting ( 2 ) , ( 3 ) , and equating coefficients , PROFESSOR A. G. GREENHILL ON THE , ( 5 ) , ( 6 ) , so that ( 7 ) , ( 8 ) .
Now denoting the elliptic argument by , where .
( 9 ) , ( 10 ) , and the preceding integrals can be utilised for the construction of solvable cases .
The chief interest is in the purely algebraical case , obtained by putting Thus we find for , putting , in ( 3 ) , S 7 , ( I 1 ) , ( 12 ) .
20 .
The expression of the pseudo-elliptic spherical catenary , discussed in L.M.S. , 27 , p. 127 , can be halved in degree by changing to the stereographic projection , with ( 1 ) ; ( 2 ) , ( 3 ) , ( 4 ) , In a pseudo-elliptic case , with ( 5 ) .
( 6 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
we put ( 7 ) , and now the expression for the catenary can be reduced to ( s ) .
Thus for ( 9 ) , for ( 10 ) ; and we can calculate , and the six 's in ( 11 ) ; this catenary has been drawn stereoscopically by the late Mr. T. I. DEWAR .
21 .
ABEL'S integrals are applicable immediately to the construction in polar co-ordinates and of algebraical orbits or catenaries under a central attraction of the form ( HUGO GILD , 'Kongl .
Svenska Vetenskaps-Akademiens Handlingar , ' 1879 ) ( 1 ) ; with the orbit can be realised by two balls , connected by an elastic thread , whirling round each other in the air ; and the addition of a term to , varying inversely as , merely has the effect of qualifying the angle by a factor .
Putting , and denoting the velocity in the orbit by and twice the rate of description of area by ( 2 ) , ( 3 ) , ( 4 ) , ( 5 ) , and now put to identify with ABEL'S results .
Thus for in , putting , and ( 6 ) , ( 7 ) .
For instance , makes , and the orbit is ( 8 ) , described under a constant central force .
For in ( 7 ) , S7 , putting ( 9 ) , ( 10 ) , and to make , put , so that with ( 11 ) , an orbit described by a body attached to an elastic thread , which is led through a fixed origin , which can be written ( 12 ) , So also for THE SECOND STAGE .
22 .
But in the dynamical applications , such as POINSOT'S herpolhode and JACOBI'S associated motion of the top , the integral of the Second Stage is required , corresponding to an even value of , and can now be resolved into its factors ( 1 ) , while in most cases of the applications ( 2 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
so that ( 3 ) , where denotes real fractions .
To obtain this resolution of we find it convenient to put ( L.M.S. , 27 , p. 449 ) , , ( 4 ) , and with ( 5 ) .
Denoting the roots of , irrespective of order of magnitude , by , we can put ( 6 ) , ( 7 ) , ( 8 ) .
With congruent to a half-period , we can take ( 9 ) , ' and this leads to a relation between and , by means of which they are expressible theoretically in terms single variable .
Also , as shown in L.M.S. , 27 , p. 450 , ( 10 ) , ( 11 ( 12 ) , ( 13 ) , ( 14 ) , where and is obt , ained from Do by a change of into , and into ( 15 ) , ; ( 16 ) , and is obtained by a change of into and into ; ( 17 ) , ; ( 18 ) , and a change of into - and into gives ; ( 19 ) , .
A simplification is effected by putting , and ( 20 ) , THIRD ELLIPTIC INTE(GRAL AND THH : ELLIPSOTOMIC PROBLEM .
and now the change of into and into is equivalent to a change of sign of , leaving unchanged .
Thus we find , with ( 21 ) , Thus , if ( llv ) , and , which replaces in a much simpler form the relation So also makes , and so enables us to connect up the results for ; and other values of can be treated in the same way .
23 .
There are three cases to consider , relative to the half-period , to which is congruent:\mdash ; I. ( 1 ) , and introducing WEBER'S function , or KIEPERT'S equivalent function , ( 2 ) , and to the complementary modulus ( 3 ) , ( 4 ) ; II .
( 5 ) , ( 6 ) , , ( 7 ) , ( 8 ) ; VOL. CCIII.\mdash ; A. 2 ( 9 ) , ( 10 ) , , .
( 11 ) .
( 12 ) .
A fourth case , where 1 \mdash ; 8 is negative and imaginary , need not be discussed here .
The Weierstrassian Sigma Functions of the Division-values ( Theilwerthe ) can now be expressed in terms of the Jacobian Theta and Eta Functions by the relations on p. .52 of SCHWARZ , ' Formeln der elliptischen Functionen , ' so that ( 13 ) , ( 14 ) , ( 15 ) , ns ( 16 ) .
( 17 ) , .
( 18 ) , ( 19 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
24 .
When is of the , so that , as required in dynamical problems , , ( 1 ) , then is congruent to or , according as is even or odd ; and the integral I ( v ) is expressible in the form ( 2 ) , and when the quantities and , and have been determined as algebraical functions of a parameter , the .
calculation of , . .
, .
, , is effected readily by a tiation and verification .
It will be found that ( 3 ) , also ( 4 ) , denoting JACOBI'S Zeta function or , while H. ( 5 ) in GLAISHER'S notation ; also .
Here already the degree of the results obtained by ABEL'S method have been halved in ( 2 ) ; and the degree can be halved again by the substitution ( 7 ) , and now I Denoting by the values of corresponding to of ' ( 9 ) , reducing to ( 10 ) ; so that , in Region I. ( 11 ) ; , , in Region .
( 12 ) .
Changing the variable to in the integral I ( 13 ) , where ( 14 ) , ( 15 ) , ( 16 ) ; ( 17 ) , ( 18 ) .
( 19 ) , ( 20 ) .
25 .
To connect up ABEL'S results for even values of we take in ( 9 ) , S 5 , so that ( 1 ) , ( 2 ) , where ( s ) , and therefore the resolving cubic becomes ( 4 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
or ( 5 ) .
Then , with ( 6 ) , Az ( 7 ) , and breaks up into two quadratic factors , and ( 8 ) , ( 9 ) ; ( 10 ) , ( 11 ) ; and by means of a homogeneity factor ( 12 ) , replacing by , we may replace and by ( 13 ) , ( 14 ) .
26 .
The resolution of in the spherical catenary for is not the same as for ; we must put ( 1 ) , and then in ( 2 ) ; so that ( 3 ) , a quartic for , having a root , which was used for ; the remaining cubic , putting , becomes or as in S 20 with ( 6 ) ; so that reduces this to as in ( 3 ) , S 8 ; and with ' ( 8 ) , ( 9 ) .
We find that CLEBSCH'S ( CRELLE , 57 , p. 105 ; L.M.S. , 27 , pp. 146 , 185 ) is given by ( 10 ) , so that ( 11 ) , and so that ( 12 ) , ( 13 ) ; and in quadratic factors ( 14 ) , ( 15 ) .
Thus , as shown in the 'Bulletin de la Societe Mathematique de France , ' 29 , the algebraical spherical catenary for , drawn stereoscopically by the late Mr. T. I. DEWAR , can be represented , by taking , in the symmetrical form THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
( 16 ) , ( 17 ) .
27 .
These expressions for I ( v ) are applicable immediately to the construction of a series of quasi-algebraical Poinsot herpolhodes , in continuation of the one invented by HALPHEN ( ' F.E. , ' 2 , p. 279 ) .
Making a digression on the motion of a rigid body about a fixed point under no forces , as illustrated by the motion of a body about its centre of gravity when tossed in the air , POINSOT'S polhode and herpolhode are obtained by rolling momental ellipsoid on the invariable plane .
The equation of the momental ellipsoid can be written ( 1 ) , where S denotes the distance of the invariable plane from the centre of the momental ellipsoid ; and then the polhode will be the intersection with the coaxial quadric ( 2 ) , and the direction cosines of the invariable line will be ( 3 ) .
Denoting by the polar co-ordinates for the herpolhode in the invariable plane , ( 4 ) , and by solution of these equations ( 1 ) , ( 2 ) , ( 4 ) , suppose , \amp ; c. ( 5 ) , an ( 6 ) , where .
Denoting the component angular velocities about the principal axes by , and about the invariable line by ( 8 ) , and DARBOUX'S equations ( DESPEYROUS , ' Course de mecanique , ' II .
, note 17 ) .
are identified with our notation by putting ( 9 ) , ( 10 ) , and ( 11 ) , ( 12 ) , are two integrals of equations ( \mdash ; ) , . .
( 13 ) , or , as they may be written , ( \mdash ; ) , . .
( 14 ) .
Then ( 15 ) , ( 16 ) , supposing ( 17 ) , and inverting the integral , , ( 18 ) .
Again , projecting on the invariable plane the areas swept out on the co-ordinate planes .
( 19 ) , and ( 20 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
so that S ( 21 ) , and this reduces to ( 22 ) ; so that ( 2S ) .
Putting ( 24 ) , and , with ( 25 ) ; so that and from ( 6 ) S 1 ( 27 ) , or ( 28 ) , where @ ( 29 ) ; also ( 30 ) .
Then ( 31 ) , and ( 32 ) , where is a homogeneity factor , ; so that as a function ( 33 ) V0L .
in the pseudo-elliptic case ; and cancelling the secular term pt by making gives a purely algebraical herpolhode .
With the degree halved by the use of the variable , as in ( 8 ) ( 34 ) , ( 35 ) .
In the associated motion of a top , the projection on a horizontal plane of a point on the axis describes the hodograph of a herpolhode traced out by the axis of resultant angular momentum of the top ; and thus EULER'S angles and for the top are connected by a relation of the form ( 36 ) Science , Annals Mathematics , Series .
We may dispense with POINSOT'S rolling surface and consider the polhode as a material line , or as the edge of a material cone , as in Dr. FR .
SCHILLING'S model , constructed by MARTIN SCHILLING , Halle , and this is rolled on a fixed plane .
According to M. DE LA GOIJRNEBIE , the polhode is also a lin of curvature , the intersection of two confocal surfaces , an ellipsoid and hyperboloid of two sheets ; and will now be the angle between the generating lines of the hyperboloid of one sheet , one generator being perpendicular to the fixed plane , so that the other generator moves parallel to the axis of the top in the associated motion ( DARBOUX in DESPEYROUS , ' Mecanique , ' note 18 ) .
28 .
To utilise for the preceding results for odd order of where ( 1 ) , ( 2 ) , put , ( 3 ) , and thence express , . .
in terms of a new parameter , and thereby express the integral I of Returning then to the of , which , when normalised to LEGENDRE'S form , can be written ( 4 ) , and replacing by its equivalent , the series of functions THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
( 6 ) , ( 7 ) , ( 8 ) , ( 9 ) , ( 10 ) , is such that , or ( 11 ) , so that JACOBI'S division-values of the Second Stage can be calculated .
The Weierstrassian Sigma Functions of the First Stage in S4 can now be converted into Jacobian Theta Functions of the Second Stage by the relations on p. 52 of ' Formeln und Lehrsatze ' ; and forming the series of functions ( 12 ) , ( 13 ) , ( 14 ) , ( 15 ) , ( 16 ) , we find that or ( 17 ) , or one of its linear transformations ; and is always rational in and , in consequence of the relation ( 18 ) .
The function is the one that should be tabulated numerically , ' , Also ( 20 ) , or one of its transformations , depending on the region .
29 .
or Derive from in ( 3 ) S 6 , by putting , so that ( 1 ) .
Next put ( 2 ) , ( 3 ) , , ( 4 ) , ( 5 ) .
Then from ( 8 ) , S23 , in the region cn , , sn ( 6 ) , so that there results the well-known relation cn ( 7 ) , and in LEGENDRE , 1 , chap .
, p. 27 , ( 8 ) .
The following table shows the six linear transformations of the division-values for , obtained in accordance with the preceding theory by the substitutions ( 9 ) ; the accents are omitted from by changing to the complementary modulus .
The numerical value gives a modular angle ; while to .
he metions ruired for 30 .
( 1 ) , ( 2 ) .
To identify with the results in L.M.S. , 27 , put ( 3 ) ; then ( 4 ) , and from ( 9 ) , S7 , I ( 5 ) .
Put ' and ) ) ( 6 ) , with , ( 7 ) , and we obtain I ( v ) in the form employed in L.M.S. , 27 , p. 601 , derived by putting , with , ( 8 ) ; and in Region II .
, or ( 9 ) , where ( 10 ) , and ( 11 ) .
TfflRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
This is connected with FRICKE'S and Diss Leipzig , 1886 ) by the relation ( 12 ) The following table , analogous to the preceding one for , gives the divisionvalues for , the six transformations being derived from the substitution ( 13 ) ; and it may be compared with the corresponding results given by F. MULLER , ' Archiv der Mathematik und Physik , ' p. 161 , 1884 , ROHDE , ' Archiv , ' 1886 , p. 138 , and by W. GoRING , 'Math .
Ann 7 , p. 311 , 1874 , working on GAUSS'S unpublished memoirs .
For the application of this integral to a case of motion of the top , consult the ' Annals of Mathematics , ' vol. 5 , 1903 .
The numerical value will serve for verification ; this corresponds to case of ten cusps at intervals of in the associated .
motion of the top ( L.M.S. , 27 , p. 602 ) ; and will make or 1 , cases of Complex Multiplication .
31 .
, 14 ; , 5 , 6 ; ( 1 ) ; and on putting ( 2 ) , the relation ( 3 ) , an elliptic-function addition equation , connecting and ( 4 ) , ( 5 ) , and ( 6 ) .
Thence ( 7 ) , ( 8 ) , ( 9 ) .
The connexion with results in ' Math. Ann. , ' 32 , p. 87 , has been given in L.M.S. , 27 , p. ; here ( 10 ) , ( 11 ) , a transformation of the Third Order ; and ( 12 ) , suppose , where is the real half-period of the elliptic function of the elliptic argument ?
, and now , ( 13 ) , ( 14 ) , THIRD ELLIPTIC INTEGPAL AND LIPSOTOMIC P the roots of which are ( 15 ) , ( 16 ) , the roots of which are derived from the preceding , by the substitution ( 17 ) .
Thence the contoul of the period parallelogram is divided into segments of , and a sinlilar table of division-values for catl be constructed for the COl'responding regions of and compared with the results of GoRING and MULLER ; this must be reserved for the present .
The associated parameters of ( Math. Ann. , ' 14 ) have simple numerical values : ( 18 ) .
The parameter , employed in S 9 ( 2 ) , is given in terms of ' by the relation ( 19 ) thence in S 9 ( 6 ) can be determined , and in factors , as required for the Second Stage of , in terms of and Changing from the variable to , in S 28 ( 1 ) , we find ( 20 ) , I ) , , ( 23 ) , ( 24 ) , ( 25 ) .
Test by For the elliptic-section values the region may be selected in which , in accordance with S28 , ( 26 ) , ( 27 ) , ( 28 ) , , ( 30 ) , ( 31 ) .
With ( 32 ) , ' ( 33 ) , ( 34 ) , ( 35 ) .
THIRD ELLIPTIC INTEGRAL AND THF IPSOTOMIC PROBLEM .
The becomes by the substitutio1ls ( 1 ) , the same as those employed with in S22 , a new .
( 2 ) , representing a in ; and putting ( 3 ) , this reduces to a in with deficiency 2 , ( 4 ) , and this with , as in L.M. , p. 463 , becomes ( 5 ) , ( 6 ) , ( 8 ) , ( 9 ) , ( 10 ) , ( 11 ) .
Now the employed in , S 10 , distinguished here as , is given by , ( 12 ) , returning for a moment to the original and of S 10 , so that ( 1.3 ) ; and ( 14 ) , ( 15 ) , ( 16 ) , ( 17 ) , ( 18 ) .
This value of , employed in the previous equations ( 7 ) , S 10 , for will lead to the functio1ls of the Second Stage for 18 ; the result , in a modified form , has been given already the ' Archiv der Mathematik und Physik , ' 3 , Reihe , 1 , p. 74 .
Put 1 ( 19 ) , ' to agree with the ults in the ' Archiv , ' and in the modified form of S28 2 .
( 20 ) , THIRD ELLIPTIC INTEGRAL ANY ) ELLIPSOTOMIC PROBLCM .
( 21 ) , ( 22 ) , ( 23 ) , ( 24 ) , ( 25 ) , ( 26 ) , ( 27 ) .
Expressed in terms of , the parameters employed by ( .
' Math. Ann 32 , p. 128 ) are given by ( 28 ) , being ( Math. Ann 14 , p. 540 ) , ( 29 ) .
( 30 ) , ( 31 ) , ( 32 ) , ( 33 ) , PROFESSOR A. .
GREENHILL ON THE ( t34 ) , ( 35 ) , ( 36 ) , ( 37 ) .
Then for the section-values ( 38 ) , ( 40 ) , ( 41 ( 42 ) , ( 43 ) .
( 44 ) , ( 45 ) .
( 46 ) , ( 47 ) , ( 48 ) , { rom the table for , 29 .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
The relation in ( 21 ) , S22 , gives a in ; but putting ( 1 ) , it becomes ( 2 ) , a in with triple point at the origin ; so putting ( 3 ) , a quartic for , which can be resolved ( 4 ) , ( 5 ) .
Writing it ( 6 ) , and putting ( 7 ) , ( 8 ) , so that , with ( 9 ) , ( 10 ) .
This is the elliptic integral of the ikosahedron irrationality of KLEIN is unity ; it is curious also that this integral occurs as one of ABEL'S numerical exercises ( ' ( Euvres , ' 1 , p. 142 ) .
Then to connect up with , as in L.M.S. , 27 , p. 469 , ( 11 ) , after reduction ; and ( 12 ) , VOL. CCIII.\mdash ; A. reducing to ( 13 ) , where and are given in terms of by equation ( 64 ) .
Substituting these values of and in terms of and in the expression for I in ( 42 ) , S 11 , for , we obtain a result applicable to , and associated functions of the Second Stage .
34 .
The next two cases of , 26 and , 30 still present analytical difficulties not yet surmounted , could be treated by the trisection method of S6 , applied to 35 .
When is of the form , so that , or ( 1 ) , then is congruent to , as in Case III .
, and I ( v ) is given by I ( 2 ) , ( 3 ) .
Here the degree is halved by putting .
( 4 ) , so that ( 5 ) , and odd even ( 6 ) , ( 7 ) , also ( 8 ) , and putting ( 9 ) , THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
and ( 10 ) , ( 11 ) .
Thus for the vector of the corresponding herpolhode , ( 12 ) , , and in the associated hodograph described by the axis of a top ( 14 ) .
36 .
With , the resolution of into factors for the spherical catenary must be such that and are quadratics in ; and thus ( 15 ) , ( 16 ) , ( 17 ) , ( 18 ) .
37 .
The simplest cases may be cited before proceeding to the general theory .
( 1 ) , I ( 2 ) ; and in the associated dynamical applications , ( \mdash ; mt ) ( 3 ) , ( 4 ) , reducing to HALPHEN'S algebraical herpolhode when the se.cular term pt is cancelled by putting In the associated motion of the top ( ' Annals of Mathematics , ' vol. 5 ) ( 5 ) .
But the quadric transformation ( 6 ) , makes ( 7 ) , thus effecting the cation with the result , p. 147 , ffluvres , ' when the sign of is changed so as to obtain the circular form .
For the division-values sn , cn , dn ( 8 ) ; zn zd 1 \mdash ; , zs zc ( 9 ) ; ( 10 ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
I ( 1 ) , where the octahedron irrationality is given in terms of by ( 2 ) , unchanged by the substitution , which interchanges and Also , in the region \mdash ; 1 , ( 3 ) , ( 4 ) dn ( 5 ) .
The result is obtained by putting ( 6 ) ; and the values of are rationalised by putting ' 1 \mdash ; 8 ( 7 ) .
To normalise the results , a homogeneity factor is introduced , such that ( 8 ) .
The division-values are shown in tabular form , and the numerical test of is required for 39 .
ABEL'S form in ( Euvres , ' 1 , p. 142 , 2 , p. 160 , for becomes the same as in 2 , p. 148 , by replacing his by , his by 2 and by But now in our notation for , equation ( 14 ) S25 becomes ( 1 ) ; and putting ( 2 ) and replacing by , where , the roots of are ( 3 ) .
A quadric transformation is required in ABEL'S result to obtain our I ( v ) , namely , with the S above , and the sign of changed to obtain the circular form .
( 4 ) , ( 5 ) , ( 6 ) , ( 7 ) , and ( 8 ) , ( 9 ) , so that is rational .
Afterwards the degree is halved , as ABEL'S integral is .
A similar quadric transformation and halving of degree will be required for all even values of ABEL'S , to reduce his results to our form , involving heavy algebraical work .
ABEL'S integral ( ' ( Euvres , ' 2 , p. 148 ) , changed into the circular form , can be utilised for the construction of an algebraical orbit described under the central attraction ( 10 ) , THIRD ELLIPTIC INTEGRAL AND THE OMIC PROBLEM .
in the form ( 11 ) , ( 12 ) , ( 13 ) , so that the condition ( 14 ) , ( 15 ) , is satisfied by taking ( 16 ) .
A degenerate circular orbit is obtained by taking ( 17 ) .
But the integral in ' ( Euyres , ' 2 , p. 147 , putting , requires so that we merely obtain a conic for 40 .
or I ( 1 ) , ( 2 ) , ( 3 ) , ( 4 ) , VOL. CCIII .
\mdash ; A. results obtained by putting ' ( 6 ) .
Treated by the trisection of , putting in equation ( 8 ) , S 14 , and putting ( 8 ) , ( 9 ) , and ' ( 10 ) , so that ( 11 ) , and is the equivalent of in -FRICKE , ' Modulfunctionen , ' 1 , p. 688 , while ( 12 ) , ( 13 ) , ( 14 ) , and GTERSTER'S ( 15 ) .
Also so that ( 17 ) .
KIEPERT'S notation , 'Math .
Anm 32 , p. 104 , , \amp ; .
( 18 ) .
Putting ( 19 ) , the section values are shown in the table:\mdash ; 41 .
In the general case a distinction must be made between odd or even .
I. , or ( 1 ) ; ( 2 ) , ( 3 ) , ( 4 ) , ( 5 ) , ( 6 ) , ( 7 ) , ( 8 ) , where denotes the octahedron irrationality ; so that if the expressions are normalised by the homogeneity factor ( 9 ) , ( 10 ) ; ( 11 ) .
Now I ( 13 ) .
THIRD ELLIPTIC INTEGRAL AND THE fOMIC PROBLEM .
Next putting ( 14 ) , ( 15 ) , and with S ( 16 ) , S positive , negative , ( 17 ) ; so that when normalised by becomes and becomes In the sequel it is convenient to put dn ( 18 ) , ( 19 ) , ( 20 ) ; so that ( 21 ) , ( 23 ) , ( 24 ) , ( 25 ) , PROFESSOR A. G. ( iREENHILL ON ( 26 ) , @ ( 27 ) , ( 28 ) .
( 29 ) , ( 30 ) , ( 31 ) , ( 32 ) , ( 33 ) .
Thus in the quadric transformation ( 34 ) , ( 35 ) , ( 36 ) , ( 37 ) , involving powers of only ; this will be useful in the sequel .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
To connect up with the order , denoting by ( 38 ) , ( 39 ) .
We have found ( 40 ) , and proceeding with the series , writing for ( 41 ) , ( 42 ) .
Working these out with the assistance of the analysis given subsequently , ( 43 ) , and generally ( 44 ) , where is a rational function of and , and derivable from it by the substitution .
Thus for .
The results in the sequel give \mdash ; \mdash ; ( 46 ) , and therefore ( 47 ) , ( 48 ) , ( 49 ) , ( 50 ) , ( 51 ) , where the 's are found in S45 ( 3 ) , and the 's are derivable by the substitution .
These expressions have been tested by putting ( 52 ) , which makes ( 53 ) , ( 54 ) , ( 55 ) , and thus yerifies ( 56 ) .
42 .
( 1 ) , ( 2 ) , ( 3 ) , ( 4 ) , ( 5 ) , Now , as in the ' Archiv , ' III .
, 1 , p. 75 , with the there replaced , and by THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
( 6 ) , ( 7 ) , ( 8 ) , ( 9 ) , ( 10 ) , ( 11 ) , ( 12 ) , \mdash ; \mdash ; ( 13 ) , while and can be calculated and the whole result verified by differentiating ( 6 ) and equating coefficients of powers of A change of into will change into 9 The employed for is , so that , putting ( 14 ) , we find ( 15 ) , so that , with ( 16 ) , ( KIEPERT , ' Math. Ann 32 , p. 107 ) , and then , with VOL. , \amp ; c. ( 18 ) .
Then , in S 28 , ' \amp ; c. ( 19 ) .
43 .
, 3 , 5 , 9 , 11 , 13 .
( 1 ) .
( 2 ) .
A change of into changes into ) .
Now , with ) , ( 3 ) ; and changing into ( 4 ) .
( 5 ) .
( 6 ) .
Substituting for and in terms of and , and cancelling a factor there results ( 7 ) , a quadratic for To obtain a rational expression for , we put ( 8 ) , and now the relation ( 7 ) becomes THIRD ELLIPTIC INTEGRAL AND PROBLEM .
( 11 ) , ( 12 ) ; so that , with , ( 13 ) , then ( 14 ) Math. .
Introducing these values into ( 35 ) , S 41 , ( 15 ) , to be compared with the expression for in , L.M.S. , 27 , p. 440 , putting iting q or for ( 16 ) .
( 17 ) , and after reduction ( 18 ) , or ( 19 ) , to be interpreted as an elliptic function relation .
Rationalising again ( 20 ) , or arranged in powers of ( 21 ) , ( 22 ) , ( 23 ) , 44 .
The.relation ( 1 ) to ( 2 ) , ( 3 ) , and putting ( 4 ) , ( 5 ) , and thereby the superfluous factor is cancelled , and finally ( 6 ) , a cubic equation for Putting as before ( 7 ) , this relation ( 63 ) becomes ( 8 ) , ( 9 ) , and putting ( 10 ) , and arranged as a quadratic in ( 11 ) , ( 12 ) , ( 13 ) , as in , S 32 ; and ( 14 ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
Here the relation to be reduced becomes ( 1 ) .
The algebraical work has been carried out by Mr. R. H. MACMAHON , and he reduces the relation to a quintic in ( 2 ) , , , , , \mdash ; 1 \mdash ; \mdash ; ( 3 ) .
As before , putting ( 4 ) , the relation can be halved in degree , and becomes ( 5 ) , a quartic in of similar structure to the one in , S 33 ( 3 ) , and capable of resolution in a similar manner into ( 6 ) , ( 7 ) .
Put ( 8 ) , and then ( 9 ) , if , ( 10 ) ; and ( 11 ) , a relation which requires interpretation , connecting , and , elliptic functions of It is now possible theoretically to determine the coefficients in ( 12 ) , and to construct an algebraical herpolhode and associated top motion complete in 44 branches .
This is as far as we can at present , as the next cases of and 60 must await the solution of and 30 , not yet accomplished .
46 .
II .
( 1 ) , ( 2 ) , ( 3 ) , ( 4 ) , ( 5 ) , ( 6 ) , and normalised by ( 7 ) , osn , ( 8 ) , ( 9 ) , and the associated integral I ( 10 ) THIRI ) ELLIPTIC INTEGBAL AND THE ELLIPSOTOMIC PROBLEM .
( 12 ) .
In this case II .
put ( 13 ) , then ( 14 ) , Ss \mdash ; ( positive ) ( 15 ) .
Now , with ( 16 ) , , suppose ( 17 ) ; ( 18 ) ; and the relation ( 13 ) becomes ( 19 ) , ( 20 ) , ( 21 ) , ( 22 ) , ( 23 ) , ( 24 ) , ( 25 ) , ( 26 ) , We now find , as in S41 , writing for ( 29 ) , and then , from the relation ( 30 ) , writing for , and putting ( 31 ) , ( 32 ) , ( 33 ) , ( 34 ) , \mdash ; 1 ) , and is obtained from by changing into 47 .
and now the condition ( 1 ) reduces to ( 2 ) , THIRD ELLIPTIC INTEGRAL AN1 ) THE ELLIPSOTOMIC PROBLEM .
( 4 ) , I \mdash ; -\mdash ; ( 6 ) , ( 7 ) , ( 8 ) , ( 10 ) , ( 12 ) .
Bisection formulas for in the region will lead to the same results , with ' ( 13 ) , ( 14 ) , ( 15 ) , having the substitution ( 16 ) .
Hence the following table of section-values for , holding in the region ; tested numerically by , as on p. 278 , region VOL. CCIIL\mdash ; A , 2 and now dn ' \amp ; c. ( 10 ) .
In notation ( ' Math. Ann 32 , p. 116 ) , , \amp ; c. ( 11 ) .
Treating by the trisection method of S14 applied to , we put , , ( 12 ) ; so that , putting in the equation ( 8 ) , S 14 , it may be written ( 13 ) .
To agree with the notation in -FRICKE , ' Modul-Functionen , ' 1 , p. 68S , put ( 14 ) , , ( 15 ) , ( A ( 16 ) , ( 17 ) ; and denoting the by , ( 18 ) , ( 19 ) , ( 20 ) ; and GIERSTER'S ( ' Math. Ann 14 ) is given by ( 21 ) .
So also can be discussed by a trisection of , by putting ( 22 ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
( 1 ) , and this leads to , ( 3 ) , so that , ( 5 ) , a in .
Put ( 6 ) , a quadratic in ( 7 ) , ( 8 ) , ( 9 ) , ( 10 ) , ( 11 ) , ( 12 ) , ( 13 ) , ( 14 ) , as in , so that the bisection formulas can be carried one stage further .
We find now ( 15 ) .
, 50 .
The relation ( 1 ) leads to ( 2 ) , or putting ( 3 ) .
With , and arranged in powers of bn ( 4 ) , and with ( 5 ) this becomes ( 6 ) .
A factor may be cancelled and ( 7 ) , ( S ) .
THIRD ELLIPTIC INTEGRAL AND THE ELLIPSOTOMIC PROBLEM .
Put ( 9 ) .
Put ( 10 ) .
Put ( 11 ) .
Put ( 12 ) .
Put ( 13 ) , ( 14 ) , ' ( 15 ) , ( 16 ) , ( 17 ) , ( 18 ) , ( 19 ) .
( 20 ) , as in ( 2 ) , S 42 , for To connect up the results , and 40 , in ( 21 ) , ( 22 ) , are equal , so that ( 23 ) , and thus , with ( 24 ) , ( 25 ) , ( 26 ) KtEPERT , Math. , and the preceding results are thus merely bisection formulas for We arrive at the conclusiqn that it is the quotient of two theta functions , and , with constant phase difference , which is required in dynamical application , the functions , S for instance employed by KLEIN in top-motion ; but the separate theta function has no mechanical interest .
This quotient , qualified by the constant factors and , is an elliptic function of when is a half-period , dn for instance when is the half-period , and the quotient is the root of an algebraical function of the elliptic functions of when is an aliquot part of a period ; in this way we express the result of ABEL'S pseudo-elliptic integral .
The formation of this algebraical function for the simplest values of has been our chief object , and in the course of the work the ellipsotomic problem has been carried out of the determination of the Division-Values of the Elliptic Function .
The Transformation problem may be considered solved at the same time by means of symmetric functions of the division-values ; but as Transformation has no dynamical utility , it has not been developed in this memoir .

