]\gt ; XII .
On the Bicircular Quartic.\mdash ; Addihon to Professor Memoir " " On a new Form of Tangential Equation By A. , LL. D. , F.R.S. , Sadlerian Professor of Pure Mathematics in the University of Cambridge .
Received January 24 , \mdash ; Read February 22 , 1877 .
PROFESSOR CASEY communicated to me the MS .
of the foregoing Memoir , and he has permitted me to make to it the present Addition , containing further developments the theory of the bicircular quartic .
Starting from his theory of the fourfold generation of the curve , Prof. CASEY shows that there exist series of inscribed quadrilaterals ABCD whereof the sides AB , , DA pass through the centres of the four circles of inversion respectively ; or ( as it is convenient to express it ) the pairs of points , , , to the four modes of generation respectively , and may be regarded as depending upon certain parameters respectively , any three of these being in fact functions of the fourth .
Considering a given quadrilateral ABCD , and giving to it an infinitesimal variation , we have four infinitesimal arcs AA ' , BB ' , ; these are differential expressions , and BB ' of the form , BB ' and of the form and of the form and AA ' of the form ; or , what is the same thing , AA ' is expressible in the two forms and , BB ' in the two forms and , \amp ; c. , the identity of the two expressions for the same arc of course depending ou the relation between the two parameters .
But any such monomial expression ) of an arc would be of a complicated form , not obviously reducible to elliptic functions ; ASEY does not obtain these monomial expressions at all , but he finds geometricaly monomial expressions for the differences and sum BB'\mdash ; AA ' , \mdash ; BB ' , , DD'-AA ' ( they cannot be all of them differences ) , and thence a quadrinomial expression AA ' his ; and that without any explicit consideration of the relations which connect the parameters .
I propose to complete the analytical theory by establishing the monomial equations AA ' , \amp ; c. , and the relations between the parameters which belong to an inscribed quadrilateral ABCD , so as to show what the process really is by which we pass from the monomial form to a quadrinomial form AA ' or wherein each term is separately expressible as the differential of an elliptic integral ; and to further develop the theory of the transformation to elliptic integrals : we require to establish for these purposes the fundamental formulae in the theory of the bicircular quartic .
I remark that in the various formulae are constants which enter only in the combinations , that are taken as current coordinates , and these letters , or the same letters with suffixes , are taken as coordinates of a point or points on the bicircular quartic ; the letters , , , are used throughout as variable parameters , viz. we have ; so that , are functions of a single parameter , and similarly .
, are functions of the parameters respectively ; we sometimes use these or similar expressions of \amp ; c. as trigonometrical functions of a single parameter , but more frequently retain the pair of quantities , considered as connected by an equation as above , and so as equivalent to a single variable parameter .
for the fourfold generation of the Bicircular Quartic.\mdash ; Art .
Nos. 1 to 5 .
1 .
We have four systems of a dirigent conic and circle of inversion , each giving rise to the same bicircular quartic : viz. the bicircular quartic is the envelope of a generating circle , having its centre on a dirigent conic , and cutting at right angles the corresponding circle of inversion ; or , what is the same thing , it is the locus of the extremities of a chord of the generating circle , which chord passes through the centre of the circle of inversion , and cuts at right angles the tangent ( at the cent1t of the generating circle ) to the dirigent conic ; the two extremities of the chord are thus inverse points in regard to the circle of inversion .
The four systems are represented by letters without suffixes , or with the suffixes 1 , 2 , 3 respectively , and we say that the system , or mode of generation , is , 1 , 2 , or 3 accordingly .
2 .
The dirigent conics are confocal , and their squared semiaxes may therefore be represented by ( which are in fact functions of the five quantities ) ; and we can in terms of these data express the equations as well of the dirigent conics as of the circles of inversion ; viz. taking X , as current coordinates , the equations are , or , or , or , or PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
where 3 .
The geometrical relations between the dirigent conics and circles of inversion are all deducible from the foregoing formulae ; in particular the conics are confocal , and as such intersect each two of them at right angles ; the circles intersect each two of them at right angles .
Considering a dirigent conic and the corresponding circle of inversion , the centres of the remaining three circles are conjugate points in regard as well to the first-mentioned conic , as to the first-mentioned circle ; or , what is the same thing , they are the centres of the quadrangle formed by the intersections of the conic and circle .
4 .
The centre of the conics and the centres of the four circles lie on a rectangular hyperbola , having its asymptotes parallel to the axes of the conics .
Given the centres of three of the circles ( this determines the centre of the fourth circle ) and also the centre of the conic , these four points determine a rectangular hyperbola ( which passes also through the centre of the fourth circle ) ; and the axes of the conics are then the lines through the centre , parallel to the asymptotes of the hyperbola .
5 .
The equation of the bicircular quartic may be expressed in the four forms the equivalence of which is easily verified by means of the foregoing relations .
Determination as to Reality.\mdash ; Art .
Nos. 6 and 7 .
6 .
To fix the ideas suppose that is positive ; then in order that the centres of the four circles of inversion may be real we must have positive , but negative ; and this will be the case if are all positive , but one of them negative , and the other three positive .
In reference to a figure which I constructed I found it convenient to take to be in order of increasing magnitude : this being so we have positive , negative ; and the other like quantities all positive : we then have and each positive , negative , positive : viz. the conics and circles are Hyperbola corresponding to real circle Ellipse , , real circle , , imaginary circle ( viz. the radius is a pure imaginary ) , , real circle and where the confocal ellipses are in order of increasing magnitude .
The centre is here a point within the triangle formed by the remaining three centres .
It will be convenient to adopt throughout the foregoing determination as to reality .
7 .
It may be remarked that a circle of a pure imaginaly radius , where is real , may be indicated by means of the concentric circle radius , which is the concentric$ orthotomic circle ; and that a circle which cuts at right angles the original circle cuts diametrally ( that is , at the extremities of a diameter ) the substituted circle radius : we have thus a real construction in relation to a circle of inversion of pure imaginary radius .
Investigation of .
Nos. 8 to 17 .
8 .
The coordinates of a point on the dirigent conic may be taken to be : and we hence prove as follows the fundamental theorem for the generation of the bicircular quartic .
Consider the generating circle , centre , which cuts at right angles the circle of inversion .
If for a moment the radius is called , then the equation of the generating circle is ; the condition for the intersection at right angles is and hence eliminating , the equation of the generating circle is ; and considering herein as variable parameters connected by the foregoing equation , we have as the envelope of this circle the required bicircular quartic .
9 .
It is convenient to write ; the equation then is ; PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
the derived equation is ; and from these two equations , together with the equation in and its derivative , we find ; from these last equations , and the equations , , eliminating , we have that is the required equation of the bicircular quartic .
10 .
We have thus , as the equations which serve to determine the bicircular quartic : if from these equations , together with , we eliminate X , , we have expressed as a function of ; and thence also X , expressed in terms of ; that is in effect the coordinates X , of a point of the bicircular quartic expressed as functions of a single variable parameter .
The process gives .
this is or putting for shortness , this is ' or say the two values are ; to preserve the generality it is proper to consider ( the positive or the negative one , as the case may be ) of the radical .
11 .
Considering the root we have , and from these equations we obtain ; but from the equation for we have that is whence , .
12 .
The differentials can be expressed in terms of a single differential , viz. writing and , then we have It is to be observed that when the dirigent conic is an ellipse , is a real angle , and is positive ( whence also is real and positive ) , but when the dirigent conic is a hyperbola , is imaginary , and is negative ; we have , however , in either case and we may therefore write where is positive ; is the increment of arc on the conic , this arc being measured in a determinate sense , and therefore being positive or negative as the case may be : has thus a real positive or negative value , even when is imaginary , and it is convenient to retain it in the formulae .
13 .
It may further be noticed that if denote the inclination to the axis of of the tangent to the dirigent conic at the point , then , where viz. we have giving , as is easily verified , ; we have therefore or which is another interpretation of 14 .
Substituting for their values , the formulae become PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
We have ; that is ; and consequently the foregoing expressions of become or finally 15 .
We have ; viz. this is , the radius of the generating circle .
Hence if , be the element of arc of the bicircular quartic , this element being taken to be positive , we have where denotes a determinate sign , , as the case may be .
16 .
I stop to consider the geometrical interpretation ; introducing , the formula may be written we have ; is the perpendicular from the centre of the circle of inversion upon the tangent to the dirigent conic , and is the half-chord which this perpendicular forms with the generating circle .
Hence perpendicular --chord -chord , the numerator being in fact the distance of the element ( or point X , Y ) from the centre of inversion : the formula thus is where is the radius of the generating , 9 the distance of the element from the centre of the circle of inversion , and the chord which this distance forms with the generating circle .
If we consider the two points on the generating circle , and write for the element at the other point , then we have ( which is CASEY'S formula ) .
17 .
The foregoing forms of are those which give most directly the required value of , but I had previously obtained them in a different form .
Writing then ; or since this is , ; that is and similarly PROFESSOR CAYLEY ON THE BICIRCULAR We have therefore , , and thence a value of which , compared with the former value , gives an equation which may be verified directly .
Formuloe for the lnscribed Quadrilateral.\mdash ; Art .
Nos. 18 to 22 .
18 .
We consider on the curve four points , , forming a quadrilateral , A D. The coordinates are taken to be ( X , Y ) , , , respectively .
It is assumed that , , , belong to the generations 1 , 2 , 3 , , and depend on the parameters , , , respectively .
We write , , , , and then , denoting as above a determinate value , positive or negative as the case may be , of the radical , and similarly denoting determinate values of these radicals respectively , each radical having its own sign at pleasure , we further write ; and this being so , we must have , , , , , , , ; PROFESSOR CAYLEY ON TIfE BICIRCULAR QUARTIC .
and then from the values of X , , we have giving ; and similarly which are the relations connecting the parameters , , , of the quadrilateral .
19 .
We have thus apparently four equations for the determination of four quantities , or the number of quadrilaterals would be finite ; but if from the first and second equations we eliminate , or from the third and fourth equations we eliminate , we find in each case the same relation between , , viz. this is found to be ; and we have thus the singly infinite series of quadrilaterals .
We have , of course , between , the like relation , 20 .
The relation between , may be expressed also in the two forms : In fact , the first of these equations is which , by virtue of the original form of relation , is ; or , in the first term , writing PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
and in the third term this is ; and in this equation the coefficients of and of are separately : in fact the coefficient of is ; and similarly the coefficient of is And in like manner the second equation may be verified .
21 .
The two equations are : ; or , substituting for their values , these are and similarly Differentiating the equation we have ; and writing herein ' we find $ ; viz. , dividing by , this becomes , that is ; or , completing the system , we have which are the differential relations between the parameters , or , , , .
22 .
From equations , we found ; the new values , give in like manner ; and in virtue of the relation just found between and these two sets of values will agree together if only These are easily verified : the first is , viz. this is , which is right ; and similarly the second equation gives , which is right .
From the first values of we have , as above , PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
and the second values give in like manner where .
It will be observed that we have in effect , by means of the relation , proved the identity of the two values of Considering the quadrilateral ABCD , and giving it an infinitesimal variation , so as to change it into , then is the element of arc AA ' ; and writing in like manner for the elements of arc , we have , of course , a like pair of values for each of the elements Formuloe for the elements of Arc .\mdash ; Art .
Nos. 23 to 27 .
23 .
The formulae are ' ' ' where the 's each denote .
Supposing as above that is negative , but that are positive ; then have opposite signs : but have the same sign , as have also and , and and .
We may take , and as each of them positive : the signs of ' - , or \mdash ; , : hence to make all positive , must have signs of ; or else the reverse signs : hence in either case ; or the equations are ' 24 .
But we have .
; and hence , putting for shortness and consequently which are the required formulae for the elements of arc .
25 .
The determination of the signs has been made by means of the particular figure ; but it is easy to see that the pairs of terms could not for instance be , or any other pairs such that it would be possible to eliminate , and thus obtain an equation such as ; this would , by virtue of the relations between , become an equation not deducible from the relations which connect ) , , and which therefore cannot be satisfied by the variable quadrilateral .
26 .
The differentials of the formulae are , it will be observed , of the form ' where is a mere constant , , and viz. the form is which is , in fact , the same as form in ( equation ( 300 ) , his being ) .
Writing as before in place of his , the differential expression becomes simply : PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
but expressed as a function of is an irrational function , and would be the root of such a function ; so that if the form originally obtained had been this form , it would have been necessary to transform it into the first-mentioned form , in which is expressed as a function of , that is of 27 .
The system of course is where \amp ; c. ; and this is the most convenient way of writing it .
Reference to Figure.\mdash ; Art .
No. 28 .
28 .
I constructed a bicircular quartic consisting of an exterior and interior oval with the following numencal data : ; \mdash ; not very convenient ones , inasmuch as the exterior oval came out too large .
The annexed figure shows , 1 , 2 , 3 , the centres of the circles of inversion , the interior oval , and a portion of the exterior oval , also the origin and axes ; it will be seen that the centres inside the interior oval , the centres 1 , 3 outside the exterior oval : I add further the values We thus see how there exists a series of quadrilaterals ABCD , where are situate on the interior oval , on the exterior oval .
Considering the sides as drawn in the senses A to to to to , and representing the inclinations measured from the positive infinity on the axis of in the sense to , by respectively , then in passing to the consecutive quadrilateral , we have and decreasing , increasing , that is , and negative , and positive ; then reckoning the elements , BB ' , , that is , as each of them positive , we have and thence which the correct signs in regard to the particular figure .
Reduction of to Elliptic Integrals.\mdash ; Art .
No. 29 .
29 .
The expression in question is CAYLEY ON IHE BICIRCULAR QUARTIC .
where where the coefficients are such that identically , that is .
30 .
It is found that are the roots of a cubic equation , which being so , we may assume , or the second condition in fact is ; and this being so , we find without difficulty the values : ( to make these positive the order of ascending magnitude must , however , be not as heretofore , but .
we must have , all positive ) .
31 .
The above are the values of the squares of the coefficients ; we must have definite relations between the signs of the products aa ' , bb ' , , \amp ; c. , viz. we may have ' bb ' \ldquo ; and further ' and also ' .
These values , in fact , satisfy the several relations which exist between the nine coefficients , viz. the original expressions of , in terms of give conversely expressions of in terms of , the two sets being and we have then the relations , , PROFESSOR CAYLEY ON THE BICIRCULAR QUARTIC .
sin giving the four sets each of six equations , or say , , , , ; all which formula are in fact satisfied by the foregoing values of the expressions \amp ; c. 33 .
We then have ; and the radical which multiplies being the differential becomes ' that is The denominator could , of course , be reduced to the form ; but the actual form seems preferable , inasmuch as it puts in evidence the linear factors , and there seems to be no advantage in further reducing the integral .

