XVIII .
On the Resolution of attractive Powers .
By Edward Waring , M. D. F. R. S. and Lucasian Professor of Mathematics at Cambridge .
Read May 28 , 1789 .
1 .
A FORCE acting at a given point may be resolved by an infinite number of ways into two , three , or more ( n ) forces acting at the same point , either in the same or different planes with the given force and each other ; and , vice versa , any number of such forces acting in the same or different planes may be reduced into one .
Ex. Fig. 1 .
Tab. III .
Let a body A be acted on by three forces AB , AC , and AD , not being in the same plane ; reduce any two of them AB and AC to one AE , by completing the parallelogram ABEC ; then reduce the two forces AE and AD to one AF by completing the parallelogram AEFD , and the three forces AB , AC , and AD , are reduced to the one AF .
2 .
If n forces act on the body A at the same time , and any ( n 1 ) of them be reduced to one , the force resulting will be situated in the same plane with the remaining , and force equivalent to the ( n ) forces .
3 .
If one force a be resolved into several others x , y , z , v , &amp;c. situated in different planes , and the sines of the angles , which the forces y , z , v , &amp;c. contain with the plane made by the direction of the forces x and a be respectively s , s ' , s ' ' , &amp;c. then will sy~s'z~s''v~ &amp;c. = 0 .
PROBLEM 1 .
Fig. 2 .
Given the law of attraction of each of the parts of a given line in terms of their distance from a given point P ; to find the attraction of the whole line ab on the point P. Find the attraction of the line ab on the point P in the two directions Pf and fb by the following method .
Draw Px from the point P to any point x of the line ab , the force acting on the point P by the particle xy will be the given function ( determined from the given law of attraction ) of the distance into the particle ; draw also Ph perpendicular from the point P to the line ab , and let Pf = a , hf = b , and fx = y ; then will the distance Px =~ ( a2 ~ 2by+y2 ) , and the function of the distance into the particle xy =~ ( ~ ( a2 ~ 2by+y2 ) ) Xy =F ( y ) Xy ; let this be denoted by lx situated in the line Px , which resolve into two others nx = yy x F : ( y)/ Px =~ ( a2 ~ 2by + y2 ) situated in the line ab , and ln ( in a direction parallel to Pf ) = ay x F : ( y)/ ~ ( a2 ~ 2by + y2 ) ; find the fluents of the fluxions yy x F : ( y)/ Px and ay x F : ( y)/ Px contained between the values af and fb of the line fx = y , which suppose Y and V respectively ; through the point p draw Pie parallel to fb = Y , and in the line Pf assume Pu = V ; complete the parallelogram Puzy ; Pz will be the force of the line ab on the point P. Cor. If F : ( y ) varies as any power or root ( 2n ) of the distance Px =~ ( a2 ~ 2by + y2 ) , and n 1/ 2 be an integer affirmative number or ~ , the fluents Y and V of both the fluxions can be found in finite algebraical terms of y ; if n 1/ 2 be an integer negative number , both the fluents can be found in the above1 mentioned mentioned finite terms together with the arc of a circle , whose radius is ~ a2 b2 and tangent y~b , unless n 1/ 2 = ~ , in which case the fluent Y involves that circular arc , and also the logarithm of y2 ~ 2by + a2 .
If n 1/ 2 denotes a fraction whose denominator is 2 , both the fluents can be expressed by the finite terms together with the log .
of y~b +~ ( y2 ~ 2by + a2 ) .
If the fluents be given , when n is a given quantity , and n 1/ 2 not a whole affirmative number , from them can be deduced the fluents of any fluxions resulting by increasing or diminishing n by a whole number , unless in the above-mentioned case of n 1/ 2 = 1 .
If b= 0 , and consequently ~ the line Pf is perpendicular to the given line ab , the fluent Y will be expressed by the finite terms , unless n 1/ 2 = 1 , in which case it will be as 1/ 2 log .
( y2 + a2 ) when properly corrected .
These fluxions Y and V may be transformed into others , whose variable quantity is Px =u the distance from P , by substituting in the fluxions for y and y their respective values ~ ( u2 a2 + b2)~b and uu/ ~ ( u2 a2 + b2 ) , and consequently for ~ ( y2 ~ 2by + a2 ) its value u. PROBLEM II .
Fig. 3 .
Given the attraction of each of the parts of a given surface in terms of their distance from a given point P , and an equation expressing the relation between an absciss Ap = x , and its correspondent ordinates pm =y of the surface ; to find the attraction of the surface on the given point P. First , by the preceding proposition find the attractions Y and V of any ordinate mp m ' in the directions of the ordinate pm and of the line Pp ; and from the equation expressing the relation between the absciss and ordinates of the given curve , find the absciss in terms of the ordinates ( pm ) = ~ : ( y ) , and thence x= ~ : ( y ) Xy and ~ ( a'2 ~ 2sa ' x+ x2 ) = ~ ' : ( y ) , where PA = a ' and s= cosine of the angle , which the absciss Ap makes with the line PA ; then find the fluents of the three fluxions xXY=yXYX ~ : ( y ) , xXVx x/ ~ ( a'2 ~ 2sa ' x+ x2 = ~ : ( y ) X yX ~ : ( y)/ ~ : ( y ) XV and xX a'V/ ~ ( a'2 ~ 2sa ' x+ x2 ) =yX a'V/ ~ ' : ( y)contained between the values of y , which correspond to the extreme values of x , which suppose Y ' , V ' , and Z ; and draw through the point P the lines Pie and Pz respectively parallel to the ordinates pm and to the absciss Ap and equal to rX Y ' and V ' ; assume Pu in the line ( PA)=t X Z , r and t denoting the sines of the angles , which the ordinates pm and line AP make with the absciss Ap : reduce these three forces Pie , Pz , and Pu , to one Pf , and Pf will be the force of the surface on the point P. Cor. 1 .
If for y and y be substituted their values in terms of x and x , deduced from the equation expressing the relation between the absciss Ap and ordinate pm of the given curve , thence will be deduced the above-mentioned fluents Y , V , Y ' , V ' , and Z , in terms of x ; and in the same manner , if for x and x be substituted in the fluxions or fluents resulting their values ~ ( u2 a'2 + 12 a'2)~sa ' , and its fluxion , there will result the above-mentioned fluxions or fluents in terms of u the distance from the point P. Cor. 2 .
Let the curve be a circle , of which A is the center , PA a line perpendicular to the plane of the circle , and the ordinate pm perpendicular to the absciss Ap ; the forces on each side of the absciss Ap will be equal , and the force in the direc6 tion tion of the absciss Ap will be equal to that in the contrary direction ; the force in the direction ( PA ) =4x~ au/ ~(u2 a2 ) x ~ uy/ ~u2 + y2 x F : ( ~(u2 + y2 ) ) = W , in which F : ( ~u2 + y2 ) is the function of the distance , according to which the given force on the particles varies ; the fluent ~ uy/ ~(u2 + y2 ) x F : ~(u2 + y2 ) is contained between the values o and ~(r2 + a2 u2 ) of the quantity y , and the fluent W is contained between the values a and ~(a2 + r2 ) of the quantity u , where a= PA and r the radius of the circle ; but the same force is =2x 3,14159 &amp;c. x~ au x F : ( u ) , where F : ( u ) denotes the given function of the distance ( u ) , and the fluent is contained between the values a and ~a2 + r2 of u. PROBLEM III .
To find the attraction of a given solid on a given point P. Find the attraction of every parallel section on that point by the preceding problem , and multiply it into the correspondent fluxion of the first abscissa AP , and also find the fluent of the resulting fluxion , which , properly corrected , multiply into the fine of the angle , which the first abscissa makes with the parallel sections , and the product will be proportional to the attraction of the solid on the given point P. 2 .
Fig. 4 .
Let the solid ABCH be generated by the rotation of a given curve round its axis AB , which passes through the point attracted P , and this solid be supposed to consist of small evanescent solids , whose bases are the surfaces OF , of , &amp;c. of spheres , of which the center is P , and altitudes Ff , &amp;c. the increments of the base AB contained between the two contiguous surfaces OF and of : from the points E and e of the curve draw ED and ed perpendicular to the axis AB , and ES perpendicular to the arc Ee of the given curve at the point E , and meeting the axis AB in S ; then will the evanescent solid EFse =px PE x FD x Ff =px FD x PS x Dd ( because Ff = PSxDd/ PE ) =px ~(z2 + y2 ) zx zz = yy , where z and y denote respectively the absciss PD , and its correspondent ordinate DE of the given curve .
The increment of the attraction of the surface OF on the point P in the direction PD will be as the increment of the surface ( p x PE x Dd ) x PD/ PE x force of each particle =px PD x Dd x given force of the particle ; but the fluent of the fluxion PD x Dd contained between the points E and F is = 1/ 2 PE2 1/ 2 PD2 = 1/ 2 ED2 , whence the attraction of the evanescent solid EFse is as 1/ 2 px ED2 x Ff x F : ( ~x2 + y2 ) force of each given particle at the distance ( PE = ~(x2 + y2 ) ) = 1/ 2p x ED2 x PS/ PE x Dd x F : ( ~z2 + y2 ) = 1/ 2 py2 x zz~yy/ ~(z2 + y2 ) x F : ( ~(z2 + y2 ) ) ; the fluent of which , properly corrected , is as the attraction of the solid on the point P ; p denotes the circumference of a circle , whose radius is 1 .
Cor. 1 .
The fluxion of this solid is 1/ 2 py2 z=Y which deduced from the preceding principles =px ( ~(z2 + y2 ) z ) x ( zz = yy ) = V , and consequently their fluents between two values of z , which correspond to two values of y= o , will be equal to each other .
Cor. 2 .
The increment of the attraction of this solid as given in this proposition 1/ 2 px y2 x zz = yy/ ~(z2 + y2 ) x F : ( ~(z2 + y2 ) = U , but in the preceding proposition the force of a circle on the point P=px~ au x F : ( u ) , where u= ~(z2 + y2 ) and a= z , and y or u the only variable quantity contained in the fluxion ; and consequently the fluxion of the attraction of the solid pxz~z yy/ ~z2 + y2 x F : ( ( z2 + y2)1/ 2 ) = W ; therefore , if for the fluent of zyy/ ~(z2 + y2 ) x F : ( ( z2 + y2)1/ 2 ) be substituted its fluent contained between the values a and the value of y , which in the given equation corresponds to z ; then the fluents of U and W contained between the two values of z , which corresponds to two values of y= o , will be equal to each other .
The difference of the fluents of Y and V , &amp;c. contained between any other two values of z , can easily be deduced from the difference of two segments of spheres .
1 .
It may not be improper to remark in this place , that from different methods of finding the sum of quantities , the fluents of fluxions , the integrals of increments , &amp;c. quantities may often be deduced equal , which otherwise cannot without some difficulty ; of which instances are contained in the Meditations , and I shall here subjoin one or two more to those already given in this Paper .
Ex. 1 .
Any curvilinear area ABC , &amp;c. may be supposed to consist of evanescent areas EFef , of which the base OF is the arc of a circle , whose radius is PE = ~(z2 + y2 ) and sine ED = y , and altitude Ff , and consequently the fluxion of the area = Ff x arc ( A ) of a circle whose radius is PE and sine ED = PS/ PE xzxA= zz~yy/ ~ ( z2+y2 ) xA= V ; the fluent of V contained between the two values of z which correspond to two values of ( y ) =0 will be equal to the fluent of yz contained between the same two values of z. Ex. 2 .
The attraction of any circular arc OF in the direction PD on a point P ( P being the center of a circle , of which OF is an arc , and ED the sine of that arc ) will be as ED x force at distance PE = ED x F : ( PE ) ; for the attraction in the direction PDll'F at the point x is as the increment of the arc xy X F : ( PE ) x Pl/ Px ( xl and yl ' being at right angles to PF ) =Ux Px/ xl x Pl/ Px x F : ( PE ) = U ' x ?
l/ ~ x F : ( PE ) = uu/ ~ PE2 u2 x F : ( PE ) , if u= Pl ; and consequently the fluent of it is as ~ PE2 u2 x F : ( PE ) = ED x F : ( PE ) , and the attraction of the surface EFef will be as ED x Ff x F : ( PE ) = ED xzx PS/ PE x F : ( PE ) =yx zz~yy/ ~ ( z2+y2 ) x F : ( ( z2+y2)1/ 2 = V ; the attraction of the curve will also vary as ~z~ zu x F : ( ( u2+z2)1/ 2)/ ( z2+u2)1/ 2 = W , in which the fluent of zu x F : ( ( z2+u2)1/ 2)/ ( z2+u2)1/ 2 is contained between u=0 and u= y ; the fluents of V and W contained between two values of z , which correspond to two values of y= 0 , will be equal to each other .
2 .
From a similar method may be deduced equalities between other like fluents , for the curve may be supposed to consist of other similar curve surfaces equally as circles , and the solid of similar segments of other solids equally as spheres .
3 .
From the same principles may innumerable serieses equal to each other be deduced ; for by different converging serieses find the sum of the same quantity or quantities , and there will result serieses equal to each other : for instance ( fig. 5 ) , if the time of falling down the arcs AC and BC and their interpolations from the principles delivered in the Meditationes Analyticae , of which the difference let be D ; find the difference between the times of a body 's falling through BC when it began to fall from A and from B by a series proceeding according to the dimensions of AB = 0~ a small quantity ; and find , by a series of the same kind , the time of falling through AB ; the sum of these two serieses will be equal to D. Similar propositions may be deduced from fluxional equations .
4 .
In some cases the ratios of the times of bodies falling through some particular distances to each other may be easily known ; for instance , let the force vary as the m1 power of the distance ( x ) , and a be the distance from which the body began to fall , then the velocity varies as ~ ( am xm ) , and the increment of the time as x/ ~ ( am xm ) ; but if the parts of different curves are proportional , then will a , x , and x vary in the same ratio as each other , and consequently the time through proportional parts of the distance will vary as a2-m/ 2 ; and if the bodies be resisted likewise by a force which varies as the 2m-2/ m power of the velocities , then the times through proportional parts will vary as before , that is , as a2-m/ 2 , where a denotes the proportional distances from the points where the forces and resistances are equal .
PROBLEM IV .
1 .
Fig. 6 .
Given an equation expressing the relation between the two abscissae z= AP and x= Pp and their correspondent ordinates y= pm of a solid , to find its solid contents contained between two values of its first abscissae z. Assume z as an invariable quantity , and from the equation resulting find the fluent Z of yx contained between the extreme values of x or y ; then find the fluent of Zz contained between the given values of z , and the fluent multiplied into the product of the sines of the angles , which the first abscissa makes with the plane of the ordinates and second absciss , and the second absciss makes with its correspondent ordinates , will be the solid content required .
2 .
Fig. 7 .
Let the first absciss z of a solid be perpendicular to the planes of the ordinates , and the second absciss Pp =x perpendicular to the ordinates themselves pm = y. First , assume the first absciss as invariable , and find the increment of the arc p 'm = ( x2 + y2)1/ 2 , then assume the second absciss Pp as constant , and let mu be the fluxion of the ordinate y or u , when the fluxion of the first absciss is z= ul , where ul is perpendicular to the plane of the ordinates p'pm , and la point of the surface of the solid ; draw uh perpendicular to the arc p 'm , and since ul is constituted at right angles to the plane pp 'm , lh will cut the arc p 'm at right angles ; but uh = um x pp'/ p 'm = ux/ ~ ( x + y2 ) , lh = ( hu2 + lu2)1/ 2 = ( ux/ x2+y2 + z2)1/ 2 the fluxion of the surface will be lh x~ ( x2 + y2 ) .
From the given equation expressing the relation between the two abscissae z and x and ordinates y find , by assuming z invariable px=y , and by assuming x invariable qz= y'=u , which being substituted for their values in the quantity lh x ~(x2+y2 ) , there will result ( q2+p2+1)1/ 2xxxz=Axz= ( q2+p2+1)1/ 2/ pxyxz=Byz ; in A and B for y and x respectively substitute their value deduced from the given equation , and let the resulting quantities be A'xz and B'yz , where A ' is a function of x and z , and B ' a function of y and z ; find the fluent of A'xz from the supposition that x is only variable contained between the extreme values of x to a given value of z , which let be Lz , then find the fluent of Lz by supposing z only variable contained between given values of z , and it will be the surface of the solid contained between those values .
The same may be deduced by finding the fluent of B'yz on the supposition that y is the only variable quantity contained between the extreme values of y as before of x to a given value of z , which let be L'z ; then will the fluent of L'z contained between the given values of z be the surface required .
If the solid be a cone generated by the rotation of a rectangular triangle round a side containing the right angle as an axis ; hue will be a given quantity , if z be given .
If the above-mentioned angles are given , but not right ones , the arc p 'm and perpendicular l'h can easily be deduced , and consequently the increment of the surface .
3 .
To define a curve of double curvature , it is necessary to have two equations expressing the relation between the abscissae z and x and their ordinates ( y ) given , and if the angles which they respectively make with each other be right ones ; the fluxion of the arc as given in the Proprietates Curvarum is ( z2+x2+y2)1/ 2 .
Find its value from the two given equations in terms of x , y , or z , multiplied into its respective fluxions , and its fluent , properly corrected , will be the length of the arc required .
If the angles are not right , they may easily be reduced to them .
4 .
The attractions of these surfaces , curves , &amp;c. on a given point P may be deduced from the preceding principles of finding the attractions of each of the parts in the directions of the first abscissa , which passes through the point P , the second abscissa , and the ordinates , and then finding the integrals of these increments .
From the method which determines the attraction of a body , surface , &amp;c. on a given point can be determined the attraction of a body , &amp;c. on any number of points , and consequently the attraction of one body , &amp;c. on another , &amp;c. It is sometimes advantageous to transform the first absciss , that it may pass through the point attracted ; and the abscissae and ordinates , that they may be at right angles to each other , &amp;c. PROBLEM V. 1 .
Fig. 8 .
Given an equation expressing the relation between the two abscissae AP and Pp of a solid , and their correspondent ordinates pm , or AP ' , P'p ' , and p 'm ' ; to transform the first abscissa into any other Lh .
Let the abscissa Lh begin from a point L of the first abscissa AP , and meet an ordinate pm in the point h ; draw hp , and let the sines of the angles Ppm , Php , and pPh ; LPh , PhL , and PLh , be denoted respectively by r , s , and t , &amp;c. r ' , s ' , and t ' ; through a point h of the line Ph draw p'h 'm ' parallel to pm , and Lh=z , hh ' = x , and h 'm ' = y : in the given equation for AP , Pp , and pm substitute respectively their correspondent respondent values s'z/ r'~AL ( a ) , st'z/ rr'~sx/ r ( for Ph = t'z/ r ' and Ph ' = Ph~hh ' = t'z/ r/ ~x ) , and y~tt'z/ rr'~tx/ r ' ; there results an equation to the same solid expressing the relation between the two abscissae z=Lh and x , and their correspondent ordinates y. 1 .
2 .
If the absciss Lh does not begin from L , a point in the first given absciss AP , but from Ma point given out of it , it may be reduced to the preceding case , by drawing from Ma line MN =c to the plane of the first and second abscissae parallel to the ordinates pm ; and from N to the first abscissa a line NO =b parallel to the second abscissae , and substituting in the equation expressing the relation between AP , Pp , and pm for AP , Pp , and pm respectively z~AO ( a ) , x~b and y~c ; and there results the equation required expressing the relation between the two abscissae z and x , and their correspondent ordinates y , of which the first abscissa z passes through the point M. 2 .
To change the second abscissa Pp into any other Lh , the first abscissa and ordinates remaining the same .
In the preceding figure let L be considered as a moveable point of the first absciss AL , and the sines of the respective angles denoted by the same letters as before , and Lh = x , AL=z , and hm = y ; in the given equation for AP , Pp , and pm , substitute z~ s'x/ r ' , st'x/ rr ' , and y~tt'x/ rr ' ; and there will result the equation required expressing the relation between z and x the abscissae , and their correspondent ordinates y. 3 .
Fig 8 .
To change the ordinates , the abscissae remaining the same , draw p 'm an ordinate transformed , p'h parallel to the first abscissa AP , and meeting a second abscissa , of which pm is an ordinate in h : for the sines of the angles p'hp , hpp ' , and hp'p ; p'pm , pmp ' , and pp 'm write r , s , and t , r ' , s ' , and t ' ; and for AP ' , P'p ' , and p 'm respectively z , x , and y ; then substitute in the given equation for AP , Pp , and pm , their respective values z ( AP ' ) ~ ss'/ rr ' X y , x ( P'p ' ~ ts'/ rr ' y , and t'/ r ' y ; and there results an equation to the solid expressing the relation between the two abscissae AP ' and P'p ' and the transformed ordinates p 'm .
From these cases , which are easily reducible to one , may be transformed any given abscissae and their correspondent ordinates into any other containing given angles , &amp;c. with the before-mentioned abscissae and ordinates .
In the properties of curve lines , first published in 1762 , is given a method of deducing the equation to any section of the solid , and in particular the case of deducing the equation to the projection of any curve on a given plane .
From the principles given in this , and the Paper on centripetal forces , which the Royal Society did me the honour to print , can be deduced the fluxional equations , whose fluents express the relations between the abscissae and their correspondent ordinates of the curves described by bodies , of which the particles act on each other with forces varying according to given functions of their distances .
Fig.1 .
Fig.2 .
BEahfnxyb Fig.3 .
Cm l AFmm p uzPuA yy DPzf Ee C Fig. 4 .
x Fig.6 .
y Fig. 5 .
Apm B PB ASDdllFfAP C H m hue Fig. 8 .
Fig. 8 .
p ' p h ' m l Fig. 7 .
p ' m ' p pm hh p ' PAP LPA

