An algebraic problem of finding four numbers, given the products of each of the numbers with the sum of the three others *   *   Originally published as Problema algebraicum de inveniendis quatuor numeris, ex datis totidem productis uniuscuiusque horum numerorum in summas trium reliquorum, Opera Posthuma 1 (1862), 282–287, and reprinted in Leonhard Euler, Opera Omnia, Series 1: Opera mathematica, Volume 6, Birkhäuser, 1992. A copy of the original text is available electronically at the Euler Archive, at http://www.eulerarchive.org. This paper is E808 in the Eneström index.

Leonhard Euler † † Date of translation: April 8, 2005. Translated from the Latin by Jordan Bell, 2nd year undergraduate in Honours Mathematics, School of Mathematics and Statistics, Carleton University, Ottawa, Ontario, Canada. Email: jbell3@connect.carleton.ca

If the four numbers that are to be searched for are given as v , x , y , z   , the following four equations will be considered:
v ( x + y + z ) = a
x ( v + y + z ) = b
y ( v + x + z ) = c
z ( v + x + y ) = d
From these equations, with basic methods three unknowns could be successively eliminated, and the fourth would be able to be dealt with according to the resolution of an equation. However, there is no reason why we would choose one as preferrable to the others to be ultimately determined. Indeed, it is suitable for none of these equations to be the final one determined, but in fact for a new unknown to be introduced that relates to each equally and by which the unknowns would be able to be defined. Therefore we sum the numbers that are to be found in this formula, v + x + y + z = 2 t ,   and so then the above equations are changed into these:
v ( 2 t − v ) = a = 2 t v − v 2 , from which v = t − ✓ ( t t − a )
x ( 2 t − x ) = b = 2 t x − x 2 x = t − ✓ ( t t − b )
y ( 2 t − y ) = c = 2 t y − y 2 y = t − ✓ ( t t − c )
z ( 2 t − z ) = d = 2 t z − z 2 z = t − ✓ ( t t − d ) .
We have the beginning of a solution now, such that from the single quantity t   we have the power to determine all four numbers that are being searched for, whereby all that remains is for us to investigate this quantity t   . With respect to the equation v + x + y + z = 2 t ,   by substituting in place of v , x , y , z   , the values formed with t   it will be: 4 t − ✓ ( t t − a ) − ✓ ( t t − b ) − ✓ ( t t − c ) − ✓ ( t t − d ) = 2 t ,   from which this equation follows: 2 t = ✓ ( t t − a ) + ✓ ( t t − b ) + ✓ ( t t − c ) + ✓ ( t t − d ) ,   which indeed could be worked through by the method of Newton for rationalization, but this would be a most tiresome labor. Therefore we will try resolving this equation with another method.
We put ✓ ( t t − a ) = p   , it will be v = t − p   ; in a similar way
✓ ( t t − b ) = q   , x = t − q  
✓ ( t t − c ) = r   , y = t − r  
✓ ( t t − d ) = s   , z = t − s  
and it will be p + q + r + s = 2 t   ; because of the irrationals p , q , r   and s   , this equation should be changed into another, in which the letters p , q , r   and s   correspond to powers of exponentials, from which by taking the place of the letters p , q , r   and s   that are to be made, a rational equation comes into existence, by which the value of the unknown t   can be defined. To this end, we form this equation: X 4 − A X 3 + B X 2 − C X + D = 0 ,   whose four roots are given by the quantities a , b , c , d   . Therefore it will be by the nature of equations:
A = a + b + c + d
B = a b + a c + a d + b c + b d + c d
C = a b c + a b d + a c d + b c d
D = a b c d .
Now it is put Y = t t − X   , that is X = − Y + t t   , so that we will have made by substituting this into the former equation:
Y 4 − 4 t t Y 3 + 6 t 4 Y 2 − 4 t 6 Y + t 8 + A Y 3 − 3 A t t Y 2 + 3 A t 4 Y − A t 6 + B Y 2 − 2 B t t Y + B t 4 + C Y − C t t + D } = 0 .  
The four roots of this equation of Y   will be: t t − a , t t − b , t t − c , t t − d .   In place of this equation, we put for the sake of brevity this: Y 4 − P Y 3 + Q Y 2 − R S + S = 0 ,   so that it will be:
R = 4 t t − A
Q = 6 t 4 − 3 A t t + B
R = 4 t 6 − 3 A t 4 + 2 B t t − C
S = t 8 − A t 6 + B t 4 − C t t + D .
Next it will be Y = Z 2   , that is Z = ± ✓ Y   , so that we will have: Z 8 − P Z 6 + Q Z 4 − R Z 2 + S = 0 .   The following will be the eight roots of this equation:
+ ✓ ( t t − a ) = + p − ✓ ( t t − a ) = − p
+ ✓ ( t t − b ) = + q − ✓ ( t t − b ) = − q
+ ✓ ( t t − c ) = + r − ✓ ( t t − c ) = − r
+ ✓ ( t t − d ) = + s − ✓ ( t t − d ) = − s .
This equation in eight dimensions is resolved into two biquadratics where for one the roots are + p , + q , + r , + s   , while for the other the roots are − p , − q , − r , − s   , which are:
Z 4 − α Z 3 + β Z 2 − γ Z + d δ = 0
Z 4 + α Z 3 + β Z 2 + γ Z + δ = 0 ,
for which it will be by the nature of equations:
α = p + q + r + s
β = p q + p r + p s + q r + q s + r s
γ = p q r + p q s + p r s + q r s
δ = p q r s .
Since the product of these two biquadratic equations should be equal to the former equation of eight dimensions, it will be:
P = α 2 − 2 β
Q = β 2 − 2 α γ + 2 δ
R = γ 2 − 2 β δ
S = δ 2 .
Also with it α = p + q + r + s   , it will be α = 2 t   , and then α 2 = 4 t t   , from which it will be: α 2 − 2 β = 4 t t − 2 β = P = 4 t t − A ,   therefore β = A 2   . The second equation Q = β 2 − 2 α γ + 2 δ   gives: 6 t 4 − 3 A t t + B = A 2 4 − 4 γ t + 2 δ ,   that is, δ = 3 t 4 − 3 2 A t t + 2 γ t − A 2 8 + B 2 .   Certainly the third equation R = γ 2 − 2 β δ   will yield 4 t 6 − 3 A t 4 + 2 B t t − C = γ 2 − A δ ,   that is, A δ = − 4 t 6 + 3 A t 4 − 2 B t t + c + γ 2 ,   then with the above it will be:
4 t 6 − 3 2 A 2 t t + 2 A γ t − A 3 8 + 2 B t t + A B 2 − C } = γ 2 .  
By having extracted the square root, it will be obtained that, γ = A t ± ✓ ( 4 t 6 + ( 2 B − 1 2 A 2 ) t t − 1 8 A 3 + 1 2 A B − C )   and then δ = 3 t 4 + 1 2 A t t − 1 8 A 2 + 1 2 B ± 2 t ✓ ( 4 t 6 + ( 2 B − 1 2 A 2 ) t t − 1 8 A 3 + 1 2 A B − C ) ,   of which the square will be
25 t 8 + 3 A t 6 − 5 2 A 2 t 4 − 5 8 A 3 t t + 1 64 A 4 + 11 B t 4 + 5 2 A B t t − 1 8 A 2 B − 4 C t t + 1 4 B 2 { ± 12 t 5 ± 2 A A t 3 ± 1 2 A 2 t ± 2 B t } ✓ ( 4 t 6 + ( 2 B − 1 2 A 2 ) t t − 1 8 A 3 + 1 2 A B − C )
which should be equal to S   itself, that is this expression: t 8 − A t 6 + B t 4 − C t t + D   , from which results this equation:
24 t 8 + 4 A t 6 − 5 2 A 2 t 4 − 5 8 A 3 t t + 1 64 A 4 + 10 B t 4 + 5 2 A B t t − 1 8 A 2 B − 3 C t t + 1 4 B 2 − D = { + 12 t 5 + 2 A t 3 − 1 2 A 2 t + 2 B t } ✓ ( 4 t 6 + ( 2 B − 1 2 A 2 ) t t − 1 8 A 3 + 1 2 A B − C ) ,
which reduced to rationals gives:
+ 3 A 4 − 12 A 2 B + 24 A C − 48 D } t 8 + 5 4 A 5 − 8 A 3 B + 7 A 2 C + 12 A B 2 − 8 A D − 12 B C } t 6 + 3 16 A 6 − 27 16 A 4 B + 7 4 A 3 C + 9 2 A 2 B 2 + 5 A 2 D − 7 A B C − 3 B 3 + 9 C 2 − 20 B C } t 4 + 3 256 A 7 − 9 64 A 5 B + 5 32 A 4 C + 9 16 A 3 B 2 + 5 4 A 3 D − 5 4 A 2 B C − 3 4 A B 3 − 5 A B D + 5 2 B 2 C + 6 C D } t t + 1 4096 A 8 − 1 256 A 6 B + 3 128 A 4 B 2 − 1 32 A 4 D − 1 16 A 2 B 3 + 1 4 A 2 B D + 1 16 B 4 − 1 2 B 2 D + D 2 } = 0
It put for the sake of brevity E = 1 4 A 2 − B   and u = 2 t   , and it will be:
+ 3 A 2 E + 6 A C − 12 D } u 8 + 2 A 3 E + 4 A 2 C + 12 A E 2 − 8 A D + 12 C E } u 6 + 9 A 2 E 2 + 28 A C E + 80 D E + 36 C 2 + 12 E 3 } u 4 + 12 A E 3 + 80 A D E + 96 C D + 40 C E 2 } u u + 4 E 4 − 32 D E 2 + 64 D 2 } = 0 .
This equation then has four positive roots and the same number of negative roots, which are equal to each other; thus the resolution of the equation can be accomplished by a biquadratic equation. Moreover, A , B , C , D   and E   are known quantities determined by the given a , b , c , d   , with it of course
A = a + b + c + d
B = a b + a c + a d + b c + b d + c d
C = a b c + a b d + a c d + b c d
D = a b c d
and likewise E = 1 4 A 2 − B .   With this having been discovered indeed, whatever the value of u   the quantities being searched for are:
v = u − ✓ ( u u − 4 a ) 2 , x = u − ✓ ( u u − 4 b ) 2
y = u − ✓ ( u u − 4 c ) 2 , z = u − ✓ ( u u − 4 d ) 2 .
Another solution
Furthermore, this problem can be solved in another way: From the first equations it is
a − b = ( v − x ) ( y + z ) b − c = ( x − y ) ( v + z )
a − c = ( v − y ) 9 x + z ) b − d = ( x − z ) ( v + y )
a − d = ( v − z ) ( x + y ) c − d = ( y − z ) ( v + x )
from the first and last of the equations is we obtain, v − x = a − b y + z , v + x = c − d y − z .   Were it a + b − c − d 2 = h   it will be h = v x − y z   , and by making it a + b + c + d 2 = k   it will be k = v x + v y + v z + x y + x z + y z ,   therefore k − h = 2 y z + ( v + x ) ( y + z ) ,   that is, k − h = 2 y z − ( c − d ) ( y + z ) y − z = c + d   , therefore 2 y z = 2 d y − 2 c z y − z   , i.e.
y y z − y z z = d y − c z   ; with it set y z = t   this equations changes into this: ( d − t ) y − ( c − t ) z = 0 .   It is not taken d y − c z = u   , and it will be y = ( c − t ) u ( c − d ) t and z = ( d − t ) u ( c − d ) t ,   for which substituting t = y z   in the values yields t = ( c − t ) ( d − t ) u u ( c − d ) 2 t t ,   from which comes u = ( c − d ) t ✓ t ✓ ( c d − ( c + d ) t + t t ) .   By substituting this into the values of y   and z   formed above we will have:
I . y = ( c − d ) ✓ t ✓ ( c d − ( c + d ) t + t t ) and
I I . z = ( d − t ) ✓ t ✓ ( c d − ( c + d ) t + t t ) ;
and then by adding and subtracting, y + z = ( c + d − 2 t ) ✓ t ✓ ( c d − ( c + d ) t + t t ) and y − z ( c − d ) ✓ t ✓ ( c d − ( c + d ) t + t t ) .   Then as formerly it is deduced, v + x = ✓ ( c d − ( c + d ) t + t t ) ✓ t and v − x = ( a − b ) ✓ ( c d − ( c + d ) t + t t ) ( c + d − 2 t ) ✓ t   from which, by once more adding and subtracting, with it put for the sake of previty b + c + d − a = m and a + c + d − b = n ,   the following values are produced for v   and x   :
I I I . v = ( n − 2 t ) ✓ ( c d − ( c + d ) t + t t ) 2 ( c + d − 2 t ) ✓ t and
I V . x = ( m − 2 t ) ✓ ( c d − ( c + d ) t + t t ) 2 ( c + d − 2 t ) ✓ t .
With what we have discovered above and also that h = v x − y z   , by now substituting in the values of v , x , y , z   and unravelling their expansion for the purpose of determining the value of t   , this equation of four dimensions is formed: 4 t ( h + t ) ( c + d − 2 t ) 2 = ( m − 2 t ) ( n − 2 t ) ( c − t ) ( d − t ) .