II .
Methods of clearing Equations of quadratic , cubic , quadratocubic , and hzgher Surds .
By William Allman , M. D. Communicated by the Right Hon , Sir Joseph Banks , K.B. P.R.S. Read July 8 , I13g .
SEVERAL years have elapsed , since my very highly esteemed friend , now Rev. Doctor MOONEY , Fellow of Trinity College , Dublin , presented to the Royal Irish Academy a paper on the Extermination of Radicals from Equations .
He has illustrated , by sundry examples , the extermination of quadratic surds .
As he has rightly observed , the method is universal .
Any number of quadratic surds , independent , or dependent , on each other , may be removed from an equation ; because , i. Any quantity , or factor of a quantity , necessarily subjected to the radical sign , is but of one dimension .
2 .
This quantity or factor being brought to one side of the equation , while the quantities unaffected with it remain at the other , may , by squaring both sides , be freed from the radical sign .
3 .
By a repetition of these reductions for each remaining independent surd quantity , any number of surd quantities may be converted into rational .
Examples .
I. Let v/ a + / b + / c = o : then ( 2 ; ) V/ a + V1/ b -/ c ; and , a+b+ 2Vab -c ( 3 ; ) a+bc= 2ab ; a+ b-c = 4ab ; free from surds .
II .
Let ^/ a + v/ b + ^/ +V / d = o ; then ( 2 ; ) a+ bt+c +2Vt/ ab+ sac + a-/ bcd ( 3 ; ) a+b+cd+ c2V/ ab =:V/ ac 22/ bc ; squared gives , abc4d+b 4.a+b+ c-d .
V/ ab= 4c .
a'+ b+ 2/ ab ( 4 ; ) a+bc -+. .
4ab-4cd==4 .
c+ d a--b .
/ ab ; squared , results free from surds .
Note , universally : the last two surd factors vanish together .
Surds , whose indices are integral powers of 2 , may be treated as quadratic surds ; and the number of them , which may be exterminated from any equation , is equally unlrimited .
Let 4/ a + 4Vb + -Vc = o : then , ( 2 ; ) V/ a + V/ b +X V/ ab6= V/ c ( 3 ; ) Va + Vb -c = *2V/ ab ; and , a+b+c+ 2V/ ab 2/ Vac 2^/ bc= 4/ ab .
The surds , now all quadratic , may be thus exterminated : ( 4 ; ) a+b+ c2A/ ab= 2o/ ac+ 2 bc ; a+ .
b +c + 4ab 4 .
aba+ c. vab -= 4c + 4bc + 8cVab .
( 5 ; ) a +b-c +ab= 4.a +b+ 3$c .
vab : puta +bc -= n ; then no + ab =n+ 4c/ ab ; and n* 42abn ' + a=b " -4'2 + i6cn + 16C ' .
ab.n-ab 16cn + 16C .
ab *a+b_.cZ_4ab 2 , e , -= 8abc .
n -.c ==8abc .
a +b +b c ' .
a-a c ' -4b ' g ac a Cbcl 8abc .
a C.4 .
a2 +b7+ 2ab-2a c-bc = 128abc.a-b + .
+ Let '/ a+ / " b + V/ c:= o : then , by the last example , ( 5 ; ) a+b+c2 / ab ac 2/ 6C2Vbc = i8 a/ bc + b/ ac + c-/ ab .
Put a+ b+cc == 2z : then n -V ab--/ ac--V/ bc , or , 'J ?
-L Vbc Vaa T =32 .
aV + bVac + cVab +ac+2a v/ bc + 2b .
A/ ac +zc VLab ~ 8~ aO ; % , a cv ++be ) S6-|-a-2n.2 .
2n +2n * kbC ( 6 ; ) Vab =Va I ac--30c = +3ob .
+30a ?
+ bc + ab ?
4-.ac+ ; c%+4ab .
n+ 15c -4 n 415c .
nIab + ac+ bc .
Vab nV=4b , +5-4b.+5+8c +15 + Vab ... ... .._ q-+ 4ab .
n+ Ig n-i ' 5c.n+ab+ac+bc .
( 7 ; ) nr*+ab-+-ac-bc b -4act n+i5/ b =4 +b .
% 4 -+I$a Z 'ab .
-4bc .
nI 5a i ' 5b.+z25ab Vab + tzzbz.6ab .
n +ab .
4. .
ab.++ + 465abc 1e.n-b2..--898a'be .
n'+ 17cn ' 3iac .
+ISa +sa ?
b , -2bwc .
-898abc + 31bc .
+ I5b ?
+9gozabc2 which being squared , an equation will result of 8 dimensions , free from surds .
In like manner may surds of the i6th , 32d , 64th , &amp;c. powersbe taken away from any equation .
Th , number which may be taken away is unlimited , as the removal of each surd quantity or factor , in all these cases , depends on the principles which direct the solution of simple equations .
In the case of cubic surds , the quantity or factor necessarily subjected to the radical sign may be of one , or of two dimensions , but not higher : since then , an universal method is known for solving quadratic equations , any number of cubic surds , independent , or dependent , on each other , may be removed from an equation .
Let 'Va + '/ b +'/ c o then ( 2 ; ) a +b +3 'Vab 3'Va ' = c : therefore , ( 3 ; ) V'/ ab + ' Vab = 3a at MDcccxIV .
E~~~~~~~~~~~~~~~~~~~~~~~~~~~ which put m : multiply by b : then , 3/ a.b4 + b'S/ ab* =bmn : this quadratic equation gives , 3la = _b +_ Vb/ z -+ -m - .
/ b -+b3bm ?
b nb ---v*lb"- ' *.* - . ?
ab= 2+b vb-4bm ; 2ab +b m ' 2b -gbm +b Vb ; 2/ i^4bim : which squared gives , 4ab2 ' 4ab ' -2b Gbabm -+ gbm2 ' b6b3m +- .
9bPm 4m3 bm ; 7r+ Vdie +'/ de = ab .
d+3e3 / de + SV/ des , + 6der ++3 ' .
+ Y1 i.e. de ++ 33edr .
'V/ de = ab .
d+e+ 3ab'V/ de+3ab3s/ de ' ; put de--ab=x , and d+e =y ; .
( 4 ) +6der + 3er.3Vde + 3dr .
3d .
+xy +3X +3x .
Substitute now , for s , r3 46der + xy ; for t , 3r 43er +4 3x ; for v , 3gr+dr + 3x ; and the equation , s3 + t3de + v3de8stvde o , will become , + 27zdey9 .
681d%e ?
..5- .
rd2eye .
r34--7de27x.Y . !
^ 9 &amp; td + 3v1^ 54de ; I r_^ r , 27+H r9-9der + 3xyr 6+54de ' + j6 ; $ they .
4 --26d7ed'.x7dx 3x2y .
9dexyr + x3y3 *o which is an equation of 9 dimensions , i. e. of 27 times the height of the given one 3V/ a + 3/ b 43/ d + 3/ e = o. If another independent cubic surd were added , the equation resulting free from surds would be of 67 dimensions , or of 81 times the height of the given equation .
Unriversally , if an equation consist of any numbet of indeC pendent surds having a common index , the equation resulting free from surds will be so many times the height of the given one , as there are units in the common index of the surds raised to the power whose index is the number of independent surds diminished by unity .
As the solution of a cubic equation is required for the extermination of some of the higher surds , it may be worth while to show the connection of the rule , called CARDAN'S , with the extermination above given of cubic surds .
If 3/ a 4+ Vb 4 3/ c = 0 , then a+ b+ c== a3/ abc , as before ; or , by substituting x , y , z , for -/ a , V/ b , / c , respectively , if x 4-y 4= o , then x3 +y3 + Z= y3Z.3 Suppose E s any one of these quantities , as x , the unkno wn quantity of a cubic equation : arranged it may stand thus , x3*3yzx +3 = o , an equation wanting the second tern .
Therefore , conversely , x= -y z : or else , by dividing the cubic equation by the simple one x+=o and solving the quote , the quadratic equation , x ' x -y = o , x= Y^ + YThen , in any cubic equation wanting the second term , v. g. 3*+ qx 4r= o , suppose , y= q ; and , y3 +3 =r then , z q3-= 3 ; and 3-5 r : therefore , 3y _ y6=ry3 4i.q 3i 3r + q3 : and , by subtract7 .
-t-rr 27 ing this from , y3 3-= r3 r , a+ -I r+ q3 : therefore , y_ z , or x=r+ v/ r+I q3 +.:o-r rar rrr If the cubic equation have but one possible root , Y+ + Y ?
V/ -S will represent the two impossible roots .
If the cubic equation have all its roots possible , the last named expression , as well as y z , implies the extraction of the cube root of an impossible binomial ; except , however , in this single case , when two of the roots of the cubic equation are equal to each other : then , the solution by the above rule is possible , tlough all the roots be possible ; for then , y = ; and the expressions Y- , and I r ' +q q3 , both vanish : then X3 nay be exterminated , any number of surds , whose indices are in any manner compounded of the factors 2 and 3 , may also be exterminated from any equation .
It may at present suffice to instance , in surds of the 6th , and in surds of the gth power .
Let " / a -+6-/ b +6/ c =o : then , as in cubic surds ( 3 ; ) a. + 3a +3a .
^/ a + / b + / c 3= 6ac ; and + 3 .
va + b. b +3b .
+3 ? .
+3c .
+ c. / c +6 v/ abc = e7^/ abc .
Put , a + 3b + Sc = n : then , ( 4 ; ) l v/ a +b Vb .
21 abc -n v/ ; and , by squaring n244a + 4a.a.n+ ; -4 bn+ 4 " .
b+ 2n 4b .
n+ 8ab Vab 44labc *I+ j2 4-cn + .
c 8 ; 42ec1n / ab .
+ 4a.._ C. --4a .
43 +4a .
( 5 ; ) q ' b. 11 ---b2 , n 2 ' +b n8.cz , ab ' by --c .
+40 .
_44Iabe --4zc .
a ; + iSa+b+ isab2+b ; -6 restitution , 3a ^ P2 + 3b^ -l5oaci 5obc v/ ab .
Put a+ bc =r .
Then r3 + l2abr 4o0abc = 6r ' 162cr .
j/ ab .
Square and transpose , r* absr 754abcr3 +48al bc*,8..28836abc : ' _i523iza~bc .
--64a3b3 _ -6g984abc r +16o569a5bcl =o , results , an equation of 6 di46656abc4 mensions , free from surds .
Let S9/ a + 9/ b +91/ c = o ; then , as in cubic surds ( 3 ; ) 'v/ a +W3Vb +3V/ = 39V/ abc ; and ++ -33/ a'b + 33V/ ab +a +c 3J/ a4c +6 ^/ v'abc +3 3V/ b'c +3 V/ ac +3 3s*/ bcs = 73v/ abc .
Put a+b+c= $n : then A+ .
V Is\ !
_ ( 4 ; ) 0Vb + / C.3/ a. + 3Vb ' --73 bc+,3yc .
3Va 3/ b2c--3/ bc-n : multiply by 3/ b +3Vc : then 3/ b +3SC .
Vat -7 .
3b 3 3V/ -bc .
vb + 3/ .
a= c+3 Vc .
3s/ b + A/ C : this quadratic equation , when solved , gives ~/ O-*. .
_a 3 -.a 3 ' ab 7 V^/ b8 ?
.4n .
3Vb+4433 Vbc2c--c18b--4n .
3 This equation , when cubed , will exhibit the quantity or factor a , free from the cubic radical sign ; thus , a. 3Vb + V3ic ; or , -b-- , ab + 349+bc + 24b .
65b .
ac + g/ bc + s3a3Vbc ' = C. -165c .
3A/ bc +24C .
3V/ b +3cn , 43 % .
i8w ... .
b. -5b../ .
_ i8b .
+ 2.-Isc.3Vb+493V/ bc +c .
3cVc -i_8c.3Ab+4333i/ ba +c .
3AV/ C -.n .
-n . ?
-.4n .
-4n .
Restore for n , a^+b , transpose , and double : then 3II --a .
--a .
33 ab ac 2a +I 71C.a --8C .
~/ --a .
--a .
33/ b. " / b +43 3/ b'c _5 -s 3/ c. 3 3\58 3 --C --C .
33 This equation squared , transposed , and divided , by 4 , be~~~~~~cotn~~me &amp; ~s.~z 7 comels % +b4 b 43 -33.2zobl8 .
53b .
+ .
+ 3c " .
+3 68bc .
+ 22obc3 + 339c .
+ 1089c .
-3C3 .
+c4 +3C .
+ 4b3 + , ,azb a 6b 3v b2c +3a9b3a +io89b .
c 3Vbc 3 o. 2zc3 lb a+7 .
--3357bc .
a +26be lbco 2 64c+6bc , +22C3 -53c q 43 Compare this with the equation , s+t 3V/ d2e +v 3/ de ' = o : the resulting equation , s 3-+ / 3d'e ?
v v/ de ' = 3stvde , being accordingly computed , will be free from surds .
It will be of 12 dimensions ; but may be depressed to one of 9 .
Instead of continuing the operation to show this , I refer to the extermination of surds of the 5th , and of the 7th power , to be given below , for the manner in which some equations , resulting on involution , are depressed .
In surds of the 5th power , the quantity or factor , necessarily subjected to the radical sign , may be of 4 dimensions , but not higher : whence , if the solution of any biquadratic equation be admitted , any number of surds of the $th power may be taken away from an equation ; and here it may be observed , that , as to the matter in hand , it is of no importance , whether the biquadratic equation may be solved in possible terms , or not ; for the value , in numbers , of any particular quantity , or factor , is not required ; it is only required to obtain the quantity , or factor , of a single dimension , in order to deprive it , by involution , of its radical sign .
When an equation consists of 3 surds of the 5th power , the biquadratic equation is virtually a quadratic .
Let5 a+ / b + '/ c =o * .
( 2 ; ) a+b+ 5'a4b + 1o Sv/ a3b ' + 0o5 / ab3 + 5s Vab4 c ; put a+b+c= Sn. .
' ( 3 ; ) 5Va4b + , 'V/ a3b ' +2 V/ a2b6 + 5/ ab4 ==m : this , 31 multiplied by b3 , gives '/ ab'6 + 2b Wa3b 2+ 2b2 V/ aPb ' 4+ bs '5/ ab4 =3m , a biquadratic equation of the quadratic form , thus , '^/ a2b8 + bsVab2 + b ' .
S'a2b ' +b 'v/ ab4 6 bAm , which solved , gives ' 6/ a2b ' +b '5ab4 = +V4bn , another b+ i/ --b ZV 4+-b3 j,.\ S/ 74 .
i ?
another quadratic : then Vab4 = + ; involve 22 to the 5th power , and the equation will be cleared of the radical sign of the surd of the 5th power : thus , ab4=-b5+ -b4m+ b2V/ 44b3m 4m4-b + b4 n divide by h3 , and transpose ; then , b2 5bm + eab = / V-b +my b2 2b VbA4bm .
Square , transpose , and divide by 2b ; then , b3 5mb2 _Ioam .
b=+62 _bm Mm Vb b ba 72a .
$mbm +m / b-4bin .
+2aZ .
Square again , transpose , divide by 4b , and arrange : --Ia3b .
-a4b then , m ' -s 5abm 3 53abraati ; +2 3a_ ' , b33oab.~ ab-a~ + 2a3b " ~then~ , m +3 oab- .
+ 2aZb3 0+ +-ab4 But , as in the case of 3 cubic surds , a simple equation supplied the place of a quadratic , so , when an equation consists of s surds of the 5th power , a quadratic may supply the place of a biquadratic , or of two quadratic equations .
Thus , 'Va +5/ b +6 Vc , = o , * .
( 2 ; ) a+b+5 Va4b + 10 S/ a3b2 + 1o 5V/ a'b3 +5 'Vab4 =C , ( 3 ; ) 'Va4b +2 V/ a3b -}-_2'V'a2ba + Vt/ ab --ab-b-c + 2V25 62b3 + 5V/ ab4 =_S a= ab : But V/ a4b+3 'va3b2 3'ab'+ 3V ab4 / = Vab .
Va-5 V5 Vabc3 5L vSab2_sVa =s va2^bZ , v/ a+5v b _-5^Vab .
-5svc = +5 / a2b c Therefore s/ a2bZc5 / abc3 = m. This , multiplied by c , gives the quadratic equation , 5Va2b2C6 -c/ abc ' = cm : which solved gives '/ abc3 _ c_ ; , -4cm involve to the 5th power ; then , abc3 = c. 5C4m . ?
... S. c4-`3cm + c2ma / V c..---IVa EI~tS c " cm Divide by c2 , and transpose ; then , C3+ 5mnc2 5a c=++ ab .
C_2 3mc + m2 Vc2 4me : squared , gives C lomc5 +35mlz C5o0r3 .
C -5m .
C= C6 lmc 35mC4 -4ab .
zoabmr .
-2oabmr .
+4.a2b .
-Som3C3+ q25mC 41 'c .
Transpose , divide by 4c , and arrange ; then , ms * gabcmn + 5abc 'm +atbc = o ; or , mn =abc ab ac b. -ac-bc ; which is an equation of 5 dimensions , free from surds .
This equation , if , instead of V/ a + '/ b + i/ c =o , were substituted , v/ a -+ 'V'b + IoVC =o , would contain no other than quadratic surds ; if , 'V/ a + " vb +'/ c = o , no higher than cubic surds ; wherefore , if the extermination of any number of surds of the 5th power from an equation be admitted , since the number of surds of any lower order which may be exterminated is unlimited , an equation consisting of any number of surds , whose indices are in any manner compounded of the factors , 2,3 , and 5 , may be totally freed from surds .
If a formula for the solution of any equation of 6 dimensions were known , any number of surds of the 7th power might be taken away from an equation : As such a formula , however , is , I suppose , at present altogether unknown , we may be contented with the extermination of 3 surds of the 7th power , which may be accomplished , because , a formula for the solution of cubic equations is known , 3 , the index of the cube , being equal to half the index of the 7th power diminished by unity .
Let a+7 Vb + 7 ' v o : then , fz ; )a+b+77Va6b+z17Vasb1+357Va94b+357Vab4+zI7Vazb4+77Vb =7 ab ( 2 ; ) a +b+77 ?
/a%b+2j , q'sb + 35 77abb3*+ 35 7a Ib*+2I 7'azbs +7 7/ b6 =-a(3 ; ) 7Va6b+3 7VaSa5b 7'b+ 5 7V + a3b4+ 3 7V'a+bs7 V ab6= ---(put ) m 77vab+ 57Va5b+ io7Va4b'3+ lo7Va3b4+57 ab5+ 7Wab67Vab Avla7Vb5 =_7^abcs 1 Butt -27V/ ab--6 7Va4b3 67V/ a3b4-^7^aabS =--b 27/ aVa'b2.7v/ a+7Vb=+27/ Vab2c3 +7V'a4b3 +7Va3b4 =+ 7vab3b.7 , ia+7Vb =.7^/ a3bC Therefore 7/ Va3b3c'+ 2 7Va2b'ZC37Vabc5 = m. Multiply by c2 ; then , 7/ a3b3c ' " 2C '/ a2b2c+ c2 77/ abc5 c2m .
Extract the square root of this cubic equation ; then , 'va3b3c5 4is't + 9lS4t .
+ sS .
+10ls4 + O002Sot2 .
364s4t .
Arranged , r abe =++343S sZ3 + gsoo3s2t .
3"/ st+300o3s3 ' 3i/ Ast3 +itoo lst'4 + 364st'4 +-2oo t3 + 14sl ' + t. +)1St ' .
Restore , first , for s , fn for t , +2i_ - ?
/in then , _ ab -I / Z731m +5461m i/ n 2V , 9 abe 2X -3348mnf .
-n 9090gomn .
/ +253Smn . ? ? ?
3+ Y729nz .
3645nl .
-729:L,37m %-- . ?
4546m .
+1-i8z5m ML V_ ?
+M / .
i/ n -9090oml .
+ i/ m -2538mn .
i/ n 2+cm_n- .
I-n .
" " " * " " ~t +3645e* +729n Restore , secondly , for n , m7 c ; divide by 13 / c , and transpose : then , t4m~ m ; .
T-'---473tcm , +2 m ?
; .
' ~/ -o , / m 4c - .
3ab .
c. c -V3 ' ?
i/ rn rn -_ m " -_==o .
--2 , ':Z -~ -m m , om '7 50+ 50 +2 Put , m " + cr + 5c = 2"x ; n^4 + ocm.+ t= =y ; l mr 4-+ 35c + 2C"-3ab= 9 : then , +-x / m-y Un . .
V2 &lt; r 2+ ^xiVm+yVn. .
Vf 2 fmc v/ z + z3/ c = o.0 Therefore , by the extermination of cubic surds , + _ ?
2/ '3 rm/ -y.n i/ n +2 '/ + i. e. x3m23Sxymn + xy2mn -y3n2 + c23 = X2m yyn c ; or , x2m + y2n .
xmn- : 3x2m+y'n * ynx'm-yn cz- .
cz = o. Restore the values of x , y , and z : divide by 2 c ; then , 200 --CS , .
14600 7850 i 6+ &gt; _ 25cm5 -2_ 74602 7 *c4 m3 -_ -4a6bc .
m2 + I4abc4 3abc .
+ 98abc^ .
++ 35a c. 4 ' I4a2b1c .
-abc5 m +2za2bcs = o : which is an equation of 7 dimensions , free +a3b3c from surds .
In like manner , the extermination of 3 surds of the 1 ith power from an equation might seem to require the solution of an equation of 5 dimensions : but in this case , the highest term , if I may so speak , vanishes ; so that an equation , consisting of 3 surds of the 1 th power , may be freed from surds , without the solution of any higher equation than a biquadratic .
The labour however is great .
As preparatory thereto , and not to refer elsewhere , the solution of a biquadratic equation may be here given .
Suppose it , as any equation may be so transformed , to waant the second term ; thus , x4 + qx2 + rx + .
s o : suppose also , x-2 + ex +f ; x ex +g=o then , 4 of .
X+ fg -- ' +4 xgX -eg +-=o : and , -ef+eg r -eC fg =S e3-+qe-r f+g e , '+q 2e._ e6+zqqe+q%e-r Tr f^S-^^qU 2e Iqeq e-I ? .
Therefore , g e'+ r =-- .
4g+_ = .
P 4s thneq* ^ &lt; ' e2 2e=o u ?
;then , e ' -n q+e1= o. Put ez q then , '"3Pr +'_2qz'+ 4q2z -q3 -e _ 3 .
3 -4 +T7 ~3 27~3 ' qz S -rs = r* J q , e : ... a- .
T ' This cubic equation gives , z= 3/ '13 S r2 27 32+ qA -2 12 + -2r4 r_A28 -s + -r4 _q S2-er r4 s_ , 63 S3 Then by hypothesis and solution of quadratics , xe+/ e2 _-f : or , x = .
e+ V4 e2 -g ; and , by substitution , rr2 Ie I/ 2eT _^/ ?3q+ 4 - &gt; , orx= e+ e-/ - , e_2q-2 = ?
iv^^~v ?
... ... q 4q 33~~~ 3 Let now the equation consisting of 3 surds of the ilth power be , " Va + " ib + " IV = o ; then ( 2 ; ) a+b+11 " va'"b + 55 Iv"a'b + 165 / a8b3 + 330 " W/ 7b4 + 462 " V/ ab5 4462 " / a5b6 + 330 V Ia4b + i65 " Va3b3 a " a2bs +11 riVab0 c. CI9 I &gt; ( 3 ; ) Valab+sX80a"b`+ l'X.va~b3+a0l'/ a7b4V+ 4IbV/ a%~b5+ 41ta v'a3b +'lVab+ 'b9 + 4Vab.-._m Va rl / Val ?
b + 91 " / a9b+ 36r1 ~/ aSb3+ 84"'V/ a7b4+ xz6l'la6b5+ iz6 v'aSb6+ 84'1 a67 a 3611 V'a3b +91T vazb9+ XvabE ?
= -ql'abc9 =-4 ' " va9bl'-zS8 q'aSb '-gb/ a b8 a s7bq4-41 / a%bS-1 4-ll Javasb6 81- ' a1 bT-z 2 'aS"4 t/ a1b9 =2 + 4 .
/ a3bYc7 : tI , + 7lV/ tasb +35x , / a4b+_ 70/ Va6bS+ 7oll Vasb6+35s/ a4b7+ 7"S/ aabs =-_7 li/ a3b3cS .
-5 V , a7b4._ SxIVa6bs._ iSxX~/ asbO.,.${Va4b =7 + s5I""a4b4c3 ' ; ( 3 ; ) ^V^^+5XV^a4-^IV^3+3oxV^4+ 4I ?
V^5+ 4IlV^6+30lxv^4^+I5SV^^8+51V^9+IV^lo== m ?
.t Therefore *5 " It/ Va4C3 7 " Va3b3 + I"^ 7 -"Vb 9= -m ( .== Z , E ) Multiply by c ; / ^V 4b-2-.c"3 bc 3I5 4C2 Iaob2c8 a C3 Vabc9 =_ c 3m Put , V"abc-9 -= + 27 c ; then , x , j7 CXa J , I4.7 C2X~ 343 C3X[ 240oI 54 +L 3L c24 3 4 " cc'3 C7 , i47 2Xz I029 3 24I +5 C2 X+ 3X +5 &gt; 9 c223 I ; -'x0 xx 6o -7 _ This , involved to the ilth power , will yield an equation , which shall have no other surds than quadratic and cubic ; and , since these may be removed , whatever be their number , it is evident , that an equation may be at length deduced free from all surds : But the accomplishment of this would require so great labour , that it may at present suffice , to have shown the possibility , and pointed out the method , of removing all surds from an equation consisting of 3 surds of the ilth power .
Far greater would be the labour to exterminate 3 surds of the 13th power .
Surds of the iothl power , it must already have sufficiently appeared , may be taken away in ally number , according to the principles of extermination of cubic and quadratic surds .
It is also sufficiently manifest , that , if an equation , consisting of 3 surds of a certain power ( v. g. the 7th ) , may be cleared of surds , an equation containing X such surds , together with any number of other surds whose extermination is unlimited , may be also cleared of surds ; and that surds , whose extermination , as to their number , is unlimited , may be exterminated from any equation containing them , however diverse they be from each other .
Thus , lias been pointed out , the extermination from equations , of surds whose indices do not exceed the number 6 , or of any combinations of such surds , in any number ; of three surds , whose common index is either of the prime numbers between 6 and 12 , or whose indices are either of these multiplied by any numbers , or powers of any numbers under 6 , provided the equation contain no other quantity ; also , of two surds , whose common index is , or , whose indices are , as last described , with an indefinite number of surds of the former description .
It only remains , however , for the complete establishment of the last observation , to note , that any surds , contained in the denominator of any fractional quantity of an equation , which cannot be transferred to the numerator , by multiplying both terms by a residual , as some have recommended to be done , may , by multiplying the whole equation by that denominatcr , be transferred to the other quantities , or numerators , of the equation .
That observation will then hold of the surds therein named , however they be situated in the equation ; whether they be in the numerators , or in the denominators of fractions .
P. S. Dr. WARING'S method of taking away surds* is very ingenious .
It is however evidently limited by the same postulate , which restricts the application of my general method ; viz. to solve an equation of the dimension next lower than the index of the surd , being prime ; for this must be effected in order to obtain the imaginary values of the surd as required by his method ; and this , and sometimes less than this , is sufficient in mine .
e. g. To obtain the imaginary roots of the 5th power of unity , the biquadratic equation a4 4a3 4 a2 4a +=o must be solved .
These roots are 7-5"-'.o+2'5 and 4 I'"X t+ 4'Icoefficients troublesome enough , especially from their variety , as WARING himself has observed .
The imaginary values of higher surds of prime indices , when found , would be still more complicated : and it is not very easy to find , for example , the imaginary values of a surd of the 1 th power.* The preceding biquadratic equation is the difference of two squares a+ a3 + -9 a+a+ 1 , and 5 a2 : consequently , its quadratic factors will be the sum and difference of tbhe roots of these squares , viz. a+'a= o , and aa a , --'5 a+ xtl*+ax9+bxs+7+ ?
dx6+mxS+dx4+ cx3+bx*+ a+ =o ; by the combination of corresponding terms , which has , in the preceding sheets , b'eu repeatedly exemplified , and by division of the roots by x , may be obtained , -----s - , X+ d ? . ?
... +m &amp; 4a .
1+b .
XT +C 3b..x x x4a .
' x +5 x +2a If this equation could be solved , the root being called n , the solution of the quadratic , x1 nx I io , would give the root of the proposed equation , which , when I +x t I~ _ every coefficient is unity , yields , x 4 .
+3x++ 3 .
++I =o .
x If this equation could be solved , then would Dr. Wa riN 's method serve for the universal exterminatiof of surds of the i th power .
I may also observe , that my method universally holds for exterminating quadratocubic surds , without the solution of any higher equation than a quadratic , as may appear from a former example , though the observation was not before made ; and , in like manner , that it universally holds for exterminating surds of the 7th power , within the condition of solving a cubic equation .
For the resulting equation of six dimen .
sions may be reduced to one of three , independently on the simplicity or composition of the third quantity of a given combination of surds .
Thus , if tfa+72/ b+7V = o , -a--b-c V/ aeb + 34V"aSb7+ 5/ a4b a+ s73a3 , + 3q'VaS+ v'ab6 - : ... ... ... .
then after 7 multiplying by bs , may be obtained this cubic equation } , 7a^z/ + bi:ab63 + 2b ' 7/ i ; ^+vab2e +b4 , Aa t-a +b 7a a 7 .
s.o Hee ter whateer may be the value of c , a and b may be freed from the radical sign of the 7th power , by the solution , first of a cubic , then of a quadratic , equation , and afterwards involution .
So that both WA IN o ' method and mine universally hold , until we arrive at surds of the nlth power ; and according to mine , three such may be exterminated .
1 = o. These quadratic equations being solved , will give the four roots of the biquadratic , as above .
So may be solved any biquadratic equation , the coefficient of whose 4th term is equal to the product of the coefficient of the ed by the square root of the last , the coefficient of the ist term being unity : and , by the intervention of a cubic equation , any biquadratic may be so transformed .

