% This paper has been transcribed in Plain TeX by
% David R. Wilkins
% School of Mathematics, Trinity College, Dublin 2, Ireland
% (dwilkins@maths.tcd.ie)
%
% Trinity College, 2000.
\magnification=\magstep1
\vsize=227 true mm \hsize=170 true mm
\voffset=-0.4 true mm \hoffset=-5.4 true mm
\def\folio{\ifnum\pageno>0 \number\pageno \else
\ifnum\pageno<0 \romannumeral-\pageno \else\fi\fi}
\font\Largebf=cmbx10 scaled \magstep2
\font\largerm=cmr12
\font\largeit=cmti12
\font\tensc=cmcsc10
\newfam\scfam \def\sc{\fam\scfam\tensc}
\textfont\scfam=\tensc
\font\largesc=cmcsc10 scaled \magstep1
\input amssym.def
\newsymbol\backprime 1038
\def\nablaperm#1#2{\mathop{\nabla_{#1}}\limits^{#2}}
\pageno=0
\null\vskip72pt
\centerline{\Largebf INQUIRY INTO THE VALIDITY OF A}
\vskip12pt
\centerline{\Largebf METHOD RECENTLY PROPOSED BY}
\vskip12pt
\centerline{\Largebf GEORGE B. JERRARD, ESQ.,}
\vskip12pt
\centerline{\Largebf FOR TRANSFORMING AND RESOLVING}
\vskip12pt
\centerline{\Largebf EQUATIONS OF ELEVATED DEGREES}
\vskip24pt
\centerline{\Largebf By}
\vskip24pt
\centerline{\Largebf William Rowan Hamilton}
\vskip24pt
\centerline{\largerm (British Association Report, Bristol 1836,
pp\ 295--348.)}
\vskip36pt
\vfill
\centerline{\largerm Edited by David R. Wilkins}
\vskip 12pt
\centerline{\largerm 2000}
\vskip36pt\eject
\pageno=-1
\null\vskip36pt
\centerline{\Largebf NOTE ON THE TEXT}
\bigskip
The text of this edition is taken from the {\it Report of the
Sixth Meeting of the British Association for the Advancement of
Science, held at Bristol in August 1836}. (London: John Murray,
Albemarle St., 1837.)
Occurrences of `$t$th' have been changed to `$t^{\rm th}$' etc.,
for consistency with analogous instances of `$m^{\rm th}$'.
The following errors have been corrected:---
\smallskip
\item{}
the final addition sign ($+$) was omitted
in equation (53.) in the original text;
\smallskip
\item{}
ellipsis ($\ldots$) was present in the original text between the
second occurrence of `$p_{m-1}$' and `$q_0$' in the text block
located between equations (59.) and (60.);
\smallskip
\item{}
the fourth summand in equation (65.) was printed `${\rm C} x^3$',
and has been corrected to `${\rm C}x$';
\smallskip
\item{}
`$h$' was printed `$k$' in the original text in the phrase
`of the dimension~$h$' in the text block between equations
(131.) and (132.);
\smallskip
\item{}
the two final occurrences of `$q_{m-2}$' were originally
printed `$q_{m-1}$' in the text block between the reprinted
equations (102.) and (126.) in article [10.] (between
equations (167.) and (168.));
\smallskip
\item{}
an unbalanced opening parenthesis was included in the original
text in equation (309.), before `$h_4 h_3$' in the third
summand on the right hand side of the equation for
${}^\backprime h_1$.
\bigbreak\bigskip
\line{\hfil David R. Wilkins}
\vskip3pt
\line{\hfil Dublin, February 2000}
\vfill\eject
\pageno=1
\null\vskip36pt
\noindent
{\largeit Inquiry into the Validity of a Method recently
proposed by {\largerm George B. Jerrard}, Esq., for Transforming
and Resolving Equations of Elevated Degrees: undertaken at the
Request of the Association by
Professor {\largerm Sir} {\largesc W. R. Hamilton}.\par}
\vskip12pt
\noindent
[{\it Report of the Sixth Meeting of the British Association
for the Advancement of Science; held at Bristol in August 1836.
(London: John Murray, Albemarle Street. 1837.) pp.\ 295--348.]}
\bigskip
[1.]
It is well known that the result of the elimination of $x$,
between the general equation of the $m^{\rm th}$ degree,
$${\rm X} = x^m + {\rm A} x^{m-1} + {\rm B} x^{m-2}
+ {\rm C} x^{m-3} + {\rm D} x^{m-4}
+ {\rm E} x^{m-5} + \hbox{\&c.} = 0
\eqno {\rm (1.)}$$
and an equation of the form
$$y = f(x),
\eqno {\rm (2.)}$$
(in which $f(x)$ denotes any rational function of $x$, or, more
generally, any function which admits of only one value for any
one value of $x$,) is a new or transformed equation of the
$m^{\rm th}$ degree, which may be thus denoted,
$$\{ y - f(x_1) \} \{ y - f(x_2) \} \cdots \{ y - f(x_m) \} = 0,
\eqno {\rm (3.)}$$
$x_1, x_2,\ldots \, x_m$ denoting the $m$ roots of the proposed
equation; or, more concisely, thus,
$${\rm Y} = y^m + {\rm A}' y^{m-1} + {\rm B}' y^{m-2}
+ {\rm C}' y^{m-3} + {\rm D}' y^{m-4}
+ {\rm E}' y^{m-5} + \hbox{\&c.} = 0
\eqno {\rm (4.)}$$
the coefficients ${\rm A}'$, ${\rm B}'$, ${\rm C}'$, \&c., being
connected with the
values $f(x_1)$, $f(x_2)$, \&c., by the relations,
$$\left. \eqalign{
- {\rm A}' &= f(x_1) + f(x_2) + \hbox{\&c.} + f(x_m),\cr
+ {\rm B}' &= f(x_1) f(x_2) + f(x_1) f(x_3) + f(x_2) f(x_3)
+ \hbox{\&c.}
+ f(x_{m-1}) f(x_m),\cr
-{\rm C}' & = f(x_1) f(x_2) f(x_3) + \hbox{\&c.}\cr}
\right\}
\eqno {\rm (5.)}$$
And it has been found possible, in several known instances, to
assign such a form to the function $f(x)$ or $y$, that the new or
transformed equation, ${\rm Y} = 0$, shall be less complex or
easier to resolve, than the proposed or original equation
${\rm X} = 0$. For example, it has long been known that by
assuming
$$y = f(x) = {{\rm A} \over m} + x,
\eqno {\rm (6.)}$$
{\it one term may be taken away\/} from the general equation (1);
that general equation being changed into another of the form
$${\rm Y}
= y^m + {\rm B}' y^{m-2} + {\rm C}' y^{m-3}
+ \hbox{\&c.} = 0,
\eqno {\rm (7.)}$$
in which there occurs no term proportional to $y^{m-1}$, the
condition
$${\rm A}' = 0
\eqno {\rm (8.)}$$
being satisfied; and Tschirnhausen discovered that by assuming
$$y = f(x) = {\rm P} + {\rm Q} x + x^2,
\eqno {\rm (9.)}$$
and by determining ${\rm P}$ and ${\rm Q}$ so as to satisfy two
equations which can be assigned, and which are respectively of
the first and second degrees, it is possible to fulfil the
condition
$${\rm B}' = 0,
\eqno {\rm (10.)}$$
along with the condition
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
and therefore to {\it take away two terms\/} at once from the
general equation of the $m^{\rm th}$ degree; or, in other words,
to change that equation (1) to the form
$${\rm Y}
= y^m + {\rm C}' y^{m-3} + {\rm D}' y^{m-4} + \hbox{\&c.} = 0,
\eqno {\rm (11.)}$$
in which there occurs no term proportional either to $y^{m-1}$
or to $y^{m-2}$. But if we attempted to take away {\it three
terms\/} at once, from the general equation (1), or to reduce it
to the form
$${\rm Y}
= y^m + {\rm D}' y^{m-4} + {\rm E}' y^{m-5} + \hbox{\&c.} = 0,
\eqno {\rm (12.)}$$
(in which there occurs no term proportional to $y^{m-1}$,
$y^{m-2}$, or $y^{m-3}$,) by assuming, according to the same
analogy,
$$y = {\rm P} + {\rm Q} x + {\rm R} x^2 + x^3,
\eqno {\rm (13.)}$$
and then determining the three coefficients ${\rm P}$, ${\rm Q}$,
${\rm R}$, so as
to satisfy the three conditions
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
$${\rm B}' = 0,
\eqno {\rm (10.)}\qquad$$
and
$${\rm C}' = 0,
\eqno {\rm (14.)}$$
we should be conducted, by the law (5) of the composition of the
coefficients ${\rm A}'$, ${\rm B}'$, ${\rm C}'$, to a system of
three equations, of the 1st, 2nd, and 3rd degrees, between the
three coefficients ${\rm P}$, ${\rm Q}$, ${\rm R}$; and
consequently, by elimination, in general, to a final equation of
the 6th degree, which the known methods are unable to resolve.
Still less could we take away, in the present state of algebra,
{\it four terms} at once from the general equation of the
$m^{\rm th}$ degree, or reduce it to the form
$${\rm Y} = y^m + {\rm E}' y^{m-5} + \hbox{\&c.} = 0,
\eqno {\rm (15.)}$$
by assuming an expression with four coefficients,
$$y = {\rm P} + {\rm Q} x + {\rm R} x^2 + {\rm S} x^3 + x^4;
\eqno {\rm (16.)}$$
because the four conditions,
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
$${\rm B}' = 0,
\eqno {\rm (10.)}\qquad$$
$${\rm C}' = 0,
\eqno {\rm (14.)}\qquad$$
and
$${\rm D}' = 0,
\eqno {\rm (17.)}$$
would be, with respect to these four coefficients, ${\rm P}$,
${\rm Q}$, ${\rm R}$, ${\rm S}$, of the 1st, 2nd, 3rd, and 4th
degrees, and therefore would in general conduct by elimination to
an equation of the 24th degree. In like manner, if we attempted
to take away the 2nd, 3rd, and 5th terms (instead of the 2nd, 3rd
and 4th) from the general equation of the $m^{\rm th}$ degree, or
to reduce it to the form
$$y^m + {\rm C}' y^{m-3} + {\rm E}' y^{m-5} + \hbox{\&c.} = 0,
\eqno {\rm (18.)}$$
so as to satisfy the three conditions (8), (10) and (17),
$${\rm A}' = 0,\quad {\rm B}' = 0,\quad {\rm D}' = 0,$$
by assuming
$$y = {\rm P} + {\rm Q} x + {\rm R} x^2 + x^3,
\eqno {\rm (13.)}\qquad$$
we should be conducted to a final equation of the 8th degree; and
if we attempted to satisfy these three other conditions
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
$${\rm C}' = 0,
\eqno {\rm (14.)}\qquad$$
and
$${\rm D}' - \alpha {\rm B}'^2 = 0,
\eqno {\rm (19.)}$$
(in which $\alpha$ is any known or assumed number,) so as to
transform the general equation (1) to the following,
$${\rm Y} = y^m + {\rm B}' y^{m-2} + \alpha {\rm B}'^2 y^{m-4}
+ {\rm E}' y^{m-5} + \hbox{\&c.} = 0,
\eqno {\rm (20.)}$$
by the same assumption (13), we should be conducted by
elimination to an equation of condition of the 12th degree. It
might, therefore, have been naturally supposed that each of these
four transformations, (12), (15), (18), (20), of the equation
of the $m^{\rm th}$ degree, was in general impossible to be
effected in the present state of algebra. {\it Yet Mr.~Jerrard
has succeeded in effecting them all}, by suitable assumptions of
the function $y$ or $f(x)$, without being obliged to resolve any
equation higher than the fourth degree, and has even effected the
transformation (12) without employing biquadratic equations.
His method may be described as consisting in {\it rendering the
problem indeterminate}, by assuming an expression for $y$ with a
number of disposable coefficients greater than the number of
conditions to be satisfied; and in employing this
indeterminateness to {\it decompose certain of the conditions\/}
into others, for the purpose of {\it preventing that elevation of
degree\/} which would otherwise result from the eliminations.
This method is valid, in general, when the proposed equation is
itself of a {\it sufficiently elevated degree\/}; but I have
found that when the exponent~$m$ of that degree is {\it below a
certain minor limit}, which is different for different
transformations, (being $= 5$ for the first, $= 10$ for the
second, $=5$ for the third, and $=7$ for the fourth of those
already designated as the transformations (12), (15), (18) and
(20),) the processes proposed by Mr.~Jerrard conduct in general
to an expression for the new variable~$y$ which is a
{\it multiple of the proposed evanescent polynome\/}~${\rm X}$ of
the $m^{\rm th}$ degree in $x$; and that on this account these
processes, although {\it valid as general transformations of the
equation of the $m^{\rm th}$ degree}, become in general
{\it illusory\/} when they are applied to {\it resolve equations
of the fourth and fifth degrees}, by reducing them to the
binomial form, or by reducing the equation of the fifth degree to
the known solvible form of De Moivre. An analogous process,
suggested by Mr.~Jerrard, for {\it reducing the general equation
of the sixth to that of the fifth degree}, and a more general
method of the same kind for resolving equations of higher
degrees, appear to me to be in general, for a similar reason,
illusory. Admiring the great ingenuity and talent exhibited in
Mr.~Jerrard's researches, I come to this conclusion with regret,
but believe that the following discussion will be thought to
establish it sufficiently.
\bigbreak
[2.]
To begin with the transformation (12), or the taking away of the
second, third and fourth terms at once from the general equation
of the $m^{\rm th}$ degree, Mr.~Jerrard effects this
transformation by assuming generally an expression with
{\it seven} terms,
$$y = f(x) = \Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''}
+ {\rm M}' x^{\mu'}
+ {\rm M}'' x^{\mu''}
+ {\rm M}''' x^{\mu'''}
+ {\rm M}^{\rm IV} x^{\mu^{\rm IV}}
\eqno {\rm (21.)}$$
the seven unequal exponents
$\lambda'$ $\lambda''$ $\lambda'''$
$\mu'$ $\mu''$ $\mu'''$ $\mu^{\rm IV}$
being chosen at pleasure out of the indefinite line of integers
$$0,1,2,3,4,\enspace \hbox{\&c.}
\eqno {\rm (22.)}$$
and the seven coefficients
$\Lambda'$ $\Lambda''$ $\Lambda'''$
${\rm M}'$ ${\rm M}''$ ${\rm M}'''$ ${\rm M}^{\rm IV}$,
or rather their six ratios
$${\Lambda' \over \Lambda'''},\quad
{\Lambda'' \over \Lambda'''},\quad
{{\rm M}' \over {\rm M}^{\rm IV}},\quad
{{\rm M}'' \over {\rm M}^{\rm IV}},\quad
{{\rm M}''' \over {\rm M}^{\rm IV}},\quad
{\Lambda''' \over {\rm M}^{\rm IV}}
\eqno {\rm (23.)}$$
being determined so as to satisfy the three conditions
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
$${\rm B}' = 0,
\eqno {\rm (10.)}\qquad$$
$${\rm C}' = 0,
\eqno {\rm (14.)}\qquad$$
without resolving any equation higher than the third degree, by a
process which may be presented as follows.
In virtue of the assumption (21) and of the law (5) of the
composition of the coefficients ${\rm A}'$, ${\rm B}'$,
${\rm C}'$, it is easy to perceive that those three coefficients
are rational and integral and homogeneous functions of the seven
quantities
$\Lambda'$ $\Lambda''$ $\Lambda'''$
${\rm M}'$ ${\rm M}''$ ${\rm M}'''$ ${\rm M}^{\rm IV}$,
of the dimensions one, two and three respectively; and therefore
that ${\rm A}'$ and ${\rm B}'$ may be developed or decomposed
into parts as follows:
$${\rm A}' = {\rm A}_{1,0}' + {\rm A}_{0,1}',
\eqno {\rm (24.)}$$
$${\rm B}' = {\rm B}_{2,0}' + {\rm B}_{1,1}' + {\rm B}_{0,2}',
\eqno {\rm (25.)}$$
the symbol ${\rm A}_{h,i}'$ or ${\rm B}_{h,i}'$ denoting here a
rational and integral function of
$\Lambda'$, $\Lambda''$, $\Lambda'''$,
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$,
which is homogeneous of the degree~$h$ with respect to
$\Lambda'$, $\Lambda''$, $\Lambda'''$, and of the degree~$i$ with
respect to ${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm
IV}$. If then we first determine the two ratios of
$\Lambda'$, $\Lambda''$, $\Lambda'''$,
so as to satisfy the two conditions
$${\rm A}_{1,0}' = 0,
\eqno {\rm (26.)}$$
$${\rm B}_{2,0}' = 0,
\eqno {\rm (27.)}$$
and afterwards determine the three ratios of
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$,
so as to satisfy the three other conditions
$${\rm A}_{0,1}' = 0,
\eqno {\rm (28.)}$$
$${\rm B}_{1,1}' = 0,
\eqno {\rm (29.)}$$
$${\rm B}_{0,2}' = 0,
\eqno {\rm (30.)}$$
we shall have decomposed the two conditions (8) and (10),
namely,
$${\rm A}' =0,\quad {\rm B}' = 0,$$
into five others, and we shall have satisfied these five by means
of the five first ratios of the set (23), namely
$${\Lambda' \over \Lambda'''},\quad
{\Lambda'' \over \Lambda'''},\quad
{{\rm M}' \over {\rm M}^{\rm IV}},\quad
{{\rm M}'' \over {\rm M}^{\rm IV}},\quad
{{\rm M}''' \over {\rm M}^{\rm IV}},
\eqno {\rm (31.)}$$
without having yet determined the remaining ratio of that set,
namely
$${\Lambda''' \over {\rm M}^{\rm IV}};
\eqno {\rm (32.)}$$
which remaining ratio can then in general be chosen so as to
satisfy the remaining condition
$${\rm C}' = 0,$$
without our being obliged, in any part of the process, to resolve
any equation higher than the third degree. And such, in
substance, is Mr.~Jerrard's general process for taking away the
second, third and fourth terms at once from the equation of the
$m^{\rm th}$ degree, although he has expressed it in his
published Researches by means of a new and elegant {\it notation
of symmetric functions}, which it has not seemed necessary here
to introduce, because the argument itself can be sufficiently
understood without it.
\bigbreak
[3.]
On considering this process with attention, we perceive that it
consists essentially of two principal parts, the one conducting
to an expression of the form
$$y = f(x) = \Lambda''' \phi(x) + {\rm M}^{\rm IV} \chi(x),
\eqno {\rm (33.)}$$
which satisfies the two conditions
$${\rm A}' = 0,\quad {\rm B}' = 0,$$
the functions $\phi(x)$ and $\chi(x)$ being determined, namely,
$$\phi(x) = {\Lambda' \over \Lambda'''} x^{\lambda'}
+ {\Lambda'' \over \Lambda'''} x^{\lambda''}
+ x^{\lambda'''},
\eqno {\rm (34.)}$$
and
$$\chi(x) = {{\rm M}' \over {\rm M}^{\rm IV}} x^{\mu'}
+ {{\rm M}'' \over {\rm M}^{\rm IV}} x^{\mu''}
+ {{\rm M}''' \over {\rm M}^{\rm IV}} x^{\mu'''}
+ x^{\mu^{\rm IV}},
\eqno {\rm (35.)}$$
but the multipliers $\Lambda'''$ and ${\rm M}^{\rm IV}$ being
arbitrary, and the other part of the process determining
afterwards the ratio of those two multipliers so as to satisfy
the remaining condition
$${\rm C}' = 0.$$
And hence it is easy to see that if we would exclude those
useless cases in which the ultimate expression for the new
variable~$y$, or the function $f(x)$, is a multiple of the
proposed evanescent polynome~${\rm X}$ of the $m^{\rm th}$ degree
in $x$, we must, in general, exclude the cases in which the two
functions $\phi(x)$ and $\chi(x)$, determined in the first part
of the process, are connected by a relation of the form
$$\chi(x) = a \phi(x) + \lambda {\rm X},
\eqno {\rm (36.)}$$
$a$ being any constant multiplier, and $\lambda {\rm X}$ any
multiple of ${\rm X}$. For in all such cases the expression
(33), obtained by the first part of the process, becomes
$$y = f(x) = (\Lambda''' + a {\rm M}^{\rm IV}) \phi(x)
+ \lambda {\rm M}^{\rm IV} {\rm X};
\eqno {\rm (37.)}$$
and since this gives, by the nature of the roots
$x_1,\ldots \, x_m$,
$$f(x_1) = (\lambda''' + a {\rm M}^{\rm IV}) \phi(x_1),\quad
\ldots \quad
f(x_m) = (\lambda''' + a {\rm M}^{\rm IV}) \phi(x_m),
\eqno {\rm (38.)}$$
we find, by the law (5) of the composition of the coefficients of
the transformed equation in $y$,
$${\rm C}' = c (\Lambda''' + a {\rm M}^{\rm IV})^3,
\eqno {\rm (39.)}$$
the multiplier $c$ being known, namely,
$$c = - \phi(x_1) \phi(x_2) \phi(x_3)
- \phi(x_1) \phi(x_2) \phi(x_4) - \hbox{\&c.}
\eqno {\rm (40.)}$$
and being in general different from $0$, because the three first
of the seven terms of the expression (21) for $y$ can only
accidentally suffice to resolve the original problem; so that
when we come, in the second part of the process, to satisfy the
condition
$${\rm C}' = 0,$$
we shall, in general, be obliged to assume
$$(\Lambda''' + a {\rm M}^{\rm IV})^3 = 0,
\eqno {\rm (41.)}$$
that is,
$$\Lambda''' + a {\rm M}^{\rm IV} = 0;
\eqno {\rm (42.)}$$
and consequently the expression (37) for $y$ reduces itself
ultimately to the form which we wished to exclude, since it
becomes
$$y = \lambda {\rm M}^{\rm IV} {\rm X}.
\eqno {\rm (43.)}$$
Reciprocally, it is clear that the second part of the process, or
the determination of the ratio of $\Lambda'''$ to ${\rm M}^{\rm
IV}$ in the expression (33), cannot conduct to this useless form
for $y$ unless the two functions $\phi(x)$ and $\chi(x)$ are
connected by a relation of the kind (36); because, when we equate
the expression (33) to any multiple of ${\rm X}$, we establish
thereby a relation of that kind between those two functions. We
must therefore endeavour to avoid those cases, and we need avoid
those only, which conduct to this relation (36), and we may do so
in the following manner.
\bigbreak
[4.]
Whatever positive integer the exponent $\nu$ may be, the power
$x^\nu$ may always be identically equated to an expression of
this form,
$$x^\nu = s_0^{(\nu)} + s_1^{(\nu)} x + s_2^{(\nu)} x^2
+ \cdots + s_{m-1}^{(\nu)} x^{m-1}
+ {\rm L}^{(\nu)} {\rm X},
\eqno {\rm (44.)}$$
$s_0^{(\nu)}, s_1^{(\nu)}, s_2^{(\nu)},\ldots \,
s_{(m-1)}^{(\nu)}$
being certain functions of the exponent $\nu$, and of the
coefficients ${\rm A}, {\rm B}, {\rm C},\ldots$ of the proposed
polynome ${\rm X}$, while ${\rm L}^{(\nu)}$ is a rational and
integral function of $x$, which is $= 0$ if $\nu$ be less than
the exponent~$m$ of the degree of that proposed
polynome~${\rm X}$, but otherwise is of the degree $\nu - m$. In
fact, if we divide the power $x^\nu$ by the polynome~${\rm X}$,
according to the usual rules of the integral division of
polynomes, so as to obtain an integral quotient and an integral
remainder, the integral quotient may be denoted by
${\rm L}^{(\nu)}$, and the integral remainder may be denoted by
$$s_0^{(\nu)} + s_1^{(\nu)} x + s_2^{(\nu)} x^2
+ \cdots + s_{m-1}^{(\nu)} x^{m-1},$$
and thus the identity (44) may be established. It may be
noticed that the $m$ coefficients
$s_0^{(\nu)}, s_1^{(\nu)},\ldots \, s_{(m-1)}^{(\nu)}$,
may be considered as symmetric functions of the $m$ roots
$x_1, x_2,\ldots \, x_m$ of the proposed equation ${\rm X} = 0$,
which may be determined by the $m$ relations,
$$\left. \eqalign{
x_1^\nu &= s_0^{(\nu)} + s_1^{(\nu)} x_1 + s_2^{(\nu)} x_1^2
+ \cdots + s_{m-1}^{(\nu)} x_1^{m-1},\cr
x_2^\nu &= s_0^{(\nu)} + s_1^{(\nu)} x_2 + s_2^{(\nu)} x_2^2
+ \cdots + s_{m-1}^{(\nu)} x_2^{m-1},\cr
\noalign{\hbox{$\cdots\cdots$}}
x_m^\nu &= s_0^{(\nu)} + s_1^{(\nu)} x_m + s_2^{(\nu)} x_m^2
+ \cdots + s_{m-1}^{(\nu)} x_m^{m-1}.\cr}
\right\}
\eqno {\rm (45.)}$$
These symmetric functions of the roots possess many other
important properties, but it is unnecessary here to develop
them.
Adopting the notation (44), we may put, for abridgment,
$$\left. \eqalign{
\Lambda' s_0^{(\lambda')}
+ \Lambda'' s_0^{(\lambda'')}
+ \Lambda''' s_0^{(\lambda''')}
&= p_0,\cr
\noalign{\hbox{$\cdots\cdots$}}
\Lambda' s_{m-1}^{(\lambda')}
+ \Lambda'' s_{m-1}^{(\lambda'')}
+ \Lambda''' s_{m-1}^{(\lambda''')} &= p_{m-1},\cr}
\right\}
\eqno {\rm (46.)}$$
$$\left. \eqalign{
{\rm M}' s_0^{(\mu')}
+ {\rm M}'' s_0^{(\mu'')}
+ {\rm M}''' s_0^{(\mu''')}
+ {\rm M}^{\rm IV} s_0^{(\mu^{\rm IV})}
&= p_0',\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm M}' s_{m-1}^{(\mu')}
+ {\rm M}'' s_{m-1}^{(\mu'')}
+ {\rm M}''' s_{m-1}^{(\mu''')}
+ {\rm M}^{\rm IV} s_{m-1}^{(\mu^{\rm IV})}
&= p_{m-1}',\cr}
\right\}
\eqno {\rm (47.)}$$
$$\Lambda' {\rm L}^{(\lambda')}
+ \Lambda'' {\rm L}^{(\lambda'')}
+ \Lambda''' {\rm L}^{(\lambda''')}
= \Lambda,
\eqno {\rm (48.)}$$
$${\rm M}' {\rm L}^{(\mu')}
+ {\rm M}'' {\rm L}^{(\mu'')}
+ {\rm M}''' {\rm L}^{(\mu''')}
+ {\rm M}^{\rm IV} {\rm L}^{(\mu^{\rm IV})}
= {\rm M},
\eqno {\rm (49.)}$$
$$\Lambda + {\rm M} = {\rm L}
\eqno {\rm (50.)}$$
and then the two parts, of which the expression for $y$ is
composed, will take the forms
$$\Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''}
= p_0 + p_1 x + \cdots + p_{m-1} x^{m-1} + \Lambda {\rm X},
\eqno {\rm (51.)}$$
$${\rm M}' x^{\mu'}
+ {\rm M}'' x^{\mu''}
+ {\rm M}''' x^{\mu'''}
+ {\rm M}^{\rm IV} x^{\mu^{\rm IV}}
= p_0' + p_1' x + \cdots + p_{m-1}' x^{m-1} + {\rm M} {\rm X},
\eqno {\rm (52.)}$$
and the expression itself will become
$$y = f(x) = p_0 + p_0' + (p_1 + p_1')x + \cdots
+ (p_{m-1} + p_{m-1}') x^{m-1} + {\rm L} {\rm X}.
\eqno {\rm (53.)}$$
At the same time we see that the case to be avoided, for the
reason lately assigned, is the case of proportionality of
$p_0', p_1',\ldots\, p_{m-1}'$ to $p_0, p_1,\ldots\, p_{m-1}$.
It is therefore convenient to introduce these new abbreviations,
$${p_{m-1}' \over p_{m-1}} = p,
\eqno {\rm (54.)}$$
and
$$p_0' - p p_0 = q_0,\quad
p_1' - p p_1 = q_1,\quad \ldots \quad
p_{m-2}' - p p_{m-2} = q_{m-2};
\eqno {\rm (55.)}$$
for thus we obtain the expressions
$$p_0' = q_0 + p p_0,\quad
p_1' = q_1 + p p_1,\quad \ldots \quad
p_{m-2}' = q_{m-2} + p p_{m-2},\quad
p_{m-1}' = p p_{m-1},
\eqno {\rm (56.)}$$
and
$$y = f(x) = (1 + p)(p_0 + p_1 x + \cdots + p_{m-1} x^{m-1})
+ q_0 + q_1 x + \cdots
+ q_{m-2} x^{m-2} + {\rm L} {\rm X};
\eqno {\rm (57.)}$$
and we have only to take care that the $m - 1$ quantities,
$q_0, q_1,\ldots\, q_{m-2}$ shall not all vanish. Indeed it is
tacitly supposed in (54) that $p_{m-1}$ does not vanish; but it
must be observed that Mr.~Jerrard's method itself essentially
supposes that the function
$\Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''}$
is not any multiple of the evanescent polynome ${\rm X}$, and
therefore that {\it at least some one} of the $m$ quantities
$p_0, p_1,\ldots \, p_{m-1}$ is different from $0$; now the
spirit of the definitional assumptions here made, and of the
reasonings which are to be founded upon them, requires only that
{\it some one\/} such non-evanescent quantity $p_i$ out of this
set $p_0, p_1,\ldots \, p_{m-1}$ should be made the denominator
of a fraction like (54),
$\displaystyle {p_i' \over p_i} = p$,
and that thus some one term $q_i x^i$ should be taken away out of
the difference of the two polynomes,
$p_0' + p_1' x + \cdots$ and
$p(p_0 + p_1 x + \cdots)$; and it is so easy to make this
adaptation, whenever the occasion may arise, that I shall retain
in the present discussion, the assumptions (54) (55), instead
of writing $p_i$ for $p_{m-1}$.
The expression (57) for $f(x)$, combined with the law (5) of
the composition of the coefficients ${\rm A}'$ and ${\rm B}'$,
shows that these two coefficients of the transformed equation in
$y$ may be expressed as follows,
$${\rm A}' = (1 + p) {\rm A}_{1,0}'' + {\rm A}_{0,1}'',
\eqno {\rm (58.)}$$
and
$${\rm B}'
= (1 + p)^2 {\rm B}_{2,0}'' + (1 + p) {\rm B}_{1,1}''
+ {\rm B}_{0,2}'';
\eqno {\rm (59.)}$$
${\rm A}_{h,i}''$ and ${\rm B}_{h,i}''$ being each a rational and
integral function of the $2m - 1$ quantities
$p_0$, $p_1,\ldots$ $p_{m-1}$, $q_0$, $q_1,\ldots$ $q_{m-2}$,
which is independent of the quantity $p$ and of the form of the
function~${\rm L}$, and is homogeneous of the dimension~$h$ with
respect to $p_0, p_1,\ldots \, p_{m-1}$, and of the dimension~$i$
with respect to $q_0, q_1,\ldots \, q_{m-2}$. Comparing these
expressions (58) and (59) with the analogous expressions
(24) and (25), (with which they would of necessity identically
coincide, if we were to return from the present to the former
symbols, by substituting, for
$p_0$, $p_1,\ldots$ $p_{m-1}$, $q_0$, $q_1,\ldots$ $q_{m-2}$,
their values as functions of
$\Lambda'$, $\Lambda''$, $\Lambda'''$,
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$,
deduced from the equations of definition (54) (55) and (46)
(47),) we find these identical equations:
$${\rm A}_{1,0}' = {\rm A}_{1,0}'';\quad
{\rm A}_{0,1} = p {\rm A}_{1,0}'' + {\rm A}_{0,1}'';
\eqno {\rm (60.)}$$
$${\rm B}_{2,0}'
= {\rm B}_{2,0}'';\quad
{\rm B}_{1,1}'
= 2p {\rm B}_{2,0}'' + {\rm B}_{1,1}'';\quad
{\rm B}_{0,2}'
= p^2 {\rm B}_{2,0}'' + p {\rm B}_{1,1}'' + {\rm B}_{0,2}'';
\eqno {\rm (61.)}$$
observing that whatever may be the dimension of any part of
${\rm A}'$ or ${\rm B}'$,with respect to the $m$ new quantities
$p$, $q_0$, $q_1,\ldots$ $q_{m-2}$, the same is the dimension of
that part, with respect to the four old quantities
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$.
The system of the five conditions (26) (27) (28) (29) (30)
may therefore be transformed to the following system,
$${\rm A}_{1,0}'' = 0,\quad {\rm B}_{2,0}'' = 0,
\eqno {\rm (62.)}$$
$${\rm A}_{0,1}'' = 0,\quad
{\rm B}_{1,1}'' = 0,\quad
{\rm B}_{0,2}'' = 0;
\eqno {\rm (63.)}$$
and may in general be treated as follows. The two conditions
(62), combined with the $m$ equations of definition (46), will
in general determine the $m + 2$ ratios of the $m + 3$ quantities
$p_0$, $p_1,\ldots$ $p_{m-1}$,
$\Lambda$, $\Lambda''$, $\Lambda'''$;
and then the three conditions (63), combined with the $m$
equations of definition (47) and with the $m$ other equations
(56), will in general determine the $2m+ 3$ ratios of the
$2m + 4$ quantities
$q_0$, $q_1,\ldots$ $q_{m-2}$, $p p_{m-1}$,
$p_0'$, $p_1',\ldots$ $p_{m-1}'$,
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$;
after which, the ratio of $\Lambda'''$ to
${\rm M}^{\rm IV}$ is to be determined, as before, so as to
satisfy the remaining condition ${\rm C}' = 0$. But because the
last-mentioned system, of $2m + 3$ homogeneous equations, (63)
(56) (47), between $2m+ 4$ quantities, involves, as a part of
itself, the system (63) of {\it three homogeneous equations\/}
(rational and integral) {\it between} $m - 1$ {\it quantities\/}
$q_0, q_1,\ldots \, q_{m-2}$, we see that it will in general
conduct to the result which we wished to exclude, namely the
simultaneous vanishing of all those quantities,
$$q_0 = 0,\quad q_1 = 0,\quad \ldots \quad q_{m-2} = 0,
\eqno {\rm (64.)}$$
{\it unless their number $m - 1$ be greater than $3$}, that is,
{\it unless the degree $m$ of the proposed equation\/} (1)
{\it be at least equal to the minor limit\/} {\sc five}. It
results, then, from this discussion, that {\it the transformation
by which Mr.~Jerrard has succeeded in taking away three terms at
once from the general equation of the $m^{\rm th}$ degree, is not
in general applicable when that degree is lower than the
$5$th\/}; in such a manner that it is in general inadequate to
reduce the biquadratic equation
$$x^4 + {\rm A} x^3 + {\rm B} x^2 + {\rm C} x + {\rm D} = 0,
\eqno {\rm (65.)}$$
to the binomial form
$$y^4 + {\rm D}' = 0,
\eqno {\rm (66.)}$$
except by the useless assumption
$$y = {\rm L} ( x^4 + {\rm A} x^3 + {\rm B} x^2
+ {\rm C} x + {\rm D} ),
\eqno {\rm (67.)}$$
which gives
$$y^4 = 0.
\eqno {\rm (68.)}$$
However, the foregoing discussion by be considered as
{\it confirming the adequacy of the method to reduce the general
equation of the $5$th degree},
$$x^5 + {\rm A} x^4 + {\rm B} x^3 + {\rm C} x^2
+ {\rm D} x + {\rm E} = 0,
\eqno {\rm (69.)}$$
{\it to the trinomial form}
$$y^5 + {\rm D}' y + {\rm E}' = 0;
\eqno {\rm (70.)}$$
{\it and to effect the analogous transformation\/} (12)
{\it for equations of all higher degrees\/}: an unexpected and
remarkable result, which is one of Mr.~Jerrard's principal
discoveries.
\bigbreak
[5.]
Analogous remarks apply to the process proposed by the same
mathematician for taking away the second, third and fifth terms
at once from the general equation (1), so as to reduce that
equation to the form (18). This process agrees with the
foregoing in the whole of its first part, that is, in the
assumption of the form (21) for $f(x)$, and in the determination
of the five ratios (31) so as to satisfy the two conditions
${\rm A}' = 0$, ${\rm B}' = 0$, by satisfying the five others
(26) (27) (28) (29) (30), into which those two may be decomposed;
and the difference is only in the second part of the process,
that is, in determining the remaining ratio (32) so as to satisfy
the condition ${\rm D}' = 0$, instead of the condition
${\rm C}' = 0$, by resolving a biquadratic instead of a cubic
equation. The discussion which has been given of the former
process of
transformation adapts itself therefore, with scarcely any change,
to the latter process also, and shows that this process can only
be applied with success, in general, to equations of the fifth
and higher degrees. It is, however, a remarkable result that it
can be applied generally to such equations, and especially that
the general equation of the fifth degree may be brought by it to
the following trinomial form,
$$y^5 + {\rm C}' y^2 + {\rm E}' = 0,
\eqno {\rm (71.)}$$
as it was reduced, by the former process, to the form
$$y^5 + {\rm D}' y + {\rm E}' = 0.
\eqno {\rm (70.)}\qquad$$
Mr.~Jerrard, to whom the discovery of these transformations is
due, has remarked that by changing $y$ to
$\displaystyle {1 \over z}$
we get two other trinomial forms to which {\it the general
equation of the fifth degree} may be reduced; so that {\it in any
future researches respecting the solution of such equations, it
will be permitted to set out with any one of these four trinomial
forms},
$$\left. \eqalign{
x^5 + {\rm A} x^4 + {\rm E} &= 0,\cr
x^5 + {\rm B} x^3 + {\rm E} &= 0,\cr
x^5 + {\rm C} x^2 + {\rm E} &= 0,\cr
x^5 + {\rm D} x + {\rm E} &= 0,\cr}
\right\}
\eqno {\rm (72.)}$$
in which the intermediate coefficient ${\rm A}$ or ${\rm B}$ or
${\rm C}$ or ${\rm D}$ may evidently be made equal to unity, or
to any other assumed number different from zero. We may, for
example, consider the difficulty of resolving the {\it general\/}
equation of the fifth degree as reduced by Mr.~Jerrard's
researches to the difficulty of resolving an equation of the form
$$x^5 + x + {\rm E} = 0;
\eqno {\rm (73.)}$$
or of this other form,
$$x^5 - x + {\rm E} = 0.
\eqno {\rm (74.)}$$
It is, however, important to remark that the coefficients of
these new or transformed equations will often be imaginary, even
when the coefficients of the original equation of the form (69)
are real.
\bigbreak
[6.]
In order to accomplish the transformation (20), (to the
consideration of which we shall next proceed,) Mr.~Jerrard
assumes, in general, an expression with {\it twelve\/} terms,
$$\eqalignno{
y = f(x)
&= \Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm M}' x^{\mu'}
+ {\rm M}'' x^{\mu''}
+ {\rm M}''' x^{\mu'''}
+ {\rm M}^{\rm IV} x^{\mu^{\rm IV}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm N}' x^{\nu'}
+ {\rm N}'' x^{\nu''}
+ {\rm N}''' x^{\nu'''}
+ {\rm N}^{\rm IV} x^{\nu^{\rm IV}}
+ {\rm N}^{\rm V} x^{\nu^{\rm V}};
&{\rm (75.)}\cr}$$
the twelve unequal exponents,
$$\lambda', \lambda'', \lambda''',
\mu', \mu'', \mu''', \mu^{\rm IV},
\nu', \nu'', \nu''', \nu^{\rm IV}, \nu^{\rm V},
\eqno {\rm (76.)}$$
being chosen with pleasure out of the indefinite line of integers
(22); and the twelve coefficients,
$$\Lambda', \Lambda'', \Lambda''',
{\rm M}', {\rm M}'', {\rm M}''', {\rm M}^{\rm IV},
{\rm N}', {\rm N}'', {\rm N}''',
{\rm N}^{\rm IV}, {\rm N}^{\rm V},
\eqno {\rm (77.)}$$
or rather their eleven ratios, which may be arranged and grouped
as follows,
$${\Lambda' \over \Lambda'''},\quad
{\Lambda'' \over \Lambda'''},
\eqno {\rm (78.)}$$
$${{\rm M}' \over {\rm M}^{\rm IV}},
{{\rm M}'' \over {\rm M}^{\rm IV}},
{{\rm M}''' \over {\rm M}^{\rm IV}},
\eqno {\rm (79.)}$$
$${{\rm N}' \over {\rm N}^{\rm V}},\quad
{{\rm N}'' \over {\rm N}^{\rm V}},\quad
{{\rm N}''' \over {\rm N}^{\rm V}},\quad
{{\rm N}^{\rm IV} \over {\rm N}^{\rm V}},
\eqno {\rm (80.)}$$
$${{\rm M}^{\rm IV} \over {\rm N}^{\rm V}},
\eqno {\rm (81.)}$$
$${\Lambda''' \over {\rm N}^{\rm V}},
\eqno {\rm (82.)}$$
being then determined so as to satisfy the system of the three
conditions
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
$${\rm C}' = 0,
\eqno {\rm (14.)}\qquad$$
$${\rm D}' - \alpha {\rm B}'^2 = 0,
\eqno {\rm (19.)}\qquad$$
by satisfying another system, composed of eleven equations, which
are obtained by decomposing the condition (8) into three, and the
condition (14) into seven new equations, as follows. By the law
(5) of the formation of the four coefficients
${\rm A}'$, ${\rm B}'$, ${\rm C}'$, ${\rm D}'$, and by the
assumed expression (75), those four coefficients are rational and
integral and homogeneous functions, of the first, second, third
and fourth degrees, of the twelve coefficients (77); and
therefore, when these latter coefficients are distributed into
three groups, one group containing
$\Lambda'$, $\Lambda''$, $\Lambda'''$,
another group containing
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$,
and the third group containing
${\rm N}'$, ${\rm N}''$, ${\rm N}'''$,
${\rm N}^{\rm IV}$, ${\rm N}^{\rm V}$,
the coefficient or function~${\rm A}'$ may be decomposed into
three parts,
$${\rm A}'
= {\rm A}_{1,0,0}' + {\rm A}_{0,1,0}' + {\rm A}_{0,0,1}',
\eqno {\rm (83.)}$$
and the coefficient or function ${\rm C}'$ may be decomposed in
like manner into ten parts,
$${\rm C}' = {\rm C}_{3,0,0}'
+ {\rm C}_{2,1,0}'
+ {\rm C}_{2,0,1}'
+ {\rm C}_{1,2,0}'
+ {\rm C}_{1,1,1}'
+ {\rm C}_{1,0,2}'
+ {\rm C}_{0,3,0}'
+ {\rm C}_{0,2,1}'
+ {\rm C}_{0,1,2}'
+ {\rm C}_{0,0,3}',
\eqno {\rm (84.)}$$
in which each of the symbols of the forms ${\rm A}_{h,i,k}'$ and
${\rm C}_{h,i,k}'$ denotes a rational and integral function of
the twelve quantities (77); which function (${\rm A}_{h,i,k}'$ or
${\rm C}_{h,i,k}'$) is homogeneous of the dimension~$h$ with
respect to the quantities
$\Lambda'$, $\Lambda''$, $\Lambda'''$,
of the dimension~$i$ with respect to the quantities
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$,
and of the dimension~$k$ with respect to the quantities
${\rm N}'$, ${\rm N}''$, ${\rm N}'''$,
${\rm N}^{\rm IV}$, ${\rm N}^{\rm V}$.
Accordingly Mr.~Jerrard decomposes the conditions ${\rm A}' = 0$
and ${\rm C}' = 0$ into ten others, which may be thus arranged:
$${\rm A}_{1,0,0}' = 0,\quad
{\rm C}_{3,0,0}' = 0;
\eqno {\rm (85.)}$$
$${\rm A}_{0,1,0}' = 0,\quad
{\rm C}_{2,1,0}' = 0,\quad
{\rm C}_{1,2,0}' = 0;
\eqno {\rm (86.)}$$
$${\rm A}_{0,0,1}' = 0,\quad
{\rm C}_{2,0,1}' = 0,\quad
{\rm C}_{1,1,1}' = 0,\quad
{\rm C}_{1,0,2}' = 0;
\eqno {\rm (87.)}$$
$${\rm C}_{0,3,0}' + {\rm C}_{0,2,1}'
+ {\rm C}_{0,1,2}' + {\rm C}_{0,0,3}' = 0;
\eqno {\rm (88.)}$$
nine of the thirteen parts of the expressions (83) and (84)
being made to vanish separately, and the sum of the other four
parts being also made to vanish. He then determines the two
ratios (78), so as to satisfy the two conditions (85); the
three ratios (79), so as to satisfy the three conditions (86);
the four ratios (80), so as to satisfy the four conditions (87);
and the ratio (81) so as to satisfy the condition (88); all
which determinations can in general be successively effected,
without its being necessary to resolve any equation higher than
the third degree. The first part of the process is now completed,
that is, the two conditions (8) and (14),
$${\rm A}' = 0,\quad {\rm C}' = 0,$$
are now both satisfied by an expression of the form
$$y = f(x) = \Lambda''' \phi(x) + {\rm N}^{\rm V} \chi(x),
\eqno {\rm (89.)}$$
which is analogous to (33), and in which the functions $\phi(x)$
and $\chi(x)$ are known, but the multipliers $\Lambda'''$ and
${\rm N}^{\rm V}$ are arbitrary; and the second and only
remaining part of the process consists in determining the
remaining ratio (82), of $\Lambda'''$ to ${\rm N}^{\rm V}$, by
resolving an equation of the fourth degree, so as to satisfy the
remaining condition,
$${\rm D}' - \alpha {\rm B}'^2 = 0.
\eqno {\rm (19.)}\qquad$$
\bigbreak
[7.]
Such, then, (the notation excepted,) is Mr.~Jerrard's general
process for reducing the equation of the $m^{\rm th}$ degree,
$${\rm X} = x^m + {\rm A} x^{m-1} + {\rm B} x^{m-2}
+ {\rm C} x^{m-3} + {\rm D} x^{m-4}
+ {\rm E} x^{m-5} + \hbox{\&c.} = 0
\eqno {\rm (1.)}\qquad$$
to the form
$${\rm Y} = y^m + {\rm B}' y^{m-2} + \alpha {\rm B}'^2 y^{m-4}
+ {\rm E}' y^{m-5} + \hbox{\&c.} = 0,
\eqno {\rm (20.)}\qquad$$
without resolving any auxiliary equation of a higher degree than
the fourth. But, on considering this remarkable process with
attention, we perceive that if we would avoid its becoming
illusory, by conducting to an expression for $y$ which is a
multiple of the proposed polynome~${\rm X}$, we must, in general,
(for reasons analogous to those already explained in discussing
the transformation (12),) exclude all those cases in which the
functions $\phi(x)$ and $\chi(x)$ in the expression (89) are
connected by a relation of the form
$$\chi(x) = a \phi(x) + \lambda {\rm X};
\eqno {\rm (36.)}\qquad$$
because, in all the cases in which such a relation exists, the
first part of the process conducts to an expression of the form
$$y = (\Lambda''' + a {\rm N}^{\rm V}) \phi(x)
+ \lambda {\rm N}^{\rm V} {\rm X},
\eqno {\rm (90.)}$$
and then the second part of the same process gives in general
$$(\Lambda''' + a {\rm N}^{\rm V})^4 = 0,
\eqno {\rm (91.)}$$
that is
$$\Lambda''' + a {\rm N}^{\rm V} = 0,
\eqno {\rm (92.)}$$
and ultimately
$$y = \lambda {\rm N}^{\rm V} {\rm X}.
\eqno {\rm (93.)}$$
On the other hand, the second part of the process cannot conduct
to this useless form for $y$, unless the first part of the
process has led to functions $\phi(x)$, $\chi(x)$, connected by a
relation of the form (36). This consideration suggests the
introduction of the following new system of equations of
definition.
$$\left. \eqalign{
{\rm N}' s_0^{(\nu')}
+ {\rm N}'' s_0^{(\nu'')}
+ {\rm N}''' s_0^{(\nu''')}
+ {\rm N}^{\rm IV} s_0^{(\nu^{\rm IV})}
+ {\rm N}^{\rm V} s_0^{(\nu^{\rm V})}
= p_0'',\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm N}' s_{m-1}^{(\nu')}
+ {\rm N}'' s_{m-1}^{(\nu'')}
+ {\rm N}''' s_{m-1}^{(\nu''')}
+ {\rm N}^{\rm IV} s_{m-1}^{(\nu^{\rm IV})}
+ {\rm N}^{\rm V} s_{m-1}^{(\nu^{\rm V})}
= p_{m-1}'',\cr}
\right\}
\eqno {\rm (94.)}$$
$${\rm N}' {\rm L}^{(\nu')}
+ {\rm N}'' {\rm L}^{(\nu'')}
+ {\rm N}''' {\rm L}^{(\nu''')}
+ {\rm N}^{\rm IV} {\rm L}^{(\nu^{\rm IV})}
+ {\rm N}^{\rm V} {\rm L}^{(\nu^{\rm V})}
= {\rm N},
\eqno {\rm (95.)}$$
$${p_{m-1}'' \over p_{m-1}} = p',
\eqno {\rm (96.)}$$
$$p_0'' - p' p_0 = q_0',\quad
p_1'' - p' p_1 = q_1',\quad \ldots \quad
p_{m-2}'' - p' p_{m-2} = q_{m-2}'
\eqno {\rm (97.)}$$
to be combined with the definitions (46), (47), (48), (49),
(54), (55), and with the following, which may now be
conveniently used instead of the definition (50),
$$\Lambda + {\rm M} + {\rm N} = {\rm L}.
\eqno {\rm (98.)}$$
In this notation we shall have, as before,
$$p_0' = q_0 + p p_0,\quad
p_1' = q_1 + p p_1,\quad \ldots \quad
p_{m-2}' = q_{m-2} + p p_{m-2},\quad
p_{m-1}' = p p_{m-1},
\eqno {\rm (56.)}\qquad$$
and shall also have the analogous expressions
$$p_0'' = q_0' + p' p_0,\quad
p_1'' = q_1' + p' p_1,\quad \ldots \quad
p_{m-2}'' = q_{m-2}' + p' p_{m-2},\quad
p_{m-1}'' = p' p_{m-1};
\eqno {\rm (99.)}$$
the expression (75) for $y$ will become
$$y = f(x) = p_0 + p_0' + p_0''
+ (p_1 + p_1' + p_1'') x + \cdots
+ (p_{m-1} + p_{m-1}' + p_{m-1}'') x^{m-1}
+ {\rm L} {\rm X},
\eqno {\rm (100.)}$$
that is, by (56) and (99),
$$\left. \eqalign{
y = f(x)
&= (1 + p + p') (p_0 + p_1 x + \cdots + p_{m-1} x^{m-1}) \cr
&\mathrel{\phantom{=}} \mathord{}
+ q_0 + q_0'
+ (q_1 + q_1') x + \cdots
+ (q_{m-2} + q_{m-2}') x^{m-2} + {\rm L} {\rm X}:\cr}
\right\}
\eqno {\rm (101.)}$$
and the excluded case, or case of failure, will now be the case
when the sums
$$p_0' + p_0'',\quad
p_1' + p_1'',\quad \ldots \quad
p_{m-1}' + p_{m-1}''$$
are proportional to $p_0, p_1,\ldots \, p_{m-1}$, that is, when
$$q_0 + q_0' = 0,\quad
q_1 + q_1' = 0,\quad \ldots \quad
q_{m-2} + q_{m-2}' = 0.
\eqno {\rm (102.)}$$
Indeed it is here tacitly supposed that $p_{m-1}$ does not
vanish; but Mr.~Jerrard's method itself supposes tacitly that at
least some one, such as $p_i$, of the $m$ quantities
$p_0,\ldots \, p_{m-1}$, is different from $0$, and it is easy,
upon occasion, to substitute any such non-evanescent quantity
$p_i$ for $p_{m-1}$, and then to make the few other connected
changes which the spirit of this discussion requires.
The expression (101) for $f(x)$, combined with the law (5) of
the composition of the coefficients ${\rm A}'$ and ${\rm C}'$,
gives, for those coefficients, expressions of the forms,
$${\rm A}'
= (1 + p + p') {\rm A}_{1,0,0}''
+ {\rm A}_{0,1,0}''
+ {\rm A}_{0,0,1}'',
\eqno {\rm (103.)}$$
and
$$\eqalignno{
{\rm C}' &= (1 + p + p')^3 {\rm C}_{3,0,0}''
+ (1 + p + p')^2
( {\rm C}_{2,1,0}'' + {\rm C}_{2,0,1}'' ) \cr
&\mathrel{\phantom{=}} \mathord{}
+ (1 + p + p')
( {\rm C}_{1,2,0}'' + {\rm C}_{1,1,1}''
+ {\rm C}_{1,0,2}'' )
+ {\rm C}_{0,3,0}'' + {\rm C}_{0,2,1}''
+ {\rm C}_{0,1,2}'' + {\rm C}_{0,0,3}'',
&{\rm (104.)}\cr}$$
${\rm A}_{h,i,k}''$ and ${\rm C}_{h,i,k}''$ being rational and
integral functions of the $3m - 2$ quantities
$p_0$, $p_1,\ldots$ $p_{m-1}$,
$q_0$, $q_1,\ldots$ $q_{m-2}$,
$q_0'$, $q_1',\ldots$ $q_{m-2}'$,
which functions are independent of $p$, $p'$, and ${\rm L}$, and
are homogeneous of the dimension $h$ with respect to
$p_0,\ldots \, p_{m-1}$, of the dimension $i$ with respect to
$q_0,\ldots \, q_{m-2}$, of the dimension $k$ with respect to
$q_0',\ldots \, q_{m-2}'$; they are also such that the sums
$${\rm A}_{0,1,0}'' + {\rm A}_{0,0,1}''
\eqno {\rm (105.)}$$
and
$${\rm C}_{2,1,0}'' + {\rm C}_{2,0,1}''
\eqno {\rm (106.)}$$
are homogeneous functions, of the 1st dimension, of the $m - 1$
sums $q_0 + q_0',\ldots \, q_{m-2} + q_{m-2}'$; while the sum
$${\rm C}_{1,2,0}'' + {\rm C}_{1,1,1}'' + {\rm C}_{1,0,2}''
\eqno {\rm (107.)}$$
is a homogeneous function, of the 2nd dimension, and the sum
$${\rm C}_{0,3,0}'' + {\rm C}_{0,2,1}''
+ {\rm C}_{0,1,2}'' + {\rm C}_{0,0,3}''
\eqno {\rm (108.)}$$
is a homogeneous function, of the 3rd dimension, of the same
$m - 1$ quantities. These new expressions (103) and (104), for
the coefficients ${\rm A}'$ and ${\rm C}'$, must identically
coincide with the former expressions (83) and (84), when we
return from the present to the former notation, by changing
$p$, $p'$, $p_0$, $p_1,\ldots$ $p_{m-1}$,
$q_0$, $q_1,\ldots$ $q_{m-2}$, $q_0'$, $q_1',\ldots$ $q_{m-2}'$,
to their values as functions of
$\Lambda'$,~$\Lambda''$,~$\Lambda'''$,
${\rm M}'$,~${\rm M}''$,~${\rm M}'''$,~${\rm M}^{\rm IV}$,
${\rm N}'$,~${\rm N}''$, ${\rm N}'''$,
${\rm N}^{\rm IV}$,~${\rm N}^{\rm V}$;
and hence it is easy to deduce the following identical equations:
$$\left. \eqalign{
{\rm A}_{1,0,0}' &= {\rm A}_{1,0,0}'';\cr
{\rm A}_{0,1,0}' &= p {\rm A}_{1,0,0}'' + {\rm A}_{0,1,0}'';\cr
{\rm A}_{0,0,1}' &= p' {\rm A}_{1,0,0}'' + {\rm A}_{0,0,1}'';\cr}
\right\}
\eqno {\rm (109.)}$$
and
$$\left. \eqalign{
{\rm C}_{3,0,0}'
&= {\rm C}_{3,0,0}'';\cr
{\rm C}_{2,1,0}'
&= 3p {\rm C}_{3,0,0}'' + {\rm C}_{2,1,0}'';\cr
{\rm C}_{2,0,1}'
&= 3p' {\rm C}_{3,0,0}'' + {\rm C}_{2,0,1}'';\cr
{\rm C}_{1,2,0}'
&= 3 p^2 {\rm C}_{3,0,0}'' + 2p {\rm C}_{2,1,0}''
+ {\rm C}_{1,2,0}'';\cr
{\rm C}_{1,1,1}'
&= 6pp' {\rm C}_{3,0,0}'' + 2p' {\rm C}_{2,1,0}''
+ 2p {\rm C}_{2,0,1}'' + {\rm C}_{1,1,1}'';\cr
{\rm C}_{1,0,2}'
&= 3p'^2 {\rm C}_{3,0,0}'' + 2p' {\rm C}_{2,0,1}''
+ {\rm C}_{1,0,2}'';\cr
{\rm C}_{0,3,0}'
&+ {\rm C}_{0,2,1}' + {\rm C}_{0,1,2}' + {\rm C}_{0,0,3}'
= (p + p')^3 {\rm C}_{3,0,0}''
+ (p + p')^2
({\rm C}_{2,1,0}'' + {\rm C}_{2,0,1}'') \cr
&\mathrel{\phantom{=}} \mathord{}
+ (p + p') ({\rm C}_{1,2,0}''
+ {\rm C}_{1,1,1}'' + {\rm C}_{1,0,2}'')
+ {\rm C}_{0,3,0}'' + {\rm C}_{0,2,1}''
+ {\rm C}_{0,1,2}'' + {\rm C}_{0,0,3}''.\cr}
\right\}
\eqno {\rm (110.)}$$
The system of the ten conditions (85), (86), (87), (88), may
therefore be transformed to the following:
$${\rm A}_{1,0,0}'' = 0,\quad
{\rm C}_{3,0,0}'' = 0;
\eqno {\rm (111.)}$$
$${\rm A}_{0,1,0}'' = 0,\quad
{\rm C}_{2,1,0}'' = 0,\quad
{\rm C}_{1,2,0}'' = 0;
\eqno {\rm (112.)}$$
$${\rm A}_{0,0,1}'' = 0,\quad
{\rm C}_{2,0,1}'' = 0,\quad
{\rm C}_{1,1,1}'' = 0,\quad
{\rm C}_{1,0,2}'' = 0;
\eqno {\rm (113.)}$$
$${\rm C}_{0,3,0}'' + {\rm C}_{0,2,1}''
+ {\rm C}_{0,1,2}'' + {\rm C}_{0,0,3}'' = 0;
\eqno {\rm (114.)}$$
and may in general be treated as follows. The two conditions
(111) may first be combined with the $m$ equations of definition
(46), and employed to determine the $m + 2$ ratios of the
$m + 3$ quantities
$p_0,\ldots$ $p_{m-1}$, $\Lambda'$ $\Lambda''$ $\Lambda'''$;
and therefore to give a result of the form
$$\Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''}
= \Lambda''' \phi(x),
\eqno {\rm (115.)}$$
the function $\phi(x)$ being known. The three conditions (112),
combined with the $2m$ equations (47) and (56), may then be
used to determine the $2m + 3$ ratios of the $2m + 4$ quantities
$q_0,\ldots$ $q_{m-2}$, $p p_{m-1}$, $p_0',\ldots$ $p_{m-1}'$,
${\rm M}'$, ${\rm M}''$, ${\rm M}'''$, ${\rm M}^{\rm IV}$, and
consequently to give
$${\rm M}' x^{\mu'} + {\rm M}'' x^{\mu''} + {\rm M}''' x^{\mu'''}
+ {\rm M}^{\rm IV} x^{\mu^{\rm IV}}
= {\rm M}^{\rm IV} \psi(x),
\eqno {\rm (116.)}$$
$\psi(x)$ denoting a known function. The four conditions (113)
may next be combined with the $2m$ equations (94) and (99), so
as to determine the $2m + 4$ ratios of the $2m + 5$ quantities
$q_0',\ldots$ $q_{m-2}'$, $p' p_{m-1}$,
$p_0'',\ldots$ $p_{m-1}''$,
${\rm N}'$, ${\rm N}''$, ${\rm N}'''$,
${\rm N}^{\rm IV}$, ${\rm N}^{\rm V}$;
and thus we shall have
$${\rm N}' x^{\nu'} + {\rm N}'' x^{\nu''} + {\rm N}''' x^{\nu'''}
+ {\rm N}^{\rm IV} x^{\nu^{\rm IV}}
+ {\rm N}^{\rm V} x^{\nu^{\rm V}}
= {\rm N}^{\rm V} \omega(x),
\eqno {\rm (117.)}$$
the function $\omega(x)$ also being known; so that, at this
stage, the expression (75) for $y$ will be reduced to the form
$$y = f(x)
= \Lambda''' \phi(x) + {\rm M}^{\rm IV} \psi(x)
+ {\rm N}^{\rm V} \omega(x),
\eqno {\rm (118.)}$$
the three functions $\phi(x)$, $\psi(x)$, $\omega(x)$ being
known, but the three coefficients
$\Lambda'''$, ${\rm M}^{\rm IV}$, ${\rm N}^{\rm V}$,
being arbitrary. The condition (114) will next determine the
ratio of any one of the quantities $q_0,\ldots \, q_{m-2}$ to any
one of the quantities $q_0',\ldots \, q_{m-2}'$, and therefore
also the connected ratio of ${\rm M}^{\rm IV}$ to
${\rm N}^{\rm V}$, and consequently will give
$${\rm M}^{\rm IV} \psi(x) + {\rm N}^{\rm V} \omega(x)
= {\rm N}^{\rm V} \chi(x),
\eqno {\rm (119.)}$$
$\chi(x)$ being another known function; and thus we shall have
accomplished, in a way apparently but not essentially different
from that employed in the foregoing article, the first part of
Mr.~Jerrard's process, namely, the discovery of an expression for
$y$, of the form
$$y = f(x) = \Lambda''' \phi(x) + {\rm N}^{\rm V} \chi(x),
\eqno {\rm (89.)}\qquad$$
which satisfies the two conditions
$${\rm A}' = 0,\quad {\rm C}' = 0,$$
the functions $\phi(x)$ and $\chi(x)$ being determined and known,
but the multipliers $\Lambda'''$ and ${\rm N}^{\rm V}$ being
arbitrary: after which it will only remain to perform the second
part of the process, namely, the determination of the ratio of
$\Lambda'''$ to ${\rm N}^{\rm V}$, so as to satisfy the remaining
condition
$${\rm D}' - \alpha {\rm B}'^2 = 0,$$
by resolving a biquadratic equation.
\bigbreak
[8.]
The advantage of this new way of presenting the first part of
Mr.~Jerrard's process is that it enables us to perceive, that if
we would avoid the case of failure above mentioned, we must in
general exclude those cases in which the ratios
$${q_0' \over q_{m-2}'},\quad
{q_1' \over q_{m-2}'},\quad \ldots \quad
{q_{m-3}' \over q_{m-2}'},
\eqno {\rm (120.)}$$
determined, as above explained, through the medium of the
conditions (113), coincide with the ratios
$${q_0 \over q_{m-2}},\quad
{q_1 \over q_{m-2}},\quad \ldots \quad
{q_{m-3} \over q_{m-2}},
\eqno {\rm (121.)}$$
determined, at an earlier stage, through the medium of the
conditions (112). In fact, when the ratios (120) coincide with
the ratios (121), they necessarily coincide with the following
ratios also,
$${q_0 + q_0' \over q_{m-2} + q_{m-2}'},\quad
{q_1 + q_1' \over q_{m-2} + q_{m-2}'},\quad \ldots \quad
{q_{m-3} + q_{m-3}' \over q_{m-2} + q_{m-2}'};
\eqno {\rm (122.)}$$
and unless the ratios, thus determined, of the $m - 1$ sums
$q_0 + q_0',\ldots \, q_{m-2} + q_{m-2}'$, are accidentally such
as to satisfy the condition (114), which had not been employed in
determining them, then that condition, which is a rational and
integral and homogeneous equation of the third degree between
those quantities, will oblige them all to vanish, and therefore
will conduct to the case of failure (102). Reciprocally, in that
case of failure, the ratios (120) coincide with the ratios (121),
because we have then
$$q_0' = - q_0,\quad
q_1' = - q_1,\quad \ldots \quad
q_{m-2}' = - q_{m-2}.
\eqno {\rm (123.)}$$
The case to be excluded, in general, is therefore that in which
the $m - 1$ quantities $q_0',\ldots \, q_{m-2}'$ are proportional
to the $m - 1$ quantities $q_0,\ldots \, q_{m-2}$; and this
consideration suggests the introduction of the following new
symbols or definitions,
$${q_{m-2}' \over q_{m-2}} = q,
\eqno {\rm (124.)}$$
$$q_0' - q q_0 = r_0,\quad
q_1' - q q_1 = r_1,\quad \ldots \quad
q_{m-3}' - q q_{m-3} = r_{m-3};
\eqno {\rm (125.)}$$
because, by introducing these, we shall only be obliged to guard
against the simultaneous vanishing of the $m - 2$ quantities
$r_0, r_1,\ldots \, r_{m-3}$; that is, we shall have the
following simplified statement of the general case of failure,
$$r_0 = 0,\quad r_1 = 0,\quad \ldots \quad r_{m-3} = 0.
\eqno {\rm (126.)}$$
Adopting, therefore, the definitions (124) and (125), and
consequently the expressions
$$q_0' = r_0 + q q_0,\quad
q_1' = r_1 + q q_1,\quad \ldots
q_{m-3}' = r_{m-3} + q q_{m-3},\quad
q_{m-2}' = q q_{m-2},
\eqno {\rm (127.)}$$
which give
$$\left. \eqalign{
q_0 + q_0' &= (1 + q) q_0 + r_0,\quad
q_1 + q_1' = (1 + q) q_1 + r_1,\quad \ldots \cr
q_{m-3} + q_{m-3}' &= (1 + q) q_{m-3} + r_{m-3},\quad
q_{m-2} + q_{m-2}' = (1 + q) q_{m-2},\cr}
\right\}
\eqno {\rm (128.)}$$
we easily perceive that the three homogeneous functions (105)
(106) (107), of these $m - 1$ sums
$q_0 + q_0',\ldots \, q_{m-2} + q_{m-2}'$, may be expressed in
the following manner:
$${\rm A}_{0,1,0}'' + {\rm A}_{0,0,1}''
= (1 + q) {\rm A}_{0,1,0}''' + {\rm A}_{0,0,1}''';
\eqno {\rm (129.)}$$
$${\rm C}_{2,1,0}'' + {\rm C}_{2,0,1}''
= (1 + q) {\rm C}_{2,1,0}''' + {\rm C}_{2,0,1}''';
\eqno {\rm (130.)}$$
$${\rm C}_{1,2,0}'' + {\rm C}_{1,1,1}'' + {\rm C}_{1,0,2}''
= (1 + q)^2 {\rm C}_{1,2,0}''' + (1 + q) {\rm C}_{1,1,1}'''
+ {\rm C}_{1,0,2}''';
\eqno {\rm (131.)}$$
the symbol ${\rm A}_{h,i,k}'''$ or ${\rm C}_{h,i,k}'''$ denoting
here a rational and integral function of the $3m - 3$ quantities
$p_0,\ldots$ $p_{m-1}$,
$q_0,\ldots$ $q_{m-2}$,
$r_0,\ldots$ $r_{m-3}$,
which is, like the function ${\rm A}_{h,i,k}''$ or ${\rm
C}_{h,i,k}''$, homogeneous of the dimension $h$ with respect to
$p_0,\ldots \, p_{m-1}$, and of the dimension~$i$ with respect to
$q_0,\ldots \, q_{m-2}$, but is homogeneous of the dimension $k$
with respect to $r_0,\ldots \, r_{m-3}$, and is independent of
$q_0',\ldots \, q_{m-2}'$ and of $p$, $p'$, $q$; whereas
${\rm A}_{h,i,k}''$ or ${\rm C}_{h,i,k}''$ was homogeneous of the
dimension $k$ with respect to $q_0',\ldots \, q_{m-2}'$, and was
independent of $r_0,\ldots \, r_{m-3}$. The three identical
equations (129) (130) (131) may be decomposed into the seven
following, which are analogous to (60) and (61):
$${\rm A}_{0,1,0}''
= {\rm A}_{0,1,0}''';\quad
{\rm A}_{0,0,1}''
= q {\rm A}_{0,1,0}''' + {\rm A}_{0,0,1}''';
\eqno {\rm (132.)}$$
$${\rm C}_{2,1,0}''
= {\rm C}_{2,1,0}''';\quad
{\rm C}_{2,0,1}''
= q {\rm C}_{2,1,0}''' + {\rm C}_{2,0,1}''';
\eqno {\rm (133.)}$$
$${\rm C}_{1,2,0}''
= {\rm C}_{1,2,0}''';\quad
{\rm C}_{1,1,1}''
= 2q {\rm C}_{1,2,0}''' + {\rm C}_{1,1,1}''';\quad
{\rm C}_{1,0,2}''
= q^2 {\rm C}_{1,2,0}''' + q {\rm C}_{1,1,1}'''
+ {\rm C}_{1,0,2}''';
\eqno {\rm (134.)}$$
and, in virtue of these, the seven conditions (112) and (113)
may be put under the forms,
$${\rm A}_{0,1,0}''' = 0,\quad
{\rm C}_{2,1,0}''' = 0,\quad
{\rm C}_{1,2,0}''' = 0,
\eqno {\rm (135.)}$$
and
$${\rm A}_{0,0,1}''' = 0,\quad
{\rm C}_{2,0,1}''' = 0,\quad
{\rm C}_{1,1,1}''' = 0,\quad
{\rm C}_{1,0,2}''' = 0.
\eqno {\rm (136.)}$$
The three conditions of the group (135) differ only in their
notation from the three conditions (112), and are to be used
exactly like those former conditions, in order to determine the
ratios of $q_0,\ldots \, q_{m-2}$, after the ratios of
$p_0,\ldots \, p_{m-1}$ have been determined, through the help of
the conditions (111); but, in deducing the conditions (136)
from the conditions (113), a real simplification has been
effected (and not merely a change of notation) by suppressing
several terms, such as $q {\rm A}_{0,1,0}'''$, which vanish in
consequence of the conditions (112) or (135). And since we have
thus been led to perceive the existence of a group (136), of
four homogeneous equations (rational and integral) between the
$m - 2$ quantities $r_0, r_1,\ldots \, r_{m-3}$, we see, at last,
that we shall be conducted, in general, to the
{\it case of failure\/} (126), in which all those quantities
vanish, {\it unless their number $m - 2$ be greater than four\/};
that is {\it unless the degree of the proposed equation in $x$ be
at least equal to the minor limit\/} {\sc seven}. It results,
then, from this analysis, that for equations of the {\it sixth\/}
and {\it lower\/} degrees, Mr.~Jerrard's process for effecting
the transformation (20), or of satisfying the three conditions
(8) (14) and (19),
$${\rm A}' = 0,\quad {\rm C}' = 0,\quad
{\rm D}' - \alpha {\rm B}'^2 = 0,$$
will, in general, become {\it illusory}, by conducting to an
useless expression, of the form (93), for the new variable~$y$;
so that it {\it fails}, for example, {\it to reduce the general
equation of the fifth degree,
$$x^5 + {\rm A} x^4 + {\rm B} x^3 + {\rm C} x^2
+ {\rm D} x + {\rm E} = 0,
\eqno {\rm (69.)}\qquad$$
to De Moivre's solvible form},
$$y^5 + {\rm B}' y^3
+ {\textstyle {1 \over 5}} {\rm B}'^2 y + {\rm E}' = 0,
\eqno {\rm (137.)}$$
except, by a useless assumption, of the form
$$y = {\rm L} (x^5 + {\rm A} x^4 + {\rm B} x^3 + {\rm C} x^2
+ {\rm D} x + {\rm E} ),
\eqno {\rm (138.)}$$
which gives, indeed, a very simple transformed equation, namely,
$$y^5 = 0,
\eqno {\rm (139.)}$$
but affords no assistance whatever towards resolving the proposed
equation in $x$. Indeed, for equations of the {\it fifth\/}
degree, the foregoing discussion may be considerably simplified,
by observing, that, in virtue of the eight conditions (112),
(113), (114), the four homogeneous functions (105) (106)
(107) (108), of the $m - 1$ sums
$q_0 + q_0',\ldots \, q_{m-2} + q_{m-2}'$, are all $=0$, and
therefore also (in general) those sums themselves must vanish
(which is the case of failure (102),) when their number $m - 1$
is not greater than four, that is, {\it when the proposed
equation is not higher than the fifth degree}. But the foregoing
discussion (though the great generality of the question has
caused it to be rather long) has the advantage of extending even
to equations of the {\it sixth\/} degree, and of showing that
even for such equations the method generally fails, in such a
manner that it will not in general reduce the equation
$$x^6 + {\rm A} x^5 + {\rm B} x^4 + {\rm C} x^3
+ {\rm D} x^2 + {\rm E} x + {\rm F} = 0
\eqno {\rm (140.)}$$
to the form
$$y^6 + {\rm B}' y^4 + \alpha {\rm B}' y^2
+ {\rm E}' y + {\rm F}' = 0,
\eqno {\rm (141.)}$$
except by the assumption
$$y = {\rm L} ( x^6 + {\rm A} x^5 + {\rm B} x^4 + {\rm C} x^3
+ {\rm D} x^2 + {\rm E} x + {\rm F} );
\eqno {\rm (142.)}$$
which gives, indeed, a very simple result, namely,
$$y^6 = 0,
\eqno {\rm (143.)}$$
but does not at all assist us to resolve the proposed equation
(140). However, this discussion may be regarded as {\it
confirming the adequacy of the method to transform the general
equation of the seventh degree},
$$x^7 + {\rm A} x^6 + {\rm B} x^5 + {\rm C} x^4
+ {\rm D} x^3 + {\rm E} x^2 + {\rm F} x + {\rm G} = 0,
\eqno {\rm (144.)}$$
to another of the form
$$y^7 + {\rm B}' y^5 + \alpha {\rm B}'^2 y^3
+ {\rm E}' y^2 + {\rm F}' y + {\rm G}' = 0,
\eqno {\rm (145.)}$$
without assuming $y = $ any multiple of the proposed evanescent
polynome $x^7 + {\rm A} x^6 + \hbox{\&c.}$; and to effect the
analogous transformation (20), for equations of all {\it higher}
degrees; a curious and unexpected discovery, for which algebra is
indebted to Mr.~Jerrard.
\bigbreak
[9.]
The result obtained by the foregoing discussion may seem, so far
as it respects equations of the {\it sixth} degree to be of very
little importance; because the equation (141), to which it has
been shown that the method fails to reduce the general equation
(140), is not itself, in general, of any known solvible form,
whatever value may be chosen for the arbitrary multiplier
$\alpha$. But it must be observed that if the method had in fact
been adequate to effect that general transformation of the
equation of the sixth degree, without resolving any auxiliary
equation of a higher degree than the fourth, then it would also
have been adequate to reduce the same general equation (140) of
the sixth degree to this other form, which is obviously and
easily solvible,
$$y^6 + {\rm B}' y^4 + {\rm D}' y^2 + {\rm F}' = 0,
\eqno {\rm (146.)}$$
by first assigning an expression of the form
$$y = f(x) = \Lambda''' \phi(x) + {\rm N}^{\rm V} \chi(x),
\eqno {\rm (89.)}\qquad$$
which should satisfy the two conditions
$${\rm A}' = 0,
\eqno {\rm (8.)}\qquad$$
$${\rm C}' = 0,
\eqno {\rm (14.)}\qquad$$
and by then determining the ratio of $\Lambda'''$ to ${\rm
N}^{\rm V}$, so as to satisfy this other condition,
$${\rm E}' = 0,
\eqno {\rm (147.)}$$
which could be done without resolving any auxiliary equation of a
higher degree than the fifth; and this {\it reduction, of the
difficulty of the sixth to that of the fifth degree}, would have
been a very important result, of which it was interesting to
examine the validity. The foregoing discussion, however, appears
to me to prove that {\it this transformation also is illusory\/};
for it shows that, because the degree of the proposed equation is
less than the minor limit~7, the functions $\phi(x)$ and
$\chi(x)$ in (89) are connected by a relation of the form (36);
on which account the expression (89) becomes
$$y = f(x) = (\Lambda''' + a {\rm N}^{\rm V}) \phi(x)
+ \lambda {\rm N}^{\rm V} {\rm X},
\eqno {\rm (90.)}\qquad$$
and the condition
$${\rm E}' = 0,
\eqno {\rm (147.)}\qquad$$
gives, in general,
$$(\Lambda''' + a {\rm N}^{\rm V})^5 = 0,
\eqno {\rm (148.)}$$
that is,
$$\Lambda''' + a {\rm N}^{\rm V} = 0;
\eqno {\rm (92.)}\qquad$$
so that finally the expression for $y$ becomes
$$y = \lambda {\rm N}^{\rm V} {\rm X},
\eqno {\rm (93.)}\qquad$$
that is, it takes in general the evidently useless form,
$$y = {\rm L} ( x^6 + {\rm A} x^5 + {\rm B} x^4
+ {\rm C} x^3 + {\rm D} x^2 + {\rm E} x + {\rm F} ).
\eqno {\rm (142.)}\qquad$$
\bigbreak
[10.]
Mr.~Jerrard has not actually stated, in his published
Researches, the process by which he would effect in general
the transformation (15), so as to {\it take away four terms\/}
at once from the equation of the $m^{\rm th}$ degree, without
resolving any auxiliary equation of a higher degree than the
fourth; but he has sufficiently indicated this process, which
appears to be such as the following. He would probably assume an
expression with {\it twenty-one\/} terms for the new variable,
$$\eqalignno{
y = f(x)
&= \Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm M}' x^{\mu'}
+ {\rm M}'' x^{\mu''}
+ {\rm M}''' x^{\mu'''}
+ {\rm M}^{\rm IV} x^{\mu^{\rm IV}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm N}' x^{\nu'}
+ {\rm N}'' x^{\nu''}
+ {\rm N}''' x^{\nu'''}
+ {\rm N}^{\rm IV} x^{\nu^{\rm IV}}
+ {\rm N}^{\rm V} x^{\nu^{\rm V}}
+ {\rm N}^{\rm VI} x^{\nu^{\rm VI}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ \Xi' x^{\xi'}
+ \Xi'' x^{\xi''}
+ \Xi''' x^{\xi'''}
+ \Xi^{\rm IV} x^{\xi^{\rm IV}}
+ \Xi^{\rm V} x^{\xi^{\rm V}}
+ \Xi^{\rm VI} x^{\xi^{\rm VI}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ \Xi^{\rm VII} x^{\xi^{\rm VII}}
+ \Xi^{\rm VIII} x^{\xi^{\rm VIII}},
&{\rm (149.)}\cr}$$
and would develop or decompose the coefficients ${\rm A}'$,
${\rm B}'$, ${\rm C}'$, of the transformed equation in $y$,
considered as rational and integral, and homogeneous functions of
the twenty-one coefficients,
$$\Lambda,\enspace
\Lambda',\enspace
\Lambda'',
\eqno {\rm (150.)}$$
$${\rm M}',\enspace
{\rm M}'',\enspace
{\rm M}''',\enspace
{\rm M}^{\rm IV},
\eqno {\rm (151.)}$$
$${\rm N}',\enspace
{\rm N}'',\enspace
{\rm N}''',\enspace
{\rm N}^{\rm IV},\enspace
{\rm N}^{\rm V},\enspace
{\rm N}^{\rm VI},
\eqno {\rm (152.)}$$
$$\Xi',\enspace
\Xi'',\enspace
\Xi''',\enspace
\Xi^{\rm IV},\enspace
\Xi^{\rm V},\enspace
\Xi^{\rm VI},\enspace
\Xi^{\rm VII},\enspace
\Xi^{\rm VIII},
\eqno {\rm (153.)}$$
into the following parts:
$${\rm A}' = {\rm A}_{1,0,0,0}' + {\rm A}_{0,1,0,0}'
+ {\rm A}_{0,0,1,0}' + {\rm A}_{0,0,0,1}';
\eqno {\rm (154.)}$$
$$\eqalignno{
{\rm B}' &= {\rm B}_{2,0,0,0}' + {\rm B}_{1,1,0,0}'
+ {\rm B}_{1,0,1,0}' + {\rm B}_{1,0,0,1}' \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm B}_{0,2,0,0}' + {\rm B}_{0,1,1,0}'
+ {\rm B}_{0,1,0,1}' + {\rm B}_{0,0,2,0}'
+ {\rm B}_{0,0,1,1}' + {\rm B}_{0,0,0,2}';
&{\rm (155.)}\cr}$$
$$\eqalignno{
{\rm C}'
&= {\rm C}_{3,0,0,0}' + {\rm C}_{2,1,0,0}'
+ {\rm C}_{2,0,1,0}' + {\rm C}_{2,0,0,1}'
+ {\rm C}_{1,2,0,0}' + {\rm C}_{1,1,1,0}'
+ {\rm C}_{1,1,0,1}' \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm C}_{1,0,2,0}' + {\rm C}_{1,0,1,1}'
+ {\rm C}_{1,0,0,2}' + {\rm C}_{0,3,0,0}'
+ {\rm C}_{0,2,1,0}' + {\rm C}_{0,2,0,1}' \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm C}_{0,1,2,0}' + {\rm C}_{0,1,1,1}'
+ {\rm C}_{0,1,0,2}' + {\rm C}_{0,0,3,0}'
+ {\rm C}_{0,0,2,1}' + {\rm C}_{0,0,1,2}'
+ {\rm C}_{0,0,0,3}';
&{\rm (156.)}\cr}$$
each part ${\rm A}_{h,i,k,l}'$ or ${\rm B}_{h,i,k,l}'$ or
${\rm C}_{h,i,k,l}'$ being itself a rational and integral
function of the twenty-one quantities (150) (151) (152) (153),
and being also homogeneous of the degree~$h$ with respect to the
three quantities (150), of the degree~$i$ with respect to the
four quantities (151), of the degree $k$ with respect to the six
quantities (152), and of the degree~$l$ with respect to the
eight quantities (153). He would then determine the two ratios
of the two first to the last of the three quantities (150) (that
is, the ratios of $\Lambda'$ and $\Lambda''$ to $\Lambda'''$) so
as to satisfy the two conditions
$${\rm A}_{1,0,0,0}' = 0,\quad
{\rm B}_{2,0,0,0}' = 0;
\eqno {\rm (157.)}$$
the three ratios of the first three to the last of the four
quantities (151), so as to satisfy the three conditions
$${\rm A}_{0,1,0,0}' = 0,\quad
{\rm B}_{1,1,0,0}' = 0,\quad
{\rm B}_{0,2,0,0}' = 0;
\eqno {\rm (158.)}$$
the ratio of the last of the quantities (150) to the last of the
quantities (151), so as to satisfy the condition
$${\rm C}_{3,0,0,0}' + {\rm C}_{2,1,0,0}'
+ {\rm C}_{1,2,0,0}' +{\rm C}_{0,3,0,0}'
= 0;
\eqno {\rm (159.)}$$
the five ratios of the five first to the last of the six
quantities (152), so as to satisfy the five conditions,
$$\left. \eqalign{
& {\rm A}_{0,0,1,0}' = 0,\cr
& {\rm B}_{1,0,1,0}' + {\rm B}_{0,1,1,0}' = 0,\cr
& {\rm B}_{0,0,2,0}' = 0,\cr
& {\rm C}_{2,0,1,0}' + {\rm C}_{1,1,1,0}'
+ {\rm C}_{0,2,1,0}' = 0,\cr
& {\rm C}_{1,0,2,0}' + {\rm C}_{0,1,2,0}' = 0;\cr}
\right\}
\eqno {\rm (160.)}$$
the seven ratios of the seven first to the last of the eight
quantities (153), so as to satisfy the seven conditions
$$\left. \eqalign{
& {\rm A}_{0,0,0,1}' = 0,\cr
& {\rm B}_{1,0,0,1}' + {\rm B}_{0,1,0,1}' = 0,\cr
& {\rm B}_{0,0,1,1}' = 0,\quad
{\rm B}_{0,0,0,2}' = 0,\cr
& {\rm C}_{2,0,0,1}' + {\rm C}_{1,1,0,1}'
+ {\rm C}_{0,2,0,1}' = 0,\cr
& {\rm C}_{1,0,1,1}' + {\rm C}_{0,1,1,1}' = 0,\cr
& {\rm C}_{1,0,0,2}' + {\rm C}_{0,1,0,2}' = 0;\cr}
\right\}
\eqno {\rm (161.)}$$
and the ratio of the last of the quantities (152) to the last of
the quantities (153), so as to satisfy the condition
$${\rm C}_{0,0,3,0}' + {\rm C}_{0,0,2,1}'
+ {\rm C}_{0,0,1,2}' + {\rm C}_{0,0,0,3}'
= 0:
\eqno {\rm (162.)}$$
all which determinations could in general be successively
effected, without its being necessary to resolve any equation of
a higher degree than the fourth. The first part of the process
would now be completed; that is, the assumed expression (149)
for $y$ would be reduced to the form
$$y = f(x) = {\rm M}^{\rm IV} \phi(x) + \Xi^{\rm VIII} \chi(x),
\eqno {\rm (163.)}$$
the functions $\phi(x)$ and $\chi(x)$ being determined and known,
but the multipliers ${\rm M}^{\rm IV}$ and $\Xi^{\rm VIII}$ being
arbitrary, and this expression (163) being such as to satisfy
the three conditions (8), (10) and (14),
$${\rm A}' = 0,\quad {\rm B}' = 0,\quad {\rm C}' = 0;$$
nineteen out of the twenty ratios of the twenty-one coefficients
(150) (151) (152) (153) having been determined so as to
satisfy the nineteen equations (157) (158) (159) (160) (161)
(162), into which those three conditions had been decomposed.
And the second and only remaining part of the process would
consist in then determining the remaining ratio of
${\rm M}^{\rm IV}$ to $\Xi^{\rm VIII}$, so as to satisfy the
remaining condition
$${\rm D}' = 0,
\eqno {\rm (17.)}\qquad$$
and thereby to reduce the general equation of the $m^{\rm th}$
degree,
$${\rm X}
= x^m + {\rm A} x^{m-1} + {\rm B} x^{m-2}
+ {\rm C} x^{m-3} + {\rm D} x^{m-4}
+ {\rm E} x^{m-5} + \hbox{\&c.} = 0,
\eqno {\rm (1.)}\qquad$$
to the form
$${\rm Y} = y^m + {\rm E}' y^{m-5} + \hbox{\&c.} = 0.
\eqno {\rm (15.)}\qquad$$
It is possible, of course, that this may not be precisely the
same as Mr.~Jerrard's unpublished process, but it seems likely
that the one would not be found to differ from the other in any
essential respect, notation being always excepted. It is, at
least, a process suggested by the published researches of that
author, and harmonizing with the discoveries which they contain.
But by applying to this new process the spirit of the former
discussions, and putting for abbreviation,
$$\left. \eqalign{
\Lambda' s_0^{(\lambda')}
+ \Lambda'' s_0^{(\lambda'')}
+ \Lambda''' s_0^{(\lambda''')}
+ {\rm M}' s_0^{(\mu')}
+ \cdots
+ {\rm M}^{\rm IV} s_0^{(\mu^{\rm IV})}
&= p_0,\cr
\noalign{\hbox{$\cdots\cdots$}}
\Lambda' s_{m-1}^{(\lambda')}
+ \Lambda'' s_{m-1}^{(\lambda'')}
+ \Lambda''' s_{m-1}^{(\lambda''')}
+ {\rm M}' s_{m-1}^{(\mu')}
+ \cdots
+ {\rm M}^{\rm IV} s_{m-1}^{(\mu^{\rm IV})}
&= p_{m-1},\cr}
\right\}
\eqno {\rm (164.)}$$
$$\left. \eqalign{
{\rm N}' s_0^{(\nu')} + \cdots
+ {\rm N}^{\rm VI} s_0^{(\nu^{\rm VI})}
&= p_0',\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm N}' s_{m-1}^{(\nu')} + \cdots
+ {\rm N}^{\rm VI} s_{m-1}^{(\nu^{\rm VI})}
&= p_{m-1}',\cr}
\right\}
\eqno {\rm (165.)}$$
$$\left. \eqalign{
\Xi' s_0^{(\xi')} + \cdots
+ \Xi^{\rm VIII} s_0^{(\xi^{\rm VIII})}
&= p_0'',\cr
\noalign{\hbox{$\cdots\cdots$}}
\Xi' s_{m-1}^{(\xi')} + \cdots
+ \Xi^{\rm VIII} s_{m-1}^{(\xi^{\rm VIII})}
&= p_{m-1}'',\cr}
\right\}
\eqno {\rm (166.)}$$
$$\eqalignno{
\Lambda' {\rm L}^{(\lambda')}
+ \Lambda'' {\rm L}^{(\lambda'')}
+ \Lambda''' {\rm L}^{(\lambda''')}
+ {\rm M}' {\rm L}^{(\mu')}
+ \cdots
+ {\rm M}^{\rm IV} {\rm L}^{(\mu^{\rm IV})} \cr
\mathord{}
+ {\rm N}' {\rm L}^{(\nu')}
+ \cdots
+ {\rm N}^{\rm VI} {\rm L}^{(\nu^{\rm VI})}
+ \Xi' {\rm L}^{(\xi')}
+ \cdots
+ \Xi^{\rm VIII} {\rm L}^{(\xi^{\rm VIII})}
&= {\rm L},
&{\rm (167.)}\cr}$$
we may change the expression (149) to the form (100), through
the theorem and notation (44); and in order to avoid the case of
failure, in which the functions $\phi(x)$ and $\chi(x)$ in (163)
are connected by a relation of the form (36), we must avoid, as
in the discussion given in the seventh article, the case where
the $m$ sums
$p_0' + p_0'',\ldots \, p_{m-1}' + p_{m-1}''$
are proportional to the $m$ quantities
$p_0,\ldots \, p_{m-1}$, that is, the case
$$q_0 + q_0' = 0,\quad \ldots \quad
q_{m-2} + q_{m-2}' = 0,
\eqno {\rm (102.)}\qquad$$
if we adopt the definitions (54) (55) and (96) (97), so as to
introduce the symbols $p$, $q_0$, $q_1,\ldots$ $q_{m-2}$, and
$p'$, $q_0'$, $q_1',\ldots$ $q_{m-2}'$. With these additional
symbols it is easy to transform the conditions (160) into
others, which (when suitably combined with the equations of
definition, and with the ratios of $p_0,\ldots \, p_{m-1}$
already previously determined through the help of the conditions
(157) (158) (159),) shall serve to determine the ratios (121)
of $q_0,\ldots \, q_{m-2}$; and then to determine, in like
manner, with the help of the conditions (161), the ratios (120)
of $q_0',\ldots \, q_{m-2}'$; after which, the condition (162)
may be transformed into a rational and integral and homogeneous
equation of the third degree between the sums
$q_0 + q_0',\ldots, q_{m-2} + q_{m-2}'$, and will in general
oblige those sums to vanish, if their ratios (122) have already
been determined independently of the condition (162), which will
happen when the ratios (120) coincide with the ratios (121),
that is, when the quantities $q_0',\ldots,q_{m-2}'$ are
proportional to the quantities $q_0,\ldots, q_{m-2}$. We must,
therefore, in general avoid this last proportionality, in order
to avoid the case of failure (102); and thus we are led to
introduce the symbols $q, r_0, r_1,\ldots \, r_{m-3}$, defined by
the equations (124) (125), and to express the case of failure by
the equations
$$r_0 = 0,\quad r_1 = 0,\quad \cdots \quad r_{m-3} = 0.
\eqno {\rm (126.)}\qquad$$
With these new symbols we easily discover that the seven
conditions (161) may be reduced to seven rational and integral
and homogeneous equations between the quantities
$r_0, r_1,\ldots \, r_{m-3}$, which will in general oblige them
all to vanish, and therefore will produce the case of failure
(126) {\it unless the number $m - 2$ of these quantities be
greater than the number seven}, that is, {\it unless the
exponent~$m$ of the degree of the proposed equation be at least
equal to the minor limit\/} {\sc ten}. It results, then, from
this discussion, that the process described in the present
article {\it will not in general avail to take away four terms at
once, from equations lower than the\/} {\sc tenth} {\it degree},
and, of course, that {\it it will not reduce the general equation
of the fifth degree},
$$x^5 + {\rm A} x^4 + {\rm B} x^3 + {\rm C} x^2
+ {\rm D} x + {\rm E} = 0,
\eqno {\rm (69.)}\qquad$$
{\it to the binomial form}
$$y^5 + {\rm E}' = 0,
\eqno {\rm (168.)}$$
except by the useless assumption
$$y = {\rm L} (x^5 + {\rm A} x^4 + {\rm B} x^3 + {\rm C} x^2
+ {\rm D} x + {\rm E} ),
\eqno {\rm (138.)}\qquad$$
which gives
$$y^5 = 0.
\eqno {\rm (139.)}\qquad$$
\bigbreak
[11.]
A principal feature of Mr.~Jerrard's general method is to
{\it avoid}, as much as possible, {\it the raising of the degree
in elimination\/}; and for that purpose to {\it decompose the
equations of condition\/} in every question {\it into groups},
which shall {\it each\/} contain, if possible, {\it not more than
one equation of a higher degree than the first\/}; although the
occurrence of {\it two equations of the second degree\/} in one
group is {\it not fatal\/} to the success of the method, because
the final equation of such a group being {\it only elevated to
the fourth degree}, can be resolved by the known rules. It
might, therefore, have been more completely in the spirit of this
general method, because it would have more completely avoided the
elevation of degree by elimination, if, in order to take away
four terms at once from the general equation of the $m^{\rm th}$
degree, we had assumed an expression with {\it thirty-three\/}
terms, of the form
$$\eqalignno{
y = f(x)
&= \Lambda' x^{\lambda'}
+ \Lambda'' x^{\lambda''}
+ \Lambda''' x^{\lambda'''} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm M}' x^{\mu'}
+ \cdots
+ {\rm M}^{\rm IV} x^{\mu^{\rm IV}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm N}' x^{\nu'}
+ \cdots
+ {\rm N}^{\rm V} x^{\nu^{\rm V}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ \Xi' x^{\xi'}
+ \cdots
+ \Xi^{\rm VI} x^{\xi^{\rm VI}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {\rm O}' x^{o'}
+ \cdots
+ {\rm O}^{\rm VII} x^{o^{\rm VII}} \cr
&\mathrel{\phantom{=}} \mathord{}
+ \Pi' x^{\varpi'}
+ \cdots
+ \Pi^{\rm VIII} x^{\varpi^{\rm VIII}};
&{\rm (169.)}\cr}$$
and had determined the six ratios of
$\Lambda', \Lambda'', \Lambda''',
{\rm M}',\ldots \, {\rm M}^{\rm IV}$,
and the twenty-five ratios of
${\rm N}',\ldots, \Pi^{\rm VIII}$, so as to satisfy the
thirty-one conditions
$${\rm A}_{1,0,0,0,0,0}' = 0,\quad
{\rm B}_{2,0,0,0,0,0}' = 0,
\eqno {\rm (170.)}$$
$${\rm A}_{0,1,0,0,0,0}' = 0,\quad
{\rm B}_{1,1,0,0,0,0}' = 0,\quad
{\rm B}_{0,2,0,0,0,0}' = 0,
\eqno {\rm (171.)}$$
$${\rm C}_{3,0,0,0,0,0}'
+ {\rm C}_{2,1,0,0,0,0}'
+ {\rm C}_{1,2,0,0,0,0}'
+ {\rm C}_{0,3,0,0,0,0}'
= 0,
\eqno {\rm (172.)}$$
$$\left. \eqalign{
& {\rm A}_{0,0,1,0,0,0}' = 0,\cr
& {\rm B}_{1,0,1,0,0,0}' + {\rm B}_{0,1,1,0,0,0}' = 0,\cr
& {\rm B}_{0,0,2,0,0,0}' = 0,\cr
& {\rm C}_{2,0,1,0,0,0}' + {\rm C}_{1,1,1,0,0,0}'
+ {\rm C}_{0,2,1,0,0,0}'
= 0,\cr}
\right\}
\eqno {\rm (173.)}$$
$$\left. \eqalign{
& {\rm A}_{0,0,0,1,0,0}' = 0,\cr
& {\rm B}_{1,0,0,1,0,0}' + {\rm B}_{0,1,0,1,0,0}' = 0,\cr
& {\rm B}_{0,0,1,1,0,0}' = 0,\cr
& {\rm B}_{0,0,0,2,0,0}' = 0,\cr
& {\rm C}_{2,0,0,1,0,0}' + {\rm C}_{1,1,0,1,0,0}'
+ {\rm C}_{0,2,0,1,0,0}'
= 0,\cr}
\right\}
\eqno {\rm (174.)}$$
$${\rm C}_{1,0,2,0,0,0}'
+ {\rm C}_{1,0,1,1,0,0}'
+ {\rm C}_{1,0,0,2,0,0}'
+ {\rm C}_{0,1,2,0,0,0}'
+ {\rm C}_{0,1,1,1,0,0}'
+ {\rm C}_{0,1,0,2,0,0}'
= 0,
\eqno {\rm (175.)}$$
$$\left. \eqalign{
& {\rm A}_{0,0,0,0,1,0}' = 0,\cr
& {\rm B}_{1,0,0,0,1,0}' + {\rm B}_{0,1,0,0,1,0}' = 0,\cr
& {\rm B}_{0,0,1,0,1,0}' + {\rm B}_{0,0,0,1,1,0}' = 0,\cr
& {\rm B}_{0,0,0,0,2,0}' = 0,\cr
& {\rm C}_{2,0,0,0,1,0}' + {\rm C}_{1,1,0,0,1,0}'
+ {\rm C}_{0,2,0,0,1,0}'
= 0,\cr
& {\rm C}_{1,0,1,0,1,0}' + {\rm C}_{1,0,0,1,1,0}'
+ {\rm C}_{0,1,1,0,1,0}' + {\rm C}_{0,1,0,1,1,0}'
= 0,\cr}
\right\}
\eqno {\rm (176.)}$$
$$\left. \eqalign{
& {\rm A}_{0,0,0,0,0,1}' = 0,\cr
& {\rm B}_{1,0,0,0,0,1}' + {\rm B}_{0,1,0,0,0,1}' = 0,\cr
& {\rm B}_{0,0,1,0,0,1}' + {\rm B}_{0,0,0,1,0,1}' = 0,\cr
& {\rm B}_{0,0,0,0,1,1}' = 0,\cr
& {\rm B}_{0,0,0,0,0,2}' = 0,\cr
& {\rm C}_{2,0,0,0,0,1}' + {\rm C}_{1,1,0,0,0,1}'
+ {\rm C}_{0,2,0,0,0,1}'
= 0,\cr
& {\rm C}_{1,0,1,0,0,1}' + {\rm C}_{1,0,0,1,0,1}'
+ {\rm C}_{0,1,1,0,0,1}' + {\rm C}_{0,1,0,1,0,1}'
= 0,\cr}
\right\}
\eqno {\rm (177.)}$$
$${\rm C}_{1,0,0,0,2,0}'
+ {\rm C}_{1,0,0,0,1,1}'
+ {\rm C}_{1,0,0,0,0,2}'
+ {\rm C}_{0,1,0,0,2,0}'
+ {\rm C}_{0,1,0,0,1,1}'
+ {\rm C}_{0,1,0,0,0,2}'
= 0,
\eqno {\rm (178.)}$$
$$\eqalignno{
& {\rm C}_{0,0,3,0,0,0}'
+ {\rm C}_{0,0,2,1,0,0}'
+ {\rm C}_{0,0,2,0,1,0}'
+ {\rm C}_{0,0,2,0,0,1}'
+ {\rm C}_{0,0,1,2,0,0}'
+ {\rm C}_{0,0,1,1,1,0}' \cr
& + {\rm C}_{0,0,1,1,0,1}'
+ {\rm C}_{0,0,1,0,2,0}'
+ {\rm C}_{0,0,1,0,1,1}'
+ {\rm C}_{0,0,1,0,0,2}'
+ {\rm C}_{0,0,0,3,0,0}'
+ {\rm C}_{0,0,0,2,1,0}' \cr
& + {\rm C}_{0,0,0,2,0,1}'
+ {\rm C}_{0,0,0,1,2,0}'
+ {\rm C}_{0,0,0,1,1,1}'
+ {\rm C}_{0,0,0,1,0,2}'
+ {\rm C}_{0,0,0,0,3,0}'
+ {\rm C}_{0,0,0,0,2,1}' \cr
& + {\rm C}_{0,0,0,0,1,2}'
+ {\rm C}_{0,0,0,0,0,3}'
= 0,
&{\rm (179.)}\cr}$$
into which the three conditions
$${\rm A}' = 0,\quad {\rm B}' = 0,\quad {\rm C}' = 0,$$
may be decomposed; the symbol ${\rm A}_{f,g,h,i,k,l}'$, or
${\rm B}_{f,g,h,i,k,l}'$, or ${\rm C}_{f,g,h,i,k,l}'$ denoting
here a rational and integral function of the thirty-three
coefficients $\Lambda',\ldots \, \Pi^{\rm VIII}$, which is
homogeneous of the degree~$f$ with respect to $\Lambda'$,
$\Lambda''$, $\Lambda'''$, of the degree~$g$ with respect to
${\rm M}',\ldots, {\rm M}^{\rm IV}$,
of the degree~$h$ with respect to
${\rm N}',\ldots \, {\rm N}^{\rm V}$,
of the degree~$i$ with respect to $\Xi',\ldots \, \Xi^{\rm VI}$,
of the degree~$k$ with respect to
${\rm O}',\ldots \, {\rm O}^{\rm VII}$,
and of the degree~$l$ with respect to
$\Pi',\ldots \, \Pi^{\rm VIII}$:
while the remaining ratio of ${\rm M}^{\rm IV}$ to
$\Pi^{\rm VIII}$, should afterwards be chosen so as to satisfy
the remaining condition
$${\rm D}' = 0.$$
But, on putting, for abridgment,
$$\left. \eqalign{
{\rm N}' s_0^{(\nu')} + \cdots
+ {\rm N}^{\rm V} s_0^{(\nu^{\rm V})}
+ \Xi' s_0^{(\xi')} + \cdots
+ \Xi^{\rm VI} s_0^{(\xi^{\rm VI})}
&= p_0',\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm N}' s_{m-1}^{(\nu')} + \cdots
+ {\rm N}^{\rm V} s_{m-1}^{(\nu^{\rm V})}
+ \Xi' s_{m-1}^{(\xi')} + \cdots
+ \Xi^{\rm VI} s_{m-1}^{(\xi^{\rm VI})}
&= p_{m-1}',\cr}
\right\}
\eqno {\rm (180.)}$$
$$\left. \eqalign{
{\rm O}' s_0^{(o')} + \cdots
+ {\rm O}^{\rm VII} s_0^{(o^{\rm VII})}
&= p_0'',\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm O}' s_{m-1}^{(o')} + \cdots
+ {\rm O}^{\rm VII} s_{m-1}^{(o^{\rm VII})}
&= p_{m-1}'',\cr}
\right\}
\eqno {\rm (181.)}$$
$$\left. \eqalign{
\Pi' s_0^{(\varpi')} + \cdots
+ \Pi^{\rm VIII} s_0^{(\varpi^{\rm VIII})}
&= p_0''',\cr
\noalign{\hbox{$\cdots\cdots$}}
\Pi' s_{m-1}^{(\varpi')} + \cdots
+ \Pi^{\rm VIII} s_{m-1}^{(\varpi^{\rm VIII})}
&= p_{m-1}''',\cr}
\right\}
\eqno {\rm (182.)}$$
$$\eqalignno{
\Lambda' {\rm L}^{(\lambda')}
+ \Lambda'' {\rm L}^{(\lambda'')}
+ \Lambda''' {\rm L}^{(\lambda''')}
+ {\rm M}' {\rm L}^{(\mu')}
+ \cdots
+ {\rm M}^{\rm IV} {\rm L}^{(\mu^{\rm IV})} \cr
+ {\rm N}' {\rm L}^{(\nu')}
+ \cdots
+ {\rm N}^{\rm V} {\rm L}^{(\nu^{\rm V})}
+ \Xi' {\rm L}^{(\xi')}
+ \cdots
+ \Xi^{\rm VI} {\rm L}^{(\xi^{\rm VI})} \cr
+ {\rm O}' {\rm L}^{(o')}
+ \cdots
+ {\rm O}^{\rm VII} {\rm L}^{(o^{\rm VII})}
+ \Pi' {\rm L}^{(\varpi')}
+ \cdots
+ \Pi^{\rm VIII} {\rm L}^{(\varpi^{\rm VIII})}
&= {\rm L},
&{\rm (183.)}\cr}$$
$${p_{m-1}''' \over p_{m-1}} = p'',
\eqno {\rm (184.)}$$
$$p_0''' - p'' p_0 = q_0'',\quad \ldots \quad
p_{m-2}''' - p'' p_{m-2} = q_{m-2}'',
\eqno {\rm (185.)}$$
$${q_{m-2}'' \over q_{m-2}} = q',
\eqno {\rm (186.)}$$
$$q_0'' - q' q_0 = r_0',\quad \ldots \quad
q_{m-3}'' - q' q_{m-3} = r_{m-3}',
\eqno {\rm (187.)}$$
$${r_{m-3}' \over r_{m-3}} = r,
\eqno {\rm (188.)}$$
$$r_0' - r r_0 = t_0,\quad \ldots \quad
r_{m-4}' - r r_{m-4} = t_{m-4},
\eqno {\rm (189.)}$$
and retaining the analogous expressions (164) (54) (55) (96)
(97) (124) (125), we find, by a reasoning exactly analogous to
that employed in the former discussions, that the final
expression for $y$ will in general be of the useless form
$$y = {\rm L} {\rm X},
\eqno {\rm (190.)}$$
in the following case of failure,
$$t_0 = 0,\quad t_1 = 0,\quad \ldots \quad t_{m-4} = 0;
\eqno {\rm (191.)}$$
and on the other hand that the seven conditions (177) may be
reduced to the form of seven rational and integral and
homogeneous equations between these $m - 3$ quantities
$t_0, t_1,\ldots, t_{m-4}$; so that the case of failure will in
general occur in the employment of the {\it present\/} process,
{\it unless the number $m - 3$ be greater than seven}, that is,
{\it unless the degree~$m$ of the proposed equation in $x$ be at
least equal to the minor limit eleven}.
It must, however, be remembered that the less complex process
described in the foregoing article, (since it contained no
condition, nor group of conditions, in which the dimension, or
the product of the dimensions, exceeded the number four,) agreed
sufficiently with the spirit of Mr.~Jerrard's general method; and
was adequate to take away four terms at once from the general
equation of the {\it tenth}, or of any higher degree.
\bigbreak
[12.]
The various processes described in the 2nd, 5th, 6th and 11th
articles of the communication, for {\it transforming the general
equation of the $m^{\rm th}$ degree}, by satisfying certain
systems of equations of condition, are connected with the
solution of this far more general problem proposed by
Mr.~Jerrard, ``to discover $m - 1$ ratios of $m$ disposable
quantities,
$$a_1, a_2,\ldots\, a_m,
\eqno {\rm (192.)}$$
which shall satisfy a given system of $h_1$ rational and integral
and homogeneous equations of the first degree,
$${\rm A}' = 0,\quad {\rm A}'' = 0,\quad \ldots \quad
{\rm A}^{(h_1)} = 0;
\eqno {\rm (193.)}$$
$h_2$ such equations of the second degree,
$${\rm B}' = 0,\quad {\rm B}'' = 0,\quad \ldots \quad
{\rm B}^{(h_2)} = 0;
\eqno {\rm (194.)}$$
$h_3$ of the third degree,
$${\rm C}' = 0,\quad {\rm C}'' = 0,\quad \ldots \quad
{\rm C}^{(h_3)} = 0;
\eqno {\rm (195.)}$$
and so on, as far as $h_t$ equations of the $t^{\rm th}$ degree
$${\rm T}' = 0,\quad {\rm T}'' = 0,\quad \ldots \quad
{\rm T}^{(h_t)} = 0,
\eqno {\rm (196.)}$$
{\it without being obliged in any part of the process, to
introduce any elevation of degree by elimination}.'' Mr.~Jerrard
has not published his solution of this very general problem, but
he has sufficiently suggested the method which he would employ,
and it is proper to discuss it briefly here, with reference to
the extent of its application, and the circumstances under which
it fails; not only on account of the importance of such
discussion in itself, but also because it is adapted to throw
light on all the questions already considered.
If we assume
$$a_1 = a_1' + a_1'',\quad
a_2 = a_2' + a_2'',\quad \ldots \quad
a_m = a_m' + a_m'',
\eqno {\rm (197.)}$$
that is, if we decompose each of the $m$ disposable quantities
$a_1, a_2,\ldots \, a_m$ into two parts, we may then accordingly
decompose every one of the $h_1$ proposed homogeneous functions
of those $m$ quantities, which are of the first degree, namely
$${\rm A}', {\rm A}'',\ldots\, {\rm A}^{(\alpha)},\ldots \,
{\rm A}^{(h_1)};
\eqno {\rm (198.)}$$
every one of the $h_2$ proposed functions of the second degree,
$${\rm B}', {\rm B}'',\ldots\, {\rm B}^{(\beta)},\ldots \,
{\rm B}^{(h_2)};
\eqno {\rm (199.)}$$
every one of the $h_3$ functions of the third degree,
$${\rm C}', {\rm C}'',\ldots\, {\rm C}^{(\gamma)},\ldots \,
{\rm C}^{(h_3)};
\eqno {\rm (200.)}$$
and so on, as far as all the first $h_t - 1$ functions of the
$t^{\rm th}$ degree,
$${\rm T}', {\rm T}'',\ldots\, {\rm T}^{(\tau)},\ldots \,
{\rm T}^{(h_t - 1)}
\eqno {\rm (201.)}$$
(the last function ${\rm T}^{(h_t)}$ being reserved for another
purpose, which will be presently explained,) into other
homogeneous functions, according to the general types,
$$\left. \eqalign{
{\rm A}^{(\alpha)}
&= {\rm A}_{1,0}^{(\alpha)} + {\rm A}_{0,1}^{(\alpha)}, \cr
{\rm B}^{(\beta)}
&= {\rm B}_{2,0}^{(\beta)} + {\rm B}_{1,1}^{(\beta)}
+ {\rm B}_{0,2}^{(\beta)}, \cr
{\rm C}^{(\gamma)}
&= {\rm C}_{3,0}^{(\gamma)} + {\rm C}_{2,1}^{(\gamma)}
+ {\rm C}_{1,2}^{(\gamma)} + {\rm C}_{0,3}^{(\gamma)}, \cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm T}^{(\tau)}
&= {\rm T}_{t,0}^{(\tau)} + {\rm T}_{t-1,1}^{(\tau)}
+ \cdots + {\rm T}_{0,t}^{(\tau)}: \cr}
\right\}
\eqno {\rm (202.)}$$
each symbol of the class
$${\rm A}_{p,q}^{(\alpha)},\enspace
{\rm B}_{p,q}^{(\beta)},\enspace
{\rm C}_{p,q}^{(\gamma)},\ldots \enspace
{\rm T}_{p,q}^{(\tau)},
\eqno {\rm (203.)}$$
denoting a rational and integral and homogeneous function of the
$2m$ new quantities,
$$a_1', a_2',\ldots\, a_m',
\eqno {\rm (204.)}$$
and
$$a_1'', a_2'',\ldots\, a_m'',
\eqno {\rm (205.)}$$
which function is homogeneous of the degree $p$ with respect to
the quantities (204), and of the degree $q$ with respect to the
quantities (205). By this decomposition, we may substitute,
instead of the problem first proposed, the system of the three
following auxiliary problems. First to satisfy, by ratios of the
$m$ quantities (204), an auxiliary system of equations,
containing $h_1$ equations of the first degree, namely,
$${\rm A}_{1,0}' = 0,\quad
{\rm A}_{1,0}'' = 0,\quad
{\rm A}_{1,0}^{(h_1)} = 0;
\eqno {\rm (206.)}$$
$h_2$ equations of the second degree,
$${\rm B}_{2,0}' = 0,\quad
{\rm B}_{2,0}'' = 0,\quad
{\rm B}_{2,0}^{(h_2)} = 0;
\eqno {\rm (207.)}$$
$h_3$ of the third degree,
$${\rm C}_{3,0}' = 0,\quad
{\rm C}_{3,0}'' = 0,\quad
{\rm C}_{3,0}^{(h_3)} = 0;
\eqno {\rm (208.)}$$
and so on, as far as the following $h_t - 1$ equations of the
$t^{\rm th}$ degree,
$${\rm T}_{t,0}' = 0,\quad
{\rm T}_{t,0}'' = 0,\quad
{\rm T}_{t,0}^{(h_t - 1)} = 0.
\eqno {\rm (209.)}$$
Second, to satisfy, by ratios of the $m$ quantities (205), a
system containing
$h_1 + h_2 + h_3 + \cdots + h_t - 1$
equations, which are of the first degree with respect to those
$m$ quantities, and are of the forms
$${\rm A}_{0,1}^{(\alpha)} = 0,\quad
{\rm B}_{1,1}^{(\beta)} = 0,\quad
{\rm C}_{2,1}^{(\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t-1,1}^{(\tau)} = 0;
\eqno {\rm (210.)}$$
$h_2 + h_3 + \cdots + h_t - 1$ equations of the second degree,
and of the forms
$${\rm B}_{0,2}^{(\beta)} = 0,\quad
{\rm C}_{1,2}^{(\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t-2,2}^{(\tau)} = 0;
\eqno {\rm (211.)}$$
$h_3 + \cdots + h_t - 1$ equations of the third degree,
and of the forms,
$${\rm C}_{0,3}^{(\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t-3,3}^{(\tau)} = 0;
\eqno {\rm (212.)}$$
and so on, as far as $h_t - 1$ equations of the $t^{\rm th}$
degree, namely,
$${\rm T}_{0,t}' = 0,\quad
{\rm T}_{0,t}'' = 0,\quad \ldots \quad
{\rm T}_{0,t}^{(h_t - 1)} = 0.
\eqno {\rm (213.)}$$
And third, to satisfy, by the ratio of any one of the $m$
quantites (205) to any one of the $m$ quantities (204), this
one remaining equation of the $t^{\rm th}$ degree,
$${\rm T}^{(h_t)} = 0.
\eqno {\rm (214.)}$$
For if we can resolve all these three auxiliary problems, we
shall thereby have resolved the original problem also. And there
is this advantage in thus transforming the question, that whereas
there were $h_t$ equations of the highest (that is of the
$t^{\rm th}$) degree, in the problem originally proposed, there
are only $h_t - 1$ equations of that highest degree, in each of
the two first auxiliary problems, and only one such equation in
the third. If, then, we apply the same process of transformation
to each of the two first auxiliary problems, and repeat it
sufficiently often, we shall get rid of all the equations of the
$t^{\rm th}$ degree, and ultimately of all equations of degrees
higher than the first; with the exception of certain equations,
which are at various stages of the process set aside to be
separately and singly resolved, without any such combination with
others as could introduce an elevation of degree by elimination.
And thus, at last, the original problem may doubtless be
resolved, {\it provided that the number $m$, of quantities
originally disposable, be large enough}.
\bigbreak
[13.]
But that some such condition respecting the magnitude of that
number $m$ is necessary, will easily appear, if we observe that
when $m$ is not large enough to satisfy the inequality,
$$m > h_1 + h_2 + h_3 + \cdots + h_t,
\eqno {\rm (215.)}$$
then the original $h_1 + h_2 + h_3 + \cdots + h_t$ equations,
being rational and integral and homogeneous with respect to the
original $m$ quantities (192), will in general conduct to null
values for all those quantities, that is, to the expressions
$$a_1 = 0,\quad a_2 = 0,\quad \ldots \quad a_m = 0,
\eqno {\rm (216.)}$$
and therefore to a result which we designed to exclude; because
by the enunciation of the original problem it was by the $m - 1$
{\it ratios\/} of those $m$ quantities that we were to satisfy,
if possible, the equations originally proposed. The same
excluded case, or case of failure (216), will in general occur
when the solution of the second auxiliary problem gives ratios
for the $m$ auxiliary quantities (205), which coincide with the
ratios already found in resolving the first auxiliary problem for
the $m$ other auxiliary quantities (204); that is, when the two
first problems conduct to expressions of the forms
$$a_1'' = a a_1',\quad
a_2'' = a a_2',\quad \ldots \quad
a_m'' = a a_m',
\eqno {\rm (217.)}$$
$a$ being any common multiplier; for then these two first
problems conduct, in virtue of the definitions (197), to a
determined set of ratios for the $m$ original quantities (192),
namely,
$${a_1 \over a_m} = {a_1' \over a_m'},\quad \ldots \quad
{a_{m-1} \over a_m} = {a_{m-1}' \over a_m'};
\eqno {\rm (218.)}$$
and unless these ratios happen to satisfy the equation of the
third problem (214), which had not been employed in determining
them, that last homogeneous equation (214) will oblige all those
$m$ quantities (192) to vanish, and so will conduct to the case
of failure (216). Now, although, when the conditions (215) is
satisfied, the first auxiliary problem becomes indeterminate,
because
$$m - 1 > h_1 + h_2 + h_3 + \cdots + h_t - 1,$$
so that the number $m - 1$ of the disposable ratios of the $m$
auxiliary quantities (204) is greater than the number of the
homogeneous equations which those $m$ quantities are to satisfy,
yet whatever system of $m - 1$ such ratios
$${a_1' \over a_m'},\quad
{a_2' \over a_m'},\quad \ldots \quad
{a_{m-1}' \over a_m'},
\eqno {\rm (219.)}$$
we may discover and employ, so as to satisfy the equations of the
first auxiliary problem, it will always be possible to satisfy
the equations of the second auxiliary problem also, by employing
the same system of $m - 1$ ratios for the $m$ other auxiliary
quantities (205), that is, by employing expressions for those
quantities of the forms (217); and reciprocally, it will in
general be impossible to resolve the second auxiliary problem
otherwise, unless the number of its equations be less than $m -
1$. For if we put, for abridgment,
$${a_m'' \over a_m'} = a,
\eqno {\rm (220.)}$$
and
$$a_1'' - a a_1' = b_1,\quad
a_2'' - a a_2' = b_2,\quad \ldots \quad
a_{m-1}'' - a a_{m-1}' = b_{m-1},
\eqno {\rm (221.)}$$
we shall have, as a general system of expressions for the $m$
quantities (205), the following,
$$a_1'' = a a_1' + b_1,\quad
a_2'' = a a_2' + b_2,\quad \ldots \quad
a_{m-1}'' = a a_{m-1}' + b_{m-1},\quad
a_m'' = a a_m';
\eqno {\rm (222.)}$$
and therefore by (197),
$$a_1 = (1 + a) a_1' + b_1,\quad \ldots \quad
a_{m-1} = (1 + a) a_{m-1}' + b_{m-1},\quad
a_m = (1 + a) a_m';
\eqno {\rm (223.)}$$
so that the homogeneous functions
${\rm A}^{(\alpha)}, {\rm B}^{(\beta)},\ldots \, {\rm
T}^{(\tau)}$ may be, in general, decomposed in this new way,
$$\left. \eqalign{
{\rm A}^{(\alpha)}
&= (1 + a) {\rm A}_{1,0}^{\backprime (\alpha)}
+ {\rm A}_{0,1}^{\backprime (\alpha)}; \cr
{\rm B}^{(\beta)}
&= (1 + a)^2 {\rm B}_{2,0}^{\backprime (\beta)}
+ (1 + a) {\rm B}_{1,1}^{\backprime (\beta)}
+ {\rm B}_{0,2}^{\backprime (\beta)}; \cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm T}^{(\tau)}
&= (1 + a)^t {\rm T}_{t,0}^{\backprime (\tau)}
+ (1 + a)^{t-1} {\rm T}_{t-1,1}^{\backprime (\tau)}
+ \cdots + {\rm T}_{0,t}^{\backprime (\tau)}: \cr}
\right\}
\eqno {\rm (224.)}$$
each symbol of the class
$${\rm A}_{p,q}^{\backprime (\alpha)},\quad
{\rm B}_{p,q}^{\backprime (\beta)},\quad \ldots \quad
{\rm T}_{p,q}^{\backprime (\tau)}
\eqno {\rm (225.)}$$
denoting a rational and integral function of the $2m - 1$
quantities $a_1',\ldots$ $a_m$, $b_1,\ldots$ $b_{m-1}$, which is
homogeneous of the dimension $p$ with respect to the $m$
quantities
$$a_1',\ldots\, a_m',
\eqno {\rm (204.)}\qquad$$
and of the dimension $q$ with respect to the $m - 1$ quantities
$$b_1,\ldots\, b_{m-1},
\eqno {\rm (226.)}$$
but is independent of the multiplier $a$. And the identical
equations obtained by comparing the expressions (202) and
(224), resolve themselves into the following:
$$\left. \eqalign{
{\rm A}_{1,0}^{(\alpha)}
&= {\rm A}_{1,0}^{\backprime (\alpha)};\quad
{\rm A}_{0,1}^{(\alpha)}
= a {\rm A}_{1,0}^{\backprime (\alpha)}
+ {\rm A}_{0,1}^{\backprime (\alpha)};\cr
{\rm B}_{2,0}^{(\beta)}
&= {\rm B}_{2,0}^{\backprime (\beta)};\quad
{\rm B}_{1,1}^{(\beta)}
= 2a {\rm B}_{2,0}^{\backprime (\beta)}
+ {\rm B}_{1,1}^{\backprime (\beta)};\cr
{\rm B}_{0,2}^{(\beta)}
&= a^2 {\rm B}_{2,0}^{\backprime (\beta)}
+ a {\rm B}_{1,1}^{\backprime (\beta)}
+ {\rm B}_{0,2}^{\backprime (\beta)};\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm T}_{t,0}^{(\tau)}
&= {\rm T}_{t,0}^{\backprime (\tau)};\quad
{\rm T}_{t-1,1}^{(\tau)}
= ta {\rm T}_{t,0}^{\backprime (\tau)}
+ {\rm T}_{t-1,1}^{\backprime (\tau)};\cr
{\rm T}_{t-2,2}^{(\tau)}
&= {t(t-1) \over 2} a^2 {\rm T}_{t,0}^{\backprime (\tau)}
+ (t-1) a {\rm T}_{t-1,1}^{\backprime (\tau)}
+ {\rm T}_{t-2,2}^{\backprime (\tau)};\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm T}_{0,t}^{(\tau)}
&= a^t {\rm T}_{t,0}^{\backprime (\tau)}
+ a^{t-1} {\rm T}_{t-1,1}^{\backprime (\tau)}
+ \cdots + {\rm T}_{0,t}^{\backprime (\tau)};\cr}
\right\}
\eqno {\rm (227.)}$$
so that the first system of auxiliary equations, (206) $\ldots$
(209), which are of the forms
$${\rm A}_{1,0}^{(\alpha)} = 0,\quad
{\rm B}_{2,0}^{(\beta)} = 0,\quad
{\rm C}_{3,0}^{(\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t,0}^{(\tau)} = 0,
\eqno {\rm (228.)}$$
may be replaced by the system
$${\rm A}_{1,0}^{\backprime (\alpha)} = 0,\quad
{\rm B}_{2,0}^{\backprime (\beta)} = 0,\quad
{\rm C}_{3,0}^{\backprime (\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t,0}^{\backprime (\tau)} = 0,
\eqno {\rm (229.)}$$
the change, so far, being only a change of notation; and after
satisfying this system by a suitable selection of the ratios of
the quantities (204), the second system of auxiliary equations,
(210) $\ldots$ (213), may then be transformed, with a real
simplification, (which consists in getting rid of the arbitrary
multiplier $a$, and in diminishing the number of quantities
whose ratios remain to be disposed of,) to another system of
equations of the forms
$$\left. \eqalign{
{\rm A}_{0,1}^{\backprime (\alpha)} &= 0,\quad
{\rm B}_{1,1}^{\backprime (\beta)} = 0,\quad
{\rm C}_{2,1}^{\backprime (\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t-1,1}^{\backprime (\tau)} = 0;\cr
{\rm B}_{0,2}^{\backprime (\beta)} &= 0,\quad
{\rm C}_{1,2}^{\backprime (\gamma)} = 0,\quad \ldots \quad
{\rm T}_{t-2,2}^{\backprime (\tau)} = 0;\cr
{\rm C}_{0,3}^{\backprime (\gamma)} &= 0,\quad \ldots \quad
{\rm T}_{t-3,3}^{\backprime (\tau)} = 0;\cr
\noalign{\hbox{$\cdots\cdots$}}
{\rm T}_{0,t}^{\backprime (\tau)} &= 0;\cr}
\right\}
\eqno {\rm (230.)}$$
which are rational and integral and homogeneous with respect to
the $m - 1$ quantities (226), and are independent of the
multiplier $a$. Unless, then, the number of the equations
of this transformed system (230), which is the same as the
number of equations in the second auxiliary problem before
proposed, be less than the number $m - 1$ of the new auxiliary
quantities (226), we shall have, in general, null values for all
those quantities, that is, we shall have
$$b_1 = 0,\quad
b_2 = 0,\quad \ldots \quad
b_{m-1} = 0;
\eqno {\rm (231.)}$$
and therefore we shall be conducted, by (222), to expressions of
the forms (217), which will in general lead, as has been already
shown, to the case of failure (216). We have therefore a
{\it new condition of inequality}, which the number~$m$ must
satisfy, in order to the general success of the method, namely
the following,
$$m - 1 > h_1' + h_2' + h_3'+ \cdots + h_t';
\eqno {\rm (232.)}$$
in which $h_1', h_2', h_3,\ldots, h_t'$ denote respectively the
numbers of the equations of the first, second, third,$\ldots$
and $t^{\rm th}$ degrees, in the second auxiliary problem; so
that, by what has been already shown,
$$\left. \eqalign{
h_t' &= h_t - 1,\cr
h_{t-1}' &= h_{t-1} + h_t - 1,\cr
h_{t-2}' &= h_{t-2} + h_{t-1} + h_t - 1,\cr
\noalign{\hbox{$\cdots\cdots$}}
h_2' &= h_2 + \cdots + h_t - 1,\cr
h_1' &= h_1 + h_2 + \cdots + h_t - 1.\cr}
\right\}
\eqno {\rm (233.)}$$
These last expressions give
$$h_1' + h_2' + h_3' + \cdots + h_t'
= h_1 + 2 h_2 + 3 h_3 + \cdots + t h_t - t;
\eqno {\rm (234.)}$$
so that the new condition of inequality, (232), may be written
as follows,
$$m - 1 > h_1 + 2 h_2 + 3 h_3 + \cdots + t (h_t - 1);
\eqno {\rm (235.)}$$
and therefore also thus,
$$m > h_1 + h_2 + h_3 + \cdots + h_t + h_2 + 2 h_3 + \cdots
+ (t - 1)(h_t - 1).
\eqno {\rm (236.)}$$
It {\it includes}, therefore, in general, the old inequality
(215); and may be considered as comprising in itself all the
conditions respecting the magnitude of the number~$m$, connected
with our present inquiry: or, at least, as capable of furnishing
us with all such conditions, if only it be sufficiently
developed.
\bigbreak
[14.]
It must, however, be remembered, as a part of such development,
that although, when this condition (232) or (235) or (236) is
satisfied, the three auxiliary problems above stated are, in
general, theoretically capable of being resolved, and of
conducting to a system of ratios of the $m$ original quantities
(192), which shall satisfy the original system of equations, yet
each of the two first auxiliary systems contains, in general,
more than two equations of the second or higher degrees; and
therefore that, in order to avoid any elevation of degree by
elimination (as required by the original problem), the process
must in general be {\it repeated}, and each of the two auxiliary
systems themselves must be decomposed, and treated like the
system originally proposed. These new decompositions introduce,
in general, new conditions of inequality, analogous to the
condition lately determined; but it is clear that the condition
connected with the decomposition of the first of the auxiliary
systems must be included in the condition connected with the
decomposition of the second of those systems, because the latter
system contains, in general, in each of the degrees
$1,2,3,\ldots,t-1$, a greater number of equations than the
former, while both contain, in the degree~$t$, the same number of
equations, namely, $h_t - 1$. Conceiving, then, the second
auxiliary system to be decomposed by a repetition of the process
above described into two new auxiliary systems or groups of
equations, and into one separate and reserved equation of the
$t^{\rm th}$ degree, we are conducted to this new equation of
inequality, analogous to (232),
$$m - 2 > h_1'' + h_2'' + h_3'' + \cdots + h_t'';
\eqno {\rm (237.)}$$
$h_1'', h_2'', h_3'',\ldots\, h_t''$ denoting, respectively, the
numbers of equations of the first, second, third,$\ldots$ and
$t^{\rm th}$ degrees, in the second new group of equations; in
such a manner that, by the nature of the process,
$$\left. \eqalign{
h_t'' &= h_t' - 1,\cr
h_{t-1}'' &= h_{t-1}' + h_t' - 1,\cr
h_{t-2}'' &= h_{t-2}' + h_{t-1}' + h_t' - 1,\cr
\noalign{\hbox{$\cdots\cdots$}}
h_1'' &= h_1' + h_2' + \cdots + h_t' - 1.\cr}
\right\}
\eqno {\rm (238.)}$$
Repeating this process, we find, next, the condition,
$$m - 3 > h_1''' + h_2''' + h_3''' + \cdots + h_t''',
\eqno {\rm (239.)}$$
and generally
$$m - i > h_1^{(i)} + h_2^{(i)} + h_3^{(i)} + \cdots + h_t^{(i)};
\eqno {\rm (240.)}$$
each new condition of this series including all that go before
it, and the symbol $h_p^{(i)}$ being such that
$$h_p^{(0)} = h_p,
\eqno {\rm (241.)}$$
$$h_t^{(i+1)} - h_t^{(i)} = -1,
\eqno {\rm (242.)}$$
and
$$h_{t-n}^{(i+1)} - h_{t-n}^{(i)} = h_{t-n+1}^{(i+1)}.
\eqno {\rm (243.)}$$
Integrating these last equations as equations in finite
differences, we find
$$\left. \eqalign{
h_t^{(i)}
&= h_t - i;\cr
h_{t-1}^{(i)}
&= h_{t-1} + i \left( h_t - {i+1 \over 2} \right);\cr
h_{t-2}^{(i)}
&= h_{t-2} + i h_{t-1}
+ i \mathbin{.} {i+1 \over 2}
\mathbin{.} \left( h_t - {i+2 \over 3} \right);\cr
h_{t-3}^{(i)}
&= h_{t-3} + i h_{t-2}
+ i {i+1 \over 2} h_{t-1}
+ i {i+1 \over 2} {i+2 \over 3}
\left( h_t - {i+3 \over 4} \right);\cr
\noalign{\hbox{$\cdots\cdots$}}
h_1^{(i)}
&= h_1 + i h_2
+ i {i+1 \over 2} h_3
+ i {i+1 \over 2} {i+2 \over 3} h_4
+ \cdots \cr
&\mathrel{\phantom{=}} \mathord{}
+ i {i+1 \over 2} {i+2 \over 3} \cdots {i+t-2 \over t-1}
\left( h_t - {i+t-1 \over t} \right).\cr}
\right\}
\eqno {\rm (244.)}$$
And making, in these expressions,
$$i = h_t,
\eqno {\rm (245.)}$$
so as to have
$$h_t^{(i)} = 0,
\eqno {\rm (246.)}$$
and putting, for abridgment,
$$h_1^{(h_t)} = {}^\backprime h_1,\quad
h_2^{(h_t)} = {}^\backprime h_2,\quad \cdots \quad
h_{t-1}^{(h_t)} = {}^\backprime h_{t-1},
\eqno {\rm (247.)}$$
we find that at the stage when all the equations of the
$t^{\rm th}$ degree have been removed from the auxiliary groups
of equations, we are led to satisfy, if possible, by the ratios
of $m - h_t$ auxiliary quantities, a system containing
${}^\backprime h_1$ equations of the first degree,
${}^\backprime h_2$ of the second,
${}^\backprime h_3$ of the third, and so on as far as
${}^\backprime h_{t-1}$ of the degree $t - 1$; in which
$$\left. \eqalign{
{}^\backprime h_{t-1}
&= h_{t-1}
+ {\textstyle {1 \over 2}} h_t (h_t - 1),\cr
{}^\backprime h_{t-2}
&= h_{t-2}
+ h_t h_{t-1}
+ {\textstyle {1 \over 3}} (h_t + 1) h_t (h_t - 1),\cr
{}^\backprime h_{t-3}
&= h_{t-3}
+ h_t h_{t-2}
+ {\textstyle {1 \over 2}} (h_t + 1) h_t h_{t-1}
+ {\textstyle {1 \over 8}}
(h_t + 2) (h_t + 1) h_t (h_t - 1),\cr
{}^\backprime h_1
&= h_1
+ h_t h_2
+ {\textstyle {1 \over 2}} (h_t + 1) h_t h_3
+ {\textstyle {1 \over 6}} (h_t + 2) (h_t + 1) h_t h_4
+ \cdots \cr
&\mathrel{\phantom{=}} \mathord{}
+ {1 \over 2.3.4\ldots (t - 2)t}
(h_t + t - 2) (h_t + t - 3) \cdots h_t (h_t - 1);\cr}
\right\}
\eqno {\rm (248.)}$$
so that, at this stage, we arrive at the following condition of
inequality,
$$m - h_t > {}^\backprime h_1 + {}^\backprime h_2
+ {}^\backprime h_3 + \cdots + {}^\backprime h_{t-1},
\eqno {\rm (249.)}$$
${}^\backprime h_1, {}^\backprime h_2,\ldots\,
{}^\backprime h_{t-1}$
having the meanings (248). In exactly the same way, we find the
condition
$$m - h_t - {}^\backprime h_{t-1}
> {}^{\backprime\backprime} h_1
+ {}^{\backprime\backprime} h_2
+ {}^{\backprime\backprime} h_3 + \cdots
+ {}^{\backprime\backprime} h_{t-2},
\eqno {\rm (250.)}$$
in which,
$$\left. \eqalign{
{}^{\backprime\backprime} h_{t-2}
&= {}^\backprime h_{t-2}
+ {\textstyle {1 \over 2}} {}^\backprime h_{t-1}
({}^\backprime h_{t-1} - 1),\cr
{}^{\backprime\backprime} h_{t-3}
&= {}^\backprime h_{t-3}
+ {}^\backprime h_{t-1} {}^\backprime h_{t-2}
+ {\textstyle {1 \over 3}}
({}^\backprime h_{t-1} + 1) {}^\backprime h_{t-1}
({}^\backprime h_{t-1} - 1),\cr
\noalign{\hbox{\&c.,}}}
\right\}
\eqno {\rm (251.)}$$
by clearing all the auxiliary systems from all equations of the
degree $t - 1$; and again by clearing all such auxiliary groups
from equations of the degree $t - 2$, we obtain a condition of
the form
$$m - h_t - {}^\backprime h_{t-1}
- {}^{\backprime\backprime} h_{t-2}
> {}^{\backprime\backprime\backprime} h_1 + \cdots
+ {}^{\backprime\backprime\backprime} h_{t-3},
\eqno {\rm (252.)}$$
in which
$${}^{\backprime\backprime\backprime} h_{t-3}
= {}^{\backprime\backprime} h_{t-3}
+ {\textstyle {1 \over 2}}
{}^{\backprime\backprime} h_{t-2}
({}^{\backprime\backprime} h_{t-2} - 1),
\quad \hbox{\&c.}
\eqno {\rm (253.)}$$
so that at last we are conducted to a condition which may be
thus denoted, and which contains the ultimate result of all the
restrictions on the number $m$,
$$m - h_t - {}^\backprime h_{t-1}
- {}^{\backprime\backprime} h_{t-2}
- {}^{\backprime\backprime\backprime} h_{t-3}
- \cdots
- {}^{(t-2)} h_2
> {}^{(t-1)} h_1,
\eqno {\rm (254.)}$$
that is,
$$m > h_t + {}^\backprime h_{t-1}
+ {}^{\backprime\backprime} h_{t-2}
+ {}^{\backprime\backprime\backprime} h_{t-3}
+ \cdots
+ {}^{(t-2)} h_2
+ {}^{(t-1)} h_1.
\eqno {\rm (255.)}$$
{\it The number~$m$, of quantities originally disposible, must
therefore in general be at least equal to a certain minor limit},
which may be thus denoted,
$$m(h_1, h_2, h_3,\ldots\, h_t)
= h_t + {}^\backprime h_{t-1}
+ {}^{\backprime\backprime} h_{t-2}
+ \cdots
+ {}^{(t-2)} h_2
+ {}^{(t-1)} h_1
+ 1,
\eqno {\rm (256.)}$$
{\it in order that the method may succeed\/}; and reciprocally,
{\it the method will in general be successful when $m$ equals or
surpasses this limit}.
\bigbreak
[15.]
To illustrate the foregoing general discussion, let us suppose
that
$$t = 2;
\eqno {\rm (257.)}$$
that is, let us propose to satisfy a system containing $h_1$
equations of the first degree,
$${\rm A}' = 0,\quad \ldots \quad
{\rm A}^{(\alpha)} = 0,\quad \ldots \quad {\rm A}^{(h_1)} = 0,
\eqno {\rm (193.)}\qquad$$
and $h_2$ equations of the second degree,
$${\rm B}' = 0,\quad \ldots \quad
{\rm B}^{(\beta)} = 0,\quad \ldots \quad {\rm B}^{(h_2)} = 0,
\eqno {\rm (194.)}\qquad$$
(but not containing any equations of higher degrees than the
second,) by a suitable selection of the $m - 1$ ratios of $m$
quantities,
$$a_1,\ldots\, a_m,
\eqno {\rm (192.)}\qquad$$
and without being obliged in any part of the process to introduce
any elevation of degree by elimination. Assuming, as before,
$$a_1 = a_1' + a_1'',\quad \ldots \quad
a_m = a_m' + a_m'',
\eqno {\rm (197.)}\qquad$$
and employing the corresponding decompositions
$${\rm A}'
= {\rm A}_{1,0}' + {\rm A}_{0,1}',\quad \ldots \quad
{\rm A}^{(h_1)}
= {\rm A}_{1,0}^{(h_1)} + {\rm A}_{0,1}^{(h_1)},
\eqno {\rm (258.)}$$
and
$${\rm B}'
= {\rm B}_{2,0}' + {\rm B}_{1,1}'
+ {\rm B}_{0,2}',\quad \ldots \quad
{\rm B}^{(h_2 - 1)}
= {\rm B}_{2,0}^{(h_2 - 1)}
+ {\rm B}_{1,1}^{(h_2 - 1)} + {\rm B}_{0,2}^{(h_2 - 1)},
\eqno {\rm (259.)}$$
we shall be able to resolve the original problem, if we can
resolve the system of the three following.
\noindent
First: to satisfy, by ratios of the $m$ auxiliary quantities
$$a_1',\ldots\, a_m',
\eqno {\rm (204.)}\qquad$$
an auxiliary system, containing the $h_1$ equations of the first
degree
$${\rm A}_{1,0}' = 0,\quad \ldots \quad
{\rm A}_{1,0}^{(h_1)} = 0,
\eqno {\rm (206.)}\qquad$$
and the $h_2 - 1$ equations of the second degree
$${\rm B}_{2,0}' = 0,\quad \ldots \quad
{\rm B}_{2,0}^{(h_2 - 1)} = 0.
\eqno {\rm (260.)}$$
Second: to satisfy, by ratios of the $m$ other auxiliary
quantities
$$a_1'',\ldots\, a_m'',
\eqno {\rm (205.)}\qquad$$
another auxiliary system, containing $h_1 + h_2 - 1$ equations of
the first degree,
$${\rm A}_{0,1}' = 0,\quad \ldots \quad
{\rm A}_{0,1}^{(h_1)} = 0,\quad
{\rm B}_{1,1}' = 0,\quad \ldots \quad
{\rm B}_{1,1}^{(h_2 - 1)} = 0,
\eqno {\rm (261.)}$$
and $h_2 - 1$ equations of the second degree,
$${\rm B}_{0,2}' = 0,\quad \ldots \quad
{\rm B}_{0,2}^{(h_2 - 1)} = 0.
\eqno {\rm (262.)}$$
Third: to satisfy, by the ratio of any one of the $m$ quantities
(205) to any one of the $m$ quantities (204), this one
remaining equation of the second degree
$${\rm B}^{(h_2)} = 0.
\eqno {\rm (263.)}$$
The enunciation of the original problem supposes that
$$m > h_1 + h_2;
\eqno {\rm (264.)}$$
since otherwise the original equations (193) and (194) would in
general conduct to the excluded case, or case of failure,
$$a_1 = 0,\quad \ldots \quad a_m = 0.
\eqno {\rm (216.)}\qquad$$
In virtue of this condition (264) the first auxiliary problem is
indeterminate, because
$$m - 1 > h_1 + h_2 - 1.
\eqno {\rm (265.)}$$
But, by whatever system of ratios
$${a_1' \over a_m'},\quad \ldots \quad
{a_{m-1}' \over a_m'}
\eqno {\rm (219.)}\qquad$$
we may succeed in satisfying the first auxiliary system of
equations, (206) and (260), we may in general transform the
second auxiliary system of equations, (261) and (262), into a
system which may be thus denoted,
$$\left. \eqalign{
{\rm A}_{0,1}^{\backprime\prime} &= 0,\quad \ldots \quad
{\rm A}_{0,1}^{\backprime (h_1)} = 0,\cr
{\rm B}_{1,1}^{\backprime\prime} &= 0,\quad \ldots \quad
{\rm B}_{1,1}^{\backprime (h_2 - 1)} = 0,\cr
{\rm B}_{0,2}^{\backprime\prime} &= 0,\quad \ldots \quad
{\rm B}_{0,2}^{\backprime (h_2 - 1)} = 0,\cr}
\right\}
\eqno {\rm (266.)}$$
and which contains $h_1 + h_2 - 1$ equations of the first degree,
and $h_2 - 1$ equations of the second degree, between $m - 1$ new
combinations, or new auxiliary quantities following,
$$b_1 = a_1'' - {a_1' \over a_m'} a_m'',\quad \ldots \quad
b_{m-1} = a_{m-1}'' - {a_{m-1}' \over a_m'} a_m'';
\eqno {\rm (267.)}$$
so that the solution of the second auxiliary problem will give,
in general,
$$b_1 = 0,\quad \ldots \quad
b_{m-1} = 0;
\eqno {\rm (231.)}\qquad$$
and will therefore give, for the $m$ auxiliary quantities (205),
a system of ratios coincident with the ratios (219),
$${a_1'' \over a_m''} = {a_1' \over a_m'},\quad \ldots \quad
{a_{m-1}'' \over a_m''} = {a_{m-1}' \over a_m'},
\eqno {\rm (268.)}$$
unless
$$m - 1 > h_1 + 2 (h_2 - 1).
\eqno {\rm (269.)}$$
When, therefore, this last condition is not satisfied, the two
first auxiliary problems will conduct, in general, to a system of
determined ratios for the $m$ original quantities (192), namely
$${a_1 \over a_m} = {a_1' \over a_m'},\quad \ldots \quad
{a_{m-1} \over a_m} = {a_{m-1}' \over a_m'};
\eqno {\rm (218.)}\qquad$$
and unless these happen to satisfy the equation of the third
auxiliary problem, namely
$${\rm B}^{(h_2)} = 0,
\eqno {\rm (263.)}\qquad$$
which had not been employed in determing them, we shall fall back
on the excluded case, or case of failure, (216). But, even when
the condition (269) is satisfied, and when, therefore, the
auxiliary equations are theoretically capable of conducting to
ratios which shall satisfy the equations originally proposed, it
will still be necessary, in general, to decompose each of the two
first auxiliary systems of equations into others, in order to
comply with the enunciation of the original problem, which
requires that we should avoid all raising of degree by
elimination, in every part of the process. Confining ourselves
to the consideration of the second auxiliary problem, (which
includes the difficulties of the first,) we see that the
transformed auxiliary system (266) contains $h_1'$ equations
of the first degree, and $h_2'$ of the second, if we put, for
abridgment,
$$\left. \eqalign{
h_2' &= h_2 - 1,\cr
h_1' &= h_1 + h_2 - 1;\cr}
\right\}
\eqno {\rm (270.)}$$
which new auxiliary equations are to be satisfied, if possible,
by the ratios of $m - 1$ new auxiliary quantities; so that a
repetition of the former process of decomposition and
transformation would conduct to a new auxiliary system,
containing $h_1''$ equations of the first degree, and $h_2''$ of
the second, in which
$$\left. \eqalign{
h_2'' &= h_2' - 1,\cr
h_1'' &= h_1' + h_2' - 1,\cr}
\right\}
\eqno {\rm (271.)}$$
and which must be satisfied, if possible, by the ratios of
$m - 2$ new auxiliary quantities; and thus we should arrive at
this new condition, as necessary to the success of the method:
$$m - 2 > h_1' + 2(h_2' - 1);
\eqno {\rm (272.)}$$
or, more concisely,
$$m - 2 > h_1'' + h_2''.
\eqno {\rm (273.)}$$
And so proceeding, we should find generally,
$$m - i > h_1^{(i)} + h_2^{(i)},
\eqno {\rm (274.)}$$
the functions $h_1^{(i)}$, $h_2^{(i)}$ being determined by the
equations
$$h_2^{(0)} = h_2,\quad h_1^{(0)} = h_1,
\eqno {\rm (275.)}$$
$$h_2^{(i+1)} - h_2^{(i)} = -1,
\eqno {\rm (276.)}$$
$$h_1^{(i+1)} - h_1^{(i)} = h_2^{(i+1)};
\eqno {\rm (277.)}$$
which give, by integrations of finite differences,
$$\left. \eqalign{
h_2^{(i)} &= h_2 - i;\cr
h_1^{(i)} &= h_1 + i \left( h_2 - {i+1 \over 2} \right).\cr}
\right\}
\eqno {\rm (278.)}$$
Thus, making
$$i = h_2,
\eqno {\rm (279.)}$$
and putting, for abridgment,
$${}^\backprime h_1
= h_1^{(h_2)}
= h_1 + {\textstyle {1 \over 2}} h_2 (h_2 - 1),
\eqno {\rm (280.)}$$
we arrive at last at a stage of the process at which we have to
satisfy a system of ${}^\backprime h_1$ equations of the first
degree by the ratios of $m - h_2$ quantities; and now, at length,
we deduce this final condition of inequality, to be satisfied by
the number~$m$, in order to the general success of the method (in
the case $t = 2$),
$$m - h_2 > {}^\backprime h_1;
\eqno {\rm (281.)}$$
that is,
$$m > h_1 + {\textstyle {1 \over 2}} (h_2 + 1) h_2;
\eqno {\rm (282.)}$$
or, in other words, $m$ must at least be equal to the following
{\it minor limit},
$$m(h_1, h_2) = h_1 + 1 + {\textstyle {1 \over 2}} (h_2 + 1) h_2.
\eqno {\rm (283.)}$$
For example, making $h_1 = 1$, and $h_2 = 2$, we find that a
system containing {\it one} homogeneous equation of the first
degree, and {\it two} of the second, can be satisfied, in
general, without any elevation of degree by elimination, and
therefore without its being necessary to resolve any equation
higher than the second degree, by the ratios of $m$ quantities,
provided that this number~$m$ is not less than the minor limit
{\it five\/}: a result which may be briefly thus expressed,
$$m(1,2) = 5.
\eqno {\rm (284.)}$$
\bigbreak
[16.]
Indeed, it might seem, that in the process last described, an
advantage would be gained by stopping at that stage, at which,
by making $i = h_2 - 1$ in the formulae (278), we should have
$$\left. \eqalign{
h_2^{(h_2 - 1)}
&= 1,\cr
h_1^{(h_2 - 1)}
&= h_1 + {\textstyle {1 \over 2}} h_2 (h_2 - 1),\cr}
\right\}
\eqno {\rm (285.)}$$
and
$$m - i = m - h_2 + 1;
\eqno {\rm (286.)}$$
that is, when we should have to satisfy, by the ratios of
$m_2 - h_2 + 1$ quantities, a system containing only {\it one\/}
equation of the second degree, in combination with
$h_1 + {1 \over 2} h_2 (h_2 - 1)$ of the first. For, the
ordinary process of elimination, performed between the equations
of this last system, would not conduct to any equation higher
than the second degree; and hence, without going any further, we
might perceive it to be sufficient that the number $m$ should
satisfy this condition of inequality,
$$m - h_2 + 1
> h_1 + {\textstyle {1 \over 2}} h_2 (h_2 - 1) + 1.
\eqno {\rm (287.)}$$
But it is easy to see that this alteration of method introduces
no real simplification; the condition (287) being really
coincident with the condition (282) or (283). To illustrate
this result, it may be worth observing, that, in general, instead
of the ordinary mode of satisfying, by ordinary elimination, any
system of rational and integral and homogeneous equations,
containing $n$ such equations of the first degree,
$${}^\backprime {\rm A}' = 0,\quad
{}^\backprime {\rm A}'' = 0,\quad \ldots \quad
{}^\backprime {\rm A}^{(n)} = 0,
\eqno {\rm (288.)}$$
and one of the second degree
$${}^\backprime {\rm B}' = 0,
\eqno {\rm (289.)}$$
by the $n + 1$ ratios of $n + 2$ disposable quantities,
$$a_1, a_2,\ldots\, a_{n+2},
\eqno {\rm (290.)}$$
it is permitted to proceed as follows. Decomposing each of the
first $n + 1$ quantities into two parts, so as to put
$$a_1 = a_1' + a_1'',\quad
a_2 = a_2' + a_2'',\quad \ldots \quad
a_{n+1} = a_{n+1}' + a_{n+1}'',
\eqno {\rm (291.)}$$
we may decompose each of the given functions of the first degree,
such as ${}^\backprime {\rm A}^{(\alpha)}$, into two
corresponding parts, ${}^\backprime {\rm A}_{1,0}^{(\alpha)}$ and
${}^\backprime {\rm A}_{0,1}^{(\alpha)}$, of which the former,
${}^\backprime {\rm A}_{1,0}^{(\alpha)}$ is a function of the
first degree of the $n + 2$ quantities,
$$a_1', a_2',\ldots\, a_{n+1}', a_{n+2}',
\eqno {\rm (292.)}$$
while the latter, ${}^\backprime {\rm A}_{0,1}^{(\alpha)}$, is a
function of the first degree of the $n+1$ other quantities
$$a_1'', a_2'',\ldots\, a_{n+1}'';
\eqno {\rm (293.)}$$
and then, after resolving in any manner the indeterminate
problem, to satisfy the $n$ equations of the first degree,
$${}^\backprime {\rm A}_{1,0}' = 0,\quad
{}^\backprime {\rm A}_{1,0}'' = 0,\quad \ldots \quad
{}^\backprime {\rm A}_{1,0}^{(n)} = 0,
\eqno {\rm (294.)}$$
by a suitable selection of the $n + 1$ ratios of the $n + 2$
quantities (292), (excluding only the assumption $a_{n+2} = 0$,)
we may determine the $n$ ratios of the $n + 1$ quantities (293),
so as to satisfy these $n$ other equations of the first degree,
$${}^\backprime {\rm A}_{0,1}' = 0,\quad
{}^\backprime {\rm A}_{0,1}'' = 0,\quad \ldots \quad
{}^\backprime {\rm A}_{0,1}^{(n)} = 0;
\eqno {\rm (295.)}$$
after which it will only remain to determine the ratio of any one
of these latter quantities (293) to any one of the former
quantities (292), so as to satisfy the equation of the second
degree (289), and the original problem will be resolved.
\bigbreak
[17.]
Again, let
$$t = 3;
\eqno {\rm (296.)}$$
that is, let us consider a system containing $h_1$ equations of
the first degree, such as those marked (193), along with $h_2$
equations of the second degree (194), and $h_3$ equations of the
third degree (195), to be satisfied by the ratios of $m$
disposable quantities (192). After exhausting, by the general
process already sufficiently explained, all the equations of the
third degree in all the auxiliary systems, we are conducted to
satisfy, if possible, by the ratios of $m - h_3$ quantities, a
system containing ${}^\backprime h_1$ equations of the first, and
${}^\backprime h_2$ of the second degree, in which,
$$\left. \eqalign{
{}^\backprime h_2
&= h_2 + {\textstyle {1 \over 2}} h_3 ( h_3 - 1 ),\cr
{}^\backprime h_1
&= h_1 + h_3 h_2
+ {\textstyle {1 \over 3}} ( h_3 + 1 ) h_3 ( h_3 - 1 );\cr}
\right\}
\eqno {\rm (297.)}$$
and after exhausting, next, all the equations of the second
degree in all the new auxiliary systems, we are conducted to
satisfy, by the ratios of $m - h_3 - {}^\backprime h_2$
quantities, a system of ${}^{\backprime\backprime} h_1$ equations
of the first degree, in which,
$${}^{\backprime\backprime} h_1
= {}^\backprime h_1
+ {\textstyle {1 \over 2}} {}^\backprime h_2
({}^\backprime h_2 - 1).
\eqno {\rm (298.)}$$
We find, therefore, that the number~$m$ must satisfy the
following condition of inequality,
$$m - h_3 - {}^\backprime h_2
> {}^{\backprime\backprime} h_1,
\eqno {\rm (299.)}$$
that is,
$$m > h_3 + {}^\backprime h_2 + {}^{\backprime\backprime} h_1.
\eqno {\rm (300.)}$$
On substituting for ${}^{\backprime\backprime} h_1$ its value
(298), this last condition becomes,
$$m > h_3 + {\textstyle {1 \over 2}} {}^\backprime h_2
({}^\backprime h_2 + 1)
+ {}^\backprime h_1;
\eqno {\rm (301.)}$$
that is, in virtue of the expressions (297),
$$\left. \eqalign{
m &> h_1 + {\textstyle {1 \over 2}} (h_2 + 1) h_2
+ {\textstyle {1 \over 2}} (h_2 + 1) (h_3 + 1) h_3 \cr
&\mathrel{\phantom{>}} \mathord{}
+ {\textstyle {1 \over 3}} (h_3 + 1) h_3 (h_3 - 1)
+ {\textstyle {1 \over 8}} (h_3 + 1) h_3
(h_3 - 1) (h_3 - 2).\cr}
\right\}
\eqno {\rm (302.)}$$
The number~$m$ must therefore equal or surpass a certain minor
limit, which, in the notation of factorials, may be expressed as
follows:
$$m < (h_1 + 1)
+ {\textstyle {1 \over 2}} [h_2 + 1]^2
+ {\textstyle {1 \over 2}} (h_2 + 1) [h_3 + 1]^2
+ {\textstyle {1 \over 3}} [h_3 + 1]^3
+ {\textstyle {1 \over 8}} [h_3 + 1]^4;
\eqno {\rm (303.)}$$
the symbol $[\eta]^n$ denoting the continued product,
$$[\eta]^n = \eta (\eta - 1) (\eta - 2) \cdots (\eta - n + 1).
\eqno {\rm (304.)}$$
So that, when we denote this minor limit of $m$ by the symbol
$m(h_1, h_2, h_3)$, we obtain, in general, the formula
$$m(h_1, h_2, h_3) = \eta_1
+ {\textstyle {1 \over 2}} [\eta_2]^2
+ {\textstyle {1 \over 2}} \eta_2 [\eta_3]^2
+ {\textstyle {1 \over 3}} [\eta_3]^3
+ {\textstyle {1 \over 8}} [\eta_3]^4,
\eqno {\rm (305.)}$$
in which,
$$\eta_1 = h_1 + 1,\quad
\eta_2 = h_2 + 1,\quad
\eta_3 = h_3 + 1.
\eqno {\rm (306.)}$$
For example,
$$m(1,1,1) = 5.
\eqno {\rm (307.)}$$
\bigbreak
[18.]
When
$$t = 4,
\eqno {\rm (308.)}$$
that is, when some of the original equations are as high as the
fourth degree, (but none more elevated,) then
$$\left. \eqalign{
{}^\backprime h_3
&= h_3 + {\textstyle {1 \over 2}} h_4 (h_4 - 1),\cr
{}^\backprime h_2
&= h_2 + h_4 h_3
+ {\textstyle {1 \over 3}} (h_4 + 1) h_4 (h_4 - 1),\cr
{}^\backprime h_1
&= h_1 + h_4 h_2
+ {\textstyle {1 \over 2}} (h_4 + 1) h_4 h_3
+ {\textstyle {1 \over 8}} (h_4 + 2) (h_4 + 1)
h_4 (h_4 - 1);\cr}
\right\}
\eqno {\rm (309.)}$$
$$\left. \eqalign{
{}^{\backprime\backprime} h_2
&= {}^\backprime h_2
+ {\textstyle {1 \over 2}} {}^\backprime h_3
({}^\backprime h_3 - 1),\cr
{}^{\backprime\backprime} h_1
&= {}^\backprime h_1 + {}^\backprime h_3 {}^\backprime h_2
+ {\textstyle {1 \over 3}} ({}^\backprime h_3 + 1)
{}^\backprime h_3 ({}^\backprime h_3 - 1);\cr}
\right\}
\eqno {\rm (310.)}$$
$${}^{\backprime\backprime\backprime} h_1
= {}^{\backprime\backprime} h_1
+ {\textstyle {1 \over 2}} {}^{\backprime\backprime} h_2
({}^{\backprime\backprime} h_2 - 1);
\eqno {\rm (311.)}$$
and the minor limit of $m$, denoted by the symbol
$m(h_1, h_2, h_3, h_4)$, is given by the equation
$$m(h_1, h_2, h_3, h_4)
= h_4
+ {}^\backprime h_3
+ {}^{\backprime\backprime} h_2
+ {}^{\backprime\backprime\backprime} h_1
+ 1;
\eqno {\rm (312.)}$$
which may be thus developed,
$$\left. \eqalign{
m(h_1, h_2, h_3, h_4)
&= \eta_1
+ {1 \over 2} [\eta_2]^2
+ {1 \over 2} \eta_2 [\eta_3]^2
+ {1 \over 3} [\eta_3]^3
+ {1 \over 8} [\eta_3]^4 \cr
&\mathrel{\phantom{=}} \mathord{}
+ \eta_2 \left\{
{1 \over 2} \eta_3 [\eta_4]^2
+ {1 \over 3} [\eta_4]^3
+ {1 \over 8} [\eta_4]^4
\right\} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {1 \over 4} [\eta_3]^3 [\eta_4]^2
+ [\eta_3]^2 \left\{
{1 \over 2} [\eta_4]^2
+ {2 \over 3} [\eta_4]^3
+ {3 \over 16} [\eta_4]^4
\right\} \cr
&\mathrel{\phantom{=}} \mathord{}
+ \eta_3 \left\{
[\eta_4]^3
+ {7 \over 4} [\eta_4]^4
+ {2 \over 3} [\eta_4]^5
+ {1 \over 16} [\eta_4]^6
\right\} \cr
&\mathrel{\phantom{=}} \mathord{}
+ {3 \over 2} [\eta_4]^4
+ {5 \over 2} [\eta_4]^5
+ {79 \over 72} [\eta_4]^6
+ {1 \over 6} [\eta_4]^7
+ {1 \over 128} [\eta_4]^8,\cr}
\right\}
\eqno {\rm (313.)}$$
if we employ the notation of factorials, and put for abridgment,
$$\eta_1 = h_1 + 1,\quad \ldots \quad \eta_4 = h_4 + 1.
\eqno {\rm (314.)}$$
In the notation of powers, we have
$$\eqalignno{
m(h_1, h_2, h_3, h_4)
\hskip-6em & \cr
&\, = 1 + h_1 \cr
&\mathrel{\phantom{= 1}} \mathord{}
+ {1 \over 24} h_2
( 12 + 10 h_4 + 9 h_4^2 + 2 h_4^3 + 3 h_4^4 ) \cr
&\mathrel{\phantom{= 1}} \mathord{}
+ {1 \over 2} h_2 h_3 ( 1 + h_4 + h_4^2)
+ {1 \over 2} h_2 h_3^2
+ {1 \over 2} h_2^2 \cr
&\mathrel{\phantom{= 1}} \mathord{}
+ {1 \over 48} h_3
( 20 + 22 h_4 + 25 h_4^2 + 9 h_4^3
+ 8 h_4^4 + 5 h_4^5 + 3 h_4^6 ) \cr
&\mathrel{\phantom{= 1}} \mathord{}
+ {1 \over 48} h_3^2
( 18 + 10 h_4 + 15 h_4^2 + 14 h_4^3 + 9 h_4^4) \cr
&\mathrel{\phantom{= 1}} \mathord{}
+ {1 \over 12} h_3^3 ( 1 + 3 h_4 + 3 h_4^2 )
+ {1 \over 8} h_3^4 \cr
&\mathrel{\phantom{= 1}} \mathord{}
+ {1 \over 1152}
( 432 h_4 + 364 h_4^2 + 108 h_4^3 + 169 h_4^4
+ 24 h_4^5 + 34 h_4^6 + 12 h_4^7 + 9 h_4^8 ).
&{\rm (315.)}\cr}$$
As examples, whichever formula we employ, we find
$$m(1,0,1,1) = 7;
\eqno {\rm (316.)}$$
$$m(1,1,1,1) = 11;
\eqno {\rm (317.)}$$
$$m(1,1,1,2) = 47;
\eqno {\rm (318.)}$$
$$m(5,4,3,3) = 922.
\eqno {\rm (319.)}$$
\bigbreak
[19.]
In general (by the nature of the process explained in the
foregoing articles) the minor limit (256) of the number~$m$,
which we have denoted by the symbol
$$m(h_1, h_2,\ldots\, h_t),$$
is a function such that
$$m(h_1, h_2,\ldots\, h_t)
= 1 + m(h_1', h_2',\ldots\, h_t'),
\eqno {\rm (320.)}$$
$h_1',\ldots\, h_t'$ being determined by the formulae (233).
This equation in finite differences (320) may be regarded as
containing the most essential element of the whole foregoing
discussion; and from it the formulae already found for the cases
$t = 2$, $t = 3$, $t = 4$, might have been deduced in other ways.
From it also we may perceive, that whenever the original system
contains only one equation of the highest or $t^{\rm th}$ degree,
in such a manner that
$$h_t = 1,
\eqno {\rm (321.)}$$
then, whatever $t$ may be, we have the formula
$$m(h_1, h_2,\ldots\, h_{t-1},1)
= m(h_1 + h_2 + \cdots + h_{t-1},\,
h_2 + \cdots + h_{t-1},\ldots\,
h_{t-1});
\eqno {\rm (322.)}$$
so that, for example,
$$m(1,1,1,1,1) = 1 + m(4,3,2,1);
\eqno {\rm (323.)}$$
$$m(4,3,2,1) = 1 + m(9,5,2) = 46;
\eqno {\rm (324.)}$$
$$m(1,1,1,1,1,1) = 1 + m(5,4,3,2,1);
\eqno {\rm (325.)}$$
$$m(5,4,3,2,1) = 1 + m(14,9,5,2) = 922;
\eqno {\rm (326.)}$$
and therefore
$$m(1,1,1,1,1) = 47,
\eqno {\rm (327.)}$$
$$m(1,1,1,1,1,1) = 923.
\eqno {\rm (328.)}$$
\bigbreak
[20.]
The formula
$$m(1,1,1) = 5,
\eqno {\rm (307.)}\qquad$$
may be considered as expressing, generally, that in order to
satisfy a system of three homogeneous equations, rational and
integral, and of the forms
$${\rm A}' = 0,\quad {\rm B}' = 0,\quad {\rm C}' = 0,
\eqno {\rm (329.)}$$
that is, of the first, second, and third degrees, by a system of
ratios of $m$ disposable quantities
$$a_1,\ldots\, a_m,
\eqno {\rm (192.)}\qquad$$
which ratios are to be discovered by Mr.~Jerrard's method of
decomposition, without any elevation of degree by elimination,
the number~$m$ ought to be at least equal to the minor limit
{\it five\/}; a result which includes and illustrates that
obtained in the 4th article of the present communication,
respecting Mr.~Jerrard's process for {\it taking away three
terms\/} at once from the general equation of the $m^{\rm th}$
degree: namely that this process is not generally applicable when
$m$ is less than {\it five}. Again, the process described in the
eleventh article, for taking away, on Mr.~Jerrard's principles,
{\it four terms\/} at once from the general equation of the
$m^{\rm th}$ degree, without being obliged to eliminate between
any two equations of condition of higher degrees than the first,
was shown to require, for its success, in general, that $m$
should be at least equal to the minor limit {\it eleven\/}; and
this limitation is included in, and illustrated by, the result
$$m(1,1,1,1) = 11,
\eqno {\rm (317.)}\qquad$$
which expresses generally a similar limitation to the analogous
process for satisfying any four homogeneous equations of
condition,
$${\rm A}' = 0,\quad
{\rm B}' = 0,\quad
{\rm C}' = 0,\quad
{\rm D}' = 0,
\eqno {\rm (330.)}$$
of the first, second, third, and fourth degrees, by the ratios of
$m$ disposable quantities, $a_1, a_2,\ldots, a_m$. In like
manner it is shown by the result
$$m(1,1,1,1,1) = 47,
\eqno {\rm (327.)}\qquad$$
that Mr.~Jerrard's general method would not avail to satisfy the
five conditions
$${\rm A}' = 0,\quad
{\rm B}' = 0,\quad
{\rm C}' = 0,\quad
{\rm D}' = 0,\quad
{\rm E}' = 0,
\eqno {\rm (331.)}$$
and so to take away {\it five terms\/} at once from the equation
of the $m^{\rm th}$ degree, without any elevation of degree being
introduced in the eliminations, unless $m$ be at least $= 47$,
that is, unless the equation to be transformed be at least of the
47th degree; and the result
$$m(1,1,1,1,1,1) = 923,
\eqno {\rm (328.)}\qquad$$
shows that the analogous process for taking away {\it six
terms\/} at once, or for satisfying the six conditions
$${\rm A}' = 0,\quad {\rm B}' = 0,\quad {\rm C}' = 0,\quad
{\rm D}' = 0,\quad {\rm E}' = 0,\quad {\rm F}' = 0,
\eqno {\rm (332.)}$$
is limited to equations of the 923rd and higher degrees.
Finally, the result
$$m(1,0,1,1) = 7,
\eqno {\rm (316.)}\qquad$$
and the connected result
$$m(1,0,1,0,1) = 7,
\eqno {\rm (333.)}$$
show that it is not in general possible to satisfy, by the same
method, a system of three conditions of the first, third, and
fourth degrees, respectively, such as the system
$${\rm A}' = 0,\quad
{\rm C}' = 0,\quad
{\rm D}' - \alpha {\rm B}'^2 = 0,
\eqno {\rm (334.)}$$
nor a system of 3 conditions of the first, third, and fifth
degrees,
$${\rm A}' = 0,\quad
{\rm C}' = 0,\quad
{\rm E}' = 0,
\eqno {\rm (335.)}$$
unless $m$ be at least $= 7$; which illustrates and confirms the
conclusions before obtained respecting the inadequacy of the
method to reduce the general equation of the fifth degree to
De~Moivre's solvible form, or to reduce the general equation of
the sixth to that of the fifth degree.
\bigbreak
[21.]
Indeed, if {\it some} elevation of degree be admitted in the
eliminations between the auxiliary equations, the minor limit of
the number~$m$ may sometimes be advantageously depressed. Thus,
in the process for satisfying the system of equations (330), we
first reduce the original difficulty to that of satisfying, by
the ratios of $m - 1$ quantities, a system containing three
equations of the first degree, two of the second, and one of the
third; and we next reduce this difficulty to that of satisfying,
by the ratios of $m - 2$ quantities, a system containing five
equations of the first, and two of the second degree. Now, at
this stage, it is advantageous to depart from the general method,
and to have recourse to ordinary elimination; because we can thus
resolve the last-mentioned auxiliary system, not indeed without
{\it some\/} elevation of degree, but with an elevation which
conducts no higher than a biquadratic equation; and by avoiding
the additional decomposition which the unmodified method
requires, we are able to employ a lower limit for $m$. In fact,
the general method would have led us to a new transformation of
the question, by which it would have been required to satisfy, by
the ratios of $m - 3$ new quantities, a system containing six new
equations of the first, and one of the second degree; it would
therefore have been necessary, in general, in employing that
method, that $m - 3$ should be greater than $6 + 1$, or in other
words that $m$ should be at least equal to the minor limit {\it
eleven\/}; and accordingly we found
$$m(1,1,1,1) = 11.
\eqno {\rm (317.)}\qquad$$
But when we dispense with this last decomposition, we need only
have $m - 2 > 5 + 2$, and the process, by this modification,
succeeds even for $m =$ {\it ten}. It was thus that
Mr.~Jerrard's principles were shown, in the tenth article of this
paper, to furnish a process for taking away four terms at once
from equations as low as the tenth degree, provided that we
employ (as we may) certain auxiliary systems of conditions, (160)
and (161), of which each contains two equations of the second
degree, but none of a degree more elevated. But it appears to be
impossible, by any such mixture of ordinary elimination with the
general method explained above, to depress so far that lower
limit of $m$ which has been assigned by the foregoing discussion,
as to render the method available for {\it resolving\/} any
general equation, by reducing it to any known solvible form.
This {\it Method of Decomposition\/} has, however, conducted, in
the hands of its inventor Mr.~Jerrard, to several general {\it
transformations\/} of equations, which must be considered as
discoveries in algebra; and to the solution of an extensive class
of problems in the analysis of {\it indeterminates}, which had
not before been resolved: the {\it notation}, also, of
{\it symmetric functions}, which has been employed by that
mathematician, in his published
researches\footnote*{{\it Mathematical Researches}, by
George B. Jerrard, A.B., Bristol; printed by William Strong,
Clare Street; to be had of Longman and Co., London.}
on these subjects, is one of great beauty and power.
\bye