Leonhard Euler · 1770 (1828 English ed.) · Internet Archive
Elements of Algebra
All 627 leaves · leaf 535
Transcription
498
ADDITIONS.
CHAP. II.
and are $p$ in number, it is evident that $0$ must be among those remainders, and, consequently, that there is one of the numbers $q \pm 1$, $2q \pm 1$, $3q \pm 1$, &c. $pq \pm 1$, which is divisible by $p$. Now, it is evident that this cannot be the last; so that there is certainly a value of $r$ less than $p$, which will make $rq \pm 1$ divisible by $p$; and it is evident, at the same time, that the quotient will be less than $q$; therefore there will always be an integer and positive value of $r$ less than $p$, and another similar value of $s$, and less than $q$, which will satisfy the equation $s = \frac{qr \pm 1}{p}$, or $ps - qr = \pm 1$.
24. The question is therefore now reduced to this; to find four positive whole numbers, $p, q, r, s$, the last two of which may be less than the first two; that is, $r < p$, and $s < q$, and such, that $ps - qr = \pm 1$; farther, that the quantities $p-aq$, and $r-as$, may have different signs, and, at the same time, that $r-as$ may be a quantity greater than $p-aq$, abstracting from the signs.
In order to simplify, let us denote $r$ by $p'$, and $s$ by $q'$, so that we have $pq' - qp' = \pm 1$; and as $q > q'$ (hyp.), let $\mu$ be the quotient that would be produced by the division of $q$ by $q'$, and let the remainder be $q''$, which will consequently be $< q'$; also, let $\mu'$ be the quotient of the division of $q'$ by $q''$, and $q'''$ the remainder, which will be $< q''$; in like manner, let $\mu''$ be the quotient of the division of $q''$ by $q'''$, and $q^{iv}$ the remainder $< q'''$, and so on, till there is no remainder; in this way, we shall have
$q = \mu q' + q''$
$q' = \mu' q'' + q'''$
$q'' = \mu'' q''' + q^{iv}$
$q''' = \mu''' q^{iv} + q^v$, &c.
where the numbers $\mu, \mu', \mu''$, &c. will all be integer and positive, and the numbers $q', q'', q'''$, &c. will also be integer and positive, and will form a series decreasing to nothing.
In like manner, let us suppose
$p = \mu p' + p''$
$p' = \mu' p'' + p'''$
$p'' = \mu'' p''' + p^{iv}$
$p''' = \mu''' p^{iv} + p^v$, &c.
And as the numbers $p$ and $p'$ are considered here as given, as well as the numbers $\mu, \mu', \mu''$, &c. we may determine from these equations the numbers $p'', p''', p^{iv}$, &c. which will evidently be all integer.
This text was produced by machine and has not been corrected by hand. Where it matters, read the scan. Confidence is the model's own, not an external check.