Steps / Historical AI

Leonhard Euler · 1770 (1828 English ed.) · Internet Archive

Elements of Algebra

All 627 leaves · leaf 554

Transcription

CHAP. II. ADDITIONS. 517 the equation $Q^v = \mu^{iv} P^{iv} + Q^{iv}$, and then $P^v$ from this, $\frac{(Q^v)^2 - P^{iv} P^v}{P^{iv}} = \frac{1}{2}E$. Whence it is easy to draw this general conclusion, that, if $p^\lambda$ and $p^{\lambda+1}$ are the leading terms of the series $p', p'', p''', \ldots$ which are successively found with different signs, the term $p^{\lambda+1}$, and the following, will all return, after a certain number of intermediate terms, and it will be the same with the term $p^\lambda$, if we have $\frac{\pm p^\lambda}{p^{\lambda+1}}$ = or < 1. For let us imagine, as in Art. 34, that we have found $p^{\pi+2\xi} = p^\pi$, and $Q^{\pi+2\xi} = Q^\pi$, and suppose that $\pi$ is $>\lambda$, that is to say, $\pi = \lambda + \nu$; wherefore we may go back, on the one hand, from the term $p^\pi$ to the term $p^{\lambda+1}$, or $p^\lambda$, and on the other, from the term $p^{\pi+2\xi}$ to the term $p^{\lambda+2\xi+1}$, or $p^{\lambda+2\xi}$; and, as the terms from which we set out are equal on both sides, all the terms derived from them will likewise be respectively equal; so that we shall have $p^{\lambda+2\xi+1} = p^{\lambda+1}$, or even $p^{\lambda+\xi} = p^\lambda$, if $\frac{\pm p^\lambda}{p^{\lambda+1}}$ = or < 1. We may, therefore, judge beforehand of the beginning of the periods in the series $p^0, p', p'', p''', \ldots$ and consequently in the other series also, $q^0, q', q'', q''', \ldots$ $\mu, \mu', \mu'', \mu''', \ldots$ but as to the length of the periods, that depends on the nature of the number E, and entirely on the value of that number, as I could demonstrate, were I not afraid of being led into too long a detail. 37. Corollary 3. What we have demonstrated in the preceding corollary, may serve to prove the following theorem: Every equation of the form $p^2 - \kappa q^2 = 1$, (in which $\kappa$ is a positive integer number, but not a square, and $p$ and $q$ two indeterminate numbers) is resolvible in integer numbers. For, by comparing the formula $p^2 - \kappa q^2$ with the general formula, $Ap^2 + Bpq + cq^2$, we have $A=1$, $B=0$, $c=-\kappa$; wherefore $E = B^2 - 4AC = 4\kappa$, and $\frac{1}{2}\sqrt{E} = \sqrt{\kappa}$ (Art. 33). Wherefore, $p^0=1$, $q^0=0$; likewise $\mu < \sqrt{\kappa}$, $q'=\mu$, and $p' = \mu^2 - \kappa$; whence we see first, that $p'$ is negative, and consequently has a different sign from $p^0$; secondly, that $-p'$ is = or > 1, because $\kappa$ and $\mu$ are integer numbers; so that we shall have $\frac{p^0}{-p'}$ = or < 1; whence we shall find, from the preceding Article, $\lambda=0$, and $p^\xi = p^0 = 1$;
confidence 0.88 50 expressions source scan at the Internet Archive

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.