Preface xv
makes the basic ideas of Markoff look more natural. With Jean Bernoulli sequences
and the formulas for Jean Bernoulli and Markoff periods one can easily calculate
numerically as many points of the Lagrange spectrum as necessary.
Markoff’s theory completes the algebraic part of this book. The analytic part begins
in Chapter 3 and is followed in Chapter 4 by Euler’s research. To a great extent
this chapter covers Euler (1750b) but with the difference that Brouncker’s method
recovered in Chapter 3 is applied. The method can be extended from the unit circle
considered by Wallis to the class of sinusoidal spirals introduced into mathematics
in 1718 by another great British mathematician, Colin Maclaurin. Chapter 4 covers
the forgotten Euler differential method of summation of some continued fractions of
hypergeometric type. It turns out that, for instance, approximately half the continued
fractions discovered later by Stieltjes can be easily summed up by this method of
Euler. However, the differential method does have some limitations. Attempting to
overcome them Euler arrived at the beautiful theory of Riccati equations. In (1933)
Sanielevici presented Euler’s method in a very general form. Later Khovanskii (1958)
using Sanielevici’s results developed as continued fractions many elementary functions.
Still, I think that Euler’s method as stated by Euler makes everything more clear. In
this part I filled some gaps in the proofs while trying not to violate Euler’s arguments.
The central result here is the continued fraction for the hyperbolic cotangent. I collect
in Chapter 5 some results which were or could be directly or indirectly influenced by
Euler’s formulas. Such an approach sheds new light on the subject.
Chapter 6 presents results either obtained by Wallis interpolation or by a direct
transfer from the regular continued fractions of number theory to polynomial con-
tinued fractions, i.e. to P-fractions. Euler’s results on hypergeometric functions play
a significant role here. Another interesting topic of this chapter is the periodicity
of P-fractions. As before, the first results were obtained by Euler. Using contin-
ued fractions of the radicals of quadratic polynomials as guidance, Euler found his
now well-known substitutions for integration. This was extended by Abel in one
of his first papers, which incidentally preceded his discoveries in elliptic functions.
I include in this chapter a beautiful result of Chebyshev on integration in finite
terms.
Chapter 7 indicates how Euler’s ideas eventually led to the discovery of orthogonal
polynomials. Finally, I present in Chapter 8 my own research on the convergence of
Schur’s algorithm.
A few words on the title. It varied several times but since the essential part of the
book is related to Euler I believe that finally I made a good choice. Moreover, in 2007
the tercentenary of Euler was celebrated.
Following Euler, I have split the book into small numbered “paragraphs” (sub-
sections). This was an old tradition in mathematics, now almost forgotten. It makes
the book easier to read. The difference from what Euler did is that almost all the
“paragraphs” also have titles.