On the metric theory of nearest integer continued fractions

Abstract

The problems empirically studied by the author [1] respectively, where G = (1+x/5)/2. The numerical valu.es given for re(t) are quite accurate; the largest error is 0.64.10-5, occurring for t = 0.14. The values for m'(t) differ by up to 3 units in the last decimal, except the value for m'(0), which differs by 4 units. The exact value of the constant In K' in (38) and (41) is u2/121n G = 1.709...The ratio 70 ~ at the end ofsubsection 17 is In G/ln 2 = 0.6942... H. Riese••