This article has two purposes. First, it proposes a more satisfactory solution to an old problem: the apparently serious inaccuracy of Thucydides¿ measurements for the length of Sphakteria island and the width of the channels dividing it from the mainland. Second, it offers some more general observations on Thucydides¿ measures of distance and the light they can shed on an important aspect of his historiographic method. The solution proposed by R. Bauslaugh (¿The text of Thucydides IV 8.6 and the south (...) channel at Pylos¿, JHS 99 (1979) 1-6) to the problem of measurements is rejected. Bauslaugh had emended two of the three figures on the ground that they were so seriously inaccurate as to require assumption of manuscript corruption. It is here contended that his argument is misconceived, and the emendations unnecessary. The counter-argument is based on a close study of Thucydides¿ idiom and practice in giving measurements of distance, particularly his use of qualifying expressions with numbers of this kind. The second half of the article uses data compiled in an ongoing study of the use of numbers by Greek historians to make some comparisons between Thucydides¿ practice and that of several other historians in giving measurements of distance. It is suggested that careful attention to the nuances of Thucydides¿ practice, especially his use of different qualifying expressions with these numbers, may enable one to draw some interesting inferences about his sources of information and how he used them. (shrink)