THE SPHERICS OF THEODOSIUS Theodosius, Sphaerica: Arabic and Medieval Latin Translations. Edited by Paul Kunitzsch and Richard Lorch (Boethius, lxii; Franz Steiner Verlag, Stuttgart, 2010). Pp. 431. euro64. ISBN 978-3-515-09288-3.Theodosius's classic treatise on the geometry of circles on the surface of a sphere, The Spherics, offered its Greek readers a systematic, logical development of all the basic theorems in spherical geometry as well as a number of theorems applicable mainly to spherical astronomy. Theodosius wrote the work, probably, in the first part of the first century b.c., and its adoption as a teaching text by Alexandrian scholars ensured its survival into the Islamic, and then Latin, Middle Ages. The numerous references to Arabic manuscripts of all or part of the work in vol. ? of F. Sezgin's Geschichte des arabischen Schrifttums testify to its continuing study in the Islamic world well into the eighteenth century.In 2000, Claire Czinczenheim produced a definitive edition of the Greek text, together with French translation and commentary. Now, two noted scholars and long-time collaborators have combined their efforts to produce editions of Arabic and medieval Latin texts of Theodosius's Sphaerica. As a result, we now have reliable editions of a major scientific work in all languages in which it appeared while it was still of scientific importance.The editors begin with an account of the still-murky history of the Arabic translation of the Spherics. They point out that there were two different translations into Arabic, one represented by three mss (A, N, and H - the last-mentioned being Arabic written in Hebrew script) and the other represented by manuscripts F and C (both of these also in Hebrew script). FC specifically ascribes the translation to Qusta ibn Luqa, but this simple ascription is qualified, or contradicted, by other versions of the text. Other than giving what evidence there is, Kunitzsch and Lorch take no definite position on this matter.For reasons that they do not explain, the editors chose to edit the Arabic text represented by three manuscripts, denoted collectively as ANH, which was done by comparing manuscripts A (Istanbul) with N (private library) and, in doubtful cases, with the Greek text. (Only in such cases - most of which, it seems, occur early in Book I - have the editors recorded the reading of H.) The editorial work has been very carefully done. With the text and its critical notes it is possible to know exactly the readings of texts of A and N.The Latin text presented here is a transcription of the text found in one manuscript, P (Paris), the shorter of two that circulated in medieval times. It is an indication of the scholarly care that has gone into the production of this work that the authors collated P with ten other manuscripts. The editors note that P is clearly a translation of the ANH text, and they argue on stylistic grounds that it was done by Gerard of Cremona who, in any case, is known to have done a translation.As the editors point out (p. …