Far-reaching Hellenistic geographical knowledge hidden in Ptolemy's data (Q1785703)

This article deals with Ptolemy's \textit{Geographia} and the famous errors of calculation of the dimension of the Earth and longitudes (otherwise both related). In order to explain these errors, the author summarizes and discusses the main ideas presented in his previous book [L'America dimenticata: I rapporti tra le civiltà e un errore di Tolomeo (Italian). 2nd ed. Milano: Mondadori Università (2013)], incorporating more recent results on Ptolemy's sources. He shows an excellent knowledge of the subject introducing the problem in its historical, scientific and historiographical context, relying on his arguments in [Math. Mech. Complex Syst. 1, No. 1, 67--79 (2013; Zbl 1341.01006)] in particular. The text under review is widely documented with personal English translations from Greek (\textit{Geographia} from Ptolemy) and Latin (\textit{Naturalis historia} from Pliny) texts following the editions of reference. First, the author questions the links between Ptolemy's both errors. To achieve his goal, he comes back on the path that Ptolemy would enable to follow to arrive at his results, especially from the estimation of the longitudinal amplitude of the \textit{oikoumene} (ancient Greek term for the inhabited part of the world), spread over \(180^\circ\) of longitude, from the westernmost locations to the easternmost ones. Thus, the author strives to demonstrate the following proposition: ``Ptolemy's error on the amplitude in longitude of the oikoumene was the cause of his systematic dilation of the differences of longitude and of his deduction, from the known data for the distances along circles of latitude, of an underestimation of the dimension of the Earth.'' (p. 187) The following part ``Global and local errors'' is interesting. The author shows that Ptolemy used a series of data ``interconnected by a very complex chain of relations'' (p. 189). Thereby, the longitudes computed by Ptolemy are not all the result of the same interpolation but they would come from interpolations corrected, according to the places, taking into account astronomical data (i.e. Ptolemy worked with local regression coefficients), especially for areas well-known in Hellenistic times. Thus, he grouped several cities/places into different regional subsets. After that, the author focuses on the westernmost locations of the \textit{oikoumene} considered by Ptolemy, that is to say the ``Islands of the Blessed'' (also called Fortunate Islands). This location is traditionally identified with the Canary Islands. Nevertheless, the latitude given by Ptolemy for this location is really different from the latitude of the Canary Islands. As a result, the author questions this identification. And, for him, the error of identification of the Islands of the Blessed would be the origin of both aforesaid errors: that of the measurement of the Earth, and that of the longitudes. He convincingly (computation of coordinates, references to geographical descriptions of the time) proposes a new identification: this time with Caribbean Islands! (so far away from the Canary Islands). And, finally, if so, it is an argument to propose Hipparchus as a source for Ptolemy's work.