20 Introduction 



My account of the Archimedian tradition is incompleted^ but suffi- 

 cient to illustrate many features, the various contingencies, riskiness and 

 at best the complexity of such traditions. A Greek text is known to us 

 by a MS. preserving it, or by extracts from it or references to it by later 

 writers; or by Arabic, Hebrew or Latin versions, commentaries, extracts; 

 or by references in each ( or all ) of these languages. The paradoxical 

 aspects of tradition are evidenced by the fact that the study of Arabic is 

 now, all considered, the most promising method to increase our knowl- 

 edge of Greek science! 



Thoughtful readers may well ask themselves two questions: (I) If 

 the tradition is so full of risks and adventures, how were any texts pre- 

 served, especially mathematical texts which could never interest more 

 than a few people? (2) Considering those risks and vicissitudes, how 

 can we be sure that the texts which have survived are really what they 

 are claimed to be? 



The two questions are pertinent and sufficiently ticklish to be stimu- 

 lating. If one bears in mind the number of wars, conflagrations and 

 other calamities which have occurred in the Mediterranean world since 

 Archimedes' death, how did any one of his writings escape destruction 

 and oblivion? When Archimedes composed one of them, say the Epho- 

 dion or the Ochumena, the number of students directly interested in it 

 must have been exceedingly small and that number remained small 

 throughout the ages. It is unlikely that the "first edition" issued by the 

 Master himself included many copies. Perhaps a dozen or even less. 

 Some of those copies found their way to the libraries of Alexandria and 

 Pergamon, but those libraries were destroyed. We have relearned quite 

 recently that the safest libraries are not absolutely safe, and the greater 

 they are, the greater the loss in case of destruction. Other copies were 

 preserved in private libraries, e.g., in the libraries of Archimedes himself, 

 of the king of Syracuse Hieron and his son Gelon, of Archimedes' 

 friends, Dositheos of Pelusion, Conon of Samos and Eratosthenes 

 OF Gyrene (III-2 B.C.), but how insecure they were! Did a copy pre- 

 served by the tyrant of Syracuse have a great chance of survival? And 

 as to Archimedes himself and his friends, these men were probably poor, 

 they were certainly not rich, but even if they had been rich enough to 

 live in palaces, what of it? Are any of the private palaces of antiquity 

 extant? Have their contents come down to us? How then did the 



edited in Arabic and German by Max Simon (2 vols. Leipzig 1906). Galen on 

 medical experience was first published in Arabic and English by Richard Walzer 

 (London 1944; Isis, 36, 251-55). 



" Complete accounts of the tradition of a text are generally given by the mod- 

 ern editors. Such accounts include a discussion of the relative trust which may be 

 placed in each MS. of the original text or of its translations, and in the early edi- 

 tions. The filiation of those MSS is symbolized by a genealogical tree or stemma. 

 For Archimedes see Heiberg's edition (2nd ed., 3 vols. Leipzig 1910-15) or the 

 English translation by T. L. Heath (Cambridge 1897), with supplement (Cam- 

 bridge 1912). 



