36 racteristics of the Sanskrifc work ; but he too wrote to me that he could not trace the original* I then addressed Mahamaho- p&dhystya Sudhslkara Dvivedi, Professor of Mathematics, Government Sanskrit College, Benares, and author of Qanaka- taraiigini and other works, requesting him to let me know whether he had any arguments in support of what he advanced in the Ganakatarangini— * ^<4)4»md: «^ ^qi l Hgq> ^^[1^- In reply the S'&stri wrote to me that he had ap Arabic work which seemed to him to be the original of the Bekh&ganita and on my requesting him to lend me the work, he was good enough to send it to me. The Arabic workf contains all the fifteen books. On comparing it with the Kekhfiganita I find in it all the striking features of the latter noticed above. The forty- seventh proposition of the First Book is proved in seventeen ways4 The book contains all the propositions preparatory to the 29 th Prop, of the first book that are found in the Sanskrit text. No authorities are given in the proof of proposi- tions. There is not a shadow of doubt that this Arabic work is the original of the Rekh&ganita. The work does not mention the name of the author, but the Preface, which runs as under, makes it clear that the author is the same scholar that has composed 'M&jisti': —
- The foUowing is the reply received: —
Bbitish Museum, London: W, C. March 8, 1899, Dear Sir, I am afraid that we can not help you in your search for the original of the Rekh&ganita. We have nothing in Arabic which appears to be at all likely, and we do not possess a copy of James Williamson's work. Yours faithfully, ( Signed ) ROBERT K. DOUGLAS. t I found another Arabic work containing all the fifteen books with Prof. Isfah&ni, but it contains no alternative proofs and is not the original of the Rekhdganita. I PerigaVs * Messenger of Mathematics ' New Series, Vol. III. p. 104 contains alternative proofs of Book I. Prop. 47.
