Those who take it upon themselves to lay down that there "must have been" an edition of that place and date are bound to establish the fact. They are not entitled to defy every rule of evidence, and to call on the other side to prove a negative. The burden of proof lies wholly with them. But, by a rare and happy accident, it is possible to prove a negative in the present case. In view of recent researches, the theory that the _princeps_ of the _Galatea_ was issued at Madrid in 1584 is absolutely untenable. All doubts or hesitations on this head are ended by the opportune discovery, due to that excellent scholar and fortunate investigator, Dr. Pérez Pastor, of the original contract between Cervantes and the Alcalá publisher, Blas de Robles. By this contract Blas de Robles binds himself to pay 1336 _reales_ (£29. 13s. 9d. English) for the author's entire rights.[25] This legal instrument is decisive, for it would be ridiculous--not to say impertinent--to suppose that Cervantes sold his interest twice over to two different publishers in two different cities. There can, therefore, be no further controversy as to when and where the _Galatea_ appeared. It is now placed beyond dispute that Cervantes had not found a publisher before June 1584, and that the book was issued at Alcalá de Henares in 1585--probably not before the month of April. The first intention was to entitle the volume _Los seys libros de Galatea_ but (perhaps with a view to emphasizing the promise of a sequel) it was actually published as the _Primera Parte de la Galatea, dividida en seys libros_.[26] On June 14, 1584, Cervantes received 1116 _reales_ in advance, and, by a deed of the same date, Blas de Robles undertook to pay the balance of 250 _reales_ at the end of September:[27] the very period when, as already conjectured, the printing was begun.[28]