Page:On the Difficulty of Correct Description of Books - De Morgan (1902).djvu/16

 lande cites it (p. 334 ). We then turned to Weidler's Bibliography, and here we really find that the Nuremberg quarto of 1522 is said to contain five treatises, the three given by Lalande, with two others by Werner, not any of those yet named. And Weidler refers to p 334 of his own book, in which, as already seen, he gives a very different and more correct account. So that the confusion is as follows. Weidler describes the book in his History with nothing but an omission. In his Bibliography he gives a totally wrong description, for which he refers to his own more correct History. Lalande adopts the account given in the Bibliography, and joins to it the reference to the History, without stating that his reference to the History is only a copy from the Bibliography. No one, without the book before him, could have unravelled this skein of mistakes. We took the word of Lalande because it is decidedly the best piece of scientific bibliography which, at its appearance, had ever been in existence, and therefore gave the best chance of a correct description. But, like other descriptive works which make a commencement of correctness upon books which the authors had examined for themselves, it relies in a great degree upon works prior to the introduction of any effort at minute description.

In the last instance, it happens that the mistake can be traced to its source in a manner which leaves no doubt that it is a mistake. But the unpracticed reader must not come to such a conclusion too rapidly. If Lalande had not named his authority, as often happens with him, we should have had three alternatives to consider. 1. A mere mistake. 2. The circumstance of his having happened [8] to fall in with a book in which someone had bound together some astronomical tracts of Werner with a copy of Regiomontanus's epistle. 3. The possibility that Werner made two distinct publications at Nuremberg in 1522, one