of Mezzofanti, but of all other special pursuits as well. And the instance of which we have spoken is a better illustration than any blunder which detects itself by its own absurdity. The English or the Gaelic scholar will not be deceived by the cases we have quoted: the worst that can happen is, that the inquirer who looks under Ceylon for Cingalese misses a book. But if, as might possibly happen, Schooten's Descartes of 1649 were entered as what it really is, the second edition; and if that of 1659 made part of a set of Descartes's works, as it often does; and if, as is very common, the Opera omnia were insufficiently detailed in the catalogue; and if, as generally has happened, a mathematical historian were somewhat easy on the point of bibliography—the works of Hudde, &c. might disappear from history, as other works have done in a similar way, without disappearing from libraaries.
The fourth book in question is another work of Maurolycus, the Opuscula Mathematica, Venice, 1575, 4to. On the title it appears that this collection was then published for the first time. It consists of a collection of tracts, and a work on arithmetic with a second title page of the same place and date as the first. Here, as often happens, is a source of confusion: in binding, these works are separated, each title page being made the beginning of a separate work. Two things are very common at the date now before us, the binding up of different publications in one, and the distribution of one publication under different title pages, often without any mark by which to know that all the titles belong to one work. Hence catalogues sometimes represent different publications as one, and sometimes represent one publication under several heads; the binder being the authority in both cases. On looking (14) more narrowly, to see whether the work itself gives any information on this point, we find, on the verso (15) of the first title page,
12
