^{page 146 note 1} The process is particularly clear in places where the scribe, by a not uncommon inadvertency, has recopied half a page or so of his own copy under the impression that it was his exemplar. In these cases the cancelled repetition still shows the uncorrected text.

^{page 147 note 1} From a noteworthy sentence in the dedicatory epistle of Aldus' Thucydides we learn that at that period he insisted on having at least three MSS. of an author before printing: ‘eramdaturus …τ⋯ τε Ξενοφ⋯ντος, κα⋯ Πλ⋯θωνος, … sed quia non habebam minimum tria exemplaria, distulimus in aliud tempus.’

^{page 147 note 2} Some time in the last century X was taken to pieces and rebound in three volumes with a blank sheet of paper between every two leaves. The original binding and guard-leaves were discarded. Such a procedure cannot be condemned too strongly. Binding and guard-leaves are often the only source for information about a manuscript's past.

^{page 147 note 3} See Baschet, : Aldo Manuzio, Venice 1867 no. XIXGoogle Scholar.

^{page 147 note 4} It is no. 40 in the list of his MSS. which Pantin sent to Meursius in a letter of March 1st 1611 (reprinted in H. Omont: Cat. des MSS. de la Bibl. roy. de Bruxelles, Ghent 1885 p. 46.)

^{page 147 note 5} Meursii opera XI, ep. 312, p. 223.

^{page 147 note 6} In Schott's classified catalogue (Omont, l.c. pp. 47–50)Google Scholar I appears three times under different contents: nos. 1, 22 and 50.

^{page 149 note 1} The progress of the error is as follows. κρατο⋯ντες τῇ ναυμαχ⋯ᾳ]οἱ II. Aldine; κρατο⋯ντες] τῇ ναυμ. οἱ II. Stephanus;^.1 κρατο⋯ντες οἱ πολ⋯μιοι] τῇ ναυμ. οἰ II. Poppo.