It must be observed, however, that in some cases a more limited reference seemed desirable, especially on matters of form and spelling, points about which it would be idle to adduce any evidence but that of a few copies. Where selection of this kind is employed, the manuscripts on both sides are cited: thus such notes as ‘3691 set AJ, S, F sette C, B,’ ‘4307 all S, F alle AJ, B’ must not be taken to imply the reading of any copy except those mentioned. In a few cases this form is used to avoid misunderstanding in passages where the record of readings is for some reason incomplete, as i. 2300, viii. 566, 1713, 1927. In citing a variation as given by a class or group of MSS. no attempt is made to give the spelling of each one separately. The form cited is that given either by the majority or by a leading MS. with variations sometimes added in parentheses. Attention should be paid also to the following points: (1) It was not found possible to complete the collation of the Glasgow MS. (G) before the text was printed, and consequently its readings must not be taken as implied, when not mentioned, any further than v. 1970. The collation has since been completed and some of the results are noted in the account of the MS. (2) K takes the place of H₃ in vi. 1671-vii. 1405, and vii. 3594 to the end, where H₃ is defective. (3) Before assuming the evidence of any MS. _ex silentio_ it is necessary that the reader should assure himself that it is not defective in the part concerned. The means of doing this are fully afforded by the accounts given of the separate MSS., where their imperfections are noted, and it must be remembered that J and Ad are for the most part defective as regards the Latin summaries, and that this is the case with T also in certain parts. The readings of S on f. 50 are for the most part passed over, as not originally belonging to that MS. (4) A few abbreviated Latin terms are used in the critical notes, as _in ras._ to indicate that the text is written over an erasure, or _p.m._ to denote the reading of the first hand. The lines are numbered in each book (for the first time), and the numbers with an asterisk attached are those of the lines in other recensions than that of the text. In addition to this it should be observed that as nearly all references to Gower for the last forty years have been made by Pauli’s edition, it has been thought advisable to place in the margin of this text indications of the volumes and pages of that edition: thus =P. 1. 153= stands for ‘Pauli, vol. i. p. 153.’ Setting aside matters of spelling, punctuation and grammatical form, we may note that the material differences of reading between the text of this edition and that of Pauli are in number about two thousand.