^{page 42 note 1} See Klussmann, Excerpta Tertullianea. Hamburg, 1892; Kiibler in Herm. 25, 496; Dressel in Riv. Filol. 3, 207.Google Scholar

^{page 43 note 1} ‘Isidorus Junior’ is generally supposed to be Isidore himself. May it not be rather to be identified with Julianus Toletanus, the re-caster of some of his predecessor′s works ? One of the earliest extant Spanish MSS. (Autun 27) contains the Quaestiones of Isidore in half-uncial script, followed by extracts from the Expositione whose author was previously unknown to us. They are in minuscule script (presumably later than the half-uncial part) and are headed ‘Isidori Junioris sententias intexuimus ’ (cf. Athenaeum of April 2, 1910). The re-casting of Etym. I xxxiii sq. in an eighth-century Fulda MS. at Basle (F III I5d) is entitled ‘ Liber ysidori iunioris de uiciis.’

^{page 43 note 2} For the provenience of the Glossarium Ansleubi this clue is useful. In the two MSS. (of Saec. viii-ix) written in what Traube calls ‘ the old script of Corbie’ (Paris 11529–11530;Cambrai 653) I found (to borrow another phrase of Traube′s) ‘ Spanish symptoms’ only in the Isidore glosses (e.g. in the Paris MS. dais ‘ dominus’ under the lemmas ‘Accubuit’ and ‘ Acyrologia’ on fol. 5, epscpi ‘episcopi’ on foil. 42V., 49r,, nsi ‘nostri’ on foil. 27V., giv., sedes ‘sedens’ and appellat ‘-ant’ on fol. 73V., the Spanish ligatured ‘ it ’ in ‘Constituz′i’ fol. 71r.;in the Cambrai MS. nsi on fol. 38V., constas ‘-ans’ on fol. 43V., the Spanish spellings nicli, quaquumque on foil. 23r., 32r.). Therefore the Isidore portions were taken by Ansileubus from a Spanish codex;but the Glossary itself is not of Spanish provenience.

^{page 43 note 3} Inperitiam must surely be right. Then ‘dicite labdae’ in Varro Men. 48 (Bue.) will mean ‘call them blockheads,’ like ‘ stultos contemnite docti’ in the following line.

^{page 44 note 1} Another Greek ‘ghost-word’ is (12 1. 351 15. 8. 5)- The source of the error appears when we read Paul. Fest, 30, 37 Th.

^{page 44 note 2} The evidence seems strong against the theory that dictation was ever practised in a monastery scriptorium. Silence was the rule there.

^{page 45 note 1} Both were written from the same original at the same time, perhaps the beginning of the eighth century, and in the same place, the scriptorium of Bobbio. Although the Vatican transcript has an eleventh-century entry stating that Boniprandus presented it to Bobio, the presentation (the word is ‘obtulit’) must have been rather a restoration; for a palimpsest in this North Italian script can hardly have been written anywhere else. Similarly the Vatican Papyrus no. 21, which passed into a private library, but was restored to the Vatican in 1821, s called a ‘gift’ in the recording entry (‘nobilium Gualteriorum donum’).

^{page 45 note 2} At Madrid, Bibl. Nac, in Tol. 15, 8 and Tol. 15, 9; Bibl. Acad. Hist, in 25 and 76. At the Escurial, in & I. 14, P I.7, P I. 8, T (formerly Q) II. 24. And presumably in the others.

^{page 45 note 3} Let me mention here its interesting gloss at 1. 39. 18 scenis id est laubia. This is Ital. ‘;loggia’ (cf. Goetz, This. Gloss., s. v.).

^{page 46 note 1} Dr, Holder, the Carlsruhe Librarian, de-scribes this MS. in Melanges Chatelain, pp. sqq. He regards the weird misspellings as evidence that MSS. were occasionally copied to dictation. But many of them occur in other MSS. of this family, and some (e.g. the persistent -ct for -e) may be referred to the unfamiliar script of the original.

^{page 48 note 1} Visigothic script has not been adequately studied by palaeographers, and sound date-tests are not yet wholly available. But Dr. Loew seems right in refusing to Escor. T (formerly Q II 24 an earlier date than the tenth century, although its original was written in the year 743.

^{page 48 note 2} Both Arevalo and Otto print the corrupt reading ‘ Petra(e) et rupes.’ Could there be clearer proof of the necessity of a new edition ?

^{page 49 note 1} Probably words were often abbreviated by ‘suspension.’ In 1, 27, 26 the divergences of our MSS. (scribendum, scribenda, scribitur) may all be referred to the suspension scrlb (probably for ‘scribendum’) in the parent MS. So that an ungrammatical case of a noun or person or tense of a verb may not be always an error of Isidore.

^{page 49 note 2} Not, e.g. at 15, 12, 4 where N comes nearest of all MSS. to the truth (magalia magaria) with ‘magaria et magaliam.’

^{page 49 note 3} Did the clever emendation in the original of T at 16, 4, 1 come from a MS. or from the scribe {or corrector) himself? The true reading is in Ida. All our MSS. have ‘in India,’ but there was a marginal query in the original of T ‘ an Ida ?’

^{page 49 note 1} Escor. P I 7 incurs suspicion by its re-casting of the opening parts of Book I. From 1,2, 2 fin. it passes to 1, 3, 3, thence to j,4,2.It places 1, 3, 1–2 and 1, 3, 4–4, 1 (followed by 1, 4, 16 med. –18 fin.) after 1, 4, 15. (St. Gall. 237, a tenth-century MS. of this family places chaps. 3–4 of Book I after chap. xiv.). The MS. belonged to the library of King Alfonso (usually identified with Alfonso II ‘el Casto,’ 791–842, but by Dr. Loew with Alfonso III, 866–910). Was the re-casting done by one of the King′s teachers? Many of the scholia added on its margins refer to matters of orthography, e.g. fol. 281V. ‘ cedrus um a scribi oportet.’ On fol. 276V. Isidore′s pis aurum dicitur′ is challenged: ‘Require alibi utrum pis aurum dicatur quia apud grecos aurum crusapi dicitur.’