to use Sweet's expression. And for this reason it seems extraordinary that he should speak of Sievers and his intonation-theory in a manner that is to me insufficiently respectful. When he finds comfort in the thought that most people in Holland will agree with him, I must say that a scholar should be the first to advocate plural voting, in matters of scholarship; and if it be urged that Sievers would be in a minority even if he got 100 votes to one for every Dutch scholar, the question would still remain to be answered if the arrangement would be fair.... to Sievers. Do I advocate the introduction of Sievers's results into a book of this sort? No, we must leave it to time to decide what is to be permanent in Sievers's results. In the mean time there can be no harm if young beginners learn that among scholars, too, there are unapproachable Olympians.

In the chapter on Accidence I noted the definition of an aorist-present as a present with the vowel that belonged 'properly' to the aorist. The question is, no doubt, difficult; but if it is mentioned at all, it seems to me that it should be explained less mechanically. Perhaps the Greek verbs that have a different stem for the present and the aorist, like paw, eidov, would be helpful in realizing a stage when Indogermanic had not yet any tenses, but distinguished aspect only. The chapter on syntax is short; the writer declares that it is exclusively meant to serve a practical purpose. But as Gotic is not studied for its own sake, and is always the introduction to the study of some other Germanic language, it would have been useful and practical if syntax had been treated a little more fully. The study of the use of the definite article in Gotic makes it easy for the student of Old English to see in what respects its use has increased. The treatment of the optative in Gotic explains the use in Present English. I believe that the author here, as elsewhere, shows that he is not really a student of any living language. The texts are not accompanied by the Greek original; this would have been a great boon to those students who know Greek; it would have been useful, too, for those who study Gotic without knowing Greek: it would remind them that no independent study of Gotic is possible without some knowledge of Greek. The addition of the original would also have shown that the heading of the Gospels as on p. 200: aiwaggeljo pairh Marku anastodeip is a rendering of the Latin incipit, so that Wulfila may have used a copy with the Latin and the Greek text in parallel columns.

The get-up of the book is admirable, and worthy of the firm that publishes the series. The photographic reproduction of a page of the Codex Argenteus is welcome, but only a facsimile would have given an idea of the beauty of the manuscript. We may hope that a second edition will give an opportunity for this improvement.

E. KRUISINGA.

The Corpus, Epinal, Erfurt and Leyden Glossaries. By W. M. LINDSAY. (Publications of the Philological Society VIII). Oxford University Press, 1921. 15/--.

The Corpus Glossary edited by W. M. LINDSAY. With an Anglo-Saxon Index by HELEN MCM. BUCKHURST. Cambridge University Press, 1921. 40/-.

In 1890 our countryman Dr. J. H. Hessels published an edition of the Corpus Glossary, which not only contained an accurate reprint of the MS., but also corrected a great many of its corrupt readings. It did not, however,

fall within the scope of that edition to mention the texts or the collections from which the glosses had come and although in the mean time many of these were traced to their origin, by the investigations of Gruber, Goetz, Schlutter, Napier, Glogger, Hessels and others, the want of a book in which everything was collected that was known about them remained to be felt. Professor Lindsay, the famous Scottish Latinist, who was eminently qualified for supplying this want, has done so in the second of the books under discussion, which is meant to be a companion-volume to Dr. Hessels' previous edition, and together with it has published a separate volume to serve the purpose of Prolegomena. He here shows how the Corpus Glossary came into existence and sketches its relation to other early English glossaries, the more primitively arranged Epinal and Erfurt and the still ruder collection known as the Leyden Glossary. In this collection the items are arranged, not in alphabetical order, but mostly in sections according to the writings to which they refer, a favourable circumstance that enabled Prof. Glogger and Dr. Hessels to point out, in the majority of cases, the passages in which the lemmata occur. Now Prof. Lindsay made two important discoveries, firstly, that in Epinal and Erfurt (which are arranged partly in simple alphabetical order, partly in AB-order) in several instances the batches of Phocas-glosses, etc. may still be recognised and that the same is to some extent possible in Corpus (arranged in AB-order throughout), and secondly, that the compilers of Epinal, Erfurt and Corpus availed themselves of existing Continental glossaries, notably Abstrusa and Abolita, of which MSS. have been edited by Goetz in his Corpus Glossarum Latinarum. He further detected that the Leyden Glossary contains glosses from a source of the Hermeneumata type, batches of which glosses proved again to be still discernible in the more perfect glossaries. The way in which all those glosses were incorporated in the Corpus Glossary is described in full in the first book mentioned above, where it is also shown that other sources, such as Aldhelm, were freely used in the making of it.

Consequently Prof. Lindsay has been able to bring out an edition of Corpus that labels the bulk of its Latin and Greek glosses and many of the Old English ones, though it stands to reason that the assignment of glosses to such sources as Abstrusa or Abolita cannot in itself rank in importance with their ascription to a definite passage in Latin literature. For the Old English glosses he had the invaluable help of the late Dr. Henry Bradley and of Mr. W. H. Stevenson and could derive much benefit from the discoveries of the late Prof. Napier and of Prof. Schlutter, while as to the Latin glosses it was especially Gruber's, Goetz' and Hessels' previous work in this quarter that had paved the way.

In abundant footnotes and copious elaborate notes at the end of the volume the author gives a host of information about the various items, though something is left to be done, especially for a great many Old English glosses (often appearing in clusters), whose sources have even now remained obscure or doubtful, as Prof. Lindsay can often only point out a possible source (Cp. Ep. Erf. Leyd. Gl., p. 105 sqq.).

The editing of the Corpus Glossary has been done in this way that the glosses have been punctuated more consistently than is done in the MS. and when necessary have been corrected by bracketing superfluous or missing letters and words, while the source or possible source, if found, has been added after the glosses. Some of the author's silent changes, I must confess, seem to me quite unnecessary, viz. the distinction between i and j, u and v and the exclusive use of ae for ae, œ and ę. In school

books it may be helpful, though not perhaps wise, to print abjecit or avus instead of abiecit or auus, but what is the good of such spellings in a book meant for scholars? A mistake like Uiscellum for Iuscellum (U 208) is made less obvious by printing Viscellum instead. Moreover in quoting from the Epinal Glossary it would have been better to use Prof. Schlutter's facsimile than Dr. Sweet's, which is manifestly less reliable.

In the matter of the elucidation of the Old English glosses, which here concern us most of all, the progress is less than in the case of the Latin ones. We are thankful, it is true, to get rid of such ghostwords as maffa O 166 (v. Cp. Ep. Erf. Leyd. Gl., p. 47), or to learn that faertyhted C 471 does not bear on the lemma Clinici, but translates Latin illectus; that meadro, the lemma to bordan M 155a, stands for Maeandro; that Suffocacium S 698 is probably a kind of cake (Bradley), hence is rightly glossed cecil = cēcil (which Ferdinand Dieter identified with coecil = cecil as early as 1885); that the lemma to agnidine D 78 was originally Detrita (ablative); that Exilia gestinccum E 421 may be a mistake for Exugia gescincium, and that aac C 648 is a gloss to Rubor (mistaken for Robur), not to Color; but on the other hand we cannot help perceiving that the author is less familiar with Old English than we might desire. A few instances will show this. For the gloss Accintu denetle. A 172 someone (was it Schlutter?) has ingeniously suggested Acanthus blinde netle, which seems possible, as Acanthus denotes a foreign plant and so may well have been glossed by a native plant name, although blindnetele, blinde netele as a rule translates archangelica, which may be either the yellow weasel-snout, or archangel (Galeobdolon luteum, or Lamium Galcobdolon), or some kind of dead-nettle. Lindsay, leaving denetle as it is and comparing Acantum: semen urticae in another glossary, alters Accintu into Acanthum and in this manner makes it more difficult to understand the Old English gloss. In gebinumini, the gloss to Ademto A 206, ge is cancelled, but it is far more probable that the gloss is a double one meant for ge-, binumini, as it corresponds to both adepto ginumini and adepta binumni in Epinal (and similarly in Erfurt). To Antulus. caecbora. A 659 there is a note (p. 194) that shows Prof. Lindsay to have misunderstood the Old English word, which is identical with cẽacbora. The Index hesitatingly identifies it with c@gbora; but of course caec- = ceac- has just as little to do with 'key' as with 'keg' (suggested by the editor) and 'jug-' or 'vessel-bearer' is the correct equivalent of the Latin. - E 328 Euiscerata. athed. The gloss Sub.iugatis. gededum should have retained the editor from altering athed (= ādēd) into athied, which, besides, is a form not to be expected in Corpus. - quidam from Epinal (Cp. Ep. Erf. Leyd. Gl. p. 4, 13 and p. 6, 1) ought to be quida; the MS. has quida with a long horizontal stroke crossing the d. The etymology given of frysca B 227, as if derived from frosc (but Corp. has forsc), is purely fantastical. Ludarius L 298 is considered as possibly a mistake for Lucar, lucaris and its gloss stēor as possibly Germ. steuer 'tax'. But O.E. stēor never has that meaning and in derived glossaries the item correctly occurs among the names of bovines or at any rate of animals, with the lemma Laudaris or Ludares. The so-called O.E. lopostum S 174 instead of Vulgar Lat. lopostris is due to an oversight on Dr. Hessels' part; the MS. seems to have lopost (or lopost?), s. Glogger, Leid. Gloss., III a, p. 51.

=

Many of Bradley's suggestions mentioned in the footnotes or the Index are specious rather than probable, e.g. the following. For Adplaudat. onhliorrouuit. he would read Adplicauit: on hlion rouuit 'rows to shelter',

but what form may *hlion be? His change of cli into oli (!) ‘oil' (p. 271) in C 974 Cyprinus. forneted cli. does not take us any nearer to the explanation of this mysterious item (Ep. fornaeticli, Erf. fornetiali). In 1 474 Iota. sochtha. (Ep. icta soctha, Erf. iota soctha) the gloss is explained as a mistake for se ehtotha 'the eighth letter', but apart from the fact that such a gloss to iota (in the well-known passage of the Gospel of S. Matthew) would be unintelligible and that the reading of the earlier MS. (Epinal) is graphically farther away from Br.'s conjecture, iota happens to be the ninth letter of the Greek alphabet, not the eighth. Bradley's explanation of cearricge S 277 as a derivative from cirran with the suffix -icge does not take into account either cear ruccae Ep., cearricae Erf., or the lemma Senon; no translation is vouchsafed. — Quite impossible is his guess (known before from Toller's Supplement, p. 141) that suacenlic P 203, which glosses percommoda (for -do), stands for sua cenlic = swa cynlic. 1°. this would not be a translation of percommodo (the passage is e mollissimis stratis cubiculoque percommodo matutinus egrediens Oros. IV, praef. 7), 2°. not sua is the word used in Corp. Gl. but suae, 3°. the mutation of u never appears as e in Corp. Gl. (nor, for that matter, in Epinal), 4°. the d in Ep. Erf. suacendlic is not accounted for. Latin per- suggests that the gloss originally consisted of the adverb suide and an adjective ending in -lic, but I cannot think of an adjective meaning 'comfortable' that is graphically nearer to the MS. reading than get@slic or zehyolic, neither of them satisfactory. Instead of gierende T 41 Bradley proposes giernde, "from geornan 'desire'" (p. 276), but leaving aside the ie, what connexion is there between 'desire' and the lemma Taxauerat 'had esteemed'? The translation of gronuisc A 160, 'moustached fish' (p. 277), is wrongly fathered on Bradley, s. Köhler, Die ae. Fischnamen, p. 39 sq. In my opinion, both u in the Corpus gloss and w in gronwisc Wr.-Wülck. 344, 39 preclude the possibility of a compound of fisc being meant.

However, these are minor blemishes, not even all of them due to Prof. Lindsay, who may be said to have achieved his principal aim of publishing a sound critical edition of the Glossary containing all the available matter for the elucidation of the Latin glosses, and of giving a wonderfully clear insight into the way in which the composition of the Glossary was accomplished. It is a great pity that an equal praise cannot be bestowed upon the endeavours of the lady who has composed the 'Anglo-Saxon Index'. It is only fair to take it at its own valuation of being meant 'to enable a reader ignorant of Anglo-saxon to know the meaning of the words and to find them in Bosworth and Toller's Dictionary (last edition 1)) and Supplement,' but the Index does not even come up to this moderate requirement. The best that may be said of it is that its quotations seem to be accurate, for the rest it is composed negligently and unsatisfactorily. If the Dictionary has the wrong form of a word and the Supplement the correct one the Index is sure to give the former; the meaning of words is not always added, even in cases where there is no difficulty, such as bune 'cup', cian 'gills'; mistakes, grammatical and other, are frequent, wrong quantities abound. Some instances may illustrate this. The gloss aam (cautere) C 16 (Ep. haam, Erf. fam) is (with some diffidence) referred to am 'weaver's reed', abunden (expeditus) E 364 identified with onbundaen (quoted as if really occurring in Corp.); the component parts of goodneppel (citonium) C 439 are mentioned as if separate words; aera S 574 is quoted as an Old English

1) whatever that means.

E. S. VI 1924.

3

=

α

word without any attempt at an explanation; of the three forms aetgaere (ansatae) A 603, cgtero (falarica) F 67 and at gaeru (framea) F 344, the first is called a singular, the other two plurals; aforht (in dies) I 195 ford is given as a compound, agnidine (detrita) D 78 is called a dative, ahaefd (suspensus) S 636 derived from ahebban 'raise'; to ampre (uarix) U 8 neither the meaning nor the corresponding Mod. E. amper is added, and mitta (exagium) E 416 is translated 'measure', an ege (luscus) L 272 combined with late aneage instead of early anīege, anmood (contumax) C 597 identified with ānmōd; anoða (formido) F 277 and the imperative ansceat (exintera) E 411 are left without an explanation; ascaeltte for aslaectae or -te (disoluerat) D 336 is derived from aslacian, to aseodenne (expendere) È 542 from aseodan (sic); asoedan (satiare) S 104 is called a compound of both sedan and sadian, astyntid (hebitatus) H 56 derived from astintan, athed (euiscerata) E 328 quoted as athied (s. above) and derived from ādīedan; the p. partic. apoht (commentum) C 779 is considered as a compound of the noun doht (the same mistake is found in the edition, which adds excogitatio> and compares E 178 Ementum. excogitatio., instead of C 760 Comminiscitur commentum. uel comentat); bencselma (sponda) S 470 is first regarded as a compound, then (correctly) as two words; bercae (latratus) L 90 is called the nom. plur. of beorc; bebītan (Mordicos. bibitne. M 251) is translated 'bite' instead of 'gnaw'; on caecbora s. above; the non-existent cenlic is given as a genuine word meaning 'convenient' (s. above); chroa and croha (citropodes) C 382 are said to be the nom. pl. of crocc; cionecti (rimosa) R 175 is described as the "nom. sg. or pl. masc. or fem. (of) cinint", instead of nom. sg. (masc. or ntr.) of a form corresponding to later cinihte; cocas, i.e. Latin cloacas (culinia) Ć 953 is called the nom. acc. pl. of coc (i.e. coc) 'cook', though the correction is indicated in the edition; cornuurma (uemiculus) U 148 is translated 'cornworm, weevil', instead of 'scarlet dye'; faedran in f. sunu (patruelis) P 95 is called the gen. sg. of fæder 'father'; the Latin flecta (cratem) C 891, (graticium) G 174 is considered as an O.E. word; foruuened (insolens) I 221 is derived from forwēnan (which means 'suspect') instead of forwenian, in the teeth of fer uuaenid Ep., oberuuenide (insolesceret) I 209 (Ep. ober uuaenidae) and forwana 'presumption' (C. Past. 465, 16); gemaad (uecors) U 122 is identified with *gemad, gescroepnis (conpedium) C 781 with *gescropenys; gestalum, i.e. gestālum (obiectionibus) O 105 is defined as the dat. pl. of *gestal 'obstacle'; the later form of haeb (salum) S 59 is said to be heaf; hlyte (portio) P 506 is called a variant form of hlyt (i.e. hlīet) ‘portion' and the accusative scildenne (tutellam) T 329 (= Ep. sclindi nnae, Erf. scildinnae) is taken for the gerund of scieldian (1. scildan) 'to shield'.

This list of sins, although a long one, is not even complete, and yet by glancing into Toller's Supplement most of the mistakes might have been avoided. For experts it is not a serious thing that the Index is unreliable, but Latin scholars ignorant of Old English should be warned against putting trust in this bungle work, which is far beneath the high scientific standard of Prof. Lindsay's part of the volume.

Groningen.

J. H. KERN.