be remembered that we found that da is used only with the preterite indicative. On the other hand, Jonne occurs with both indicative and optative, as well as with both the present and the preterite tenses. However, naturally enough when we consider the meaning, the optative occurs, for the most part, only in connection with the imperative, after another optative, or in indirect discourse. The normal difference in use between da and donne is well brought out by this example: John 21.18 Soð ic segge þe, þa þu gingre wære þu gertest þe, and eodest þær þu woldest. Witodlice ponne pu ealdest pu strecst þine hande and oþer þe gyrt and læt þider pe pu nelt. Of the two thousand clauses with this connective which I have examined, fewer than two hundred are in the preterite tense. Ælfric also in Gram, when translating the Latin cum, uses da with the preterite, but donne when the tense of the Latin changes to the present. For citations see da1.

As we found to be the case with da, clauses introduced by Jonne frequently carry a causal or concessive or conditional force, in addition to the primary temporal signification. However, the temporal element is in all cases prominent enough to justify my inclusion of them in the lists.

Since most of my discussion of these secondary and in a sense incidental meanings of da is equally applicable to donne, I forbear any extended consideration of the matter here. The reader is referred to the excellent dissertation2 by Frank J. Mather, Jr. for a discussion of the conditional use of donne.

Occasionally the verb, the time of the activity of which is determined by the temporal clause, is omit1 p. 14.

The Conditional Sentence in Anglo-Saxon, Munich, 1893, p. 49.

ted, as in this example: Mart. 4. 18 pa was hire ansyn swa reod ond swa fæger swa þære wynsumestan fæmnan, bonne heo fægerost bid. Likewise I have noticed one or two instances in which the auxiliary in the temporal clause is omitted: Lch. 2, 290. 14 ponne þ fyr of þam stane aslegen, hit is god wio ligetta 7 wið þunorrada; LS. 2, 38. 555 þonne þyllice geþohtas on astigan, ponne astrehte ic me sylfe on eorðan. In this case a note informs us that agunnon is supplied in a later hand. Probably all such cases as these two are to be explained as mere accidental omissions of the scribes.

I have not been able to make out the form of the verb in this sentence: Lch. 1, 390. 18 ĐƉis man sceal cweðan donne his ceapa hwilcne man fortsolenne. Cockayne translates: 'A man must sing this when one hath stolen any one of his cattle'. I judge it is a mere corruption for the accusative singular of the past participle of forstelan, with the auxiliary omitted.

In this example, donne seems to be equivalent to hwonne: OET. Vesp. Psalms 118. 84 hu monge sind dægas diowes dines donne du doest oehtendum mec dom. The Vulgate reads: Quot sunt dies servi tui: quando facies de persequentibus me judicium? It will be remembered that this is a mere gloss, and perhaps donne or da was the only rendering the glossator knew for the Latin quando.

The following sentence is the only one of the sort I have noted in all OE. This peculiar construction is also doubtless due to the wish to preserve the construction of the Latin original: L. 13. 35 Soðlice ic eow secge þ ge me ne geseoð ærþam þe cum se ponne ge cweðað gebletsod sy se de com on drihtnes naman. The Latin is Dico autem vobis, quia non videbitis me donec veniat cum dicetis: Benedictus,

qui venit in nomine Domini. This curious construction does not go back to the Greek.

Sometimes donne seems to be used as a sort of temporal relative, as in this sentence: John 9.4 niht cymþ, þonne nan man wyrcan ne mæg. In such sentences the clause introduced by donne does not determine the time of the action of the main verb, but rather is used to characterize or describe the subject of that verb. It is therefore rather an adjective relative clause than an adverbial temporal clause in any ordinary sense. The Latin idiom is precisely similar venit nox, quando nemo potest operari. So also the Greek: ἔρχεται νὺξ ὅτε οὐδεὶς δύναται ἐρyasɛodai. In this sentence donne translates the Latin cum: John 16. 25 Seo tid cymð, þænne ic eow ne sprece on bigspellum. Latin: venit hora cum jam non in proverbiis loquar vobis. Greek, orɛ. See also L. 17. 22; John 4. 21; 5. 25; LS. 1. 510, 384; Wulf. 118. 9; 208. 30. The relative nature of these clauses appears from a comparison with such cases as this M. 9. 15 soolice þa dagas cumað þæt se brydguma byð afyrred fram him. For this passage the Hatton MS. reads: þe se bredgume byþ aferred fram heom. Latin: cum auferetur ab eis sponsus. Greek: ὅταν απαρθῆ ἀπ' αὐτῶν ὁ νυμφίος. Modern English when is of course used in just the same way, so there is really nothing peculiar about the construction in OE. But in OE., donne is so often used to introduce a real adverbial temporal clause that this relative use seems at first strange. Further evidence of the fact that donne occasionally had this relative use in OE. is afforded by these examples: Lch. 1. 324.6 heo hafap wæstm sinewealtne 7 byterne, se ys to nymenne to pam timan ponne he æfter his grennysse fealwað. Mk. 24. 25 Soðlice ic eow secge pic

heonon forð ne drince of þyses wingeardes cynne, oð þone dæg, þonne ic hine niwne drince on godes rice. In the following example, don is doubtless for Jonne, since MS. O. has ponne, and MS. Ca. ponn: BH. 340.7 da gehyrde heo semninga in þære lyfte uppe cuone sweg 7 . . ., þær heo wunedon to gebedum gecegde 7 awehte beon, pon heora hwylc of worulde geleored wæs.

NOTE 1. Jonne is as common in the poetry as in the prose, and, so far as I have noticed, its uses here are parallel to those of the prose. It would therefore be super

fluous to quote examples.

NOTE 2. None of the Middle English forms of donne seem to be used in Chaucer as temporal conjunctions, and I have only one example from Piers Plowman: B. 16.69 Thanne contenence is nerre the croppe as calewey bastarde, Thanne bereth the croppe kynde fruite and clenneste of alle, Maydenhode, angeles peres.

NOTE 3. Then, the Modern English representative of OE. donne, is not used to introduce temporal clauses.

I

NOTE 4. Examples of pan in Gothic are numerous. quote one or two, selected at random: M. 6.6 Iþ þu þan bidjais, gagg in heþjon þeina; Mk. 4. 15 jah pan gahausjand unkarjans, suns qimiþ Satanas jah usnimiþ waurd. Numerous examples might be adduced from the Heliand, but one will suffice: 5603 Than thu an thin riki kumis, wis mi than ginaðig. This also will illustrate the use of the OHG. cognate to OE. donne: Otfrid 5. 19. 34 Wer ist manno in lante, ther thanne witharstante, thanne er iz zi thiu gifiarit, thaz sih ther himil ruarit.

2b. donne donne.

The doubled form donne donne is rare in OE. prose, occurring only in Bo. (5 times) and in CP. (33 times), and not at all in the poetry, so far as I have been able to discover.

Originally, in all probability, one element was felt as an adverb, the other as a conjunction; but it is impossible to determine which is which. To all intents and purposes, the two form one conjunction, not differing in meaning or use from the simple donne.

Since the doubled form is comparatively rare, I subjoin an example, though it presents no peculiarities of use: CP. 311. 12 Ɖæt is, donne donne sio wamb bið aðened mid fylle for giefernesse, donne towierpo hio ðurh fierenlustas da mægenu dære saule.

2c. donne . . . donne.

This divided form of the connective is even more rare than the donne donne form. It occurs six times in Bo., once in Sol, and once in CP. This form needs no extended discussion, since all that has been said concerning the meaning and use of donne applies here. I submit an example: Bo. 33. 14 ponne þu bonne orsorg wære, 7 þa þeofas þe from gewiten weron, bonne meahtes pu bismrian þæs andweardan welan. So far as my observation goes, this form does not occur in the poetry.

3. da de.

This somewhat rare connective is, of course, made up of the common temporal conjunction da and the relative particle de. Da itself is from the root of the demonstrative se, so the particle arose in a manner perfectly logical. Though the connective is by no means found only in early texts, its use may be a survival from a syntactical period in which this was the common form, rather than the simple đa, which may possibly have developed out of it. Though the cognate of Ja is not used as a temporal conjunction in Gothic, the closely related þan is; and occasionally the relative particle ei, corresponding to OE. de, is