This idea of time is always present, though sometimes less prominent, perhaps, than some of the other possible relations of the da-clause to its main verb.

So much for the meaning of da; now let us consider its use. My study has established the fact that Ja is used only with the preterite tense of the indicative mode.

There are some exceptions, real or apparent, which we shall now consider.

I have examined about thirty-three hundred clauses with this connective or its variants, dada and da... da, and have noticed only seventeen instances in which either the present tense or the optative mode appears. Most of these are in late or corrupt texts, about ten of them being in the late entries of the Chronicle.

We shall first consider the cases in which the present tense appears: Chron. 261. 25 God hit bete pa his wille bed. This seems to be a perfectly clear case, and we should rather expect the optative as well. Chron. 266. 12 þer efter in þe lengten þestrede þe sunne 7 te dæi, abuton non tid dæies, þa men eten, me lihtede candles to æten bi. This appears in the entry for the year 1140; and the phonology is so changed, that it is difficult to say what the mode and tense are. Does the clause mean while men were eating, or at the time that men (habitually) eat? The verb in the following sentence is probably preterite indicative, with a for e: Cart. 2, 290. 14 hit hiera yrfe is þ hit swa umbe sæccen gange into þære Cyrican swa hit pa on dæg wes, þa hit man him to læt. Lch. 3, 82. 11 þæt syndon sa ysene, pa man mid cnifun hæle menn. Neither the tense nor the mode is hard to understand here, but the text is very corrupt and late. HL. 122. 184 And wite þu þæt heo is

of ðinum sæde geeacnod, swa swa þu nytest, þa þa þu hi ana forlate. The sense here requires the preterite. May not this be regarded as a preterite of a weak verb lætan?

Of the five apparent exceptions, then, to the rule in regard to the tense with da, only two seem reasonably certain.

We pass now to the question of mode. Chron. 215. 6 hi ferdon æfter heom into þam mynstre 7 woldon hig ut dragan, þaða hig ne dorsten na utgan; Chron. 264. 4 þa þe castles uuaren maked pa fylden hi mid deoules 7 yuele men (entry for 1137); Chron. 266. 37 þa hi per inne waren, pa com þe kinges cuen mid al hire strengthe. See also Chron. 161. 26; 218. 1; 259.37; 264. 28; 267.1: also Chron. 266. 12; Lch. 3 82. 11; HL. 122. 184, as quoted above. In these cases the optative form is to be accounted for by the weakening of the a of the ending, and there is really no violation of the rule that the indicative is the mode used after Ja, except in the case of those examples in which the present tense appears also.

In O. 56.17 þa æt nihstan hie hæfden getogen eal Creca folc to dæm gewinnum, þa Læcedemonia besætan þa burg Mæs[ian]e X winter; the reading hæfdon of MS. C. is probably correct.

BH. 162. 21 secgað men, pa Oswald se cyning of Scotta ealonde biscopes bede, . . . þa wæs him sended ærest oder biscop redes modes monn. BH. 198.31 þa eode se mæssepreost to Aidane þæm biscope; bæd hine þæt he for hine gebæde 7 for his geferan, 7 for heora gesyntu to Gode þingade, pa heo swa micelne siðfæt feran scolden; Rood. 11.2 Da hio pus hiom betweonan spræcen, þa cliopodan þare cwene campan pider. In all these cases, also, we most probably have the weakening of the termination, and not really

an optative. Or, in the case of the examples from BH., the mode of the verbs of the temporal clauses may be due to the fact of their being in indirect discourse.

In any case, the number of exceptions to the principle laid down is insignificant, granting even that all the cases cited are bona fide optatives, which is certainly by no means beyond dispute.

It is interesting to note that Ælfric in Gram., in giving equivalents for Latin modes and tenses, always uses donne to translate cum when used with the present tense, but changes to a when the Latin changes to the past tense. For example, 132. 14 subivnctivo mode under-deodendlicum gemete Tempore Praesenti cum amem þonne ic nu lufige, &c., eodem modo Tempore Praeterito inperfecto cum amarem þaða ic lufode hwat hwega, cum amares ðaða ðu lufodest, &c.

I have not been able to discover whether or not this distinctive syntactical function of da obtains in all the cognate languages in which it appears. But all the examples I have found indicate at least that this use is the prevailing one, as will be seen from the examples cited below. Erdmann1 says: "tho steht im Nachsatze nach so I, 22.42 u. öfters; im temporalen Nebensatze steht es ebenso wie so und in gleicher Bedeutung nur beim Ind. Prät." Therefore in one important OHG. text, at least, tho has the same meaning and use as da in OE.

Very often a correlative da stands at the beginning of the main clause, as in this example: O. 19, 25 Đa he þiderweard seglode fram Sciringes heale, þa wæs him on þæt bæcbord Denamearc. But the number of cases in which this correlative da does not appear are very numerous, even in what we may call the 1 Syntax der Sprache Otfrids, pp. 120, 204.

formal temporal period. I have been careful in my indexes to distinguish the cases in which ða is used from those in which it is not; but I am not prepared to make any statement as to the uniformity of the use or omission of the da in such cases. However, when the da-clause is drawn in parenthetically, there is likely to be no da; nor does da usually appear if the clause determines the time of the action of a verb which is itself in a subordinate clause.

The use of such balancing adverbs is a very noticeable feature of OE. prose, and it may be that a more minute study than I have been able to make of this matter would yield more definite and valuable results.

Occasionally the da- clause depends on a verb omitted, as in this example: BH. 106. 21 Onhyrede he on þon þa bysene þæs ærestan heordes Godes cirican Scē Petres pas apostoles, pa he æt Rome ærest Cristes cirican staðolode. Since clauses introduced by da are so common, it would be superfluous to quote examples here. Almost any page of OE. would yield one or more.

Of the three connectives, da, dada, and da . . . đa, da is the most common in all the texts except Bo., in which the form da. da predominates, Dial., and the works of Ælfric, in which dada is by far the most common. However, in his Old Testament translations the simple da is more frequent. In general, ðaða seems to be more common in the late texts, although appearing in very early ones also. Perhaps the consistent use of the simple da, although there are several examples of the divided form, itself most common in early texts, may be considered as another of the archaic peculiarities of BIH. The divided form is most common in BH., Bo., and Guth. For statistics as

to the number of examples of each sort in each text, see the tables and index-lists1.

NOTE 1. Of the OE. poetry I have examined only the Christ. I find that da is used here as in the prose.

NOTE 2. The Middle English form of the OE. đa is tho, and I find it used both as adverb and as conjunction in Piers Plowman. In Chaucer, however, it seems to occur only as an adverb. By this time hwen has become the ordinary temporal connective in such uses as those of da and donne in OE. I find the form do first in Cart. 3, 217.7 Her switelep on þis write ihu Sifled uthe hire ait he po sche ouer se ferde. I have noticed it also in the Winteney-Version of the Regula S. Benedicti (circa 1200 A. D.), for example, 5. 17; 7.27. In the first of these cases the present tense is used, and in the second an optative form occurs. I cite one or two examples from Piers Plowman: Pro. 176 Ac tho the belle was ybougt, and on the beize hanged, There ne was ratoun in alle the route for alle the rewme of Fraunce, That dorst have ybounden the belle aboute the cattis nekke. Passus 21.243 'By godes body' quath this Book 'ich wole bere wyttnesse, Tho this barn was ybore, ther blased a sterre.'

NOTE 3. So far as I have observed, no cognate of the OE. da occurs in Gothic. However, there are cognates in most of the Germanic dialects, and I cite such examples as I have noted. Since OF. is most nearly related to OE., we begin with that: Lesebuch2 95. 16 Tha use drochten ebern warth, tha warther alle brekanden to boden ebern. 96. 22 Tha mat alra erest sette thet ield, tha slochma enre frowa hire brother. In reading the OS. Heliand3 I noted a number of examples of tho, one or two of which I quote: 794 Tho sie that geld habdun, erlos an them alaha, so it an iro ewa gibod, gilestid an iro land-wisum, tho forun im eft thie liudi thanan. Numerous illustrations, might be quoted from OHG., but I shall cite only one or two. Erdmann notes that tho 1 pp. 162-177.

2 Wilhelm Heuser, Altfriesisches Lesebuch, Heidelberg, 1903. Heyne's 3 te Auflage.