To the above root I also put a form in ed, corresponding with Welsh eddain, to flow, Ang.-Sax. edre, a water-course, &c.

With the above may perhaps also be classed the Celtic and or ant,[10] to which Mone, (Die Gallische sprache), gives the meaning of water.

To the Celt. dubr, Welsh dwfr, water, are by common consent referred the names in the second division of the undermentioned. But the forms dub, duv, which in accordance with the general system here advocated, I take to be the older and simpler form of the word, are, by Zeuss (Gramm. Celt.), as well as most English writers, referred to Welsh du, Gael. dubh, black.

Another Celtic word for water is dur, which, however, seems more common in the names of towns (situated upon waters) than in the names of rivers. Is this word formed by syncope from the last, as duber = dur? Or is it directly from the root of the Sansc. drâ or dur, to move?

In this chapter is to be included the root ar, respecting which I quote the following remarks of Förstemann. "The meaning of river, water, must have belonged to this wide-spread root, though I never find it applied as an appellative, apart from the obsolete Dutch word aar, which Pott produces. I also nowhere find even an attempt to explain the following river-names from any root, and know so little as scarcely to make a passing suggestion; even the Sanscrit itself shows me no likely word approaching it, unless perhaps we think of ara, swift (Petersburger Wörterbuch)."

The root, I apprehend, like that of most other river-names, is to be found in a verb signifying to move, to go—the Sansc. ar, ir or ur, Lat. ire, errare, &c. And we are not without an additional trace of the sense we want, as the Basque has ur, water, errio, a river, and the Hung. has er, a brook. The sense of swiftness, as found in Sansc. ara, may perhaps intermix in the following names. But there is also a word of precisely opposite meaning, the Gael. ar, slow, whence Armstrong, with considerable reason, derives the name of the Arar (or Saone), a river noted above all others for the slowness of its course. Respecting this word as a termination see page 11.

From ar and ur, to move, the Sanscrit forms arch and urj, with the same meaning, but perhaps in a rather more intense degree, if we may judge by some of the derivatives, as Lat. urgeo, &c. In two of the three appellatives which I find, the Basque erreca, brook, and the Lettish urga, torrent, we may trace this sense; but in the third, Mordvinian (a Finnish dialect), erke, lake, it is altogether wanting. And on the whole, I cannot find it borne out in the rivers quoted below. Perhaps the Obs. Gael. arg, white, which has been generally adduced as the etymon of these names, may intermix.

From the Sansc. ri, to flow, Gr. ῥεω, Lat. rigo (often applied to rivers—"Qua Ister Getas rigat," Tibullus), Sansc. rinas, fluid, Old Sax. rîha, a torrent, Ang.-Sax. regen, Eng. rain, Slav. rêka, a stream, Welsh rhe, rapid, rhean, rhen, a stream, &c., we get the following group. The river Regen Berghaus derives from Germ. regen, rain, in reference to the unusual amount of rain-fall which occurs in the Böhmer-wald, where it has its source. Butmann derives it from Wend. and Slav. rêka, a stream, connecting its name also with that of the Rhine. Both these derivations I think rather too narrow.

With respect to the Rhine I quote the following opinions. Armstrong derives it from Celt. reidh-an, a smooth water, than which nothing can be more unsuitable—the characteristic of the river, as noticed by all observers, from Cæsar and Tacitus downwards—being that of rapidity. Donaldson compares it with Old Norse renna, fluere, and makes Rhine = Anglo-Saxon rin, cursus aquæ. Grimm (Deutsch. Gramm.) compares it with Goth. hrains, pure, clear, and thinks that "in any case we must dismiss the derivation from rinnan, fluere." Zeuss and Förstemann support the opinion of Grimm; nevertheless, all three agree in thinking that the name is of Celtic origin. The nearest word, as it seems to me, is Welsh rhean, rhen, a stream, cognate with Sansc. rinas, fluid, Old Norse renna, fluere, and (as I suppose), with Goth. hrains, pure.

The Sansc. lî, to wet, moisten, spreads into many forms through the Indo-European languages. I divide them for convenience into two groups, and take first Lat. liqueo, Old Norse leka, Ang.-Sax. lecan (stillare, rigare), Gael. and Ir. li, sea, Gael. lia, Welsh lli, llion, a stream. Most of the following names, I take it, are Celtic. I am not sure that the sense of stillness or clearness does not enter somewhat into the two following groups.

For the second group I take Lat. lavo, luo, Old Norse lauga, lavare, Anglo-Saxon lagu, water, Gael. lo, water, Gael. and Ir. loin, stream. In this group there may perhaps be something more of the Germain element, e.g., in the rivers of Scandinavia.

To the above root I also place the following, corresponding more distinctly with Welsh llifo, to pour.

More remotely with the Sansc. lî, liquere, and directly with Welsh lleithio, to moisten, llyddo, to pour, Gael. lith, a pool, smooth water, Goth. leithus, Ang.-Sax. lidh, liquor, poculum, potus, I connect the following. The rivers themselves hardly seem to bear out the special idea of smoothness, which we might be apt to infer from the root, and from the character of the mythological river Lethe.

From the Sansc. nî, to move, comes nîran, water, corresponding with the Mod. Greek νερόν of the same meaning. And that the Greek word is no new importation into that language, we may judge by the name of Nereus, a water-god, the son of Neptune. The Gr. ναω, fluo, the Gael. nigh, to bathe, to wash, and the Obs. Gael. near, water, a river, show a close relationship; the Heb. nhar, a river, also seems to be allied. Compare the Nore, a name given to part of the estuary of the Thames, with the Narra, the name of the two branches by which the Indus flows into the sea. Also with the Nharawan, an ancient canal from the Tigris towards the Persian Gulf. And with the Curische Nehrung, a strip of land which separates the lagoon called the Curische Haf in Prussia from the waters of the Baltic. On this name Mr. Winning remarks,[17] "I offer the conjecture that the word nehrung is equivalent to our break-water, and that it is derived from the Sabine (or Old Prussian) term neriene, strength, bravery." I should propose to give it a meaning analogous, but rather different—deriving it from the word in question, nar or ner, water, and some equivalent of Old Norse engia, coarctare, making nehrung to signify "that which confines the waters" (of the lake). In all these cases there is something of the sense of an estuary, or of a channel communicating with the sea—the Curische Haf being a large lagoon which receives the river Niemen, and discharges it by an outlet into the Baltic. The following names I take to be for the most part of Celtic origin.

From the Sansc. nî, to move, Gael. nigh, to bathe, to wash, comes, I apprehend, the Welsh nannaw, nennig, nant, a small stream.

Closely allied to nî, to move, I take to be Sansc. niv, to flow, Welsh nofio, to swim, to float, whence the names undermentioned. The Novius of Ptolemy, supposed to be the Nith, if not a false rendering, might come in here.

From the same root, nî, to move, and closely connected with the last group, I take to be Sansc. nis, to flow, to water. Zeuss (Die Deutschen) takes the word, as far as it relates to the rivers of Germany, to be of Slavonic origin. It appears to be the word found as the second part of some Slavonic river-names, as the Yalomnitza. But it is also both Celtic and Teutonic, for the Armorican has naoz, a brook, and the German has nasz, wet, nässen, to be wet.

From the Greek ναω, fluo, comes νᾶμα, a stream, ναματιᾶιον ὕδωρ, running water. Hence seems to be Namadus, the name given by the Greek geographers to the Nerbudda of India.

Another form which I take to be derived from the above Sanscrit root nî, by the prefix s, is Sansc. snu, fluere, stillare, (whence Germ. schnee, Eng. snow, &c.)

A derivative form is the Gael. and Ir. snidh or snith, to ooze through, distil, Obs. Gael. and Ir. snuadh, to flow, and snuadh, a river, whence I take the following. Förstemann refers to Old High German snidan, Modern German schneiden, to divide, in the sense of a boundary, which is a root suitable enough in itself, though I think it ought to yield the preference to the direct sense of water.

The form snid or snith introduces the form nid or nith, and suggests the enquiry whether that may not also be a word signifying water. Donaldson, (Varronianus), referring to a word Nethuns, "found on a Tuscan mirror over a figure manifestly intended for Neptune," observes that "there can be little doubt that nethu means water in the Tuscan language." Assuming the correctness of the premises, I think that this must be the case; and that as the Naiades (water-nymphs), contain the Greek ναω; as Nereus (a water-god), contains the word ner before referred to; as Neptune contains the Greek νίπτω, in each case involving the signification of water, so Nethuns (=Neptunus) must contain a related word neth or nethun of the same meaning. Also that this word comes in its place here, as a derivative of the root nî, and as a corresponding form to the Celtic snidh or snith.

There are, however, two other meanings which might intermix in the following names; the one is that suggested by Baxter, viz., Welsh nyddu, to turn or twist, in the sense of tortuousness; and the other is Old Norse nidr, fremor, strepitus.

There can hardly be a doubt that the words sar, sor, sur, so widely spread in the names of rivers, are to be traced to the Sansc. sar, sri, to move, to go, sru, to flow, whence saras, water, sarit, srôta, river. The Permic and two kindred dialects of the Finnic class have the simple form sor or sur, a river, and the Gaelic and Irish have the derived form sruth, to flow, sroth, sruth, river. In the names Sorg, Sark, Sarco, I rather take the guttural to have accrued.