The root sal Förstemann takes to be Celtic, and to mean salt water. No doubt saltness is a characteristic which would naturally give a name to a river. So it does in the case of the "Salt River" in the U.S., and of the Salza in the Salzkammergut. But I can hardly think that all the many rivers called the Saale are salt, and I am inclined to go deeper for the meaning. The Sansc. has sal, to move, whence salan, water. The first meaning then seems to be water—applied to the sea as the water—and then to salt as derived from the sea. So that when the Gr. άλς, the Old Norse salt, and the Gael. sal, all mean both salt, and also the sea, the latter may be the original sense. From the above root, sal, to move, the Lat. forms both salire and saltare, as from the same root come sal and salt. I take the root sal then in river-names to mean, at least in some cases, water. In one or two instances the sense of saltness comes before us as a known quality, and in such case I have taken the names elsewhere. But failing the proper proof, which would be that of tasting, I must leave the others where they stand.

It is possible that the root als, ils, found in the name of several rivers, as the Alz, Elz, Ilse, may be a transposition of the above, just as Gr. άλς = Lat. sal. But upon the whole I have thought another derivation better, and have included them in a preceding group.

From the Sansc. var or vars, to bedew, moisten, whence var, water, varsas, rain, Gr. ἐρση, dew, Gael. and Ir. uaran, fresh water, I get the following, dividing them into the two forms, var and vars.

In the above Sansc. var, to moisten, to water, is contained, as I take it, the root of the Finnic wirta, a river, the only appellative I can find for the following.

The following names have been generally supposed to be derived from Welsh cledd or cleddeu, sword, and to be applied metaphorically to a river. But I think it will be seen from the Sansc. klid, to water, whence klaidan, flux, Gr. κλύδων, fluctus, unda, Ang.-Sax. glade, a river, brook, that the meaning of water lies at the very bottom of the word. Perhaps, however, as the senses of a running stream and of a sharp point often run parallel to each other, there may be in this case a relationship between them.

There are two Sanscrit roots from which the word ag, ang, ing, in river-names might be deduced. One is the verb ag or aj, to move, whence anjas, movement, (or the verb ac or anc, to traverse), and the other is the verb ag or ang, to contract, whence Latin anguis, snake, anguilla, eel, Eng. angle, &c. The sense then might be either the ordinary one of motion, the root-meaning of most river names, or it might be the special sense of tortuousness. But as the only appellative I can find is the word anger, a river, in the Tcheremissian dialect of the Finnic (Bonaparte polyglott), I think it safer to follow the most common sense, though the other may not improbably intermix. The derivation of Mone, from Welsh eog, salmon, I do not think of.

From the Sansc. pî, to drink, also to give to drink, to water, Gr. πιω, πινω, we may get a form pin in river-names.

From the above Sansc. pi we may also derive the form pid. The only appellative I find, (if it can be called one), is the Ang.-Sax. pidele, a thin stream, given by Kemble in the glossary to the Cod. Dip.; and hence the name Piddle, of several small streams. The only name I find in the simple form, and that uncertain, is the Pindus of Greece. Then there is a form peder, which seems to be from a definite word, and not from the simple suffix er.

Also from the Sansc. root pi, to drink, to water, we get the form bib or pip, as found in Lat. bibo, and in Sansc. pipâsas, toper. Here also in the simple form I only find one name—the Beuve in France, Dep. Gironde. In the form biber there are many names, particularly in Germany. Graff (Sprachschatz), seems to refer the word to biber, beaver, but Förstemann, with more reason, as I think, suggests a lost word for water or river.

Perhaps also from the root pi we may derive the Ir. buinn, river, bual, biol, water. From the former Mr. Charnock derives the name of the Boyne, a derivation which I think suitable, even if we take the ancient form Buvinda, (Zeuss, Gramm. Celt.,) which might be more properly Buvinna, as Gironde for Garonne in France. For the Bunaha in Germany, the Old Norse buna, scaturire, might also be suggested.

From the Ir. biol, buol, I derive the following, keeping out the rivers of the Slavonic districts, which may be referred to the Slav. biala, white.

A very obscure root in river-names is gog or cock. The only appellatives I find are in the Celtic, viz., Gael. caochan, a small stream, Arm. goagen, wave; unless we think also of the word jokk, jöggi, which in the Finnic dialects signifies a river; and in that case the most probable root would be the Sansc. yug, to gush forth. To the river Coquet, in Northumberland, something of a sacred character seems to have been ascribed; an altar having been discovered bearing the inscription "Deo Cocidi," and supposed to have been dedicated to the genius of that river. Again, we are reminded of the Cocytus in Greece, a tributary of the river Acheron, invested with so many mysterious terrors as supposed to be under the dominion of the King of Hades. Possibly, however, it might only be the similarity, or identity, of the names which transferred to the one something of the superstitious reverence paid to the other. At all events, I can find nothing in the etymology to bear out such a meaning.

From the Sansc. mid, to soften, to melt, (perhaps formed on the root mi, p. 59), come Sansc. miditas, fluid, Lat. madidus, wet. Herein seems a sufficient root for river-names, but there is another which is apt to intermix, Sansc. math, to move, whence, I take it, and not from the former is Old Norse môda, a river. I separate a form med or mid, in which the sense of medius, and also that of mitis, is in some cases clearly brought out; and another, muth or muot, which, though from the same root, as I take it, as môda, a river, (math, to move), has more evidently the sense of speed.

The only appellative for a river which I find derived from its sound is the Sanscrit nadi, Hind. nuddy, from nad, sonare. Whether the following names should come in here may be uncertain; I can find no links between them and the Sanscrit; perhaps the root nid, p. 54, may be suitable.

The only words I can find at all bearing upon the following river-names are the Serv. jezor, Bohem. and Illyr. jezero, lake, wherein may probably lie a word jez, signifying water. But respecting its etymology I am entirely in the dark.

Another word, of which the belongings are not clearly to be traced, is the Armorican houl, houlen, unda, to which we may put the following.

In the Irish and Obs. Gael. dothar, water, Welsh diod, drink, diota, to tipple—with which we may perhaps also connect the Lapp. dadno, river, Albanian δέτ, sea, and Rhæt. dutg, torrent, we may find the root of the following.

From the Welsh wyl, Ang.-Sax. wyllan, Eng. well, to flow or gush, (Sansc. vail, to move?), we got the following.

A word which appears to have the meaning of water or river, but respecting the etymology of which I am quite ignorant, is asop or asp. That it has the above meaning I infer only from finding it as the second part of the word in the ancient river-names Cho(aspes), Hyd(aspes), and Zari(aspis). In an independent form it occurs in the following. Lhuyd, (in the appendix to Baxter's glossary), referring to Hespin as the name of sundry small streams in Wales, derives it from hespin, a sheep that yields no milk, because these streams are almost dry in summer. This derivation is unquestionably false so far as this, that the two words are merely derived from the same origin, viz., Welsh hesp or hysp, dry, barren. But whether this word has anything to do with the following names is doubtful; it seems at any rate unsuitable for the large rivers, such as the Hydaspes, (the Jhylum of the Punjaub). From the derivation of Mone, who finds in Isper, as in Wipper, p. 64, a word per, mountain, I entirely dissent.

CHAPTER V.

THAT WHICH RUNS RAPIDLY, FLOWS GENTLY, OR SPREADS WIDELY.

In the preceding chapter I have included the words from which I have not been able to extract any other sense than that of water. As I have before mentioned, it is probable that in some instances there may be fine shades of difference which would remove them out of that category, but whenever I have thought to have got upon the trace of another meaning, something has in each case turned up to disappoint the conditions.

In the present chapter, which comprehends the words which describe a river as that which runs rapidly, that which flows gently, that which spreads widely, there may still in some cases be something of an appellative sense, because there may be a general word to denote a rapid, a smooth, or a spreading stream.

Among the rivers noted for their rapidity is the Rhone. This is the characteristic remarked by all the Latin poets—