The names thus given at the outset were of the utmost simplicity, rarely, if ever, containing a compound idea. They were indeed for the most part simple appellatives, being most commonly nothing more than words signifying water. But these words, once established as names, entered into a different category. The words might perish, but the names endured. The words might change, but the names did not follow their changes. Inasmuch as they were both subject to the same influences, they would most probably in the main be similarly affected by them. But inasmuch as the names were independent of the language, they would not be regulated in their changes by it. Moreover, in their case a fresh element came into operation, for, being frequently adopted by races speaking a different language, they became subject to the special phonetic tendencies of the new tongue. The result is that many names, which probably contained originally the same word, appear in a variety of different forms. The most important phonetic modifications I take to be those of the kind referred to in the next chapter. There is no branch of philological enquiry which demands a wider range than that of the origin of the names of rivers. All trace of a name may be lost in the language in which it was given--we may have to seek for its likeness through the whole Indo-European family--and perhaps not find it till we come at last to the parent Sanscrit. Thus the name of the Humber is probably of Celtic origin, but the only cognate words that we find are the Lat. _imber_ and the Gr. ὄμβρος, till we come to the Sansc. _ambu_, water. Celtic also probably are the names of the Hodder and the Otter, but the words most nearly cognate are the Gr. ὕδωρ and the Lith. _audra_, (fluctus), till we come to the Sansc. _ud_, water. Again, there are others on which we can find nothing whatever to throw light till we come to the Sanscrit. Such are the Drave and the Trave, for which Bopp proposes Sansc. _dravas_, flowing. And the Arve in Savoy, which I cannot explain till I come to the Sansc. _arb_ or _arv_, to ravage or destroy, cognate with Lat. _orbo_, Eng. _orphan_, &c. And--far as we have to seek for it--how true the word is, when found, to the character of that devastating stream; and how it will come home to the frequenters of the vale of Chamouni, who well remember how, within the last few years, its pretty home-steads were rendered desolate, and their ruined tenants driven out like "orphans" into the world! With such fury does this stream, when swollen by the melted snows, cast its waters into the Rhone, that it seems to drive back the latter river into the lake from whence it issues. And Bullet relates that on one occasion in 1572, the mills of Geneva driven by the current of the Rhone were made for some hours to revolve in the opposite direction, and to grind their corn backwards.