The question then arises--supposing these endings to be phonetic--were they given in the first instance, or have they accrued in after times? It is probable that both ways might obtain; indeed we have some evidence to shew that the latter has sometimes been the case. Thus the Medina in the Isle of Wight was once called the Mede, and the Shannon of Ireland stands in Ptolemy as the Senus. On the other hand cases are more frequent in which the ending has been dropped. Thus the Yare is called by Ptolemy the Garrhuenus, _i.e._, the Garron or Yarron. And the Teme appears in Anglo-Saxon charters as the Taméde or Teméde. Indeed the Thames itself would almost seem, by having become a monosyllable, to have taken the first step of a change which has been arrested for ever. So in Germany the Bille, Ohm, Orre, and Bordau, appear in charters of the 8th and 9th cent., as the Bilena, Amana, Oorana, and Bordine. And in France the Isara and the Oscara have in modern times become respectively the Oise and the Ousche; in both these two cases the ending _er_ has been dropped; for Oise=_is_, not _isar_; and Ousche=_osc_, not _oscar_. This latter principle is indeed only in accordance with the general tendency of language towards what Max Müller terms "phonetic decay"--a principle which seems less active in the rude than in the cultivated stages of society. It would appear as if civilization sought to compensate itself for the increased requirements of its expression, by the simplification of its forms, and the rejection of its superfluous sounds. Upon the whole then I think that as the general rule these endings have been given in the first instance, and that they have but rarely accrued in after times. Such being the case, though in one point of view they may be called phonetic, as adding nothing to the sense, yet in another point of view they may be called formative, as being the particles by means of which words are constructed out of simple roots. And of the names in the following pages, a great part, in some language, or in some dialect, are still living words. And those that are not, are formed regularly upon the same principle, common to the Indo-European system. CHAPTER III. ON THE MEANING OF RIVER-NAMES. The names of rivers may be divided into two classes, appellative and descriptive--or in other words, into those which describe a river simply as "the water" or "the river," and those which refer to some special quality or property of its own. In the case of a descriptive name we may be sure that it has been given--not from any fine-drawn attribute, but from some obvious characteristic--not from anything which we have to seek, but from something which, as the French say, "saute aux yeux." If a stream be very rapid and impetuous--if its course be winding and tortuous--if its waters be very clear or very turbid--these are all marked features which would naturally give it a name.