With respect to the ending in _a_, found in some English rivers, there is reason to think that it is a word signifying water--the Old Norse _â_, Goth. _ahva_, Lat. _aqua_, &c. So that the _a_ in Rotha may be the same as the _a_ in the Norwegian Beina and the Swedish Tornea--as the _au_ in the Germ. Donau (Danube)--and as the _ava_ in the Moldava of Austrian Poland. Others of these endings have by different writers been supposed to be also words signifying water. Thus Donaldson (_Varronianus_), takes the ending _es_ to have that meaning. And Förstemann, though more cautiously, makes the same suggestion for the termination _ar_ or _er_. "I allow myself here the enquiry whether possibly the river-names which contain an _ar_ as the concluding part of the word may not be compounded with this unknown word for a river; to assume a simple suffix seems to me in this case rather niggardly." So also the ending _en_ has been supposed by some of our own Celtic scholars, as Armstrong and O'Brien, to be the same as the Welsh _aven_, Gael. _amhainn_, water or river, an opinion which has also, though to a more limited extent, received the sanction of Pott. There are various minor objections to the above theories which I forbear to urge, because I think that the main argument against them is to be found in the manner in which these endings run through the whole European system of river-names. And it seems to me therefore more reasonable to refer them to a general principle which pervades the Indo-European languages, than to a particular word of a particular language. The principle I refer to is that of phonetic accretion, and it is that upon which the above word _aven_ or _amhainn_, is itself formed from a simple root, by one of the very endings in question, that in _en_. Instead then of explaining--as the followers of the above system have done--the Saone (Sagonna) by the Celt. _sogh-an_, "sluggish river", I prefer to point to the general principle upon which the root _sogh_ has the power, so to speak, of making itself into _soghan_ (_e.g._, in Lat. _segn-is_.) Not but that the principle contended for by the above writers may obtain in some cases: the Garumna, ancient name of the Garonne, looks like one of them, though even in this case I think that the latter may be the proper form, and the former only a euphonism of the Latin poets: the geographers, as Ptolemy, call it Garunna.