Cylchgronau Cymru

names seem to illustrate this view. The most obvious feature which gives rise to affiliated names is a confluence. The names of the two streams are paired by calling one fawr and the lesser fach. With four tributaries of Evyrnwy in Llanwddyn an unusually logical pattern of names has evolved: Afon Eiddew being the combined waters of Nant Eiddew Fawr and Nant Eiddew Fach, and so with Eunant, Hirddu and Nadroedd. (It should however be added that in the older map, OS 1836, the abnormally perfect formality is not found). In contradiction to the tendency illustrated in these names the north and south branches of the Rhiw are both known simple as Afon Rhiw. The affinity of the two branches is acknowledged in the place name Dwyrhiw for the territory around the confluence. Less obvious is the pairing through a common factor in two names. The factor which pairs Nant Iwrch (stag) and its tributary Nant Ewig (hind) is related animals; in all probability it also accounts for Twrch (boar) and Banw (young pig). Less probable perhaps is the affinity based on colour as the common factor in Nant Goch and Nant Ddu at the source of Afon Clywedog. More problematical is the existence of the common initial root in Miwl and its tributary Meheli. Any investigation of the affinity and interdependence of stream names should extend to the placenames of the neighbourhood of the streams. The people who gave names to streams were also the people who named the hills and valleys, lakes, bogs, peaks and hollows, cairns, habitations, fields and ruins. Stream-names did not come into being in a mechanical way or in accordance with Linnean- like logic. There were no rules that those above a certain length should be called afon, the smaller ones nant and the minute left nameless. What has actually happened is that some very small ones have names while other longer water courses though marked on maps bear no names on them. Is this cartographical anonimity due to the streams not having names or to the failure on the part of the mapmakers in ascertaining and recording them? I have obtained the impression that unnamed streams of significant length are more in evidence in the lowlands than in the uplands of the region. A likely explanation is that in the hills there was a greater need to be able to refer to a stream by name, be it to indicate a boundary or the location of a sick sheep, than in the lowlands where there were many other convenient referents in communications, such as buildings and fields with distinctive names. Many streams are shown without names on maps, even on large-scale ones, though they have names in common use. One can find examples by consulting the Parish Histories in Mont. Coll. in which streams are listed in the section usually entitled Drainage. In the most comprehensive list of this kind, the one in the Parish History of Llanwddyn (Mont Coll. VI, 403), eighteen tributaries of Afon Cedig are named. This list was published in 1873, a decade before the surveying which resulted in the Ordnance maps which appeared in 1903 was started. Only eleven tributaries are named on the 1903 sheets. The current SH 92 and SJ 02 name ten. The early 1836 OS map has seven of these names. Several streams bear the same name. Silvan Evans called this reduplication and observed that the reason for it would be "an interesting subject for investigation." For an examination of duplicated names to be valid it would have to cover places beyond the bounds of, in this study, the Severn basin. It would be necessary to look at the names of streams down river in England, north towards the Dee, on the Teme and Eithon and the Upper Wye in Radnorshire, on the upper reaches of Rheidiol and throughout the Dyfi basin. People's knowledge of places is not limited to what lies within an area prescribed as theirs. The weakness of a study confined to an area is well illustrated by the incidence of Clywedog as a river-name. In the Severn basin it only occurs once but there is a Clywedog to the east in Radnorshire and another to the west in the dyfi