Cylchgronau Cymru

Constant mention will be made of facts and suggestions contained in Machreth Ellis's thesis. My indebtedness, however, is far greater than indicated by the numerous references. The fact that at least two hundred and fifty five stream-names3 appear in his work points to the extent of my debt. Another source for names and later for guidance on derivations has been R. J. Thomas's work, regretably unfinished, on the names of streams in Wales. (E.A. in Bibliography). He classified the names according to their endings and since several endings remained for treatment in a second volume, many Montgomeryshire names are not covered. The Ordnance Survey maps which have been constantly in use are detailed in the Bibliogra- phy. The earliest OS map, which for Montgomeryshire was published in 1836, has been particularly useful. One has relied on the current OS maps in locating streams, some unnamed on the sheets. With a few exceptions I have in the list reproduced the form and spelling of the Ordnance Survey even when I suspected misrepresentations. Many erroneous forms have been entered by the map-makers but not to the extent implied in R. G. Thomas's observation, (E.A., xv, translated), "often it is not possible to rely at all on the form of the name on the Government maps. In spite of this, it is of value as it is"4. The names in the list can be divided into two groups. The first consists of those which are associated with an existing placename, e.g. Nant Llyn Mawr5, Cwm Harry Brook, while the second consists of those names which are apparently independent of any other placename e.g. Afon Gam, Nant Felen. These groups can be referred to for convenience as associative and independent respectively. Names which refer to unspecified locations, e.g. Esgair, Maes, Carnedd, are treated as belonging to the independent group. In names of the first group the association can be with a physical feature (e.g. the cwm through which the stream flows); a locality (e.g. Guilsfield Brook); habitation (e.g. Pen- rhuddlan Brook) or some other fabrication (e.g. Nant Melin y Grug). The most common association is with settlements and particularly habitations. Physical features are rarer except for the inevitable cwm. Contrary to one's expectation, association with territorial divisions, such as commot, manor or township, is also rarely found. In a name such as Sylfaen Brook which contains a territorial division, the probability is that the association is with a habitation of that name and not with the township. The traffic in names has been in the opposite direction: it is the territorial divisions that have the names of streams, e.g. the townships of Cownwy, Cynorion, Dwyffrwd, Nantfforch, Nantffyllon, Llifior, Nant y Meichiad, Rhysnant. The inter-relations of names of settlements and streams often result in back-formations6. The 3On the number of streams in the county there is an interesting note in the first volume of Byegones62, "What county in England or Wales can rival Montgomeryshire for the number of rivers? Nearly 200 rivers, brooks and streamlets have the whole or a portion of their course within the boundaries of the county." 4This view is confirmed by the account in the article on "Welsh Orthography and Ordnance Survey Mapping 1820-1905 by J. B. Hurley and Gwyn Walters (Arch. Cant. CXXXI (1982), 98) of the attention paid by the map-makers to the problems of collecting names, getting them right and spelling them correctly. The pedantic malformations may in many instances be the result of their well-meaning and zealous concern to be "right". (for examples v. Gobannau and Clydfannau). In spite of the instruction, "When names have assumed a corrupted form, which is thoroughly established their orthography should not be altered even when they are known to be etymologically wrong," the map makers were not always likely to dispute the recommendations of reputed experts either during the survey or later at the revisionary stage concerning the names of farms and streams. 'When a stream is mentioned in this Introduction, the reader is advised to refer to the List. hv. E.R. N., xii of Introduction. for an explanation of this term.