Welsh Journals

when large numbers of people come to the island. The effects of this are difficult to estimate since few details of T. pisana at this locality have been published (Stubbs 1900; Wintle 1925). The Priory Bay colony appeared to be thriving; but, perversely since it is much less subject to the depredations of the day-trippers, the Sandtop Bay colony was extremely sparse and may be unlikely to survive. The vegetation on the south side of the island (T. pisana recorded there by Stubbs) is largely short turf, unsuitable for T. pisana. The two tiny colonies at Saundersfoot are also in danger of extinction, being at the edge of a very popular beach, and in small isolated patches of vegetation. The larger colony at Stackpole Warren seems less subject to human interference, despite being close to a popular beach, and may not be in any imminent danger from this source. In Tenby itself, the overall distribution of the species has remained much as indicated by Stubbs, but the colony has been fragmented by the development of housing, roads, etc. During the course of the survey a part of the colony was destroyed when the 'fundrome' was demolished, and presumably part was also destroyed during its construction. During the summer the Esplanade and paths down to the beach are in constant use by large numbers of people. In order to keep the area looking tidy, and to allow unhindered views, the alexanders in this area are cut down each year in about July. At this time of year the upper stems are dying, but at places where there is no cutting, they remain standing and provide positions for T. pisana to sit above ground. Many specimens are found doing this. Despite these pressures, T. pisana has remained abundant in this area. The drastic reduction of the Narberth Road population and the restriction of the Marsh Road population seem likely to be due to the destruction of habitat resulting from the metalling and widening of the roads, increased traffic, and the building of houses alongside parts of the roads. The banks along the Narberth Road are now largely overhung by trees and appear quite damp; as such they do not provide the kind of habitat in which T. pisana is usually found, that is, an unshaded, more exposed habitat; and this must also be important. The distribution of the species is probably restricted by temperature (Cowie 1982; 1984). In Britain it is found in open, sunny habitats only in regions with high winter temperatures. This, combined with its striking restriction on Castle Hill and St. Catherine's Rock at Tenby to the south and west sides only, strongly suggests that climatic changes would also be important in reducing its range, or indeed in allowing it to expand. It is difficult to estimate how rapidly a species would respond to an improvement in climate by widening its range (Cowie 1980, 1982), but certainly its response to a worsening climate would be faster (Lamb 1970). The slight warming of the British climate between about 1890 and 1940 (Lamb 1982) could, therefore, have allowed an increase in abundance and a widening of the distribution of T. pisana around Tenby. The present reduction in abundance and range may be the result not only of human pressure, but also of the reversal of this climatic trend and the increased climatic variability, both described by Lamb (1982). Furthermore, small isolated colonies at the edge of a species' climatically-determined range are susceptible to short term fluctuations such as an especially cold winter. This may be another factor involved in the extinction of T. pisana at Manorbier in particular. In conclusion, since the local distribution seems strongly influenced by climate, changes in the range of the species may well have partly resulted from climatic fluctuations. The overwhelming influence, however, has been the activities of Man. Although T. pisana is still abundant at Tenby in the areas in which it remains, availability of suitable habitat has decreased since the turn of the century as a result of urban development and increased pressure from tourism. Changes in vegetation, in particular the planting of sea buckthorn on the Burrows, but also increased overshading on the Narberth Road, have also been important. Elsewhere in Wales, development of the embankment along the coast from Swansea to Mumbles has resulted in the extinction of T. pisana in this area, leaving the two sites near Porthcawl (Cowie 1982) as the only remaining Welsh localities not described in this paper. Despite its ability to survive in small insolated patches of suitable vegetation, T. pisana is under considerable pressure from human activity in South Wales. Acknowledgements. I am grateful to Professor A. J. Cain for discussion and for reading earlier versions of this paper. Professor R. J. Berry and Dr. G. M. Davis both offered valuable advice, and Mr. J. Humphreys and Mr. L. R. Noble made useful comments on the manuscript. Mrs. J. E. Hunter assisted with the survey.