207 NOTES Watsonia 25 (2004) Notes TOFIELDIA PUSILLA (MICHX.) PERS. AT THE SCOTTISH BORDER

from Berwick-upon-Tweed on the North Sea coast. Given the northern and montane distribution of this plant, as it is now known, this is a remarkable locality and this note presents evidence on its exact whereabouts and botanical character and considers whether the question of whether the record belongs to England or Scotland can be resolved. John Ray (1677) records ‘about two miles from Berwick, by the side of a rivulet, in boggy ground, not far from the road leading to Edinburgh, we found a sort of Pseudo-asphodelus which I had never before seen, much less than that common in England, having, as I guess, white flowers in a spike, to which succeed roundish seed-vessels. The stalk of the spike is naked, or not having above one leaf, the spike itself short, the root fibrous, as that of the common’. As the Scottish border lies a little under three miles from Scotch Gate, the northern exit from the walled town of Berwick, this record nominally falls in England (v.c. 68). Ray gave an abbreviated account of the same find in his ‘Synopsis ’ (1724) as ‘juxta rivulum non procul Bervico in Scotia’. There is an ambiguity as to whether Ray wrongly thought Berwick to be in Scotland or whether he sought to localise the record in Scotland. As this detail is not in the earlier account, which presumably was drawn from a field notebook, its authority is weakened. Ray may not have been clear just where the Scottish border lay as it may have been marked less well than by the later Lamberton Toll. No subsequent botanist has refound the plant in this area, though many have looked for it, notably George Johnston and Charles C. Babington in 1834. Johnston (1829, 1853) quotes Ray but does not attempt to allocate the record to a country. George Swan in his recent ‘Flora of Northumberland ’ (1993) follows Johnston (1829) who wrote ‘Ray seems to have found Tofieldia palustris by the side of that stream which rises on this hill (Lamberton) and runs by the Shields’, which would place it in Scotland (v.c. 81), despite Lamberton Shiels lying a full four miles from Berwick. This interpretation has been followed by Stace et al. in the ‘Vice-county Censu