In conclusion, some inferences may be made regarding the ecological status of Nardus on these Southern Scottish sheep gratings. In relation to the grassland communities and to heather, the role of Nardus is quite certainly that of invader. Under conditions of sheep grazing that have prevailed on these hills for at least two hundred years, the species is able to make an entry into almost any community and, with the exception of a few of the most competitive, eventually to reach a state of dominance. It may be argued from this that since the effect of selective grazing is to modify the competitive balance of the species within the community, that it is likely that this factor is of importance. However, it is equally probable that while this may be true for some communities, in others, the competitive resistance may be low enough - even in the absence of grazing - for Nardus to be able to invade successfully. Amongst the communities studied, those dominated by Deschampsia flexuosa appeared to offer the least resistance to Nardus and it is at least possible that selective grazing is not an important factor in the invasion of these swards, although it may be so in other communities. The manner by which invasion takes place varies from one vegetation type to another, but one characteristic that they have in common is the extreme slowness of the succession. This is particularly well brought out in the invasion of Callunetum, where an advance by Nardus seems to be co- incident with the burning of heather at excessively long intervals, intervals of perhaps fifteen or twenty years or more. It is probable that it may require four, five or more such cycles of growth and burning before Nardus becomes wholly dominant, so that the succession may take anything from sixty to a hundred years(or more to complete. Nardus stricta is not a species of woodland or shade conditions and it is possible that the present tendency towards Nardus dominance is part of a succession that was initiated when the original woodland of the area was destroyed and grazing became general. In this respect, the succession may be parallel to the spread of bracken at the present time in many areas, which is another example of a succession originating in the destruction of woodland. As has already been shown in the account of the past history of land use of the area (p.26), the greater part of the timber on Sourhope must have been destroyed at the latest, by the year 1550 and, in the same locality, many areas have a grazing history going back to the twelfth century. Sheep farming in something like its present day form has been in existence since about 1700. It seems, therefore, that the abundance of Nardus in the vegetation of the Southern Uplands, may be related not so much to the intensity of sheep grazing, as has been suggested by Fenton 1937, as to the length of time that has elapsed since the forest disappeared and grazing, together with its associated management practices, became widespread.