ON THE PLANT REMAINS OF THE SCOTTISH PEAT MOSSES. 799 



a thin pasture of Nardus striata, Festuca ovina, J uncus squarrosus, Galium saxatile, 

 with stunted plants of Calluna vulgaris and Erica cinerea. Peat does not appear to 

 form again, although some of the areas have been in this condition for at least more 

 than a hundred years. Account must, however, be taken of the fact that the ground 

 is thinly stocked with sheep. 



It is obvious to anyone visiting the district that peat is being removed by natural 

 agencies, amongst which wind plays a prominent part ; but it is by no means so easy to 

 ascertain the exact way in which the process begins. No denudation can take place 

 until the close covering of mosses, Calluna, and Ericas is broken. This may be accom- 

 plished by wind or heavy rain acting upon some spot in which the vegetation is 

 temporarily thin, and so causing a slight depression. Water collects in such a place, 

 and, freezing in the winter, disintegrates the surface peat and so renders it easily washed 

 away by heavy rain, or when dry in the summer, easily blown away by wind. Once 

 such a bare hollow is formed, it is easy to see how the peat is disintegrated and carried 

 away by frost, wind, and rain, the result being the formation of a channel ever widening 

 and deepening year by year. But the initial stages require investigation, as it is a 

 phenomenon met with in most peaty districts in Britain, and which in its ultimate stages 

 produces the most striking changes in the flora and appearance of the country. In the 

 Walls district the various stages can be seen to perfection. On Sandness Hill, Stane- 

 vatsoe Hill, and the Dale of Burn, the older channels are 8 and 10 feet in depth and 

 15 feet across, whilst there are all intermediate conditions down to furrows not more 

 than 2 inches in depth and a few inches across. Curiously enough, these smaller 

 furrows generally point downhill, suggesting that heavy rain plays some part in their 

 formation. The newer peat deposits, with a dry surface and a close sward of vegetation, 

 are usually very little denuded ; but the older series, with a vegetation of Scirpus 

 c&spitosus, Eriophorum vaginatum, Erica Tetralix, Rhacomitrium lanuginosum, and 

 Drosera rotundifolia, are usually very furrowed. The characteristic appearance of 

 some of these areas is shown in PL II. fig. 4. One interesting feature in this con- 

 nection observed in the district is the fact that denudation appears to be more active 

 below a certain elevation. In stormy and windy weather the cloud-belt often lies very 

 low in these islands, the hills being covered with cloud down to 500 or 600 feet above 

 sea-level. Above this limit, denudation is not going on very rapidly ; at lower eleva- 

 tions the denudation channels are deeper and more numerous. But there are many 

 exceptions to this rule, such as the case of hillsides in an exceptionally exposed position, 

 as, for example, the summit and western slope of Sandness Hill, about 800 feet, within 

 half a mile of the west coast, where the peat is being bodily blown away, and where 

 large mats of the vegetation are rolled back by the wind during storms. 



Where any considerable portion of the vegetation has died away, the wind plays a 

 foremost part in removing the peat, as can be seen by the steeper sides of the furrows, 

 which all face to the westward. In the Dale of Burn the peat is much denuded, and 

 the floor of the valley, viewed from the westward, appears to be bare peat ; seen from the 



