The Scottish Naturalist. 267 



upon by the corresponding pebble-bed. Two other curves, a 

 necessary result of the formation of the first, are formed at the 

 junction of its extreme ends with the direct course of the river. 

 Here, according to our rule, two new pebble-beds should form 

 on the inner side of the curves. This is invariably the case, and 

 the result is that we have three growing promontories encroach- 

 ing upon corresponding banks that are being denuded. The 

 first promontory extends in an outward direction into the major 

 curve ; the two smaller promontories encroach from the opposite 

 side in a lateral direction into the minor curves, that is, upon 

 the first formed promontory. The efforts of these three growing 

 points give the river more and more a circular form until, 

 finally, the lateral promontories meet. The water naturally 

 takes the more direct course, leaving the circular channel a 

 " dead-water." That such is the river's mode of working one 

 has but to turn to the accompanying map (Plate III.) to believe 

 possible, or to pay a visit to the valley to be convinced. If we 

 refer to the map, a fine example of a forming circle maybe seen 

 at B. If the survey map ( 1 8 7 1 ) may be relied on, the form of this 

 extending promontory has materially altered since its publica- 

 tion. Indeed an alteration is so constantly occurring that the 

 inhabitants of the district describe in positive terms how 

 the river has " gone south " within their remembrance ; 

 but the accounts of the rate of its movements are so distress- 

 ingly conflicting that I refrain from giving any. I believe, from 

 my own observation that the growth of the pebble-beds and con- 

 sequent denudation may in some cases be several yards per annum. 

 At A is a portion of an ancient dead-water, i.e., of a discarded 

 channel. There are others leading to B. B is a dead-water of 

 comparatively modern construction. It is partially filled, but 

 still contains in parts a foot or two of water. C is an immense 

 dead-water, through which the river flowed prior to the incursion 

 of the Caledonian Railway. But the engineers cut through the 

 neck of land at I, and gave the river a more direct course, 

 simply anticipating the intention of the stream. The neck of 

 land 'west of C is represented too wide, it is I think not 

 more than ten or twelve yards across at this point. The 

 short distance cut through by the railway reduced the 

 river's journey more than a mile. The stream at F is 

 working at a much slower pace. This will be understood 

 when it is asserted that the tide is experienced as far up the 

 stream as D. D became a dead-water in 1853. Previous to 



