350 T. S. Ellis— River Curves. 



We have, however, in this occurrence a clear case of the occur- 

 I'ence of corundum (and spinel) in a member of the ' Charnockite 

 Series ' in Ceylon, in a thin zone interbanded with normal varieties 

 of granulite. It seems more likely that in this case the presence of 

 corundum has resulted from a local variation in the constitution of the 

 consolidating magma than that the magma should have absorbed some 

 ■rock rich in alumina of which no trace remains here or elsewhere. 



The additional alumina can hardly have been obtained from the 

 yellow micaceous band, for in that case the accompanying iron and 

 magnesia would probably have prevented the formation of corundum ; 

 this band seems rather to have consolidated from a 'magmatic streak' 

 which, though (like the corundiferous band) rich in alumina, contained 

 too much iron and magnesia to allow of the formation of corundum. 



YI. — KiVEE, Curves kound Alluvial Plains. 



By T. S. Ellis. 



(PAGE-PLATE XIX.) 



N the Geological Magazine for October, 1902, Dr. Callaway 

 mentions the explanation of these curves that I gave in 

 Sk paper printed twenty-one years ago. His quotation should be 

 read with the immediate context — " These [the tributary streams] 

 keep open a channel into which the larger stream falls." This is 

 the essence of my argument. 



Professor Phillips and Sir A. Geikie have remarked that an 

 alluvial plain in the course of a river may be regarded as an old 

 lake-bottom, now drained ; the lake-like appearance being renewed 

 in times of flood. Let us suppose ABCD (Fig. 1) to represent such 

 an area with a river flowing through it in a straight line, and, on one 

 side, a tributary stream, or, to use a shorter and more expressive 

 term, an affluent, coming in at an angle. Such a condition, if it 

 existed, would not continue even in consolidated alluvial soil ; it is 

 still less likely to have existed in the soft mud when the area was 

 first drained. A succession of floods would certainly wash away 

 the bank where the affluent, coming through it, had caused a break 

 in its continuity. By this the river-bed opposite the affluent would 

 be expanded beyond its requirements when at low watei*. At every 

 flood the whole of the bed of the river and the adjoining area of 

 land will be covered with water, the ordinary river channel being 

 effaced. From this water suspended matter falls and forms 

 a deposit, visible after the flood has subsided, but in greater 

 quantity on the banks by the sides of the low- water channel than on 

 the adjoining laud. The difference is, in my view, thus explained. 

 The water deep down in the river-bed is subjected to great pressure 

 by that above it, so as to be, relatively to the freely moving current 

 on the surface, stagnant. Any suspended matter in the current that 

 falls into this relatively stagnant water sinks to the bottom, and, in 

 the absence of current there, remains ; whereas, in the flood over 

 the land adjoining the river, thei'e is less depth of water, and 

 therefore less liability of the suspended matter to fall below the 

 line of current necessary to keep it in suspension. 



