PltysiograpJty of Yarra, etc. 509 



takes its place, and by this means the island is gradually re- 

 moved, and the rocks beneath are reduced to the general level. 

 If a temporary grade has been reached, the same process may 

 take place several times, until the stream arrives at a perma- 

 nent grade, when the islands finally disappear. Hence the 

 absence of islands in the lower well-graded parts of the 

 Yarra is accounted for. 



In well graded streams however, as in the Yarra near 

 Heidelberg, numerous trunks and boughs of trees collect. At 

 first sight, this fact would appear to be favourable to the for- 

 mation of snuill islands, but the latter are rarely present. 

 The explanation is apparently that a trunk or bough of a 

 tree does not usually offer the continuous bar across the sti'eam 

 as a whole or in part that a band of rock does. The latter 

 retains the silt and sand collected, whilst the former, owing to 

 its generally tilted, uneven and often movable jjosition, allows 

 the current to wash around it, and so remove the material that 

 might otherwise have formed an island. 



An interesting eixample, however, of the formation of an 

 island through driftw^ood was recorded in the " Age " news- 

 paper on the 20th of April last. This occurred at Watts Gulch, 

 in the Snowy River, below Orbost. The island " appears to 

 have found a beginning in some floating debris which lodged 

 against an embedded log at very low water. It grew gradually 

 in magnitude with fresh accumulations of silt and debris, and 

 is now covered wdth a prolific growth of vegetation, including 

 scrub and grass, binding the soil into a compact mass. An- 

 other small island has since grown up under similar condi- 

 tions." The channel was becoming blocked, and the Govern- 

 ment agreed to expend £200 in removing the islands. 



The islands in the Warrandyte Gorge grow faster in the 

 directions parallel to the stream than across it. This is 

 obviously due to the river keeping for itself on each side of 

 the island as wide a passage as possible, and is somewhat 

 analogous to silt jetties in lakes. 



The formation of islands nmst also deflect the stream in 

 places, causing it at times to impinge on one bank more than 

 it otherwise would do, and thus tending to increase the curves 

 of a stream. Tlie course of the river is often directly influenced 



