292 Charles Fennev : 



railway runs. The basalt sheet may here be regarded as con- 

 tinuous, and yet we have evidence of a '' tongue," the latter being 

 exposed by the dissection at Dog Traj) Gully. Along th« right 

 bank of this gully we see the basalt as a thin sheet covering thei ter- 

 tiary materials. On the left bank, basalt continues right to the 

 bottom of the gully, and it is. ^lear that Dog Trap Gully has cut 

 its valley just alongside a large ancient basalt-filled valley. This 

 old river flowed right across where the scarp now is, and it is diffi- 

 cult to know its exact course on the other side of the fault line. 



The Parwan has cut right through the basalt sheet in the- 

 lowest part of its course, and the writer knows of no exposed 

 basalt-filled sections there. There remains only the hard bar of 

 basalt that the Parwan is still engaged in cutting through, just 

 below the scarp, and in the reconstruction (Fig. 35), it is sugofest^d 

 that the ancient and important stream now being discussed turned 

 southward there as shown by the arrows. 



. We have now, for comparison, two basalt-filled rivers crossing 

 fault lines : (i.) the ancient Korweinguboora Creek, as already 

 described, and (ii.) the case mentioned in the preceding para- 

 graph. Whiles in the former case the surface of the basaltic tongue 

 maintains an even gentle grade right across the fault line, in the 

 latter the surface of the basalt descends steeply for over 300 feet. 



The second method of discovering ancient river beds is perhaps 

 of greater interest, namely, to piece together the known sections of 

 old valleys exposed in present river cuttings. The most complete 

 evidence of this nature is to be found in the area of deep dissection 

 in the neighbourhood of Werribee Gorge, and Fig. 36 represents 

 that locality on a larger scale. Twelve separate old river sections 

 are shown in this diagram, numbered 1 to 12. 



The writer is greatly indebted to Mr. C. C. Brittlebank for a 

 knowledge of the majority of these sections. That gentleman gene- 

 rously spent a part of his vacation in the field with the writer, con- 

 ducting him to such places of interest as might be of value from 

 the physiographic standpoint, and giving freely of his intimate 

 know^ledge of the whole Werribee area. 



Nos. 1 and % (Fig. 36) are exposed in the valley of the Werribee 

 about a mile above its junction Avith Pyke's Creek. The Werribee 

 valley is here about 200 feet deep, but it has not yet cut quit-e 

 through the basalt tongue. The old river valley section is well 

 shown; it is cut in granodiorite, is probably about 150 feet deep, 

 and with moderately sloping sides. A little higher up a small 

 tributary valley came in from the west. 



