298 PROF. T. W. EDGEWORTBZ DAVID ON [May 1 896, 



(iii) Sandstones. — These vary from hard to soft, from 

 fine to coarse, and are frequently laminated, the laminae 

 occasionally showing distinct evidence of contortion, espe- 

 cially in the neighbourhood of the irregular pockets of 

 conglomerate. Individual beds vary in thickness from a 

 few feet to nearly 100, being mostly about 30 feet thick. 

 "Well-preserved plant-remains are present on at least 

 two horizons ; on the lower horizon occur the three species 

 of Gangamopteris already referred to, and on the higher 

 specimens of Zeugophyllites, Schizoneura, etc. The total 

 thickness of the glacial beds seen in the upper portion of 

 Korkuperrimal Creek, as measured last December, proved to 

 be 1427 feet. To this, Mr. Brittlebank estimates, a thickness 

 of about 700 feet of strata should be added to carry the 

 section from the top of the Gangamopteris-he&s to the top 

 of the strata seen above the ScJiizoneura-honzon. (The 

 order of succession of the beds is illustrated in the accom- 

 panying horizontal and vertical sections, PI. XII.) 



The altitude of the glacial beds varies from 600 to about 1400 

 feet above sea-level. The source of the erratics is not known, but 

 evidence shows that the ice which furrowed the rocks came, as 

 already stated, from the south. 



The most northerly point to which drift containing undoubtedly 

 glaciated boulders has been traced is Springhurst, on the main 

 railway-line between Melbourne and Sydney, at the boundary 

 between Victoria and New South Wales. 



III. Correlation of the Glacial Deposits. 

 (a) Australasian. 



It is extremely probable that the glacial beds of Bacchus Marsh, 

 Wild Duck Creek, and Springhurst in Victoria were of homotaxial, 

 if not of contemporaneous origin. They may probably be correlated 

 with the glacial conglomerates of Mount Eeid and Mount Tyndall 

 in Tasmania. 



The above correlations are based chiefly on lithological evidence ; 

 there is, however, good palaeontological evidence for the correlation 

 of the Bacchus Marsh glacial beds with the erratic-bearing Permo- 

 Carboniferous mudstones of Maria Island, One Tree Point, and 

 Bruni Island in Tasmania, the similar beds at Maitland, Branxton, 

 and Grasstree in New South Wales, and those of the Bowen River 

 Coalfield in Queensland, as the genus Gangamopteris is distributed 

 abundantly throughout the formations at all these localities. The 

 glacial evidences at Hallett's Cove and at Curramulka in South 

 Australia may safely be correlated one with another, and are very 

 likely homotaxial with the above-mentioned glacial deposits in East 

 Australia and Tasmania. 



