300 PKOF. T. W. EDGEWOKTH DATID ON [May 18967 



in addition typical roches moutonnees (R. D. Oldham, op. at. p. 160). 

 The marine fauna associated with the Boulder-beds of the Salt 

 Kange, and partly underlying, partly overlying them, suggests that 

 they may be homotaxial with those of Eastern Australia and 

 Tasmania. 



This short summary of references to evidence of Permo-Car- 

 boniferous ice-action in Africa and India shows that the glacial 

 phenomena in those regions may probably be correlated on physical 

 and palseontological grounds with those already described as- 

 occurring in Australia, and the recent discovery of rocks of Lower 

 Gondwana age at Bajo de Velis, in Argentina, 1 renders it not 

 improbable that the evidences of glacial action which are recorded 

 as having been observed in Brazil may be homotaxial with those of 

 Southern Africa, India, Australia, and Tasmania. 



IY. Pkovisional Deductions. 



If the correlation suggested for the Australasian glacial beds be 

 admissible, it is probable that ice-action of some kind was taking 

 place over a large area of the Australasian region in Permo- 

 Carboniferous time. This region extended at least from Zeehan 

 in Tasmania, in lat. 42° S., to the Bo wen River Coalfield in 

 Queensland, in lat. 20° 30' S., and from long, about 137° 30' E. 

 (Curramulka) to about 151° 30' E. (Maitland). In Victoria many 

 hundreds certainly, and probably several thousands, of square miles 

 are still occupied by glacial beds, and it is likely that very large 

 areas once glaciated have had all traces of glaciation effaced by 

 denudation. 



The fauna and flora associated with the Permo-Carboniferous 

 glacial beds of Australia render it probable that these glacial beds 

 are homotaxial with the Dwyka glacial beds of Southern Africa and 

 the Talchir Boulder-beds of India. In the case of Australia, Southern 

 Africa, and India, the general direction in which the ice moved 

 appears to have been from south to north. In Australia the thick- 

 ness of the glacial beds (unparalleled, so far as the author is aware,, 

 in any other part of the world, being about 2000 feet, if the inter- 

 calated beds of sandstone and conglomerate are included in the 

 estimate) implies that the Permo-Carboniferous glacial epoch in the 

 Southern Hemisphere was of prolonged duration. This inference 

 is supported by the fact that in New South Wales a group of Coal 

 Measures, over 230 feet thick, and comprising from 20 to 40 feet in 

 thickness of coal (the Greta Coal Measures), is sandwiched in 

 between the erratic-bearing horizon of the Lower Marine Series 

 and the similar horizon of the Upper Marine Series. 



As will be seen from the vertical section accompanying this paper 

 (PI. XII.), there is evidence at Bacchus Marsh of at least from nine to 

 ten distinct boulder-bed horizons, separated one from another by thick 

 deposits of sandstone and conglomerate. The exact significance of 



1 Eec. Geol. Surv. Ind. vol. xxviii. pt. iii. (1895) pp. 111-117, and Revista del 

 Museo de la Plata, vol. vi. pp. 117 et seqq. 



