occurrence this conglomerate was pronounced to he of Palaeozoic age, although 
up to that time it was considered to he, prohahly Cretaceous. But, in its 
physical aspect, the conglomerate of the Salt Bange agreed with a similar 
conglomerate, or honldcr hed, in the Talcliirs, the hasal group of the Gond- 
wana System, and thus this also Avas declared to he of Palaeozoic (Upper 
Carboniferous) age. By a subsequent inspection on the spot, by Mr. B. D. 
Oldham, the above observations could not be verified, so that I did not feel 
justified in accepting the Salt Bange conglomerate as a proof of the Palaeozoic 
age of the Talchir boulder bed, and I Avas, therefore, forced to look out for 
other evidence. 
I turned first to Africa, Avhere the Karoo formation, with tlie under- 
lying Dwyka Conglomerate, is an analogous formation to our Gondwana 
System. As in Africa, this conglomerate is underlaid by strata, containing 
fossil plants of Upper Carboniferous age, I thought it natural that it could 
not be older than this ; that the beds immediately above it (the so-called 
Upper Ekka beds) might be Permian, the Karoo formation proper Triassic, 
and so on. Both the Talchir and Dwyka Conglomerate are considered to 
have been formed by the action of floating ice. 
With regard to Australia, important observations have been lately 
made also. 
There is no doubt that the Bacchus Marsh conglomerate in Victoria, 
Avhich also shows unmistakable signs of ice action in its formation, is contem- 
poraneous Avith, or, at least, analogous with, the mentioned conglomerates in 
the Talcliirs in India, and at the base of the Karoo formation (Dwyka 
Conglomerate) in South Africa. The next loAver beds in Victoria are the 
Avon Sandstones of LoAver Carboniferous age. 
In X cav South Wales the relations are somewhat different. Up to 
1885 the conglomerate* in the Hawkesbury rocks was believed to be the 
equivalent of that in the Bacchus Marsh beds and in the Talchir Group, the 
more so, as it also showed signs of having been brought together by drift ice. 
More recent observations (1885) ha\'e, however, shown that in the marine 
beds below the XeAvcastle Coal Measures there is another conglomerate which 
is to be correlated with that in the Bacchus Marsh beds, and, consequently, 
also Avith that in the Talchir Group of India, while the conglomerate of the 
Hawkesbury rocks is of later date. These observations placed, therefore, the 
* [See note on p. 2S — C.S.AA T . 
