OKICIX OF RAND 11A\KI:TS. 59 



In the Lower Witwatersraiid bevls the various strata are 

 sufficiently well characterised to allow their recognition over 

 considerable areas, and the series seems to be a normal succession 

 of slates and sandstones ; but the Upper Witwatersrand series, 

 with its 10,000 feet of quartzites and conglomerates, presents 

 difficulties of deposition which are harder to explain. The Table 

 Mountain sandstone consists of 5,000 feet of fairly normal sand- 

 stones, cemented, it is true, with a certain amount of silica. 

 and in the fresh state with a considerable amount of iron sul- 

 phide, which disappears on the surface owing to the coarse 

 texture of the rock allowing free circulation of oxygenated water. 

 This can he explained as deposit in a channel near inshore, but 

 even here the conditions were unable to last the whole time, 

 and there is a big bed of slates near the to]5. Ten thousand feet 

 of inshore deposit — and this compacted and compressed so that 

 it re])resented probably only half the original thickness of the series 

 — is a very great thickness, and seems beyond what could happen 

 under conditions of deposition such as we know them: this is 

 the state of the case with regard to the Upper Witwatersrand 

 beds. The sand is derived from the weathering of a continent, 

 and must be laid down near the shore to account for the gravelly 

 l)atches, now banket ; one has to assume that the deposit was 

 always a shallow one, sometini'es even being exposed above the 

 water-level, so that the floor of the ocean in which the sand was 

 being laid down would have to sink continually, the land being 

 supposed to rise in compensation and to provide fresh material 

 for disintegration. To assume that the Upper Witwatersrand 

 beds is a single continuous series, we must also assume that the 

 crust sank continuously 15,000 to 20,000 feet, and that the shore 

 remained in the same place all the time. 



It seems therefore that, although the deductions from field- 

 survey must always carry more weight than theoretical specula- 

 tions, nevertheless, there are certain facts about the stratigraphy 

 of the Witwatersrand which leave one with curious false con- 

 clusions if tested from a causal point of view. We know so 

 little about the conditions of the earth in remote ages that it 

 is the custom to say that any reference to them is superfluous, 

 and the only thing that matters is the correct plotting of the 

 boundaries of the formations. We now know, however, so much 

 about the earth, and the exploration in so many lands enables 

 us to bring to the elucidation of any one part so many facts 

 from similar countries elsewhere, that it seems to be time to 

 discard the older methods, safer though they be, and try and 

 raise the science of geology from a simple cataloguing of facts 

 into its proper domain as a causal science. 



