CONTINENTAL PROBLEMS OF GEOLOGY. 171 



and possibly there should be no dissent — but the evidence on which it 

 is founded appears to nic so far from conclusive that I venture to 

 doubt. 



The evidence employed consists partly in the geueral distribution of 

 formations as shown by the oeoloyic map and i)artly iu inferences drawn 

 from certain formations which contain internal evidence that they origi- 

 nated on coasts. With the aid of such data are drawn the outlines of 

 ancient ocean and land at ^■arious geologic dates, and from the com 

 parison of these outlines contineutal growth is infened. In passing 

 from the formation boundaries of the geologic map to the oceanic limits 

 of the charts of aiuuent geography, allowance is made for the former 

 extent of non-littoral formations beyond their present boundaries. 

 This allowance is largel.y conjectural and the range of possible error is 

 confessedly great. In i)assing from the observed limits of littoral for- 

 mations to the coast lines of ancient geography little or no allowance 

 is usually made for the former extent of the formations, and I conceive 

 that great possibility of error is also thus admitted, louring a period 

 of oceanic transgression over the land, all portions of the transgressed 

 surface are successively coastal, and tlie coastal deposits they receive 

 are subsequently buried by off-shore deposits. When tlierefore littoral 

 beds are found in remnants of strata surviving the processes of deg- 

 radation, it is indeed proper to infer the ])roximity of ancient coasts 

 during their formation, but the inference that they represent the limit 

 of transgression for that epoch may be far from the truth. For these 

 reasons it appears to nie that the specific conclusions which have been 

 reached with reference to the original extent of various formations are 

 subject to wide uncertainties; and, if this be granted, then but brief 

 attention to a simple law of (Unuidation is necessary to show that the gen- 

 eral conclusion maybe illusory. The process of degradation by aque- 

 ous agencies is chiefly regulated, not by the thickness of formations, 

 but by the height to whicli they are uplifted. Thus the present extent 

 of most formations is determined in large part by crustal oscillations 

 subsequent to their deposition. As formations are progressively eroded, 

 the under-lying and older can not be attacked until the over-lying and 

 younger have been carried away, and so the outcro})s of the older of 

 necessity project beyond the boundaries of the younger. The progress 

 of vague inferencte, making indefinite allowance for tlieuidcnown quan- 

 tity of eroded strata, nearly always assigns to the older formation, 

 wlii<;h projects visildy beyond the newer, a greater original extent. It 

 ap])ears to me thus possible that the greater part of the data from which 

 contineutal growth is inferred may be factitious and misleading. 



Furthermore, inference, such as it is, deals with only one phase of 

 the problem. It is applied to the incursions of the sea upon the land, 

 but it is not ai)plied to the excursions of the lan<l upon the sea. Just 

 as we infer from stratified rocks the presence of the sea, so also we 

 infer from unconformities the .sea's absence; and to the student of 



