304 A. J. Jukes-Broiciie — Tico Stages in Upper Chalk. 



derived from the rainfall caught on the bare rock surface, the top 

 of the gneiss boss is yet not too high above the stream bed to be 

 beyond the reach of the flood waters arising from the few torrential 

 storms in which the whole annual rainfall (12 inches) of the district 

 falls. No other features bearing on the problem were observed ; 

 still, these facts and the attendant climatic conditions are sufficient 

 to indicate a feasible explanation of the phenomenon. 



The shell of gneiss, except beneath and at the side bordering on 

 the trough, is in continuity with the mass of the rock. The diui'nal 

 range of temperature on the rock surface is considerable in the 

 summer — certainly not less than 100° F. During the day the shell 

 of gneiss, prevented from horizontal expansion on three sides, springs 

 to an arch, and the water in the trough retreats — probably entirely 

 by gravity — through the rift. At night, with a considerable fall in 

 the rock temperature, the shell contracts, and settling down, forces, 

 firstly, air through the narrow irregular rift (hence the whistling 

 and hissing sounds that for half an hour precede the appearance of 

 water in the trough) and finally water, which reaches its maximum 

 height in the trough about daybreak. The first appearance of the 

 water generally occurs about 8 o'clock at night. The expulsion of 

 air prior to that of the water indicates that the water is withdrawn 

 within the rock by gravity and not by any siphon effect. Since the 

 linear expansion of gneiss is approximately only one part in 200,000 

 for every degree f ahr., the cavity must be large and the adjustment 

 of level fairly delicate. This, indeed, is also indicated by the fact 

 that, for some time after heavy rains, water remains during the 

 day in the outer trough. In the absence of any evidence of displace- 

 ment of the outer shell of gneiss, it may be assumed that the internal 

 cavity has been assisted in reaching its obviously large dimensions 

 by disintegration of its walls in the same manner that a gnamma 

 hole grows. 



V. — On the Recognition of Two Stages in the Upper Chale. 

 By A. J. Jukes-Bkowne, F.E.S., F.G.S. 



WHEN writing a general account of the Upper Chalk of England 

 for vol. iii of the Cretaceous Rocks of Britain (Mem. Geol. 

 Survey, 1904), I stated the belief of Mr. W. Hill and myself that 

 the British Upper Chalk comprised more than one division of the 

 value of a stage. We grouped the zones then recognized into two 

 divisions, pointing out that the lower zones were essentially the 

 Chalk with Micrasters and the upper zones the Chalk with Belemnites ; 

 but we did not systematically adopt these divisions as stages with 

 definite names. 



It was remarked (op. cit., p. 5) that " Our knowledge of the 

 several faunas is not yet sufficiently complete to enable us to say 

 precisely where a line could best be drawn between the two divisions 

 or stages, and it is possible that some modifications of the zones will 

 have to be made before two such stages can be established. Moreover, 

 if such a division be found natural and desirable the existing 



