180 MR. B. SMITH ON THE [vol. lxxiv„ 



probability, had a smooth upper surface, but its lower boundary 

 may have been irregular. 



Each isolated mass of gypsum grew from the base upwards, 

 concurrently with the deposition of ordinary sediment 

 around it, and each thin deposit of crystals overlapped the one 

 below; thus the pillar was built up of successive layers of increased 

 diameter. A certain amount of thickening probably occurred about 

 the centre, so as to give to the upper surface a slightly domed 

 contour, but this is by no means certain. During crystallization 

 impurities may have been expelled to the margin of each layer. 



The final result was a pillar of gypsum filling a ' pot ' in the 

 horizontally-bedded marls. 



Formation of the Breccia. 



After the deposition of a certain amount of green and red marly 

 cap on the white gypsum, it appeai-s that the weight of the cap, 

 which is of higher specific gravity than gypsum, caused the under- 

 lying semi-consolidated white gypsum to break up into irregular 

 masses, some of which foundered and slid down beneath their 

 fellows into the lower layer of red gypsum. The upper masses- 

 were thrown into confusion and packed together, or half turned 

 over, erecting their broken corners as irregular pinnacles and bosses, 

 which probably remain as such to this day, for some of the pin- 

 nacles are extremely sharp. Solution, however, has taken effect at 

 their bases. 



The green colouring matter of the cap has crept downwards along 

 the cracks between the lumps in the upper layer and formed the 

 nearly vertical colour stains, while in places thin columns or bands 

 of pale buff-coloured dolomitic mud or marl descended quite 2 feet 

 into, and even through, the upper Avhite gypsum. Some of the red 

 marl, associated with or overlying the green cap, has also penetrated 

 downwards a short distance. 



The break-up of the white gypsum accounts for the irregularity 

 of the under-surface of the upper layer, and the turning on their 

 sides of broken fragments would close up the gaps left by the 

 foundered masses. 



In the next layer — the red and white breccia — the different 

 degrees of angularity, or of rounding, of the lumps of white 

 gypsum presented a difficulty until it became noticeable that the 

 roundest fragments are usually the lowest and have obviously 

 suffered most during their descent. The foundered white masses 

 are penetrated irregularly by veins of the green colouring-matter, 

 and call to mind crystals of olivine undergoing serpentinization. 

 The veins occur along incipient cracks, which in many eases have 

 become actual fractures during descent. 



It is interesting to note that the green coloration is more closely 

 related to the white fragments than to their red setting, and in one 

 pillar at least this connexion was so evident that the whole process 

 of brecciation was clearly revealed. In this instance a quantity of 



