British Pillow-lavas. 245 



found in the Middle and Upper Devonian ; the attendant sediments 

 are shales and grits, hut Ussher (43) found that between the pillows 

 of the lava on the shore at Saltash a few thin beds of chert occur, and 

 in microscopic section they appear to be of radiolarian origin. No 

 cherts are found in the Devonian rocks at lower levels, but as soon as 

 the spilites are met with the cherts also make their appearance, and 

 at first are only found on the upper surfaces of the flows. This fact 

 awakens a strong suspicion that there must be some genetic connexion 

 between these two kinds of rock. At higher horizons in Upper 

 Devonian and Culm the pillow-lavas abound, and there are great 

 masses of chert, but we cannot be sure that this is more than 

 a coincidence, since in South Wales at the same time radiolarian 

 cherts were being laid down in great sheets, and there are no pillow- 

 lavas. Again, in the Lower Palaeozoic rocks of Cornwall, as Fox (44) 

 has shown, there are sometimes a few lenticles of chert among the 

 black shales, but when the pillow-lavas flowed out on the sea-bottom 

 in Arenig or Llandeilo time the cherts appeared at once in great 

 quantity, and on Mullion Island they fill up the gaps between the 

 sack-shaped masses of the igneous rock. 



Recent researches on the plankton of the northern seas have proved 

 that the modern organisms which form siliceous tests are dependent 

 on the supply of silica in the water which they inhabit (45). At 

 certain seasons a large increase in the amount of dissolved silica takes 

 place, and this is followed almost immediately by rapid proliferation 

 of the diatoms, etc., that inhabit the water. There seems no inherent 

 reason why the ' law of the minimum ', as it has been called, may not 

 also have held good in Palaeozoic times ; if so, it explains the occurrence 

 of organic cherts with pillow-lavas. The igneous rocks as they cooled 

 down exhaled vapours or solutions of magmatic origin, rich in dissolved 

 silicates of soda and other bases. These were the agencies which 

 albitized and decomposed the lavas, and any excess must have escaped 

 into the sea-water. In this way precisely those conditions were 

 provided which are most favourable to the rapid multiplication of 

 siliceous protozoa such as the radiolaria. As the spilitic rocks ai - e 

 generally found at some distance from the shore, in quiet waters 

 where there was little sediment and current action was at a minimum, 

 the radiolaria, though they may have been of slow growth, were not 

 carried away. The dead shells fell to the bottom and rested on the 

 surface of the lava, while there was comparatively little clastic or 

 land-derived sediment to mingle with the siliceous organic deposit. 

 Where there are no cherts with the spilites, as for example at 

 Tayvallich, Argyllshire, we may infer either that the dissolved mineral 

 substances were not in excess of those required to albitize the lava, or 

 more probably, that the local action of currents bore away the minute 

 siliceous shells and brought other sediment in their place. 



Conclusions. 



1. The pillow-lavas are members of a natural family of igneous 

 rocks, the spilitic suite, that can be clearly distinguished from the 

 Atlantic and Pacific suites. 



2. This family comprises a great variety of types — picrites, diabase, 



