Geuli)ijical Suciettj. 239 



is a great impedinieat to further progress. Instead of all these 

 immense thicknesses of beds belonging to the Miocene, they have 

 their base somewhere in the so-called Cretaceous series ; 400 feet 

 higher up we are about the horizon of the Thanet Beds ,' while at 

 1000 feet up the flows were contemporaneous with the Bracklesham 

 and Barton deix)sits. The lirst acid eruptions were Miocene, as 

 shown by the floras preserved in Iceland. 



The Author described the various exposures from his own 

 observations and Mr. Cole's notes. At Ardtun the Traps are sur- 

 mised to be parts of once continuous flows, still represented at Stafla 

 and Burgh, but broken through by an intrusive sheet. The leaf- 

 beds are varied in composition, the richest being ver}^ friable, while 

 the best matrix is a limestone as fine as lithographic stone, in which 

 plant-remains are few, but exquisitely preserved. They are over- 

 Liin by thick deposits of river-gravel, chiefly composed of flint or 

 silicified chalk, but in which Mr. Cole has detected fragments of 

 sanidine like that of Ischia or the Rhine, and of trachyte. At Carsaig 

 the gravels are coarser and less evenly bedded, and the sandy matrix 

 apparently is entirely made up of flint. The coarse gravels are 

 flanked by sands and indicate a rapid flow of water, occupying a 

 valley not less than a mile across. The Ardtun gravels indicate a 

 less rapid but more extensive river. The section at Burgh is very 

 like that at Ardtun, with the addition of an extensive ash-bed at 

 the base, with sand instead of gravel, and with many hundred feet 

 of Trap above. In the Wilderness there is a small outcrop of Chalk- 

 rubble, less than 300 square feet in extent, and evidently redeposited. 

 {Some distance under this is Greensand in situ, then Lower Lias, and 

 lastly Poikilitic sands. This descriptive part of the paper concluded 

 with some remarks on the estuarine formation between the Chalk 

 and Upper Greensand at Beinn Jadain in Morven, which the Author 

 investigated in the hope of finding plant-remains belonging to that 

 interesting age. He doubts that the Chalk is in situ, and considers 

 the evidence of age to be not quite conclusive. 



The second part of the paper dealt with the Palseontological 

 evidence. The evidence, if confined to the plants of Ardtun, was 

 said to be scarcely worth serious discussion, and the analysis was 

 extended to the far richer plant-bed at Atanekerdluk. The identi- 

 fications of these with Miocene plants of Europe were discussed 

 seriatim, and shown to be groundless, or only applied to such pre- 

 vailing types of leaves as are common to widely distinct genera, and 

 occur in floras recent as well as fossil, and which cannot be super- 

 ficially distinguished in even a living state. The strong resemblance 

 and even identity of the best-characterized forms with the older 

 Eocene plants has been, on the other hand, hitherto ignored. The 

 most strongly marked types of Greenland, and which recur in 

 Antrim, are met with in the Heersian of Gelinden and at no other 

 horizon, and amply sufiice to fix the date of the Antrim floras. The 

 Mull flora, as its aspect indicates, is still older, and consequently 

 earlier than the Thanet Beds of England. lndei)endently of positive 

 evidence, the absence of any late Tertiary tylle:^, even of the Legu- 



