The Pliocene Flora of Transcaucasia. 295 



runs parallel to the left ridge of tliis valley. In the middle of its 

 course the valley broadens somewhat, but further on it becomes 

 narrower and cuts through the western slope of the Godersky summit, 

 by a rather narrow canyon. The upper part of the valley is formed 

 by two ravines where the River Dzundza has its source. The 

 Godersky summit has an altitude of 6,500 feet ; on the eastern side 

 the slope is very gentle, but its western slopes descend steeply into 

 the valley of the Adjariszhali. The valley of the River Dzundza 

 throughoiit its whole length cuts through a region of volcanic 

 strata, the lowest of which are andesite lavas. Higher up above the 

 village of Uthisuban the valley becomes much narrower and the 

 bed of the river cuts into volcanic deposits, for the most part Andesite 

 tuffs mixed with lava of a light grey colour. The artificial cornice 

 of the road afforded an opportunity of ascertaining not only the 

 stratigraphical features of these volcanic beds, but it also led to the 

 discovery of rich collections of fossil plants. On examining the 

 tuffs lying above the line of the road one realizes that the strata are 

 not uniform in their formation. The tuffs are almost horizontal, 

 dipping slightly to the west, with an inclination of some of the 

 strata towards the south ; they are intersected by many cuttings. 

 The rock is far from being homogeneous ; it consists of rather 

 irregular strata made up of light grey and light violet-grey tuffs 

 cemented together with light yellow clay. 



The engineer, A. P. Guerassimoff, v/ho kindly investigated the 

 rocks, found they were fine-grained, rather soft, and comparatively 

 loosely cemented, containing small black carbonized remains of 

 plants with volcanic glass of irregular form and of light yellow colour. 

 The clay cementing the tuffs is sometimes covered with oxide of iron. 

 Small fragments of minerals are common ; the mineral most 

 frequently met with is a colourless, quite fresh labradorite ; felspar 

 is also not uncommon ; prismatic crystals and small fragments of 

 colourless or light green augite are more rarely met with. Pyroxene 

 not infrequently contains irrcrustations of magnetite. Another 

 characteristic mineral is the brown hornblende. Lastly, one some- 

 times finds occasional sections of pleochroic hypersthene. There 

 are many small crystals of magnetite and, in some places, little liea])s 

 of discoloured volcanic glass and fragments of andesite lava. The 

 distribution of coloured remains in the rock is irregular ; in one 

 sample hornblende is predominant, in another pyroxene. These 

 tuffs are cemented with clay where they have fallen, and there are 

 no traces of any disintegration of the rock. It is to be presumed, 

 therefore, that their fall was the cause of the burning of the forest. 



Trunks are abundant, also fragments of trees and branches of 

 different kinds. In some places, for instance, one may see immense 

 stumps, one of them standing in situ with wide spreading roots and 

 with a diameter of 1 meter 30 centimetres. Higher up in the tuffs 

 one can distinguish the remains of the upper part of this trunk. 

 The tree is a dicotyledon. Wood is common throughout the tuffs, 



