140 Proceedings of the Koyal Society of Edinburgh. [Sess. 
less than lg — can be followed in their course through several cells, so that 
they are not in any respect, except that of size, different from the thicker 
branches. It is noteworthy that the petrified remains of threads, even less 
in thickness than that of an average bacillus, should have been preserved 
in an intact condition. In some of the host cells fragments of threads, 
measuring from lO/x to 40/x in length and about 1/x in thickness, were 
present. If these had been seen apart from the threads of the fungus it 
would have been impossible to have given a satisfactory explanation of 
their nature. In this case, however, it was possible to prove their organic 
connection with the fungus under discussion. A host cell showing these 
fragments is shown in Plate, fig. 7. A detailed examination of such frag- 
ments brought to light a feature which is of great interest because it has a 
bearing upon one of the most important problems in mycology. In some 
cases it was found that a very thin branch of the fossil fungus, after travers- 
ing perhaps four of the host cells in an unbroken condition, had broken up 
in the fifth cell, in which fragments only were seen (see Plate, fig. 7). The 
number of fragments in the fifth cell were so numerous, however, that it 
was evident that the breaking up of the thread into fragments had not 
been followed by a cessation of growth, in which case elongation, follozved by 
division and further growth, must have been carried out by the fragments. 
