188 OLAF GALLOE 
where along the sides of the fjords. Owing to their great water- 
contents they are always devoid of lichens. 
Grimmia-heaths are found, as mentioned, at all altitudes right 
up to the snow-line, but differ somewhat according to the altitude. 
The substratum which supports the Grimmia-carpet is some- 
times solid rock, sometimes loose soil. My own observations are 
derived almost exclusively from carpets upon lava. 
On mountain heights (fell-fields) the carpets are often small 
in extent, but further down in the low land they may cover large 
continuous tracts. 
The plant-carpet is a few centimetres high and the moss-shoots 
stand erect. Whatever may be the nature of the deeper-lying sub- 
stratum, at the bottom of the carpet there is always found, as a 
matter of course, a peat-formation consisting of the dead remains 
of the mosses, as soon as the moss-covering is only a few years 
old. The deeper-lying soil is evidently of no direct importance, or 
concern to the lichens; they are connected with the peat and the 
dust-particles, which always occur on it, and amongst the mosses. 
With regard to the amount of water contained amongst the 
mosses, very little is mentioned in the literature. It may, however, 
be taken for granted that all Grimmia-vegetation in Iceland is dry 
enough to bear lichens. My own observations show this distinctly 
enough. For the rest, there is, as usual, the great defect, that we 
have no fixed method to indicate the degree of dampness of the 
plant-association, as far as emergent associations are concerned. We 
are constantly reduced to the entirely relative, and consequently 
almost useless terms, “dry,” “damp,” etc., without any fixed state- 
ment as regards measured amount. 
Jénsson states that there is an essential difference in the ac- 
companying phanerogams in high and in low land, in that only a 
few occur on the rocky flat, whilst they are found far more numer- 
ously further down. He states, in addition, as a general fact, that 
lichens are found more abundantly in the Grimmia-carpets of the 
rocky flat, than in those of the low land. How far this is quite 
right can only be proved by frequency-numbers, and statements of 
mass-occurrences (in weight), and such are not found in sufficient 
numbers. I must, however, say that Jönsson’s statement sounds 
very reasonable, and is supported by Hesselbo. It can undoubtedly 
be explained by the fact that the climatic conditions are more un- 
favourable to the mosses in the high land, and relatively more 
