The Botany of Iceland - 01.12.1945, Síða 152
494
STEINDÓR STEINDÓRSSON
with this a difference in the formation of ice in the soil. I regard this
assumption as highly probable as far as the rústs are concerned, but
I believe, also, that differences in the vegetation of the mýr may exert
a similar influence to that of different snow-covering. In Illugaver I
made an observation which in my opinion points in the same direction.
When I visited this place on July 25th, 1937, I noticed that in the
wettest parts of the svamp Eriophorum polystachyum occurred ex-
clusively on the higher-lying places, while a Carex rariflora association
was found in the intermediate depressions. As the development of leaves
is always very abundant in Eriophorum, a great deal of withered
material from the previous year was found in the Eriophorum growths,
and this material formed a thick mat which covered the ground;
no such mat was found in the C. rariflora growths. On closer inspec-
tion it tumed out that while the soil in the C. rariflora growths had
thawed entirely, only a layer some 20-30 cm thick had thawed in
the Eriophorum growths. There is hardly anv doubt that the thick
mat of withered material considerably delays the thawing of the soil;
but this renders it obvious, as was also assumed by F r i e s, that it is
the different mode of thawing of the two places, situated near each
other, which exerts the greatest influence on the Palsen formation. If
thus, as assumed by F r i e s, the elevation constantly increases, it is
quite obvious that its vegetation will change. The hydrophilous vegeta-
tion disappears, giving room for another, more xerophilous vegetation.
When once the rúst has attained its maximum size (how great this is
will be determined by the external factors in each locality), its destruc-
tion will commence. The xerophilous vegetation is unable to form a
closed carpet on top of the high, snow-free tops of the rústs. The frost
will form cracks and fissures in the rúst which will make the soil even
more unfavourable for the vegetation, and in connection with the bare
spaces they will give the wind free play to sweep away the dry soil.
At the same time the evaporation from the rúst increases, the water
is, so to speak, sucked up through the porous soil, possibly due to
capillary action, and the desiccation will therefore not only affect the
rúst itself but also the surrounding depressions, so that the flá as a
whole will begin to dry up. The farther this development proceeds, the
more easily will the whole flá fall a prey to the wind erosion, so that
at last it will perhaps be swept away entirely; in several places there
is evidence of this. Possibly, however, so much moisture will remain
that the erosion does not developed beyond the flag stage, and the plants
growing there will be able to prevent further destruction, perhaps even