110 GOTE TURESSON 
at a point just south of that village, where the woods of the Stenshuvud 
region begin. The point where the two forms meet has been found, 
and the cultivated material from this place shows all kinds of inter- 
mediate types. The geographical point to the south where the variety : 
typical of the woodland region ceases has not yet been located, but the 
prostrate variety inhabiting the sandfields north of Kivik reappears in 
its most typical form on the sandfield north of Simrishamn. . It then 
dominates the coast at least down to Skillinge, a small fishing 
village, but disappears again a little farther to the south, where the sand- 
dune region of Sandhammar begins. The typical sand-dune variety 
now flourishes unrestricted. 
The successions of varieties 
here described take place 
within an area found to be 
about 35 kilometres in ex- 
/ tent. The point northwest 
of Kivik, where the prost- 
rate variety of the arena- 
cious fields meets the in- 
land variety of the woods, 
has been located. The 
prostrate variety follows 
the sandfields and is found 
quite typically about 15 ki- 
lometres from the sea. The 
sandfields are now replaced 
Fig. 6. Hieracium umbellatum, from arenacious LYCOS. Woo 
fields (about !/s nat. size). tuated about 20 kilometres 
from the sea have been 
found to harbour intermediate types. The intermediate zone is probably 
quite narrow, as cultivated material collected at random at a distance 
of 2 kilometres to the west of the latter locality contains the typical 
woodland variety only. 
The evidence furnished by the distribution of these varieties of 
Hieracium umbellatum within a limited and continuous area shows 
that the different varieties are restricted to distinct localities within 
the area, and that the reappearance of a distinct locality occasions 
the reappearance of the variety typical of that locality. These facts, 
as well as those discussed above with regard to coast plant varieties, 
are inconsistent with the view often held and defended that the kind 
