THE GENOTYPICAL RESPONSE OF THE PLANT 309 
The Torekov set (table 14, fig. 69) is very interesting in this con- 
nection because of its close proximity to the broad-leaved cliff-type of 
Hofs Hallar and Hallands Väderö, which has already been discussed. 
The collection raised from seeds includes individuals with varying 
shape and width of leaf. Some show resemblance to the cliff-type leaf, 
others to the leaf of the inland-type (broadest in the middle). When 
it is found (cf. tables 13, 15; fig. 62) that the Torekov set includes in- 
dividuals with a leaf narrowness only exceptionally found in the cliff 
series, we may well assume that the Torekov type has been differen- 
tiated locally from the H. umbellatum population of that region in 
response to dune conditions. The individuals with leaves resembling 
the inland type have the leaf-width typical of the most narrow-leaved 
individuals of that type (cf. figs. 59—61). As to leaf-width only the 
most narrow-leaved individuals seem thus to have been favoured, 
while the particular shape of the leaf, broadest in the lower portion 
or in the middle, does not seem to be of any moment. 
There are also dune types consisting almost entirely of individuals 
with leaves of a shape’ typical of the woodland type of the interior, 
e. g. with leaves broadest in the middle. These are the series from 
‘ Nyhamn and Skelderviken (figs. 70—71). Inspection in the field of 
the H. umbellatum population in these localities has also brought out 
this peculiarity. 
The narrow-leaved type of the shifting dunes is thus seen to differ 
in different places in regard to the detailed shape of the leaves. It 
may have exclusively linear leaves (Halmstad, Sandhammar), or the 
leaves broadest towards the base suggestive of the particular shape of 
the cliff type (Falkenberg), or may be made up almost exclusively of in- 
dividuals with leaves broadest in the middle and may thus resemble 
in this character the inland type (Skelderviken, Nyhamn). It may 
finally be composed of individuals some of which resemble the cliff 
type, others the inland type, in regard to this characteristic (Torekov). 
I do not think that these findings are consistent with the generally 
accepted migration theory. The differences observed, which must be 
considered quite non-essential to the existence of the plant in the va- 
rious dune localities, go to show that the dune type has on the contrary 
become differentiated separately at different points, as a response 
on the part of the species-population to dune conditions. 
The broad-leaved sea-cliff type of Stenshuvud illustrates, I think, 
much the same thing. The important characteristics of this type, 
which distinguish it from the cliff type of the west coast, viz. the 
Hereditas III. 23 
