Fróðskaparrit - 01.01.1998, Qupperneq 21

EIN OYGGJALÍVLANDAFRØÐILIG GREINING AV FLORUNI f FØROYUM
27
Koltur are among the smallest islands, and
the model is probably less reliable for small
areas. Nólsoy also deviates significantly
(Table 2) and we are unable to account for
the underlying cause.
The model is directly misleading in the
case of Koltur, which has as many as 108
species on its 2.5 km2 which according to
the null hypothesis ‘should’ have held an
estimated 51 species. In our opinion, the
underlying causes are multiple.
Koltur is a fertile island which would al-
low for a higher number of species. Bengt-
son and Bloch (1983) mention that Koltur
is priced at 17 merkur from ancient times,
whereas Hestur despite its more than dou-
ble size is priced at only 18 mercur. Mork
(pl.: merkur) is an old Faroese measure for
soil fertility.
If you study the island’s profile (Figure
3) it seems obvious that here must be many
species. There is plenty of lowland and this
is where the many species are found. But
the island also has an intermediate zone,
because Koltur reaches 479 metres despite
its very limited area. Finally, the fine acces-
sibility of the island has the effect that it
was possible to find many species within
the allocated time when the investigation
was being conducted.
One way to compare island flora
The starting point in every comparison be-
tween the flora or fauna of two regions
must always be a list of the organisms of
the regions. Subsequently, a measure is pre-
sented - based on different criteria - as to
how many different species occur in both
regions. The subject is thoroughly de-
scribed in e.g. Connor and Simberloff
(1978). By the development and selection
of a suitable similarity index, it has been a
recurring problem to assess the indexes,
and as often a biological null hypothesis
has been missing.
In the present section, we shall briefly in-
troduce a way in which to compare the flo-
rae of two regions - a similarity index, if
you will. The method is based on the previ-
ously described null hypothesis.
The nature of the species as well as the
total number of species are well-known for
each of the islands. The number of common
species for the two islands is also well-
known, let us refer to that by the letter u.
You can also calculate the number of com-
mon species on the two islands based on the
predictions of the null hypothesis in terms
of which common species ‘ought’ to be on
these two islands, considering their given
areas. We shall call this figure v.
If one island contains x species, the area
Aj, which the island ‘ought’ to have ac-
cording to the null hypothesis, is deter-
mined. This also applies for the other island
having y species and ‘area’ A2. The null hy-
pothesis now gives the number of species
anticipated, k, in a region of an area A,+A2
and v = x + y - k.
u-v is thus an expression of whether the
florae of the two islands are more or less
similar than could be anticipated.
But it is not possible to apply the expres-
sion directly for comparisons, as the vari-
ance differs from one pair of islands to an-
other. In order to compensate for this, we
have constructed a variance, see appendix.
Table 3 shows a comparison of all 153