Fróðskaparrit - 01.01.1998, Side 17

EIN OYGGJALÍVLANDAFRØÐILIG GREINING AV FLORUNI í FØROYUM
23
island of Fair (5.2 km2) completely changes
the z values. Nevertheless, if we choose to
accept the values, the z value of the Shet-
lands should be smaller than that of the
Faroes, and that of the Orkneys still small-
er yet. And if we decide to omit Fair, the z
value of the Orkney would prove dramati-
cally different from what we expected.
Adsersen (1990) points out that the vas-
cular plants form a very non-homogeneous
group in terms of biological distribution.
He suggests that it might be a good idea to
pick out groups of plants which are more
homogeneous. We have selected 27 species
all characterised by effective dispersion of
seeds or spurs, namely the 11 species of
fem (Pteropsida), 3 club moss (Lycopsida),
6 species of horsetail (Equisetaceae) as
well as 7 orchids (Orchidaceae).
Species having a very effective disper-
sion should be better equipped to immi-
grate than other species. According to the
island theory, we would be inclined to an-
ticipate that the z values of those species
would be lower than the overall Faroese
material. A calculation of z as regards the
7 good dispersers, however, is z = 0.24
(correlation coefficient R = 0.89). The val-
ue is not too different from the z value of
the overall material of 0.20 but it is certain-
ly not smaller.
A null hypothesis
n order to find something more palpable
íhan the equation S = cAz, we have pre-
pared a model on the assumption that each
species is randomly distributed, indepen-
dently of the other species.
By using such a model as a ‘base line’, it
would be easier to fmd deviations and to
explain the deviations biologically.
Different plant species have different fre-
quencies. We have been informed hereof by
Hansen (1966) and have taken into account
the frequency of the species into their prob-
abilities of being or not being at a specific
island of a specific area. The number of
species can thus be described as a function
of the area, and the function can be adapted
to the actual number of species on the is-
lands. The function - the null hypothesis -
can be seen in Figure 2. For more details on
the calculations, see appendix and Chris-
tiansen and Hansen (1983).
The null hypothesis seems to be an ade-
quate approximation to the actual number
of species on the islands. If the method of
least squares is applied as a measure for fit-
ness, the hypothesis will produce a slightly
poorer result than S = cAz (12 000 as op-
posed to 9 000), but the null hypothesis has
only one constant that can be adapted to the
points (the area unit), whereas the equation
S = cAz has both c and z.
It appears from Figure 2 that the z slope
of the line varies. It can be demonstrated
that z approximates 1 when the area ap-
proximates 0. If the area approximates an
infinite number, z approximates 0. In a
large area, z maintains a relatively homoge-
neous value close to 0.20, which we also
established in S = cAz. The variation of z
along with the area in this model explains
why small z values are observed for very
large areas, e.g. continents.
Others have worked with corresponding
null hypotheses. Adsersen (1990) has e.g.
demonstrated that ferns (Pteridophyta) in