0198566123.pdf

It was found that none of the 10 species exhibited
a maximum isolation effect (Fig. 5.3d), which was
not surprising given the relative lack of variation in
isolation in the data set. Only one species, Microtus
pennsylvanicus, a small but dispersive generalist
herbivore, showed no area or isolation effect. Six
species exhibited minimum area effects (Fig. 5.3c)
and the remaining three species exhibited compen-
satory effects (Fig. 5.3f). Blarina brevicaudais an
example of the latter group, which is consistent with
independent evidence that it is a relatively poor dis-
perser. Lomolino (1986) applied his discriminant
analyses to additional data sets for the islands of
Lake Michigan, Great Basin mountain tops, and

islands in the Bass Straits. He found general support

for the model he had developed, the relevance of

scale of isolation to the incidence patterns, and in

some cases, for patterns predicted by the compensa-

tory model (see also Peltonen and Hanski 1991).

Microtus pennsylvanicusandBlarina brevicaudain

the Thousand Islands region provide an illustra-

tion of another important effect structuring island

communities, namely predation. The carnivorous

shrew Blarina brevicaudapreys mainly on imma-

ture stages of the vole Microtus pennsylvanicus. In

the absence of the shrew on more remote islands,

Microtuswas found to undergo ecological release,

occurring in habitats atypical of the species

ISLAND ASSEMBLY THEORY 121

4

3

Area

2

1

0

4

3

Area

2

1

0

4

3

Area

2

1

0

0123

Isolation Isolation

45 012345

(a) Hypothetical archipelago (b) Random

(c) Minimum area (d) Maximum isolation

(e) ‘Block pattern’ (f) Compensatory

Mainland

Figure 5.3Patterns of insular

occurrence of a species on islands of a

hypothetical archipelago (a). Presence

and absence in (b)–(f) are indicated by

filled and open circles, respectively. The

units for isolation and area are arbitrary.

(Redrawn from Lomolino 1986, Fig. 1.)