0198566123.pdf

FUTURE DIRECTIONS 163

Island characteristics relating to the positions indicated in the figure

Island feature (a) Large, near (b) Large, far (c) Small, near (d) Small, far

Immigration High Low High Low

Speciation Low High None Low

Extinction Low Low High High

Species richness High High Moderate to low Low

Turnover Low Low High Low

Island type Continental Continental fragments or Small continental Depauperate oceanic

islands large oceanic islands islands islands

Example Newfoundland, Madagascar, Hawaii Skokholm (Wales) Easter Island,

Tasmania Tristan da Cunha

hitherto neglected: evolution. Moreover, although

Lomolino’s model recognizes the patterns of

diversity and turnover, it doesn’t assume

equilibrium.

So, is the way ahead to take a species-based

approach? Lomolino’s argument that many

biogeographic patterns derive from, not despite,

differences among species echoes many of the

points discussed throughout our examination of

assembly rules, incidence functions, and the

importance of species-level effects in the

dynamics of island biotas. A species-based

approach is in some respects more flexible, can be

more relevant to the needs of conservationists in

particular circumstances (e.g. modelling the

incidence or potential incidence of an endangered

species), and opens up a new set of hypotheses

for testing. However, we should also recall that

species responses and impacts can vary between

systems. For example, Watson et al. (2005)

demonstrated variation in species incidence

functions of woodland birds in the Canberra area

(Australia), which appeared attributable to

properties of the matrix in which the woodland

systems were embedded (see Box 5.2). Similarly,

many species considered benign and non-invasive

within their native ranges, upon being introduced

into an exotic island, turn into radical

transformers of ecosystems (Chapter 10). This

suggests that their pattern of incidence in one

area may not accurately predict their incidence

elsewhere. Hence, there are likely to be limits to

the predictive power of species-based models, just

as there are with system-level models.

representation of turnover and compositional
patterns than do simpler stochastic models based
only on parameters such as area and isolation.
These arguments amount to a case for more com-
plex models of island ecology. Building in greater
complexity is fine under two conditions: (1) you are
sure of your starting assumptions, i.e. the paradigm
you are working under, or (2) you are attempting to
develop better understanding of detailed case study
systems or areas. However, we should not abandon
the arguably more important scientific task of
searching out the simple explanation where it will
suffice, and of finding new models of general appli-
cability. Box 6.2 reviews one possible direction.

Consideration of scale has been another key

theme of the last two chapters, as we have devel-

oped the argument that although some processes

may be universal, their impact is detectable at par-

ticular scales of analysis, while at other scales

other processes or confounding variables dominate

the signal. The increasing availability and sharing

of data should allow the scale sensitivity of island

ecological models to be more thoroughly investi-

gated, allowing us to determine the applicability of

particular theories within the space/time/scale

continuum.

Throughout the island ecology section of this

book we have to a remarkable degree disregarded