to concerns over the loss of species to human-
caused extinctions (Macket al. 2000; Lockwood
et al. 2007; van der Weijdenet al. 2007). The problem
is that most of what we know about invasions
comes either from observational studies that docu-
ment the invasions that have occurred, which gives
little insight into possible mechanisms promoting
invasions, or from small-scale experimental studies
that provide insights into mechanisms, but that also
suggest patterns that are at variance with large-
scale patterns. One example of this is the difference
between patterns of invasibility and diversity
observed at small and large spatial scales.
Experimental studies conducted at small spatial
scales, of the order of a few square metres or less,
tend to show that the ability of species to invade
established communities declines with the diversi-
ty of species already present in those systems
(McGrady-Steed et al. 1997; Knops et al. 1999;
Stachowiczet al. 1999; Levine 2000; Naeemet al.
2000; Symstad 2000; Kennedyet al. 2002; Fargione
et al. 2003). The pattern occurs in terrestrial and
aquatic environments, and seems to be quite gener-
al. It is also consistent with notions of biotic resis-
tance first articulated by Elton (1958).
Observational studies that consider patterns of
invasions at much larger spatial scales tend to
show a rather different pattern, although the mea-
sure of invasibility used is fundamentally different.
The measure of invasibility used is the total number
of invasive species that have become established,
rather than the performance of any single invading
species (such as ability to increase when rare, sur-
vival of invading propagules, or some other fitness
component of the invaders). These studies rather
consistently show that the most diverse commu-
nities also house the greatest number of successful
invading species (Robinson et al. 1995; Planty-
Tabacchiet al. 1996; Wiseret al. 1998; Lonsdale
1999; Stohlgrenet al. 1999; Levine 2000; Brown and
Peet 2003; Stohlgrenet al. 2003). Obviously, this
pattern seems to be inconsistent with the mechan-
isms of biotic resistance invoked to explain reduced
invasion success at small spatial scales. This dispar-
ity has resulted in a rather heated discussion about
the merits and proper interpretations of studies
conducted on different scales (Fridleyet al. 2007).
Obviously, the studies differ in more than just the
scale considered. There is the obvious difference
between experimental and observational ap-
proaches used at these different scales, and the
potential to infer incorrect causal pathways in ob-
servational studies. There is also a fundamental
difference in the operational definitions used to
measure invasibility. In one case (small scale), in-
vasibility is measured by the actual performance of
experimentally introduced invaders. At large
scales, we know only how many invasive species
have become established, and we typically do not
know how many have failed to become established.
To this end, it would be useful to use model
systems to attempt to resolve these differences. In-
deed, Macket al. (2000) have written, ‘We need to
develop innocuous experimental releases of organ-
isms that can be manipulated to explore the enor-
mous range of chance events to which all
immigrant populations may be subjected...’. We
have used such an approach to explore one of the
possible factors that could contribute to the appar-
ent positive relation between diversity and invasi-
bility seen at large spatial scales (Jiang and Morin
2004). This approach builds on the suggestion that
successful invasion depends on productivity and
the availability of resources for invaders (Shea and
Chesson 2002).
Jiang and Morin (2004) suggested that variation
in productivity among communities could swamp
out within community effects to create an apparent,
but causally spurious, correlation between diversi-
ty and the success of invaders over an array of
different experimental communities. The approach
first documented that increased productivity gen-
erated increased diversity of pre-invasion commu-
nities. The success of model invaders added to
these established communities also increased with
productivity. These two patterns resulted in a posi-
tive correlation between ‘native’ diversity and in-
vasibility (Fig. 14.2). However, when the effects of
productivity on ‘native’ diversity were statistically
controlled by partial correlation analysis, the corre-
lation between diversity and invasion success be-
came statistically non-significant. The inference is
that productivity, and not ‘native’ diversity, is the
cause of enhanced invisibility, and that more di-
verse communities are more subject to invasion,
not because they are more diverse but becauseEMERGING FRONTIERS OF COMMUNITY ECOLOGY 197