Tropical Forest Community Ecology

Resource Niche and Trade-offs 169

applied to the latter dataset shows that the fre-
quency of rank reversals was significantly less
than the null expectation (313 out of 1225
maximum possible rank reversals,P < 0.001
by a Monte Carlo simulation of 1000 shuffles;
Kendall’sτ = 0.49,P < 0.0001; Pearson’s
r=0.52,P<0.0001). Significant concordance
of RGR between sun and shade is also demon-
strated in many other studies (Ellisonet al.1993,
Kitajima 1994, Osunkoyaet al.1994, Valladares
et al.2000, Bloor and Grubb 2003, Dallinget al.
2004, Baralotoet al.2005), with few studies
showin ga lack of a relationship (Popma and
Bongers 1988) or the opposite pattern (Agyeman
et al.1999). Thus, species switch their growth
rate ranks between the understory and gaps
less frequently than expected accordin gto null
models.
Parametric analysis demonstrates that most
rank reversals in growth rates are expected to
occur at extremely low light availability (Sack
and Grubb 2001). However, differences in growth
rates are so small in deep shade that care-
fully replicated experiments should be used to
detect rank reversals (Kitajima and Bolker 2003).
Because most rank reversals for growth rates
occur between species that are similar to each
other, growth rank reversals cannot explain the
habitat difference between pioneers and shade-
tolerant species (Kitajima 1994), and not even
between pioneers that prefer large versus small
gaps (Dallinget al.2004).Thus, performance rank
reversals in either growth rate or survival alone do
not provide a general mechanism underlying light
preference of seedlin gdistribution.
However, overall performance of species must
be evaluated in an integrative manner taking
into account both growth rates and survival,
because of the trade-off between these two per-
formance measures (Kitajima 1994, Kobe 1999).
Amon gthe six species overlappin gbetween the
two datasets shown in Figure 10.4, growth rates
are negatively correlated with survival proba-
bility in low light (r =−0.97,P = 0.007;
τ =−0.80,P = 0.05), as well as in high
light (r =−0.57,P = 0.23;τ =−0.60,
P=0.15). Such negative cross-species correla-
tion can be explained by allocation-based trade-
offs; allocation patterns that enhance growth

0.07

0.06

0.05

0.04

0.03

Seedling RGR (day

−^1

)

0.02

0.01

0

−0.01

Seedling survivorship

0 0.2 0.4 0.6 0.8 1.0

Figure 10.4 Growth–survival trade-offs for seedlings

grown under high and low light regimes (open and

closed symbols, respectively) amon gsix and five species

that overlap in each light regime between the two

datasets shown in Figure 10.3.

rates may come at the cost of reduced biomass

allocation to survival-enhancin gfunctions, such

as structural and chemical defenses and storage

(Kitajima 1996). More importantly, species’ posi-

tions alon gthe growth–survival trade-off line are

associated with their light niches in Figure 10.4;

alon geach re gression line, gap species occupy the

“fast-growth, low-survival” end, whereas shade-

tolerant species occupy the “slow-growth, high-

survival” end. A similar association between light

preference and the position alon gthe growth–

survival trade-off line has been found for naturally

recruited seedlings of 22 liana and 31 tree species

in Panama (Gilbertet al.2006), as well as for

saplings of 53 rainforest tree species in Bolivia

(Poorter and Bongers 2006).

Would species performance ranks reverse more

often between low and high light when growth

and survival, which are negatively correlated

with each other, are integrated into one per-

formance measure? In general, growth–survival

trade-offs should have an equalizin geffect, that

is, yieldin gmore similar reaction norms across

species for a performance measure integrating

both growth and survival, compared with the

reaction norms for either growth or survival