234 Walter P. Carsonetal.
well-defended leaves, low photosynthetic capac-
ity, slower growth, higher survivorship, and low
responsiveness to increased light), which confer
an advantage both in the shade and under con-
ditions of prolonged exposure to pathogens and
herbivores in the understory (Coley 1983, Coley
etal. 1985, Sagers and Coley 1995, Kursar & Coley
1999, Wright 2002, Leighetal.2004). It remains
unclear thedegreeto which traits that confer
high survivorship under low light conditions are
positively correlated with traits that confer high
survivorship under prolonged pest pressure. In the
literature, shade tolerance has come to mean the
ability to survive in the understory for lon gperi-
ods of time at relatively small stature. This trait or
strategy could be due to varying combinations of
the ability to survive at low light and the ability
to survive (tolerate, defend, or avoid) prolonged
periods of browsin gor herbivore dama ge prior
to reaching a size refuge (e.g., from browsers) or
reachin gthe canopy. In the extreme case, a species
may be able to persist in the shaded understory
only in years or locations where pest pressure is
extremely low; thus juveniles might only rarely
be encountered. This might lead to the erroneous
conclusion that the species is shade intolerant.
Here we provide two examples from temper-
ate deciduous forests in North America. Eastern
hemlock (Tsuga canadensis) and American beech
(Fagus grandifolia) can survive prolonged periods
under deep shade. Beech is also highly browse tol-
erant while hemlock is not. Consequently, when
browsersareabundant,hemlockmayfailtoregen-
erate and is restricted to refugia, whereas beech
becomes extremely abundant (Horsleyetal. 2003,
Bantaet al. 2005). We suggest that the degree of
shade tolerance for any species will vary and likely
decline with an increase in herbivore damage
(Longet al. 2007).
For instance, in the presence of browsers in the
understory of a temperate forest, we found that
saplings of sugar maple (Acer saccharum), a puta-
tively highly shade-tolerant species, had patterns
of growth and mortality similar to saplings of
black cherry (Prunus serotina), a shade-intolerant
species (Longet al. 2007). In the absence of
browsers, sugar maple had patterns of growth
and mortality consistent with its classification as
highly shade tolerant. These findings emphasize
the need to critically evaluate the relationship
between tolerance to low light and tolerance to
herbivore damage and that these attributes will
not always be highly correlated among coexist-
in gspecies. Thus we propose that there may be
another important trade-off amon gsome coexist-
in gspecies in forest understories, namely alloca-
tion to physiological and morphological traits that
confer survivorship at low light versus traits that
confer survivorship under varyin gde grees of pest
pressure. There are hundreds of shade-tolerant
species in tropical forests and they will vary in the
degree to which they are tolerant to herbivores –
and as herbivore damage increases, the degree of
their shade tolerance relative to each other may
change substantially. Testing for the existence of
thistrade-off orrigorouslyevaluating therelation-
ship between low light survival and tolerance to
damage by enemies will require studies of growth
and survivorship alon ga continuum of li ght lev-
els and simultaneously a continuum of herbivore
or pathogen damage. If hierarchies of shade toler-
ance amon gspecies shift as pest dama ge increases,
then models of forest dynamics will make different
predictions of future canopy composition depend-
in gupon pest pressure (Royo and Carson 2006,
Longet al. 2007).The impact of enemies on aggregated
adults and of outbreaks has been
neglectedThe main focus of tests of the Janzen–Connell
model is how enemies create repelled patterns
of juvenile recruitment around adults due to
density- and distance-dependent predation. Stud-
ies have paid far less attention to the effect of
natural enemies on adult plants; however, adults
in dense aggregations could, in addition to their
juveniles, be vulnerable to higher per capita rates
of pest attack and damage. In a classic paper,
Root (1973) formalized this concept and proposed
the resource concentration hypothesis: “herbi-
vores are more likely to find and remain on
hosts that are growing in dense or nearly pure
stands.” Much evidence supports this hypothesis
for plant populations (e.g., Andow 1991), yet its