Tropical Forest Ecology: Sterile or Virgin for Theoreticians? 135rarity ar eth etwo ways a tr e esp eci es can
reduce consumption by, and mortality from, pests
and pathogens. Jurassic forests were dominated
by conifers and other wind-pollinated plants
(Corner 1964). Wind pollination works only for
plants close to conspecifics (Regal 1977, Davis
et al. 2004), so wind-pollinated trees must invest
heavily in anti-herbivore defense. Modern wind-
pollinated conifers have tough, long-lived leaves
whos etoxins poison th esoil wh en th el eav es
fall (Northupet al. 1995, Reichet al. 1995,
pp. 28–29). Rare trees can survive only by attract-
ing animals that convey their pollen to distant
conspecifics. Angiosperms began in the tropics
(Wing and Boucher 1998), where pest pressure is
most intense (Coley and Barone 1996). Lineages
with flowers attracting pollinators willing to seek
floral rewards from distant conspecifics diversified
extensively (Crepet 1984). The ability to enlist
animals as faithful pollinators (Crepet 1984) and
as dispersers of large seeds (Winget al. 1993,
Tiffney and Mazer 1995) enabled a diverse set of
rare, fast-growing flowering trees to replace a less
diverse array of common, better-defended, slower-
growing gymnosperms (Regal 1977). Opting for
animal pollination and seed dispersal triggered the
evolution of flowering rainforest whose diversity
reduced the depredations of pests with less loss of
productivity.
The theory just presented predicts that the evo-
lution of a diverse forest of rare trees would
enhance forest productivity. Extending this theory
suggests that where a tropical forest’s productivity
is higher, predators contribute more to its anti-
herbivore defense (Oksanenet al. 1981). To see
this, let the biomass per hectare at timetof plants,
consumers (herbivores), and predators beN(t),
C(t), andP(t), and set
dlnN
dt=r−aC−bN=b(
r
b−
aC
b−N
)
dlnC
dt=−m+λaN−a′PdlnP
dt=−m′+λ′a′C=λ′a′(
C−
m′
λ′a′)
Herer,a,b,m, andλhav eth em eanings of
ri,ai,b,m, andλin Equations 8.1,a′Pis thedecrease from predation of the per capita increase
of herbivores, andm′andλ′a′Car eth epr eda-
tors’ per capita death and birth rates. Predators
can invad eonly if th econsum er abundanc e
Cexceeds the densitym′/λ′a′needed to main-
tain predator numbers. Ifm/λa < r/b, then,
at the equilibrium with predators absent (P =
0),N = m/λaand C = (r−bm/λa)/a =
(b/a)(r/b−m/λa). Mor eproductiv eplant popula-
tions hav ehigh err.Ifris so high that(b/a)(r/b−
m/λa)>m′/λ′a′, predators can invade. Then the
equilibrium becomesC=
m′
λ′a′<
b
a(r
b−
m
λa)
,N=
(
r
b−
a′m
λ′a′b)
,
P=
[
λa(
r
b−
a′m
λ′a′b)
−m]/
a′This theory predicts that if predators are
removed, consumer abundanceCincreases more
when r, and therefore plant productivity, is
higher. Moreover, increased r increases plant
biomass N and predator abundance P,leav-
ing consumer abundanceCunchanged. Wootton
and Power (1993) tested the latter prediction
in river-bottom enclosures. Here, algae grew
on th erocky bottom, snails, mayfly nymphs
and th elik egraz ed th ealga e, and stickl e-
backs,Gasterosteusand dragonfly nymphs ate
the grazers. Different levels of shade were
imposed on these enclosures to create dif-
ferences in algal productivity. As predicted,
increased light increased algal and predator,
but not herbivore, biomass (Wootton and Power
1993).
In dry forest canopy in Panama, excluding
birds during th eproductiv es eason of l eaf flush
increases insect populations and leaf damage,
as this theory predicts. Excluding birds from
canopy branches in wetter, more evergreen for-
est, or from understory plants in either forest,
where leaves flush at a low rate all year long,
does not increase insect abundance or leaf dam-
age, perhaps because arthropods in the exclusion
zon e“tak eup th eslack” (van Ba el and Brawn
2005). Finally, in wet and dry forest, insects are
equally common in canopy accessible to birds,
even though insect-eating birds are more common