Constraints on Succession 377in removal plots, and the survival of other species
did not differ amon gtreatments. Notably, three
of the four species with higher germination had
survival that did not differ amon gtreatments, so
for these species, removal of pasture vegetation
had a net negative effect on establishment.
In post-agriculturalSaccharumgrasslands in
Panama, Hooperet al. (2002) compared the effect
of two cuttin gtreatments (cut once or twice),
shadin gplus cuttin g(75% and 95% li ght reduc-
tion), and fire on the establishment of 20 woody
species. Cutting significantly increased light at
ground level, while cuttin gplus shadin gsi gnif-
icantly increased soil moisture while reducing
light. In general, species with larger seeds tended
to have higher germination in the presence of
denseSaccharum, although these findings were
not analyzed statistically. These results and other
studies (e.g., Hollet al. 2000, Myster 2004) sug-
gest that resident grasses suppress smaller-seeded
woody colonists more than large-seeded species,
perhaps because grasses ameliorate detrimen-
tal light and heat levels for large-seeded species
(Figure 22.2). For the 16 species that germinated,
removin g graminoids did not affect germination;
however, shadin gled to substantially improved
germination for 10 species. Thus shade may
ameliorate harsh conditions (e.g., heat or desic-
cation) that inhibit germination. For 19 of 20
species, survival was highest in one of the shaded
treatments versus the control, fire, or cutting
treatments (Hooperetal. 2002). Overall, they con-
cludedthatcuttingplusshadingeliminatedbelow-
ground interference betweenSaccharumand tree
seedlings. They suggested that the major con-
straint on survival and growth of seedlings is
below-ground interference, that it is worse for
smaller-seeded tree species, and that high light
conditions could be inimical to germination.
Findings are similarly complex for survival and
growth of transplants. We summarize the results
of four studies (Gerhardt 1993, Hardwicket al.
1997, Zimmermanet al. 2000, Hooperet al.
2002) that examined the survival of tree seedlings
transplanted into experimental plots where res-
ident vegetation was left intact and where it
had been removed (Figure 22.3). Survival was
greater within intact vegetation for 16 species and
greater in the cleared treatments for 10 species.
We arranged species in rank order of seed size
from smallest to largest (Figure 22.3). The results
suggest greater negative effects of grassy vegeta-
tion for species with smaller seeds but there was
no statistically significant relationship between
seed size and response to vegetation removal.
Thus, it is premature to draw any firm conclusions
from research to date.
Still, when takin gall studies into account,
the general trend is that high abundance of
non-woody vegetation in old fields or abandoned
pastures retards succession. For example, sev-
eral types of non-woody vegetation can hin-
der woody plant colonization in post-agricultural
sites:Bacharis trinervisin Ecuador (Zahawi and
Augspurger 1999); the fernDicranopteris linearis
in Sri Lanka (Cohen et al. 1995); the fern
Nephrolepis multiflorain the Caribbean (Rivera
et al. 2000); a weedyBidensspecies in Guatemala
(Fergusonet al. 2003);Melampodium divaricatum,
Bidens pilosa, andPaspalum conjugatumin Mexico
(Purata 1986);Phytolacca rivinoidesin Costa Rica
(Harcombe 1977); and bamboos in southeastern
Peru (Griscom and Ashton 2003).Favorablemicrositesmayameliorate
inhibitionbypioneervegetationPasture that is dominated by graminoids is often
inimical to woody species recruitment. Conse-
quently, microsites within pasture where grasses
are less abundant may be foci for tree recruitment.
For example, Peterson and Haines (2000) found
much higher densities of woody species on rotting
logs versus unbroken graminoid cover in a pas-
ture in Costa Rica (see also Lack 1991). Similarly,
Slocum (2000) found that woody recruits were
five and eight times more abundant on rotting
logs and within fern patches, respectively, versus
open pasture dominated by grasses. In southern
Venezuela, in the first year followin gslash-and-
burn agriculture, Uhlet al. (1982) found sig-
nificantly more tree seedlings on logs and slash
versus areasawayfrom these microsites. Appar-
ently, these sites reduced competition with grasses
and moderated harsh environmental conditions,
but such sites may also be a refuge from seed and
seedlin gpredators (e. g., Lon get al. 1998).