Constraints on Succession 379models of seed dispersal and propagule supply
(e.g., Nuttle and Haefner 2005) could provide a
robust conceptual foundation for understanding
early succession in the tropics.
The frequently dense graminoid vegetation in
post-agricultural sites is clearly able to influ-
ence establishment of woody species. Indeed,
many reports of arrested succession have been
attributed to inhibition caused by competition
with resident vegetation (Cohenet al. 1995,
Zahawi and Augspurger 1999, Riveraetal. 2000).
Still, as reported above, the presence of intact veg-
etation in some cases facilitates woody species
establishment (Zimmermanet al. 2000, Hooper
et al. 2002). It is likely that the effect of resident
vegetation is neither uniformly negative nor pos-
itive, but that the effect varies predictably with
the abundance of resident vegetation.Thus exper-
iments are needed that extend removal studies
to a focus on how variation in the abundance
and identity of resident herbaceous vegetation
influences woody species recruitment amon gcon-
trastin glife-history types. The de gree to which
the presence of resident vegetation facilitates
or inhibits subsequent colonization of woody
species needs to be integrated into the nucleation
model.LIMITATIONS OF EXISTING
MODELS
Two of the major functions of models and
hypotheses about succession are to organize and
make sense of divergent findings, and to predict
dynamics in new settings. We suggest that existing
succession models are to varyin gde grees inad-
equate though they remain for the most part
untested and they deserve greater consideration
even if they are rejected. Existin gmodels of sec-
ondary succession have overwhelmingly concen-
trated on life-history traits (e.g., Egler 1954, Drury
and Nisbet 1973) and competitive interactions
(Tilman 1985, Huston and Smith 1987). These
models need to be expanded or refined in order to
apply to tropical habitats. Two simple examples
are illustrative. First, the least surprisin gspa-
tial pattern documented in empirical studies is a
muchgreaterdensityandrateof colonizationnear
forest edges (e.g., Myster 2003a). Despite the ubiq-
uityof thispattern,itcannotbeexplainedviarelay
floristics, initial floristic composition, resource
ratios, shade tolerance, or interaction categories.
Second, several well-known examples illustrate
the potential for multiple constraints on tropical
succession (e.g., Aide and Cavelier 1994, Nepstad
et al. 1996, Chapman and Chapman 1999, Holl
et al. 2000), resultin gin very slow or non-existent
structural and compositional change. The expla-
nations from existin gmodels are unsatisfactory.
The gradient-in-time models would imply that
no species is adapted to do well in these sites,
a conclusion challenged by the use of trans-
plants. The resource ratio and shade tolerance
models would imply that sites had reached a
final stage with the best competitor occupying
the site, a conclusion seed predation and disper-
sal studies show to be incorrect. The inhibition
mechanism under Connell and Slatyer’s inter-
action categories would imply that the current
vegetation is simply preventing later-stage species
from establishing, when the cause is at least
partly a lack of propagules of later successional
species.AN EMPIRICAL RESEARCH
PROGRAM
We recommend research that will contribute to
layin ga foundation for improved models and
synthesis. We argue that it is important to
undertake research that more fully quantifies the
followin gfive processes.
1 Propagule input as a function of distance from
source areas.Succession requires that propagules
arrive into abandoned pastures and a lack of
propagules no doubt delays tropical succession.
Models that attempt to predict colonization based
upon interspecific differences in fecundity, seed
size, modes of dispersal, and source populations
will no doubt pay large dividends (e.g., Nuttle and
Haefner 2005).
2 The degree to which resident herbaceous vegeta-
tion inhibits or facilitates woody species recruitment.
The essential element is to evaluate the response
of potential woody colonists to natural variation
in the abundance and identity of the herbaceous