56 John R. Paul and Stephen J. Tonsor
a rare species advantage in many tropical forests,
probably resulting from lower density-dependent
or frequency-dependent mortality (e.g., Harms
et al. 2000, Volkovet al. 2005, Willset al. 2006).
Rare species that have a fitness advantage are
expected to increase in abundance much more
rapidly than predicted under neutral drift, for
example, resulting in younger species that have
large range and population sizes. Thus, if new
species do indeed start with small population
and range sizes, some of these species may be
expected to increase their population and range
sizes rapidly. Overall, the generality of a positive
age and area relationship in tropical plant species
awaits future analyses, particularly of densely
sampled, speciose clades.
Fortunately, there is considerable promise that
in the near future we can gain a broader per-
spective on age and area relationships in tropical
plants. For example, work on the diverse tropi-
cal herbaceous genusBegonia(Begoniaceae) has
provided insight into the phylogenetics and tim-
ing of diversification in this pantropical genus
(e.g., Forrest and Hollingsworth 2003, Plana
et al. 2004). Likewise, phylogenetic work on the
diverse pantropical genusPsychotria(Rubiaceae;
Nepokroeffet al. 1999, J. Paul unpublished data)
promises to provide evidence from a genus that
in many ways mirrorsPiperin its species’ ecol-
ogy, abundance, and distribution (e.g., high local
and regional species richness, numerical abun-
dance, understory and ga phabitat, etc.), although
it is phylogenetically distantly related. Interest-
ingly, Hamilton (1989) suggested that within the
Mesoamerican members ofPsychotriasubgenus
Psychotria, species groups often contained one
basal member with a large geographic range,
and putatively derived members with narrow
ranges.
AN AGE-AND-AREA HYPOTHESIS
FOR MODERN TIMES
The strong positive age and area relationship
found for neotropical Piper species warrants
further investigation into the generality of this
relationship in tropical plants. If, in general, many
rare species are found to be young species, this
information may be crucial to incorporate into
our understanding of the variation in range size
among species and, at the local scale, variation
in abundance, which often shows a positive rela-
tionshi pwith range size (Gaston 1994). In order
to effectively integrate species age information
derived from molecular inference (as most future
data promise to be) into our understanding of
tropical forest community structure, we need to
recognize the potential sources of error in these
data, as well as take a broader view on the sim-
ple age-and-area hypothesis proposed by Willis
(1922).
First, one of the obvious shortcomings of the
traditional age-and-area hypothesis (Willis 1922)
is its failure to account for old species with small
ranges. Empirical evidence suggests that in some
cases, the age and area relationshi pmay be
a hump-shaped relationship (Webb and Gaston
2000), where both old and young species have
small ranges, and intermediate age species have
the largest ranges (or the greatest degree of eco-
logical occupancy, e.g. Footeet al. 2007). Clearly,
many old species must either go through range
contraction as they age, or have their range sizes
reduced through the process of speciation. As a
result, a complete age-and-area hypothesis needs
to account for these species, recognizing that a
positive age and area relationship may be limited
to the lower end of the temporal axis. For exam-
ple, if the assumption that new species start with
small ranges is accepted, then the general positive
relationshi pbetween s pecies age and range size
can be expected to persist until some threshold,
and then the relationshi pwill become flat or neg-
ative, as older species lose range size. Almost all of
the models of post-speciation range-size transfor-
mation presented in Gaston (1998), for example,
have an initial phase in which there is a roughly
linear positive relationship between species age
and range size. The differences in these lines is
the steepness of their slope and their temporal
duration; some models, such as a cyclical and sta-
sis models, predict a rapid increase in range size
post-speciation, while the traditional age and area
model is depicted as a gradual increase. How-
ever, depending on the total age of a clade of
interest, and the rate at which transformations
occur, all of these models are similar in their initial