54
Basic biogeography: estimating biodiversity and mapping nature
commented that we do not have accurate information
on the geographical distribution of any plant species in
the Amazon (Bush & Lovejoy, 2007 ).
It has recently been calculated that 43 per cent of
the total area of Amazonia has never been surveyed by
botanists, while another 28 per cent is poorly collected
and only two per cent can be considered ‘ relatively
well ’ collected (Schulman et al ., 2007 ). Interestingly,
the well - collected part is close to Manaus, where Erwin
did his seminal work on beetle diversity. These observations highlight an important problem in conservation
biogeography, whereby those seeking to provide scientifi c guidance as to where to locate protected areas can
be tempted to base their analyses on ‘ the best available
data ’ , and in the process may pay insuffi cient attention
either to the magnitude or the geographical structure
of the Wallacean shortfall.
A paper by Hopkins (2007) shows how it is possible
to use data from herbarium (or museum) collections to
tease apart what component of diversity variation may
be an artefact of collecting intensity. Taking a lead from
a seminal paper by Nelson et al . (1990) , which showed
a strong correspondence between regions that were
considered to be ‘ centres of richness and endemism ’
and collecting intensity, Hopkins (2007) used the
known occurrences of 1,584 species of Magnoliophyta
to build models of collecting defi cit. Noting that as
many as 40 per cent of Amazonian plants in herbaria
may bear incorrect identifi cations, he placed his reliance instead on the analysis of monographed taxa,
being those for which specialists have examined the
specimens, revised the identifi cations and mapped
known distributions using grids of 1 degree latitude/
longitude.
The steps in his analysis are demonstrated in Figure
4.2 , in which he illustrates, for two species, how the
Several biologists have questioned the overall utility
of trying to name all species without, at the same time
as we delimit them, making greater efforts to describe
their biology, systematics, ecology and distribution
(Raven, 2004 ). In particular, it would be diffi cult to
imagine the utility of long lists of taxa for conservation
biogeography without information on geographical
distributions, on scales spanning from the local to the
global.
4.2.2 The Wallacean s hortfall
The Wallacean shortfall refers primarily to the
inadequacy of our knowledge of the geographical distributions of species, although Lomolino et al . (2010,
p. 740) have defi ned it slightly more broadly as ‘ the
paucity of information on species distributions and
on the geographic dynamics of extinction forces, especially the geographic dynamics of human civilizations. ’
Like the Linnean shortfall, the term ‘ Wallacean shortfall ’ pays tribute to a great naturalist, in this case Alfred
Russel Wallace (1823 to 1913), who, in addition to
being the co - discoverer of the theory of evolution by
natural selection, is also recognized as one of the
founding fathers of zoogeography.
As we have already established, there are millions of
species in the world and we continue to discover and
describe new ones. The scale of the task in determining
their distributions is comparable to the taxonomic
labour required, particularly so in many developing
countries which lack the resources for the necessary
fi eld inventory and related identifi cation work. Some
habitats, such as the Amazon rain forest, are so vast so
as to render systematic sampling an impossibility using
the technology currently available. Indeed, it has been
Figure 4.2 Steps in producing hypothetical distribution maps, illustrated (panels a – d) for a species with a restricted
distribution ( Inga plumifera ), and (panels e – h) for a species with a widespread distribution ( Inga capitata ). (a) and (e) the
degree squares with confi rmed occurrences. (b) and (f) the contours of the predicted probability of occurrence, using a
probability of occurrence in adjacent degree squares of 0.5 and allowing this effect to accumulate for 5 degree squares.
(c) and (g) the hypothetical distribution deduced by accepting a probability of occurrence of greater than 0.5 in any degree
square. (d) and (h) the degree squares for each species. Summing these values across all species modelled in the exercise
allows the estimation of the total number of species hypothetically occurring in any one degree square. From Hopkins (2007) .
(See Plate 4.2 for a colour version of these images.)
Basic biogeography: estimating biodiversity and mapping nature
commented that we do not have accurate information
on the geographical distribution of any plant species in
the Amazon (Bush & Lovejoy, 2007 ).
It has recently been calculated that 43 per cent of
the total area of Amazonia has never been surveyed by
botanists, while another 28 per cent is poorly collected
and only two per cent can be considered ‘ relatively
well ’ collected (Schulman et al ., 2007 ). Interestingly,
the well - collected part is close to Manaus, where Erwin
did his seminal work on beetle diversity. These observations highlight an important problem in conservation
biogeography, whereby those seeking to provide scientifi c guidance as to where to locate protected areas can
be tempted to base their analyses on ‘ the best available
data ’ , and in the process may pay insuffi cient attention
either to the magnitude or the geographical structure
of the Wallacean shortfall.
A paper by Hopkins (2007) shows how it is possible
to use data from herbarium (or museum) collections to
tease apart what component of diversity variation may
be an artefact of collecting intensity. Taking a lead from
a seminal paper by Nelson et al . (1990) , which showed
a strong correspondence between regions that were
considered to be ‘ centres of richness and endemism ’
and collecting intensity, Hopkins (2007) used the
known occurrences of 1,584 species of Magnoliophyta
to build models of collecting defi cit. Noting that as
many as 40 per cent of Amazonian plants in herbaria
may bear incorrect identifi cations, he placed his reliance instead on the analysis of monographed taxa,
being those for which specialists have examined the
specimens, revised the identifi cations and mapped
known distributions using grids of 1 degree latitude/
longitude.
The steps in his analysis are demonstrated in Figure
4.2 , in which he illustrates, for two species, how the
Several biologists have questioned the overall utility
of trying to name all species without, at the same time
as we delimit them, making greater efforts to describe
their biology, systematics, ecology and distribution
(Raven, 2004 ). In particular, it would be diffi cult to
imagine the utility of long lists of taxa for conservation
biogeography without information on geographical
distributions, on scales spanning from the local to the
global.
4.2.2 The Wallacean s hortfall
The Wallacean shortfall refers primarily to the
inadequacy of our knowledge of the geographical distributions of species, although Lomolino et al . (2010,
p. 740) have defi ned it slightly more broadly as ‘ the
paucity of information on species distributions and
on the geographic dynamics of extinction forces, especially the geographic dynamics of human civilizations. ’
Like the Linnean shortfall, the term ‘ Wallacean shortfall ’ pays tribute to a great naturalist, in this case Alfred
Russel Wallace (1823 to 1913), who, in addition to
being the co - discoverer of the theory of evolution by
natural selection, is also recognized as one of the
founding fathers of zoogeography.
As we have already established, there are millions of
species in the world and we continue to discover and
describe new ones. The scale of the task in determining
their distributions is comparable to the taxonomic
labour required, particularly so in many developing
countries which lack the resources for the necessary
fi eld inventory and related identifi cation work. Some
habitats, such as the Amazon rain forest, are so vast so
as to render systematic sampling an impossibility using
the technology currently available. Indeed, it has been
Figure 4.2 Steps in producing hypothetical distribution maps, illustrated (panels a – d) for a species with a restricted
distribution ( Inga plumifera ), and (panels e – h) for a species with a widespread distribution ( Inga capitata ). (a) and (e) the
degree squares with confi rmed occurrences. (b) and (f) the contours of the predicted probability of occurrence, using a
probability of occurrence in adjacent degree squares of 0.5 and allowing this effect to accumulate for 5 degree squares.
(c) and (g) the hypothetical distribution deduced by accepting a probability of occurrence of greater than 0.5 in any degree
square. (d) and (h) the degree squares for each species. Summing these values across all species modelled in the exercise
allows the estimation of the total number of species hypothetically occurring in any one degree square. From Hopkins (2007) .
(See Plate 4.2 for a colour version of these images.)