Content uploaded by Stuart L Pimm
Author content
All content in this area was uploaded by Stuart L Pimm on Mar 25, 2015
Content may be subject to copyright.
HOW MANY PLANT SPECIES ARE Stuart L. Pimm
1
and Lucas N. Joppa
2
THERE, WHERE ARE THEY, AND
AT WHAT RATE ARE THEY
GOING EXTINCT?
ABSTRACT
How many flowering plant species are there? Where are they? How many are going extinct, and how fast are they doing so?
Interesting in themselves, these are questions at the heart of modern conservation biology. Determining the answers will dictate
where and how successfully conservation efforts will be allocated. Plants form a large taxonomic sample of biodiversity. They are
important in themselves and directly determine the diversity of many other taxonomic groups. Inspired by conversations with
Peter Raven, we set out to provide quantitative answers to these questions. We argue that there are 450,000 species, two thirds
of which live in the tropics, a third of all species are at risk of extinction, and they are going extinct 1000 to 10,000 times the
background rate. In obtaining these results, we point to the critical role of dedicated taxonomic effort and biodiversity
monitoring. We will only get a good answer to the age-old question of ‘‘how many species are there?’’ when we understand the
population biology and social behavior of taxonomists. That most missing species will be found in biodiversity hotspots reaffirms
their place as the foci of extinction for decades to come. Important, but not yet addressed, are future studies of how long plant
species take to become extinct in habitat fragments. These will deliver not only better estimates of extinction rates, but also the
critical timeframe of how quickly one needs to act to prevent extinctions.
Key words: Biodiversity, discovery curves, extinction rates, flowering plants, missing species.
We ask four questions that arose in the context of only how many species were known, but also how
25 years of discussions with Peter Raven. How many many were still unknown to science. We start with
plant species are there? Where are they? How many those questions.
are going extinct, and how fast are they doing so?
Raven has considered these questions throughout his HOW MANY PLANT SPECIES ARE THERE?
career. Their extension to all species is obvious and, There are two questions in estimating the total
as such, they have an illustrious pedigree. For numbers of plant species: how many do we know, and
example, Westwood (1833) speculated ‘‘ on the how many do we not know, i.e., how many are still
probable number of species of insects in the unknown to science? For shorthand, we will call the
Creation.’’ These questions are also both hard to latter ‘‘ missing species.’’ The first question might
answer and exceptionally rich in the subsequent seem easy. The problem of synonymy (i.e., taxono-
explorations we must undertake to get the final mists giving different names to the same species
answers. We argue that there are 450,000 species, inadvertently) complicates it. There have been
two thirds of which live in the tropics, a third of all several recent estimates of the currently known
species are at risk of extinction, and they are going number of unique species of seed plants (Prance et
extinct 1000 to 10,000 times the background rate. In al., 2000; Govaerts, 2001; Bramwell, 2002; Thorne,
obtaining these results, we point to the critical role of 2002; Scotland & Wortley, 2003; Paton et al., 2008),
dedicated taxonomic effort and biodiversity monitor- with the highest estimate (422,127; Govaerts, 2001)
ing. two times the lowest one (223,300; Scotland &
We shall consider each of the questions in turn but Wortley, 2003).
not in the order that we thought about them. Pimm The Plant List (,www.theplantlist.org.)isan
and Raven (2000), in a commentary that accompa- ongoing collaboration between the Missouri Botanical
nied Myers et al. (2000), addressed the last question Garden and the Royal Botanic Gardens, Kew. In May
first by building simple scenarios of extinction based 2013, it estimated that there were 352,000 species of
on the loss of tropical forests. It quickly became flowering plants (angiosperms; the same number as
apparent that the answer would depend on how many Paton et al., 2008), with 298,900 accepted names
species lived in tropical forests and, indeed, in which and 263,925 names yet to be assessed, though surely
particular forests. In turn, that meant knowing not many of them will be synonyms. Because the
1
Nicholas School of the Environment, Duke University, Box 90328, Durham, North Carolina 27708, U.S.A.
2
Microsoft Research, 14820 NE 36th St., Redmond, Washington 98052, U.S.A.
doi: 10.3417/2012018
ANN.MISSOURI BOT.GARD. 100: 170–176. PUBLISHED ON 16 MARCH 2015.
Volume 100, Number 3 Pimm & Joppa 171
2015 How Many Plant Species Are There?
accepted names among those resolved is 38%,it Moreover, a broad exponential increase in names
seems reasonable to predict that the same proportion applies to other taxa including mammals, spiders,
of unresolved names will eventually be accepted. amphibians, and example genera of marine gastro-
This yields another ca. 100,000 species for a total pods (Joppa et al., 2011c). Under such circumstanc-
estimate of ca. 400,000 species. These are monu- es, estimates of missing species from extrapolations of
mental assessments, necessary because the individ- uncorrected discovery rates will be nonsense. (Only
ual names, identities, and relationships are the basis birds are exceptional and for which the taxonomic
of so much subsequent science. catalogue now appears almost complete.)
Yet, there are also questions that require estimat- The underlying problem is that the total number of
ing how incomplete this catalogue is. How much more plant taxonomists active in any period has increased
effort will be required to complete it? Are we years, exponentially, too, doubling about every 30 years.
decades, or centuries away from a widely accepted (We define ‘‘taxonomists,’’ simply as those who
answer? For consortia such as the Catalogue of Life describe new species.) Broadly similar rates charac-
(,http://www.catalogueoflife.org.), these are funda- terize the animal taxa. Again, excepting birds, the
mental questions. At a local scale, projects such as rates of increase are even higher over the last half
the All Taxa Biodiversity Inventory ask how many century. Given this observation, it is not particularly
species live in an area. For the Great Smoky surprising that the raw numbers of species described
Mountains National Park in the eastern United over time have increased as well. The number of
States, the answer is 7636 of all species combined taxonomists is a powerful predictor of the number of
(as of 5 October 2014 there are 7799 species, 931 of species described.
them new to science), with 923 of these (13%) new to The solution to this problem also seems obvious:
science (,http://www.dlia.org/node/204.). That so divide the number of species described by the total
many species of animals and plants are new in an number of taxonomists and expect that this ratio will
otherwise well-explored area surrounded by major decline as the pool of missing species declines. This
research universities is a testament to our ignorance also fails.
about life’s diversity. The World Checklist of Selected Plant Families
There are practical concerns, too, about knowing (WCSP; ,http://www.kew.org/wcsp.)providesa
the number of missing species. Because they are most sample of approximately 119,000 species of system-
likely rare, their number and geographical distribu- atically revised species, including all of the monocots
tion are essential to answering the second and third and selected non-monocot families. We show that for
questions we ask. both monocots and selected non-monocot families,
There have been many previous attempts using, there was an increase in the number of species
e.g., scaling laws in food webs, abundance, body size, described per taxonomist, typically for the first
and rarity to estimate the number of missing species century or so after Linnaeus (Joppa et al., 2011b).
(May, 1988, 1990, 1992). Recent attempts employ Taxonomists have likely increased the efficiency of
what surely seems like the most promising approach. their efforts since the mid 1700s when Linnaeus
As the pool of missing species declines, the numbers introduced the system of binomial nomenclature and
of discoveries will also likely decline in direct founded modern taxonomic practice. There could be
proportion (Solow & Smith, 2005; Wilson & Costello, many reasons for taxonomists becoming more effi-
2005). The obvious analogy is with predator-prey cient over time, of course. Clearly, the world is easier
interactions: taxonomists are the ‘‘predators,’’ and as to explore now than it was in the past; natural history
the supply of ‘‘prey’’ (i.e., missing species) declines, collections are more accessible; techniques are
so too should the capture rate. better; there is more collaboration, and so on. In
the predation analogy, the predators have become
TAXONOMISTS ARE PREDATORS,MISSING SPECIES ARE PREY smarter over time.
Completely counter to this expectation, the Combining these considerations suggests a model
numbers of both monocot and non-monocot species where one predicts the numbers of species described
described per five-year interval have increased per taxonomist over a given interval in terms of the
almost exponentially since 1950 (Joppa et al., continually diminishing pool of missing species,
2011b). To take an example, since 1950, there have combined with some simple increase in the efficiency
been 95 accepted species names for Onagraceae. of the taxonomic enterprise. By doing so, we
From 1900 to 1949, taxonomists named only 43 concluded that between 10%and 20%of the current
species. species total are missing, suggesting that there are
172 Annals of the
Missouri Botanical Garden
probably about 450,000 species of flowering plants description of how many missing species receive
(Joppa et al., 2011b). names each year.
We also estimated the numbers of missing species
family by family to assess which families might hold WHERE ARE THE MISSING SPECIES?
the greatest number (Joppa et al., 2011b). Then, we Obviously, we can apply these ‘‘ taxonomists as
compared our estimates to those based on a selection predators’’ models to predicting where missing
of taxonomic experts who knew the families. There species live. The WCSP gives species locations
was broad agreement between the models and the based on classifying the world into 369 regions. Only
experts. 33 of these ranges have sufficient endemics to run our
Now, the details of what constitutes a taxonomist models, but by judiciously combining regions, we
and who contributes what to a species description are grouped about 72%of the species in 50 broad areas
elsewhere (Joppa et al., 2011c), but the Onagraceae (Joppa et al., 2011a). For example, many countries in
are broadly typical of other families. In the century Central America are quite small and individually
after Linnaeus (1753), 74 taxonomists described 141 have few endemics. Yet, a combined 9%of all known
Onagraceae species and mostly did so on their own. species occur there. The remaining 28%of species
Since 2000, 34 taxonomists have contributed to occur in two or more of the regions as we defined
descriptions and revisions of 47 species of Onagra- them.
ceae. Single authors described only four of these So, where is plant diversity highest once we adjust
species; the rest were collaborations. Collaborators for where we think the missing species occur? Our
are not mere assistants, however, but those who, at analyses leave untouched the idea that plant species
least eventually, describe species in their own right. are mostly tropical; almost exactly two thirds (65.7%)
Of those 34 recent taxonomists, 25 of them were of all species, known and missing, live in the tropics
senior authors on one or more descriptions. Joppa et (Joppa et al., 2011a). That is exactly the fraction that
al. (2011c) also show that taxonomists are becoming Raven (1981) asserted were tropical.
more specialized over time. The modern tendency is We predicted that the great majority of the missing
for taxonomists to work on only one or a few families species would be in the region from Mexico to
and often in one broad geographic region. In the Panama (6%of all the predicted missing species),
predation analogy, predatory taxonomists are now Colombia (6%), Ecuador to Peru (29%), Paraguay,
more social and more specialized than they were Argentina, and Chile (5%), southern Africa (16%),
before. and Australia (8%). These areas combined have 70%
For some plant families (e.g., Onagraceae), the of all the species we predict are missing. All of these
number of species described per taxonomist does not areas are biodiversity hotspots (Myers et al., 2000),
obviously drop over time. This defies any effort to which leads directly to our next question.
predict the number of missing species from simple
description rates. Inspection of the descriptions HOW MANY PLANT SPECIES ARE GOING EXTINCT?
shows that taxonomists (in this case, often Raven
and his colleagues) are not ‘‘eating’’ species Estimating the fraction of plant species threatened
randomly but working their way through plant with extinction requires that we understand how
families, group by group. That is, taxonomists revise many species are missing from the taxonomic
a family genus by genus, doing so at a rate that catalogue (Joppa et al., 2011b). Brummitt et al.
reflects the resources and time at their disposal. (2008) suggested that 20%of known plant species
Genuinely new species as well as revisions of old are threatened. Take this estimate, and then add to
ones appear in clusters by group, sometimes a few at that our result that there are 10%–20%more missing
a time, sometimes involving many. When the supply species. Then assume that essentially all of these
of unrevised genera dries up, we can expect the rates missing species have small geographical ranges and
of description to drop dramatically. are locally rare. That is likely why they are missing,
We have yet to model this circumstance. It is clear after all! We predict that they are in the biodiversity
that as we delve deeper into particular plant families, hotspots that, by definition, have high levels of
the complex social behavior and working habits of habitat loss. Thus, the missing species are also surely
taxonomists become more important in driving the threatened with extinction. This argument predicts
rates of species’ descriptions. Nonetheless, for large that 27%–33%of all plant species are likely
sets of species (e.g., all monocots), the basic model of threatened. These estimates are based on immediate
taxonomists as predators gives a good statistical threats and do not consider further development of
Volume 100, Number 3 Pimm & Joppa 173
2015 How Many Plant Species Are There?
destructive factors, including climate disruption tinction rates requires knowing whether the areas
(Pimm, 2009), during the remainder of this century. where species with small ranges are concentrated
So, how good are the assumptions that the missing coincide with places where there is extensive habitat
species have small ranges and that taxonomists will loss.
indeed find them in the biodiversity hotspots? The unique contribution to conservation from
Understanding where missing species are likely to Myers et al. (2000) was to note not only that small-
live is vital in setting international priorities for ranged species are geographically concentrated, but
conservation and ensuring that we find and protect also that perversely most are located in areas with
the missing species before we drive them to disproportionately high levels of habitat destruction.
extinction. Cincotta et al. (2000) added high human population
Answering this question requires several of the growth to the list of threats to these biodiversity
basic ‘‘laws’’ of biodiversity. By ‘‘ law’’ we mean hotspots.
generalizations or patterns, if you will, that apply Given that we have shown the taxonomic catalogue
widely and across many taxa. The first law is that the is short by 10%–20%, then how would an under-
average geographical range size of a group of species standing of where the missing species live alter our
is very much larger than the median range. For views of hotspots? For obvious reasons, new species
vertebrate taxa, there are now global maps of species’ discoveries are overwhelmingly of species that have
ranges. The average of the geographical ranges of small geographical ranges. Will knowing where the
1684 species of mammals in the New World is 1.8 missing species reside change the way we set
2
conservation priorities? Will relative priorities
million km , but 50%of those species have ranges
smaller than 250,000 km
2
change as taxonomists complete the catalogue? Will
, a seven to one ratio. For
new priorities become apparent? Are the missing
the region’s three main bird groups (non-passerines, species in places where they are likely to be
suboscine passerines, and oscine passerines), the threatened? Will we discover them before they
ratios are between five and eight to one. For become extinct?
amphibians, they are 40 to one. Simply, there are We have already anticipated the answers to these
many species with small ranges and few with large questions: the hotspots hold the majority of the
ranges. species we predict to be missing. The missing species
We do not have comparable data for flowering are in harm’s way and possess small ranges. Myers et
plants, unfortunately. In the sample of systematically al. (2000) concluded that the hotspots are where
revised species compiled from the WCSP database species are going extinct. Our estimates of where the
(Joppa et al., 2011a), one half of the species occur in missing species occur do not contradict this conclu-
regions that individually are no larger than 1.4
2
sion (Joppa et al., 2011a). What the estimates do
million km . Of course, these species do not occur show is that extinctions are even more likely to be
over the entire extent of the region in which they concentrated in hotspots than we previously thought.
occur. Moreover, many other regions contain small-
2
Are there some places where recent habitat loss
ranged species. Therefore, 1.4 million km sets a high might cause the region to now qualify under criteria
upper limit for the median range size of a flowering from Myers et al. (2000) as a hotspot? New Guinea,
plant. There are surely many plant species with small e.g., has far greater than 1500 endemic plant species
ranges. but was not originally included as a hotspot due to
The next law of biodiversity is that species with relatively intact natural vegetation. Ongoing loss of
small ranges are geographically extraordinarily natural habitat through logging, mining, and road
concentrated. In total, 88 of the 369 regions covering construction (Shearman et al., 2009) could make the
only 5%of the ice-free land surface combined hold region a likely candidate for inclusion as a hotspot in
30%of the known plant species, while 111 regions the near future. Increased access could also lead to
covering a combined total of 10%hold 40%of the an increase in the descriptions of new species, of
species. course. New Guinea’s inaccessibility may explain its
Currently, small geographical range is overwhelm- low rates of description and the predictions of there
ingly the best statistical predictor of whether a being few missing species.
species will be in danger of extinction. Indeed,
sufficiently small geographical range combined with EXTINCTION SCENARIOS
an occurrence in areas of extensive habitat loss
classifies a species as being ‘‘threatened’’ under the Pimm and Raven (2000) used results from Myers
IUCN Red List criteria (,http://www.iucnredlist.org/ et al. (2000) to develop scenarios of how many
static/categories_criteria_3_1.). Understanding ex- species would become extinct. We used the familiar
174 Annals of the
Missouri Botanical Garden
species-area relationship (S ¼cA
z
) that relates the endangered species (Brooks et al., 1997, 1999a;
number of species (S) to area (A), using two Cowlishaw, 1999; Brook et al., 2003). Armed with
parameters, c and z, for which z is vital. The this agreement, Pimm and Raven (2000) took the
parameter c is a constant of proportionality. For estimates of endemic species in the hotspots Myers et
samples of different sizes within continuous habitat, z al. (2000) defined, with the estimates of remaining
is very small (Rosenzweig, 1995). This describes, habitat and, thus, predicted the total extinctions that
e.g., the observation that the number of tree species would occur. For example, Myers et al. estimated that
in the Great Smoky National Park is not that much only 7.5%of the Atlantic coast rainforest of Brazil
smaller than the total number across eastern remains and that this area contained 2.7%of their
deciduous forest from Florida to Maine and westward estimate of 300,000 plants as endemics to this area.
to the prairies. The continued rate of species descriptions should
Now, what happens after habitat destruction? Two eventually yield 3.6%of 450,000 species of plants
things: first, some species will go extinct immediate- (Joppa et al., 2011a) for this area. Applying the
ly, because they lived only in the destroyed area. species-area calculation, about one half of the
Metaphorically, they go extinct ‘‘overnight.’’ The endemic plant species in this region should be
second wave of extinction invokes a separate species endangered (ca. 4000 according to Myers et al., but
to area relationship, that for oceanic islands where 7200 with the missing species included).
the areas involved are isolated. Here, z is much This and similar calculations have uncertainties, of
larger, typically about one fourth (Rosenzweig, 1995), course. This example suggests that the largest
meaning in simple terms, an island one half the size uncertainty is surely that the estimate of the missing
species comes from present rates of species descrip-
of a larger one will have 85%as many species.
tions. Those rates now reflect collecting and
The difference between the two relationships systematic revisions of plants from a mere 7.5 of
comes from Preston’s (1962) classic work. He %
the region’s original forested area. There is a strong
understood that for isolated populations there is a possibility that there may have been many extinctions
minimum viable population size. Suppose we reduce across this area that are not estimated by our
the total number of individuals by one half, shrinking ‘‘taxonomists as predator’’ models of species discov-
as it were the larger island to the smaller one. Then ery. Simply, there are the species we know, some of
15%of the species would have population sizes too which are threatened and some of which are not.
small to be viable. After habitat loss, some species Second, there are the species we estimate are missing
will go extinct ‘‘overnight,’’ but many more will linger from the taxonomic catalogue. We argue above that
until chance events eventually doom them. Such surely all of these are threatened. There is a final
experiences fit well with much practical experience of class of species: those that went extinct before we
endangered species. In addition, it fits with what we could even estimate that they were missing from the
know of the risks from stochastic events and taxonomic catalogue.
inbreeding in small populations (Brooks et al., 2011). Most of the remaining Atlantic Coast forest habitat
There are now many calibrations of forest losses is in mountains. Almost all of the lowland forests are
and subsequent extinctions, at least for birds and gone. How many species occurred in the lowland
mammals, the taxa that we know best. Pimm and forests before they were destroyed is beyond our
Askins (1995) asked how many bird species we present ability to model. What we know of the
expect to go extinct following the loss of forests in region’s orchids is that one half of the records are
eastern North America. About 150 species lived in from only one location and a quarter of the 600
those forests, but only 30 of these were endemic to species in the state of Espırito
´Santo (ca. 46,000 km
2
)
those forests. Pimm and Askins applied the island are endemic to it (Pimm, 2005). Almost certainly, the
species-area relationship (because by 1870 eastern median range size of orchids (at least) was so small
North America was a series of forest islands) to the 30 that many species could have lived in areas
endemic species, because the other species survived completely destroyed before taxonomists explored
in Canada and elsewhere in North America. This them.
model predicts four and one-half extinctions. As
forests were cleared in the 18th and 19th centuries, HOW LONG DOES EXTINCTION TAKE?
there were four documented forest bird extinctions,
and one species became endangered. Apart from the small fraction of species that go
Many other studies find similar, compelling extinct ‘‘overnight’’ after habitat loss, most species
agreements between forest losses and either the linger, declining inexorably toward extinction. The
number of actual extinctions or the number of obvious question is how long does this take? For the
Volume 100, Number 3 Pimm & Joppa 175
2015 How Many Plant Species Are There?
birds in eastern North America, we know it took to be separate from those species threatened by
decades after the low point of forest cover in 1870 habitat loss, making climate disruption an added and
before the species finally expired. Pimm and Raven potentially significant extra cause of extinctions.
(2000) assumed 50 years for the half-life. This affirms DISCUSSION:WHAT WENEED TO KNOW
the biodiversity hotspots as the foci of extinction for
decades to come. Studies of how long plant species Asking how many species there are suggests that
take to be lost from habitat fragments that parallel we will only get a good answer when we understand
those for birds (Brooks et al., 1999b; Ferraz et al., the population biology and social behavior of
2003) are vital, not only to get better estimates of taxonomists (Joppa et al., 2011c). We cannot
extinction rates, but also to give an idea of how understand where species live unless we know where
quickly one needs to act to prevent extinctions. the missing species occur. Likely, we also need to
The background rate of plant and animal extinc- know where species have already become extinct
tions—those before human impacts—appear to be before taxonomists started their explorations and
about one extinction per ten million species per year classifications. How fast species become extinct
(de Vos et al., 2014). Amphibians, birds, and requires knowledge of how long species last in the
mammals have current extinction rates of approxi- habitat fragments that are presently their only homes.
mately 100 extinctions per million species per year, It also requires an understanding of how many
(that is, 1000 times faster). If currently threatened species live in tropical mountains too close for
species last a century, a most optimistic scenario, the comfort to the mountain tops given minimum
extinction rates will be ten times higher. Plants have projected global temperature rises of 28C.
a higher fraction of threatened species than do these
vertebrates so the plant extinction rates are likely to Literature Cited
be comparable or higher. Bramwell, D. 2002. How many plant species are there? Pl.
Pimm and Raven (2000) considered the dynamics Talk 28: 32–34.
of unfolding habitat losses only briefly. They Brook, B. W., N. S. Sodhi & P. K. L. Ng. 2003. Catastrophic
extinctions follow deforestation in Singapore. Nature 424:
considered two scenarios: extinctions based on how 420–426.
much habitat remained in 1990 within the biodiver- Brooks,T.M.,S.L.Pimm&N.J.Collar.1997.
sity hotspots and how much habitat was protected, Deforestation predicts the number of threatened birds
assuming that only this area would remain intact. in insular Southeast Asia. Conservation Biol. 11: 382–
394.
Since then, there have been new threats to species in Brooks, T. M., S. L. Pimm, V. Kapos & C. Ravilious. 1999a.
the form of oil-palm cultivation, especially in Threat from deforestation to montane and lowland birds
Southeast Asia, but also encouraging signs in the and mammals in insular Southeast Asia. J. Anim. Ecol.
increase of the global total of protected areas (Jenkins 68: 1061–1078.
Brooks, T. M., S. L. Pimm & J. O. Oyugi. 1999b. Time lag
& Joppa, 2009) and the success of protected areas in between deforestation and bird extinction in tropical
reducing habitat loss (Joppa et al., 2008; Joppa & forest fragments. Conservation Biol. 13: 1140–1150.
Pfaff, 2011). Exactly where protected areas are Brooks, T. M., B. W. Brook, L. P. Koh, H. M. Perreira, S. L.
established is a key variable in driving extinction Pimm, M. L. Rosenzweig & N. S. Sodhi. 2011.
rates (Rodrigues et al., 2004), and many protected Extinctions: Consider all the species. Nature 474: 284.
Brummitt, N., S. Bachman & J. Moat. 2008. Applications of
areas are not in the necessary places (Joppa & Pfaff, the IUCN Red List: Towards a global barometer for plant
2009). diversity. Endang. Sp. Res. 6: 127–135.
Finally, Pimm and Raven (2000) did not consider Cincotta, R. P., J. Wisnewski & R. Engelman. 2000.
global change. Given our predictions that the greatest Human population in the biodiversity hotspots. Nature
404: 990–992.
numbers of threatened plant species are likely to be Cowlishaw, G. 1999. Predicting the pattern of decline of
the northern Andes, the coastal forest of Brazil, African primate diversity: An extinction debt from
Mexico, and Central America, and other mountainous historical deforestation. Conservation Biol. 13: 1183–
areas, this raises an obvious question. Because the 1193.
De Vos, J. M., L. N. Joppa, J. L. Gittleman, P. R. Stephens,
latitudinal gradient of temperatures in the tropics is & S. L. Pimm. 2014. Estimating the normal background
weak, the way species will avoid the effects of climate rate of species extinction. Cons. Biol. (in press).
disruption will be to move to higher elevations (Pimm, Ferraz, G., G. J. Russell, P. C. Stouffer, R. O. Bierregaard,
2009). That almost inevitably means less suitable S. L. Pimm & T. E. Lovejoy. 2003. Rates of species loss
areas for species, some of which may simply run out from Amazonian forest fragments. Proc. Natl. Acad. Sci.
U.S.A. 100: 14,069–14,073.
of space as their ranges move higher into the Forero-Medina, G., L. N. Joppa & S. L. Pimm. 2011.
mountains, if they can get there in the first place Constraints to species’ elevational range shifts as climate
(Forero-Medina et al., 2011). These species are likely changes. Conservation Biol. 25: 163–171.
176 Annals of the
Missouri Botanical Garden
Govaerts, R. 2001. How many species of seed plants are Pimm, S. L. 2009. Climate disruption and biodiversity.
there? Taxon 50: 1085–1090. Curr. Biol. 19: 595–601.
Jenkins, C. N. & L. N. Joppa. 2009. Expansion of the global Pimm, S. L. & R. Askins. 1995. Forest losses predict bird
terrestrial protected area system. Biol. Conservation 142: extinctions in eastern North America. Proc. Natl. Acad.
2166–2174. Sci. U.S.A. 92: 9343–9347.
Joppa, L. N. & A. Pfaff. 2009. High and far: Biases in the Pimm, S. L. & P. Raven. 2000. Biodiversity: Extinction by
locations of protected areas. PLoS ONE 4: e8274. numbers. Nature 403: 843–845.
Joppa, L. N. & A. Pfaff. 2011. Global protected area Prance, G., H. Beentje, J. Dransfield & R. Johns. 2000. The
impacts. Proc. Roy. Soc. London, Ser. B, Biol. Sci. 278: tropical flora remains undercollected. Ann. Missouri Bot.
1633–1638. Gard. 87: 67–71.
Joppa, L. N., S. R. Loarie & S. L. Pimm. 2008. On the Preston, F. W. 1962. The canonical distribution of
protection of ‘protected areas.’ Proc. Natl. Acad. Sci. commonness and rarity. Ecology, 43: 185–215.
U.S.A. 105: 6673–6678. Raven, P. H. 1981. Research Priorities in Tropical Biology.
Joppa, L. N., D. L. Roberts, N. Myers & S. L. Pimm. 2011a. National Academy Press, Washington, D.C.
Biodiversity hotspots house most undiscovered plant
species. Proc. Natl. Acad. Sci. U.S.A. 108: 13,171– Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L.
13,176. Boitani, T. M. Brooks, R. M. Cowling, L. D. C. Fishpool,
Joppa, L. N., D. L. Roberts & S. L. Pimm. 2011b. How G. A. B. da Fonseca, K. J. Gaston & M. Hoffmann. 2004.
many species of flowering plants are there? Proc. Roy. Effectiveness of the global protected area network in
Soc. London, Ser. B, Biol. Sci. 278: 554–559. representing species diversity. Nature 428: 640–643.
Joppa, L. N., D. L. Roberts & S. L. Pimm. 2011c. The Rosenzweig, M. L. 1995. Species Diversity in Space and
population ecology and social behaviour of taxonomists. Time. Cambridge University Press, Cambridge.
Trends Ecol. Evol. 26: 551–553. Scotland, R. W. & A. H. Wortley. 2003. How many species
Linnaeus, C. 1753. Species Plantarum, Vol. 2. Imprensis of seed plants are there? Taxon 52: 101–104.
Laurentii Salvii, Stockholm, pp. 970–971. Shearman, P. L., J. Ash, B. Mackey, J. E. Bryan & B. Lokes.
May, R. 1988. How many species are there on earth? 2009. Forest conversion and degradation in Papua New
Science 241: 1441–1449. Guinea 1972–2002. Biotropica 41: 379–390.
May, R. 1990. How many species? Philos. Trans., Ser. B Solow, A. & W. Smith. 2005. On estimating the number of
330: 293–304. species from the discovery record. Proc. Roy. Soc.
May, R. 1992. How many species inhabit the earth? Sci. London, Ser. B, Biol. Sci. 272: 285–287.
Amer. 267: 42–48. Thorne, R. F. 2002. How many species of seed plants are
Myers, N., R. Mittermeier, C. Mittermeier, G. da Fonseca & there? Taxon 51: 511–522.
J. Kent. 2000. Biodiversity hotspots for conservation Westwood, J. 1833. On the probable number of species of
priorities. Nature 403: 853–858. insects in the creation; together with descriptions of
Paton, A., N. Brummitt, R. Govaerts, K. Harman, S. several minute Hymenoptera. Mag. Nat. Hist. & J. Zool.
Hinchcliffe, B. Allkin & E. Lughadha. 2008. Towards
Target 1 of the Global Strategy for Plant Conservation: A 6: 116–123.
working list of all known plant species—Progress and Wilson, S. & M. Costello. 2005. Predicting future
prospects. Taxon 57: 602–611. discoveries of European marine species by using a non-
Pimm, S. L. 2005. It’s a new century: Do you know where homogeneous renewal process. Appl. Statistics 54: 897–
your orchids are? Selbyana 26: 5–13. 918.
