Estimating vertebrate biodiversity using the tempo of taxonomy - a view from Hubbert's peak

Page created by Salvador Aguilar
 
CONTINUE READING
Biological Journal of the Linnean Society, 2021, XX, 1–21. With 13 figures.

Estimating vertebrate biodiversity using the tempo of
taxonomy – a view from Hubbert’s peak

                                                                                                                                       Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
BRUCE H. WILKINSON1,2,*, LINDA C. IVANY2 and CARL N. DRUMMOND3
1
  Department of Earth and Environmental Sciences, University of Michigan, Ann Arbor, MI 48109-1005,
USA
2
  Department of Earth and Environmental Sciences, Syracuse University, Syracuse, NY 13244, USA
3
  Department of Physics, Purdue University Fort Wayne, Fort Wayne, IN 46805, USA

Received 28 April 2021; revised 30 April 2021; accepted for publication 5 May 2021

Reservoirs of natural resources are finite and, with increasing exploitation, production typically increases, reaches
a maximum (Hubbert’s peak) and then declines. Similarly, species are the currency of biodiversity, and recognized
numbers are dependent upon successful discovery. Since 1758, taxonomists have exploited a shrinking reservoir of
as-yet-unnamed vertebrate taxa such that rates of species description at first rose, reached a peak and then declined.
Since about 1950, increases in research funding and technological advances have fostered a renewed increase in
rates of discovery that continues today. Many attempts to estimate global biodiversity are forecasts from data on past
rates of description. Here we show that rates of discovery of new vertebrate taxa have been dependent upon the size
(richness) of the taxonomic pool under consideration and the intensity of ‘sampling’ effected by taxonomists in their
efforts to discover new forms. Because neither the current number of as-yet-to-be-described taxa nor future amounts
of taxonomic efforts can be known a priori, attempts to produce an accurate estimate of total global biodiversity
based on past rates of discovery are largely unconstrained.

ADDITIONAL KEYWORDS: biodiversity – Hubbert – Linnaeus – taxonomy – vertebrates.

    ‘We will only get a good answer to the age-old                  numbers of described taxa, the translation of these
    question of “how many species are there?” when                  metrics to total or as-yet-undiscovered richness is
    we understand the population biology and social                 fraught with assumptions. Coming from our perspective
    behavior of taxonomists’ (Pimm & Joppa, 2015).                  as geoscientists, we posit that the realization of a more
                                                                    complete taxonomy, and thus a better appreciation of
                                                                    Earth’s total biodiversity, is similar to attempts to infer
                                                                    the sizes of global reserves of many natural resources
                    INTRODUCTION
                                                                    currently under development. Species of vertebrates,
Taxonomy is one of the oldest areas of biological                   barrels of oil and tons of copper, for example, can be
investigation. The finding, naming and grouping of                  thought of as the units that comprise reservoirs to be
taxonomic entities are generally understood to be                   produced (discovered). They are distributed within
critical pursuits in the quest for a fuller understanding           local reservoirs (higher taxa, oil fields, ore deposits)
of the current state of Earth’s biodiversity and its                that exhibit fractal-like (many small – few large)
ongoing loss. As recently discussed by Moura & Jetz                 size frequency distributions, and the probability of
(2021), an assortment of biological, environmental                  their discovery depends on the sizes of the units in
and sociological circumstances serve to influence the               question and on the intensity of efforts devoted to
probability that a new species will be discovered and               their development. This comparison offers insight into
described. Although past approaches to estimating                   our ability to project the total biodiversity of a group
group richness have often relied on knowledge of                    from the history of taxonomic discovery of the fraction
                                                                    currently known.
                                                                      In the following, we first examine the numbers of
*Corresponding author. E-mail: eustasy@umich.edu                    taxonomic units comprising the biological reservoirs

© 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21                                    1
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.
org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is
properly cited.
2   B. H. WILKINSON ET AL.

that have been ‘exploited’ during the history of              Database, 2020 (www.mammaldiversity.org). Details
vertebrate taxonomy, and then discuss several                 about dates of access and numbers of taxonomic units,
metrics that might serve to quantify the scientific           authors and publications are listed as the Supporting
effort expended to produce said taxonomies. For the           Information (Table S1).
former, we describe the numbers and size frequencies             Current names, taxonomic classification and age of
of taxonomic units at different levels in the Linnaean        discovery were tabulated for all valid species; many are

                                                                                                                               Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
hierarchy and how they have changed over the past             now considered to have a different classification than
~260 years of discovery; in the latter, we consider           originally described, and we used the most current.
change in the numbers of papers and authors and their         Dates of establishment for all higher taxonomic
association with the pace of description of new taxa over     units are taken as those of the earliest identified
that same window of time. Based on observed patterns          species within that group. We use parameters of best-
of change at different taxonomic levels, we then develop      fit skewed normal and exponential distributions in
models that reproduce observed tempos of description.         order to derive semi-quantitative descriptors of the
These confirm the supposition that observed rates of          general trends apparent in these somewhat noisy
resource extraction (rates of taxonomic description)          data on taxonomic histories. Parameters of these best-
are a function of both reservoir size (total biological       fit functions are appended in the Supporting Tables.
diversity) and intensity of exploitation (taxonomic           Many aspects of taxonomic histories are similar among
effort). While past rates of description are known            different major groups. For brevity, we have elected to
from observation, past and future expenditures of             show and/or discuss various relationships with respect
taxonomic effort reflect a complex concatenation of           to one particular group as an example, and to illustrate
factors that cannot be forecast a priori. As a result,        that same relationship among other vertebrate classes
we are left with one known – the success rate of past         in the Supporting Figures.
efforts at description – and two unknowns – the actual
amount of biological diversity that exists on the Earth’s
surface and the total effort that must be expended to
                                                                   THE OBSERVED PACE OF TAXONOMY
effect a complete census of that diversity. We therefore
suggest that, as has been maintained for reserves of          The tempo of description of vertebrate taxa with time
natural resources (e.g. Lynch, 2010), estimates of total      shows clear patterns related to hierarchical level (Fig. 1).
biodiversity are largely unconstrained and cannot be          Higher taxonomic levels (e.g. orders and families) were
calculated with confidence from information on the            differentiated by and soon after Linnaeus in 1758,
history of taxonomic description to date. As much as          while numbers at lower taxonomic levels (e.g. genera
one might like to have ‘the number’, species richness of      and species) gradually increased, reached a peak and
the planet will remain an elusive target.                     then decreased. Around 1950, numbers of species
                                                              then underwent an abrupt and generally exponential
                                                              increase in description that continues today. Of
                                                              the 27 orders of modern mammals, 16 (59.3%) were
 AVAILABILITY, SOURCES AND TREATMENT
                                                              recognized by Linnaeus (1758), and all were defined by
                 OF DATA
                                                              1894, a span of 137 years. In contrast, only 157 (2.0%)
A variety of sources of taxonomic information are             of the 6485 currently recognized mammal species
generally available for vertebrate organisms (data            were known to Linnaeus in 1758; about 40 new species
for mammals, for example, are accessible from The             are currently documented each year, and this rate of
Mammal Diversity Database, the Mammal Species of              discovery has increased by about 2% per year since
the World, and the Integrated Taxonomic Information           1950. This ‘ontogeny of taxonomy’ embodied in the
System). We have examined several databases for each          history of classification of mammals (Fig. 1) is typical
of the groups considered here and find no important           of other vertebrate classes (Supporting Information,
differences with respect to results of analyses               Fig. S1; Tables S1 and S2).
performed; those selected for analysis herein afforded          The history of exploration of oil fields and ore
the most complete and/or current tabulations. Data            deposits reflects the fact that the largest reservoirs
for fishes, amphibians, reptiles, birds and mammals           are more likely to be discovered early on. So too is the
come, respectively, from FishBase (Froese and Pauly,          case with higher taxa: vertebrate groups described
2019; www.fishbase.se), the Integrated Taxonomic              chronologically earlier are also those with the highest
Information System online database (www.itis.                 memberships. As above, more than half the current
gov), the Reptile Database (Uetz et al., 2020; www.           orders of mammals (16 of 27) were established by
reptile-database.org) the International Community of          Linnaeus in 1758 based on his initial description of only
Ornithologists (IOC) World Bird List (Gill et al., 2020;      157 mammal species. Those original 16 mammal orders
www.worldbirdnames.org) and the Mammal Diversity              currently contain 95% of the 6485 now-recognized

                             © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                               3

                   100                                                                                                       100
                         A                                                B
                                     Orders                                          Families
Number Described

                                                                                                                                  Number Described

                                                                                                                                                     Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                                            Mode = 1750
                    10                                                                                                       10
                                                                                          n = 67, r2 = 0.269
                                              Mode = 1751
                                             n = 9, r2 = 0.917
                                                                                                            Laonastidae

                    1                                                                                                        1

                         C                                               D
                                    Genera                                     Species               Mode = 1886
Number Described

                                                                                                                                  Number Described
                                                                                                  n = 187, r2 = 0.379

                   100                                                                                                       100
                                                     Mode = 1928
                                                  n = 226, r2 = 0.350

                   10                                                                                                        10

                                                                                         S = 5.61 x 10-17 e-0.0203D
                                                                                           n = 187, r2 = 0.595

                             1800     1850      1900     1950     2000        1800    1850      1900        1950      2000

  Figure 1. Rates (number/year) of named groups of living mammals. Data as filled circles; general trends as best-fit skewed
  normal distributions except post-1950 species (D, yellow) which is a best-fit exponential. With decreasing taxonomic rank,
  rates change from decreasing (orders and families) to modal (genera) to presently increasing (species). Note logarithmic
  y-axes. Similar plots for fish, amphibians, reptiles and birds as Figure S1A–M.

  species (Fig. 2). Similarly, the 68 mammal families                    genus) taxonomic–ontogenetic age are apparent among
  recognized by Linnaeus (1758) represent 41% of                         data for other classes of vertebrates (Supporting
  (the 167) currently recognized families and 74% of                     Information, Fig. S2).
  currently recognized species. Conversely, the ‘youngest                  These observations – of early recognition of
  family’ of living mammals (Laonastidae) was erected                    higher taxa, continuing increases in description
  with the description of Laonastes aenigmamus by                        at the species level, and earlier recognition of the
  Jenkins et al. (2005) (though this family might                        most speciose groups at any level – are requisite
  actually represent a ‘Lazarus’ group that was thought                  features of any effort to model the history of
  to have become extinct in the Miocene; Dawson et al.,                  taxonomy. Embedded in these metrics are the need
  2006). As currently defined, the family contains one                   to understand the rules of taxonomic membership
  genus and one species (Fig. 2B). Likewise, among                       (the numbers of subtaxa within supertaxa) as well
  finfishes, the 47 orders erected by Linnaeus (1758)                    as the pace of description of the taxa themselves.
  (60% of the 78 modern orders) contain 33 913 (94%)                     We investigate both of these in subsequent sections
  of now-recognized species. The most recently defined                   before constructing numerical models to reproduce
  order, Stylephoriformes (Miya et al., 2007), contains                  the major features of the histories of taxonomic
  one living species (Stylephorus chordates, the deep-sea                endeavour, and ultimately to evaluate the potential
  tube-eye or thread-tail). Similar relationships between                for using such a model to predict future discoveries
  numbers of species and higher group (order, family,                    of new species.

  © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
4                 B. H. WILKINSON ET AL.

                                                                                                          A
                                                                                                  200

                                                                              Species per Genus

                                                                                                                                                                          Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                                                  150
                                                                B                                                                                  Genera
                                                        800                                       100
                                   Species per Family

                                                                                                   50
                            C                           600

                    2500
                                                        400
                                                                                                                  1800        1850       1900       1950       2000
Species per Order

                    2000                                                 Families                                        Year of Species Description
                                                        200
                    1500                                                                                                     Laonastidae

                    1000
                                                                      1800                        1850          1900        1950       2000
                                   Orders                                               Year of Species Description
                    500

                                 1800                         1850     1900                        1950          2000
                                                    Year of Species Description

  Figure 2. Relationships between the first year of description of some currently-recognized mammal species (x-axes) and
  the numbers of species in that group (y-axes). Number of members (species) largely predicates probability (the date) of
  group (order, family, genus) recognition. Similar plots for fish, amphibians, reptiles and birds as Figure S2A–L.

                           SIZES AND MEMBERSHIPS OF                                                           (e.g. Fig. 1) and is requisite for the construction of
                               TAXONOMIC GROUPS                                                               model taxonomies, which provide insight into the
                                                                                                              utility of observed taxonomic data for estimating the
  Understanding the numbers of subtaxa within any
                                                                                                              numbers of remaining taxa to be described.
  higher level of taxonomic consideration has been a focus
                                                                                                                A higher taxonomic unit, such as an order, can
  of scholarship since Willis (1922) described the ‘hollow-
                                                                                                              be visualized as encompassing some amount of
  curve’ nature of membership frequencies. Within any
                                                                                                              n-dimensional morphospace whose volume is occupied
  supertaxon, relatively few groups (taxa) contain most
                                                                                                              (and defined) by some number of non-contiguous
  of the subtaxa, while many are monotypic (Supporting
                                                                                                              subspaces, each representing a group (e.g. a family)
  Information, Table S1). Different authors have variably
                                                                                                              that is a member of that order. The amount of
  interpreted these distributions as representing
                                                                                                              morphospace occupied by that order is proportional to
  hyperbolic, logarithmic, log-normal, exponential,
                                                                                                              the number of families it contains, and we know from
  geometric and/or power law functions; Anderson (1974)
                                                                                                              the frequency distribution of family sizes that a few are
  provides an excellent review. Moreover, such ‘hollow
                                                                                                              large (comprising a significant fraction of the genera
  curves’ have been ascribed to both deterministic and
                                                                                                              or species within that order) and many are small.
  stochastic processes of biological diversification, and/
                                                                                                              A two-dimensional (2-D) plane through this visualized
  or historical artefacts of taxonomic classification (e.g.
                                                                                                              morphospace transects the volumes of the contained
  Walters, 1961; Reddingius, 1971; Chu & Adami 1999;
                                                                                                              families such that each is represented by an area on
  Scotland & Sanderson, 2004). An understanding of
                                                                                                              that plane. The diameters of those areas describe an
  membership frequencies of subtaxa within supertaxa
                                                                                                              exponential density function that is the same as that
  is critical to understanding differences in taxonomic
                                                                                                              which describes the variation in the sizes of taxonomic
  histories at different levels of Linnean classification

                                                          © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                              5

subgroups (like the families comprising an order)                         through reptile morphospace. The diameters of those
(Wilkinson, 2011). Because here we presuppose that                        areas exhibit an exponential distribution (many small,
taxonomic ‘size’ equates to numbers of subgroups, ‘size                   few large) (Fig. 3B).
diameter’ approximately corresponds to the square                           Given this relationship, we can define a membership
root of the numbers of members comprising that group                      inclusion parameter p, the probability that the addition
(e.g. genera or species included in a family).                            of 1 unit of ‘size’ (in this case a species) will result

                                                                                                                                                        Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
   Consider the 11242 species partitioned among the 92                    in its inclusion in a different family. This inclusion
families that comprise the class Reptilia. The frequency                  parameter is dependent only upon the number of
distribution of memberships (species) within families                     families (F) and species (S) in this class:
defines a curvilinear trend in log membership vs. log
exceedance space (Fig. 3A) that reflects the variation                                                         πF
                                                                                                      p=
in areas of families intersected by any 2-D transect                                                           2S

                                                  Families (12)
                                                  with 1 species

                                                                               10                  100               1000
                                                     100                                                                             100
                                Number of
                                                                                                            Families (F) = 92
                               Families (92)
                                                                A                                         Species (S) = 11,242
                                                                                                      Inclusion p = 0.113 /species

                                                                                                            p=      πF/2S

                                                                                  - p2 ES/F

                                                                                                                                           Exceedance
                                                                     ES/F = O e
                                                                         r2 = 0.983

 Number of                 100                                                                                                       10
Families (92)                           B
                                                             Total Diameters (D) = 788
                                                                   Families (F) = 92
                                                            Average Diameter = D/F = 16
                                                              Inclusion p = F/D = 0.116
                  Exceedance

                               10
                                                       Sl
                                                            op

                                                            1                                                                        1
                                                              e=

                                                                                                   100               1000
                                                                    -p

                                                                                                      Species/Family
                                                  -p ES/F
                                    ES/F D = O p e
                                           2
                                          r = 0.989                                                                Family with
                                                                                      Colubridae
                                1                                                                              Most Species (1,972)
                                          10         20             30        40              50                Colubrid snakes
                                                                                                                  (Colubridae)
                                                 ~     Species/Family

                                                                                      (2 x 1,972/π)
                                                                                          = 50.1

Figure 3. Sizes of families of reptiles as measured by numbers of species per family. A, number of species per family
among the 92 families of reptiles. B, frequencies of distances across 2-D transects of ‘shape-space’ represented by each of the
families. Functions describing numbers of species (A) and ‘diameters’ of numbers of (B) both suggest that lateral boundaries
of taxonomic units occur independently (randomly). In both cases, the inclusion parameter, p (~0.113), is the likelihood of
incrementing the number of families with the addition of one species.

© 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
6   B. H. WILKINSON ET AL.

The cumulative percentage of families with                       ASSESSING EFFORT – WHAT CONTROLS
memberships equal to or greater than some number                  RATES OF TAXONOMIC DESCRIPTION?
of species (s) is the exceedance (EFs), and is given as:
                               √                              Forecasts of biodiversity also depend on attaining some
                                  2
                      EFs = Fe− p Fs                          aggregate measure of the intensity of efforts exerted
                                                              by taxonomists in order to reveal said diversity. The
This function yields good agreement with the

                                                                                                                               Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                              issue of ‘taxonomic effort’ is important (e.g. Cribb &
distribution of species memberships within families of
                                                              Bray, 2011). If the collective sampling efforts needed to
reptiles (Fig. 3A), and suggests that the distribution of
                                                              discover one new taxon were invariant, then the rate
taxon sizes (measured as the square root of numbers
                                                              of description of new taxa should decrease with time
of subtaxa within that group) is the same as that of
                                                              because the pool of remaining (undescribed) taxa will
the diameters across morphospaces whose areas are
                                                              decrease as well; a classic case of rarefaction. The fact
exponentially distributed. Agreement between this
                                                              that rates of new descriptions at lower taxonomic levels
theoretical distribution and observed data on taxon
                                                              (genera and species) have increased during of much
sizes suggests that taxonomic memberships largely
                                                              of the last 200 years (Fig. 1; Supporting Information,
represent the random subdivision of the n-dimensional
                                                              Fig. S1) requires that the intensity of sampling – the
morphospace occupied by some larger taxon (e.g. a
                                                              amount of effort devoted to measuring biodiversity –
class) where intermediate taxon sizes (e.g. families) are
                                                              has increased as well.
represented by the numbers of contained subtaxa (e.g.
                                                                 How does one measure aggregate taxonomic
species). Memberships of taxa within a larger group
                                                              effort? Rates of discovery must directly or indirectly
are therefore dependent only on their number and the
                                                              correspond to the number of opportunities that arise to
total number of subtaxa of which they are comprised.
                                                              recognize a new taxon each year, where ‘opportunity’
  Observed taxonomic memberships of vertebrate
                                                              is the observed occurrence of a species in some place
groups at all levels of Linnaean classification are
                                                              by some individual; some subset of these observations
closely approximated when assuming such a stochastic
                                                              will be that of a new taxon. Data of this nature, on
pattern of division. When considering the six possible
                                                              occurrences regardless of novelty, are recorded
taxonomic subgroupings (species per genus, per family,
                                                              in some datasets (e.g. the Paleobiology Database,
per order; genera per family, per order; and families
                                                              paleobio.org), but not consistently for modern taxa of
per order) within the class Mammalia, for example, all
                                                              vertebrates. A more tractable measure of taxonomic
model memberships are in good agreement with the
                                                              effort is perhaps the numbers of biologists engaging in
data (Fig. 4). Data on group memberships for finfishes,
                                                              descriptive research, and the number of publications
amphibians, reptiles and birds exhibit similar
                                                              in which new taxa are described per year. A number
correspondence (Supporting Information, Fig. S3,
                                                              of studies suggest a relationship between rates of
Table S3). Agreement between a stochastic function
                                                              taxonomic description and the aggregate scientific
representing the chance division of morphospace and
                                                              exertions of systematists, who together examine
taxonomic membership data observed in the real
                                                              many individuals from many populations spanning
world, for these and other animal groups, suggests
                                                              many biogeographical regions in search of something
that current taxonomic classification largely serves
                                                              different enough to be considered a new taxon. Of
to randomly subdivide a morphospace continuum
                                                              particular interest has been the nature of relationships
(Wilkinson, 2011).
                                                              between rates of species description, and numbers of
  The relationship describing the distribution of
                                                              taxonomists and papers describing new species (e.g.
subtaxa among taxa in a larger group is consistent
                                                              Joppa et al., 2011a, b; Mora et al., 2011; Bebber et al,
across taxonomic levels; that is, p is similar between
                                                              2014; Costello et al., 2014; Gómez-Daglio & Dawson,
levels with similar amounts of taxonomic separation
                                                              2019). We explore these below with respect to data on
(TS; Table S3), and generally decreases with increasing
                                                              the major classes of vertebrates.
degrees of TS (Fig. 5) as:
                                                                 Rates of new species described and rates of taxonomic
                                                              papers published (in which a new species is described)
                     p = e−1.12 TS
                                                              exhibit similar patterns of temporal variation. For
The exponent of this decrease averaged across                 example, beginning with the description of 330
vertebrate classes, −1.12, is the natural logarithm           species of finfish by Linnaeus in 1758, 33 913 species
of 32.5%, which represents the average decrease               (Supporting Information, Table S1) have been described
in p with increasing separation among Linnaean                to the end of 2019 (254 years) in 6055 publications (3.0
levels (i.e., p = 0.325TS). Taxonomic memberships –           species per paper). Over this time interval, rates of
the partitioning of subtaxa among higher taxa – are           species description and rates of publication have both
largely the same, regardless of taxonomic levels of           increased more or less exponentially. Since 1950, the
consideration.                                                rate of naming of new finfish species has increased

                             © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                                       7

                                                                                    100
                                                                                                   A                       Species = 6,485
                                                                                                                             Orders = 27
                                                                                                                              p = 0.081
                                                                                                                              r2 = 0.930

                                                                              Exceedance
                                                                                           10

                                                                                                                                                                                                                  Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                    100

                                            Exceedance        B

                                                                                             1
                                                                                                             10        100                 1000
                                                         10                                                       Species per Order
        100
                      C                                           Genera = 1,331
                                                                   Orders = 27
                                                                    p = 0.179                                                                D                      Species = 6,485
                                                                    r2 = 0.914
Exceedance

                                                                                                                                                                     Families = 167
                                                          1                                                                                                            p = 0.201
             10                                                          10                            100                           100
                                                                                                                                                                       r2 = 0.966

                                                                                                                        Exceedance
                                                                        Genera per Order
                      Families = 167
                       Orders = 27                                                                 E                                  10
                        p = 0.504
                        r2 = 0.885
              1
                  1                    10                              100                   100

                            Families per Order
                                                                                Exceedance

                                                                                                                                                  10                100       1000
                                                                                                                                                  Species per Family
                                                                                              10   Genera = 1,331
                                                                                                   Families = 167                                                         F                 Species = 6,485
                                                                                                     p = 0.444
                                                                                                                                                                                            Genera = 1,331
                                                                                                     r2 = 0.906                                                1000
                                                                                                                                                                                               p = 0.568
                                                                                                                  10                   100             Exceedance                             r2 = 0.896
                                                                                                              Genera per Family
                                                                                                                                                                    100

                                                                                                                                                                     10

                                                                                                                                                                                      10        100
                                                                                                                                                                                     Species per Genus

 Figure 4. Random division functions fit to taxonomic data on mammals plotted as number of subtaxa among each
 hierarchically higher level of supertaxa (x-axes) relative to numbers of subtaxa that are equal to or are greater than
 some x-axis membership number (y-axes). Metrics for each membership curve are listed in Table S2. Similar plots for fish,
 amphibians, reptiles and birds as Figure S3.

 by ~2.1% per year to a current pace of ~400 species                                                                   define a trend of exponentially decreasing importance
 per year (Fig. 6A). Similarly, since 1950, the number                                                                 of monographs toward the present (Fig. 7B for
 of publications containing species descriptions has                                                                   amphibians; see Supporting Information, Fig. S5,
 increased by ~2.6% per year (Table S4) to a current                                                                   Table S5 for other classes). These changes are just one
 rate of ~250 papers per year (Fig. 6B). Data on rates of                                                              ramification of the increasing specialization of modern
 description and numbers of papers related to species                                                                  scientific inquiry.
 of amphibians, reptiles, birds and mammals exhibit                                                                      Trends in numbers of described species and
 nearly identical relationships (Fig. S4).                                                                             numbers of papers designating new taxa exhibit a
    A decrease in the variance of the rate of species                                                                  rather obvious change around the early 1950s (Fig.
 description is also apparent over time in all vertebrate                                                              6; Supporting Information, Fig. S4), when rates of
 taxonomies (Fig. 7), and most probably reflects the                                                                   description and publication began renewed and
 decreasing importance of monographs as an outlet                                                                      exponential increases that continue to the present.
 for the formal recognition of new groups (e.g. Joppa                                                                  Rate metrics for all vertebrate classes are similar
 et al., 2011a). Residuals of best-fit trends through data                                                             except for birds (an order of magnitude lower; Fig.
 on the numbers of species described per publication                                                                   S4), and equate to the naming of about 500, 160,

 © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
8                       B. H. WILKINSON ET AL.

                                                                                                        1
                                                                                                                        A                                                                                               species per class

                                                                                         Inclusion p
                                                                                                                                                                     Mammals                                            genera per class
                                                                                                                                                                                                                        species per order

                                                                                                                                                                                                                        families per class
                                                                                                       0.1

                                                                                                                                                                                                                                                 Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                                                                                                                                                                        genera per order
                                                                 1                                                                                                                                                      species per family
                                                                      B
                                                                                                                                               p = 1.25 exp-1.07TS                                                      orders per class
                                                  Inclusion p

                                                                                                                                                n = 9; r2 = 0.942                                                       families per order
                                                                                                  0.0 1                                                                                                                 genera per family
                                                                                                                                1                       2                          3                  4                 species per genus
                                                                0.1                                                                                              Taxonomic Separation
                                                                                                                                                                                       1
                           1                                                                                                                     Birds                                     D
                                C
Inclusion parameter (p)

                                                                                                                                                                     Inclusion p
                                                                              p = 1.25 exp-1.14TS
                                                                               n = 9; r2 = 0.926

                                                           0.0 1                                                                                                                                                         Amphibians
                                                                      1                 2                                              3                    4                      0.1
                          0.1                                                                                  Taxonomic Separation
                                                                              Reptiles                                                     1
                                                                                                                                                E
                                                                                                             Inclusion parameter (p)

                                    p = 1.25 exp-1.21TS                                                                                                                                            p = 1.25exp-1.16TS
                                     n = 9; r2 = 0.978                                                                                                                                              n = 9; r2 = 0.944

                      0.0 1                                                                                                                                                                1                 2                3              4
                                1                2                        3                 4                                                                                                                             Taxonomic Separation
                                                                                                                                                                                                     Fishes
                                                                                                                                       0.1
                                         Taxonomic Separation
                                                                                                                                                      p = 1.25 exp-1.04TS
                                                                                                                                                       n = 10; r2 = 0.996

                                                                                                                                  0.0 1
                                                                                                                                                  1                  2                         3                 4
                                                                                                                                                            Taxonomic Separation

   Figure 5. Values of the inclusion parameter p as a function of degree of taxonomic separation among classes of vertebrates.
   Here, the slope (on average, −1.12) is the natural logarithm of 0.325, and represents the rate of decrease in p for each
   increase in level of separation (0 = 100%; 1 = 32.5%, 2 = 10.6%, 3 = 3.4%, 4 = 1.1%). This rate of change corresponds to about
   a three-fold increase in membership with each increase in Linnaean level of taxonomy.

   150, five and 40 new species of fishes, amphibians,                                                                                                papers began to increase (by ~10% per year) around
   reptiles, birds and mammals, respectively, in the year                                                                                             1950 (Supporting Information, Fig. S6). Papers with
   2020. Interestingly, current rates of discovery and                                                                                                three or more authors account for ~65% of species
   description are some several hundred times higher                                                                                                  descriptions since 2000. With more authors overall, the
   than estimated coeval rates of extinction of vertebrate                                                                                            number of papers published has grown (Fig. 8A), and
   species (Ceballos et al., 2017, 2020).                                                                                                             papers published since 1950 are also more likely to
      Such a profound change – one of both sign (negative                                                                                             describe fewer (or one) new species per paper (Fig. 8B).
   to positive) and acceleration – in species identification                                                                                          The shift to multi-authored and monospecific papers
   and publication rates seen across all vertebrate classes                                                                                           alone, however, cannot account for the increase in both
   since c. 1950 argues for a common cause (e.g. Joppa                                                                                                papers and species descriptions per year. An increase
   et al., 2011b). Publication norms have changed since                                                                                               in the latter must be associated with increasing rates
   1950 (Ioannidis et al., 2018; Gómez-Daglio & Dawson,                                                                                               of discovery of new taxa.
   2019), particularly with respect to the proportion of                                                                                                 Several factors in particular probably account
   multi-authored papers per year; might this contribute                                                                                              for the 1950s’ inflection in the rate at which new
   to the inflection? Single and dual authored papers                                                                                                 species are described. The establishment of national
   account for ~85% of all new species descriptions over                                                                                              agencies dedicated to the public support of university-
   the past 250 years, but the number of multi-authored                                                                                               based research, including taxonomic studies, surely

                                                                      © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                                       9

                                     A
                               500                                   S = 8.42 x 10-17 e-0.0214D
                                                                        n = 68, r2 = 0.802
            Species per year
                               400

                                                                                                                                                                           Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                            Mode = 1900
                               300
                                         n = 185, r2 = 0.605
                                                                                                          B
                                                                                                                                                250
                               200
                                                                                                          S = 3.80 x 10-21   e-0.0261D
                                                                                                             n = 68, r2 = 0.907

                                                                                                                                                     Papers per year
                                                                                                                                                200
                               100                             200

                                                                                                                                                150

                                         1800       1850        1900          1950          2000
                                                                                                                                                100

                                                                                       Mode = 1950
                                                                                    n = 185, r2 = 0.810                                         50

                                                                               1800          1850         1900       1950                2000

Figure 6. Relationships between naming of 33 913 new species of finfishes (A) and rates of their publication as papers per
year by unique authors of sets of authors (B). Brown curves are best-fit skewed normal distributions to data before 1950;
yellow curves are best-fit exponentials. Low numbers for 2018 and 2019 in both plots (lighter points) represent incomplete
tabulations. Since 1950, rates of newly named species and associated papers have increased at about the same rate: by ~2.14
and 2.62% per year, respectively. Similar plots for amphibians, reptiles, birds and mammals as Figure S4A–H.

contributes to the renewed increase in description                                      with which scholars in the developing world can share
rate. In the United States, for example, the National                                   their work with the broader international community
Science Foundation was created in 1950. In Canada,                                      (e.g. Grieneisen et al., 2012).
while the National Research Council was created                                            Finally, the advent of molecular techniques with
in 1916, war-related and medical research were                                          which to recognize and distinguish among genetically
handed off to newly formed organizations around                                         distinct populations must contribute to the renewed
1950, and research funding in the natural sciences                                      increase in the rate of description of new species.
has been handled through the Natural Sciences and                                       Data in Bouchet et al. (2016), for example, suggest
Engineering Research Council since 1978. Federal                                        that the description of marine molluscs based on
funding for scientific endeavour promoted growth in                                     molecular criteria has increased by ~35% per year
the number of taxonomic researchers and hence in the                                    over the past decade. In addition, the recognition
number of papers and new species described per year.                                    of cryptic species has resulted in a subdivision of
   Access to new habitats afforded by technological                                     existing taxa into multiple new taxa based on genetic
innovation has also allowed researchers to explore                                      information (e.g. Bickford et al., 2007; Pfenninger &
biodiversity in places that had been difficult or                                       Schwenk, 2007; Ladner & Palumbi, 2012); molecular
impossible to reach before (Donoghue & Alverson,                                        methods may add tens of thousands of cryptic marine
2000). As an example, the ratio of marine to terrestrial                                species (Appeltans et al., 2012). Genetically distinct
species described each year began to increase at about                                  populations recognized using molecular data, however,
this time (Costello et al., 2012), reflecting a progressive                             are not always the same as the species that would
expansion in the exploration of marine habitats,                                        have originally been defined morphologically, so the
including deep-sea settings. Similarly, the increasing                                  inclusion of both types of ‘species’ in the same dataset
globalization of science has allowed more attention to                                  largely equates to the development of new reservoirs
formerly difficult-to-access or otherwise understudied                                  that were not considered during earlier research.
biogeographical regions through the growing number                                         An intriguing correlate to the naming of vertebrate
of international collaborations and the increasing ease                                 species can be found in the nearly identical changes

© 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
10    B. H. WILKINSON ET AL.

                                                                                        100                                                            100
                                                                                               A

                                                                    Species per paper
                                                                                                                             Mode = 1892

                                                                                                                                                             Species per paper
                                                                                                                          n = 221, r2 = 0.210

                                                                                                                                                                                 Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                                        10                                                             10

                                             B
         Residual species per paper

                                       10                                                                                                              1
                                                                                                                         1900       1950        2000

                                        1

                                       0.1

                                      0.01       Res = 4.81 x 108 exp-0.0109Date
                                                       n = 221; r2 = 0.254

                                                  1800       1850                       1900       1950     2000

Figure 7. A, historical (x-axis) change in rate of description of new species of amphibians (y-axis, log scale). Best-fit skewed
normal distribution (brown line) to log rates (brown dots) suggests a maximum taxonomic ‘per paper productivity’ c. 1890.
Note decreasing scatter of rates with decreasing age. B, exponential decrease in the ‘monograph effect’ among amphibians
manifested as decreasing differences between observed rates of description of new species and that described by a longer-
term average. ‘Spikiness’ in the description of new species has decreased by ~1.1% per year since the end of the 17th century.
Similar data for fishes, reptiles, bids and mammals as Figure S5.

in the rate of discovery of new minerals and the                                                          proxies for ‘taxonomic effort’? Probably not. Linnaean
proportion of multi-authored papers describing new                                                        species are the coin of the biodiversity realm; the first
mineral discoveries that occurred at the same time.                                                       description of any vertebrate species necessitates the
Barton (2019) reports that the rate of discovery of new                                                   identification of some genus, some family, and some
minerals was relatively constant (~15 per year) from                                                      order within the class to which it belongs. Numbers
1917 to about 1950, but then increased exponentially                                                      of species descriptions, published papers and/or
(by ~1.9% per year) to a current (2020) rate of ~100 per                                                  contributing authors are all measures of the success
year. Over the same time interval, the average number                                                     of taxonomic efforts, rather than a direct measure
of authors on mineral discovery papers also underwent                                                     of the aggregate scientific exertion expended by the
an exponential (~2% per year) increase from abour two                                                     community in deriving this classification. The difficulty
in 1950 to over six in 2020. She ascribes these changes                                                   in assessing taxonomic effort is that there is no easily
to a variety of interrelated factors, including the greater                                               accessible record of ‘unsuccessful’ research; there is no
availability of instrumentation and an exponential                                                        adequate metric for measuring the cost and/or effort
growth in the funding and focus of mineralogical research                                                 expended during those expeditions that failed to locate
at universities and museums. The striking coherence                                                       and/or identify new forms.
in pattern between such disparate fields argues for                                                         Even if a satisfactory metric were available with
similar drivers, and the combination of federal research                                                  which to track the history of taxonomic effort, several
funding, globalization and technological advance would                                                    factors would complicate a straightforward forecast
affect both in the same way.                                                                              of future industry and its impact on probabilities
   All this being said, do numbers of authors or                                                          of discovering new forms. First, there are a number
publications or described taxa per year serve as valid                                                    of factors intrinsic to specific taxa that predicate

                                                         © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                                11

         Number of Papers
                                  A

                            100

                                                                                                                                                  Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                   2018          B       P = 0.288 S1.11
                                                        2019                               r2 = 0.935
                             10

                                                      P = 1.19 A0.877

                                                                                                                          Number of Papers
                                                    n = 254, r2 = 0.991                                             100

                                              10             100
                                             Number of Authors
                            2000-2025
                            1975-2000
                            1950-1975
                            1925-1950   10                                                                          10
                            1900-1925
                            1875-1900
                            1850-1875
                            1825-1850                                                          P = 0.671 S0.655
                            1800-1825                                                         n = 185, r2 = 0.684
                            1775-1800
                            1750-1775
                                                                   10                   100
                                                                   Number of Species
Figure 8. Relationships between numbers of papers and participating authors describing new vertebrate species (A), and
numbers of species described (B) since 1758. A, number of authors and resulting papers are well correlated, an increasing
proportion of multi-authorship giving rise to a slight decrease in slope. The two youngest points (2018 and 2019) are thought
to represent incomplete database tabulations. B, correlations between numbers of papers and described species are also
increasingly well correlated, with greater noise during earlier years reflecting a greater impact of monographs.

probabilities of discovery. These include things                          up only about 1.4% of the Earth’s land surface.
related to appearance such as flamboyance of colour,                      Sociopolitical infrastructures are also frequently
openness of biome and mode of mobility, as well as                        limited in high-biodiversity regions, and scientists from
abundance of members (e.g. Cribb & Bray, 2011),                           countries with more robust taxonomic infrastructures
range size (e.g. Collen et al., 2004; Krasnov et al.,                     typically require significant resources to explore in
2005) and body size (e.g. Gaston, 1991). Second, the                      other areas (e.g. Grieneisen et al., 2012).
spatial distributions of new (undescribed) taxa are                          The concatenation of all these and other factors, as
geographically heterogeneous, and in order to attain                      illustrated so dramatically by the 1950s’ inflection,
taxonomic ‘success’ through their description, even                       collectively makes projections about the success rate of
relatively abundant species will not be found until                       future taxonomic efforts based on past trajectory highly
exploration expands into their range; the discovery of                    uncertain and subject to happenstance. If, as noted by
hydrothermal vent communities in the late 1970s is                        Pimm & Joppa (2015), we can only achieve an accurate
a classic example. Geographical ranges and densities                      estimate of the size of the global biodiversity reservoir
of taxa are also often variable and uncorrelated.                         when we can also accurately forecast numbers,
More recently discovered species, for example, have                       practices and collective efforts of systematists, we are
typically been found in biodiversity hotspots of high                     left in a precarious position. The equation has one
endemism (e.g. Thompson et al., 2021); Myers et al.                       known – the success rate of past collective efforts –
(2000) estimated that 35% of all recently described                       and two unknowns – the actual amount of biological
amphibian, reptile, bird and mammal species are                           diversity that exists on the Earth’s surface, and the
restricted to some 25 hotspots that collectively make                     total effort necessary to effect a census of that diversity.

© 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
12    B. H. WILKINSON ET AL.

A better understanding of how different scenarios of            each of which represents some intermediate level of
future effort might affect estimates of total diversity is      Linnaean taxonomy, like families. The population is
afforded from simple numerical models                           structured such that that most of the marbles (e.g.
                                                                species) represent only a few colours (e.g. families),
                                                                while many other colours are represented by a few
                                                                or only one marble(s). The distribution of marbles
     MODELLING TEMPOS OF TAXONOMIES

                                                                                                                                 Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                (subtaxa) among colours (taxa) is determined by the
Thus far, we have proposed that: (1) the tempo at               stochastic function defining memberships for a given
which new Linnaean categories have been established             taxon provided by the inclusion parameter p (Section
is similar for equivalent ranks across different                IV, Figs 4, 5). The population is randomly sampled
vertebrate classes; (2) rates of definition at the family       with replacement over time, with each draw reflecting
level and higher have deceased rapidly since 1758,              the taxonomic observation of a (new or previously
when a large number of groups were first erected;               recognized) species. Because the same marble (species)
(3) rates of generic description increased until                can be drawn more than once, the number of marbles
c. 1870 before falling off; (4) rates of description of         drawn per time step reflects some measure of effort,
new species exhibit a similar peak in the mid-to-late           as some observations turn out to be taxa already
1800s (1889 ± 20), declined until about 1950, and have          described. The time step in which new marbles
exponentially increased since then, now by ~2.5% per            (subtaxa) and new colours (taxa) are first sampled is
year; (5) the most speciose orders, families and genera         the recorded year of discovery.
are historically the earliest ones defined; (6) Linnaean          If sampling intensity – the number of marbles
group memberships measured as numbers of contained              drawn in each time step – were constant over time,
subtaxa exhibit frequency distributions similar to              the number of new taxa discovered at any level per
those expected from the random division of taxonomic            time step (per year) would have to decline from the
morphospace; and (7) numbers of described taxa,                 inception, as speciose clades are discovered early and
numbers of publications and numbers of contributing             then repeatedly resampled, and progressively less
authors are all correlated manifestations of taxonomic          speciose groups are eventually encountered by chance
successes. Although their rates of appearance are               over greater intervals of time. This simple scenario,
directly related to scientific effort expended, no readily      however, is inconsistent with the observation of modes
available data exist by which to adequately quantify            in the number of new descriptions per year, which also
increases in total taxonomic effort.                            occur later in time at progressively lower taxonomic
   To better understand the processes underlying these          levels (e.g. Fig. 1). These observed trends require
observations, and hence to evaluate the potential for           that sampling intensity (taxonomic effort) must be
predicting them in the future, we construct several             increasing over time.
numerical models that incorporate parameters                      To more accurately capture the observed structure
reflecting both the inherent nature of the groups being         in taxonomic histories, we specify various changes in
sampled and described (numbers and memberships) as              the rate of sampling (the number of marbles drawn,
well as some model estimate of the effort expended in           or taxonomic observations made) over time. The first
the process by systematists over time. As demonstrated          model assumes a constant rate of increase in sampling
in the prior section, quantifying the effort involved           effort and considers how membership histories unfold
in discovering a new taxon is difficult, but one can            as a function of taxonomic rank. A second set of models
bring some clarity to the impact of different degrees of        explores the impact of each of three different rates of
effort on resulting biodiversity trajectories and tallies       increase in sampling effort on the shapes of resulting
through the use of such simple numerical models. To             membership histories at a single taxonomic rank.
this end, we combine the understanding of taxonomic             Both of these models span 350 years (1758–2108) of
memberships at different Linnaean levels with three             taxonomic history.
prescribed scenarios of taxonomic effort in order to:             The first scenario (Fig. 9) is parameterized so as to
(1) replicate taxonomic tempos at different Linnaean            loosely approximate the taxonomic histories of a clade
levels; (2) assess the impact of differing effort scenarios     of vertebrates (e.g. mammals, with 6485 described
on taxonomic history at a single Linnaean level; and            species divided among 1331 genera and 176 families;
(3) examine how changes in taxonomic effort affect the          Supporting Information, Table S1). A pool of taxonomic
tempo of taxonomic description.                                 observations comprising 6500 species is randomly
   Conceptually, we might imagine a very large                  sampled with replacement each year for 350 years,
population of marbles that represents some high                 for a total of 30 000 observations. We presuppose that
taxonomic level (like a class) of vertebrates. Marbles          sampling effort increases exponentially by ~1% per
each represent some unit at a lower taxonomic level,            year, from a rate of six observations made per year
like species, and marbles comprise different colours,           in 1758 to 154 per year in 2108 (Fig. 9A). Drawn

                               © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                               13

                                                                                             A
                                                                                                                       Samples

                                                              Taxonomic Observations
                                                                                       150

                                                                                               Samples in 1759, 2020, 2108 = 6, 63, 154
                                                                                                 Sampling increase = 1.0% per year

                                                                                                                                                                         Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                                                       100

              B          Families                                                       50
          6

          5
Famlies

                           Mode = 1759
          4                                                                                      1800         1900         2000           2100
                         n = 73, r2 = 0.514
          3

          2
                                                   C                                         Genera
                                                                                                                           5

          1
                                              10                                                 Mode = 1860
                                                                                              n = 263, r2 = 0.477
                  1800       1900
                                     Genera

                                                                                                                       D       Species
                                              5
                                                                                                                                                          30

                                                                                                                                                               Species
                                                                                                                                                          20
                                                       1800              1900                    2000           2100

                                                                                                                                                          10
                                                                                                                                    Mode = 2014
                                                                                                                                 n = 350, r2 = 0.848

                                                                                                         1800          1900          2000          2100

Figure 9. Modelling differences in taxonomic histories as a function of Linnaean level of classification employing the
same sampling regime. A, a scenario of taxonomic effort modelled as an exponential change in per-year ‘attempts’ at taxon
description, increasing from six per year in 1758 to 154 per year in 2108. B–D, rates of ‘discovery’ of new taxa as a function
of their memberships at different taxonomic levels.

observations constituting new species are apportioned                                        only 76% (4965) of species, are discovered by the year
into genera and families based on the inclusion                                              2020. By model year 2108, a time span of 350 years,
parameters (p) for modern mammals (Table S3). We                                             98% (6393) of the prescribed 6500 species have been
plot the first model year in which each species, genus                                       discovered. The lower the Linnaean taxonomic rank,
and family is first ‘discovered’, yielding histories of                                      the later the date of ‘peak taxonomy’.
discovery over 350 years at each of those taxonomic                                             The second model compares the effect of varying
levels (Fig. 9B–D).                                                                          the rate of increase in sampling on the recovered
  Such modelled taxonomic histories are qualitatively                                        taxonomic history at a single taxonomic level, in this
similar to those observed among all vertebrate groups:                                       case genera. A total of 15 000 taxonomic observations
higher levels (e.g. families) are completely censused                                        are drawn with replacement from a pool comprising
within a few decades, while rates of ‘discovery’ at lower                                    6500 species, and the membership inclusion parameter
levels (e.g. genera and species) initially increase because                                  (p; 0.57 per species) comes from modern mammals.
rates of sampling exceed rates of pool depletion, reach                                      We calculate the number of new genera discovered/
a peak rate of discovery (in model year 1860 for genera,                                     described each year given each of three different rates
year 2014 for species), and then decline as rates of pool                                    of increase of sampling intensity (Fig. 10). The results
depletion overtake rates of sampling. In this example,                                       demonstrate that taxonomic effort has a first-order
100% (170) of families and 99% (1355) of genera, but                                         influence on resultant taxonomic histories for a given

© 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
14        B. H. WILKINSON ET AL.

                                                                     A
                                                               400

                                                     Samples
                                                                                Samples          Annual
                                                               300       1759    2020    2108   increase
                                                                          22     58      90       0.5%

                                                                                                                                                     Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                               200
                                                                           6     63      153      1.0%
                                                                           1     31      443      3.5%
              B
                                                               100
         20              Mode = 1759
                         n = 211, r2 = 0.786
Genera

         15
                                                                         1800           1900       2000           2100
         10
                                                     C
                                                                                   Mode = 1860
          5                                                                        n = 263, r2 = 0.477
                                                10

                  1800   1900          2000

                                                 5                                                          D
                                       Genera

                                                                                                                                       15

                                                                                                                                            Genera
                                                                                                                                       10
                                                         1800            1900           2000         2100

                                                                                                 Mode = 2018
                                                                                                 n = 318, r2 = 0.887                   5

                                                                                                  1800          1900     2000   2100

Figure 10. Differences in simulated taxon discovery histories as a function of sampling intensity of subtaxa. A, three
different scenarios of taxonomic effort modelled as exponential increases in number of subtaxon descriptions/year resulting
from sampling a pool of 6500 subtaxa. B–D, rates of ‘discovery’ of new taxa as a function of the three subtaxon sampling
scenarios in A.

hierarchical level. Low rates of increase in sampling                           must reflect synchronous and significant changes
intensity (i.e. closer to the initial scenario of constant                      in the amount of taxonomic effort. We examine
sampling intensity over time) result in progressive,                            this supposition with a model of taxonomic history
approximately exponential depletion of the pool of                              similar to that described above but now imposing
as-yet undescribed genera, while higher rates of                                two changes in sampling effort over the 350 model
increase in taxonomic effort yield progressively more                           years. Both cases comprise a net taxonomic ‘effort’
bell-shaped ‘discovery’ histories. The model year of                            of 30 000 observations. In the first instance (Fig.
peak rate of discovery marks the time that the impact                           11A, B), the pool contains 6500 species. We make
of increasing sampling intensity is overtaken by the                            observations at an initial rate of four per year in
influence of depletion of the pool of undiscovered genera.                      1758 increasing to 35 per year in 1840, holding
With higher rates of increase in taxonomic effort, ‘peak                        steady at that rate until 1950, and then increasing
taxonomy’ occurs at progressively later dates.                                  up to ~280 per year in 2108 (Fig. 11A). This scenario
  Finally, recall that the rate of description of new                           results in the ‘discovery’ of ~5000 species (~76%) by
species in nearly all vertebrate groups experienced                             2020 and ~6400 (~98%, nearly all) by 2108 (Fig. 11B).
a renewed and exponential rise following c. 1950                                In the second instance, we again make 30 000 total
(Fig. 1D; Supporting Information, S1). Because                                  observations, but now drawn from a pool containing
(presumably) finite pools of vertebrate species are                             20 000 species. We make observations at an initial
increasingly depleted with each new discovery,                                  rate of ten per year in 1758 increasing to 40 per year
changes in sign of the slope of rates of description                            in 1840, then decreasing in rate to one per year in

                                   © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
TEMPO OF VERTEBRATE TAXONOMY                            15

          300
                A                                                               C
                                30,000 samples                                                 30,000 samples

                                                                                                                           r
                                                              r
Samples

                                                                                                                       ea
                                                           ea
          200

                                                        /y

                                                                                                                    5/y

                                                                                                                                               Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
                                                     15

                                                                                                                   0.1
                                                   0.
          100
                           ar       0.00/year                                              ar -0.015/y
                0.03/ye                                                         0.02/ye                e    ar

          120
                B               6,500 species
                                                                                D            20,000 species

          100
Species

                                                      Mode = 2022
           80
                                                   n = 148, r2 = 0.804                  Mode = 1834
                       Mode = 1862
                    n = 202, r2 = 0.427                                              n = 202, r2 = 0.764
           60

           40
                                                                                                                       Mode = 2051
           20                                                                                                       n = 148, r2 = 0.950

                    1800    1850     1900   1950    2000     2050        2100       1800    1850     1900   1950    2000    2050     2100

Figure 11. Similarities and differences in 350-year taxonomic histories resulting from slight differences in taxonomic
effort. A, hypothetical tempo of effort as two changes in rate of sampling (+0.03/year to 1840; 0.00/year to 1965; +0.15/year
to 2107). B, rates of recognition when drawing samples from a pool of 6500 hypothetical species; 76% are sampled by 2020,
98% by 2107. C, hypothetical effort as two changes in rates of sampling (+0.02/year to 1840; −0.015/year to 1965; +0.15/year
to 2107). D, rates of recognition of when drawing samples from a pool of 20 000 hypothetical species; 35% are sampled by
2020, 77% by 2107. Although the latter scenario (C and D) represents about three times the biodiversity, rates of recognition
before ~2020 (brown circles and lines in B and D) are nearly the same.

1950, and then increasing to ~350 per year in 2108                           TAXONOMY AND HUBBERT’S PEAK –
(Fig. 11C). This scenario results in the ‘discovery’                      PARALLELS WITH RESOURCE PRODUCTION
of ~7000 species (~35%) by 2020 and ~16 000 (only
                                                                          The histories of discovery and description of new
~78%) by 2108 (Fig. 11D).
                                                                          taxa embody philosophic and practical similarities to
  The rate of species ‘discovery’ in either case
                                                                          processes involved in the discovery and exploitation
comprises two cycles in which rates of sampling first
                                                                          of natural resources such as petroleum or mineral
exceed rates of pool depletion, then are overtaken
                                                                          deposits. In both instances, rate of recovery is
by it. More importantly, despite the fact that total
                                                                          dependent on the inherently finite sizes of reservoirs
numbers of species in the latter case is about three
                                                                          to be exploited and the effort devoted to exploitation.
times that in the former, histories of taxonomic
                                                                          The production of a barrel of petroleum, extraction
discovery are very similar up to ~2020, and
                                                                          of a tonne of ore and the description of a new taxon
comparable to those actually observed for classes
                                                                          are all exercises in the successful discovery/sampling
of vertebrates. That small differences in taxonomic
                                                                          of some resource reservoir that is being depleted,
effort can yield a high degree of similarity in
                                                                          and histories of taxonomic description and resource
realized histories of taxonomy, yet might represent
                                                                          exploitation therefore share a number of similarities.
grossly different numbers of total extant species,
                                                                          Memberships of taxonomic units, volumes of petroleum
casts doubt on our ability to accurately estimate
                                                                          reservoirs and tonnages of ore deposits exhibit similar
total standing biodiversity on the basis of currently
                                                                          ‘many small – few large’ size frequency distributions
known tabulations.

© 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
16                             B. H. WILKINSON ET AL.

  (e.g. Laherrere, 2000). A greater likelihood of sampling                                                   abrupt increase in effort at exploration, the application
  from the larger entities results in their earlier dates of                                                 of new technologies and the subsequent ‘discovery’ of
  discovery in comparison to small/monospecific entities.                                                    new large reservoirs that were not available for earlier
     Exponential decreases in the rate of description                                                        exploitation. In the case of petroleum production,
  of higher Linnaean levels over time reflect the early                                                      growth since c. 2010 has been almost entirely due to the
  depletion of resource pools; any subsequent increase in                                                    increasing exploitation of ‘unconventional oil’ from coal,

                                                                                                                                                                                                    Downloaded from https://academic.oup.com/biolinnean/advance-article/doi/10.1093/biolinnean/blab080/6309407 by guest on 02 September 2021
  discovery or production requires substantial renewed                                                       tar sands, biofuels and primarily shale, made possible
  effort. In each case, rates of exploitation/discovery                                                      by the technological advances of horizontal drilling
  have been a manifestation of the ‘successful sampling’                                                     and hydraulic fracturing, as opposed to ‘conventional
  of these reservoirs; increases occurred during intervals                                                   oil’ produced by drilling a vertical hole in the ground.
  when rates of sampling exceeded rates of reservoir                                                         The striking similarities in these otherwise unrelated
  depletion, while times of decrease represent intervals                                                     histories of reservoir exploration suggest that lessons
  when rates of reservoir depletion exceed effort at                                                         learned from one will apply to the other. In both cases,
  exploitation. Following the first oil well drilled near                                                    attempted forecasts for when the reservoir would be
  Titusville, Pennsylvania, in 1859, ever-increasing                                                         fully explored, and hence how big it is in total, based
  efforts at exploitation led to a more-or-less exponential                                                  on the documented histories of sampling would have
  increase in production until the mid-1970s, when                                                           been very different prior to the renewed sampling
  production peaked, followed by a subsequent drop-off                                                       enabled by the – largely unpredictable – respective
  (Fig. 12A), much like the history of species descriptions.                                                 developments in each field. We consider such efforts in
  ‘Hubbert’s peaks’ in both records demark the apices of                                                     more detail below.
  exploitation (Hubbert, 1956).
     Because resource exploitation curves can be
  described mathematically, one is then tempted to
                                                                                                                            TAXONOMIC TEMPO AND THE LIMITS OF
  extrapolate them out to their presumed time of
                                                                                                                                 MEASURING BIODIVERSITY
  depletion in the future, thereby forecasting the
  amount remaining in the reservoir at any given time                                                        Earth’s biodiversity is currently decreasing at rates
  and the total cumulative reservoir size. All else being                                                    that many have compared to the mass extinctions
  equal, this is not unreasonable. However, following                                                        in the geological past (e.g. Pimm et al., 1995). Our
  decades of decline, both histories exhibit an abrupt                                                       collective understanding of the scope and magnitude
  change in slope: oil production has increased rapidly                                                      of depletion would therefore be improved by the
  since ~2010 and description of vertebrate species has                                                      knowledge of how many species actually exist within
  increased rapidly since ~1950 (Fig. 12). In both cases,                                                    groups (e.g. Brito, 2010; Braje & Erlandson, 2013).
  the central reasons for the turnaround include an                                                          The phrase ‘how many species’ appears in the titles

                                  A                                                                                                B
New Barrels per year (106)

                                                                                                 New species per year (%)

                                                                                                                            1.0%
                             4
                                                                                                                                          Mode = 1878
                                                                                                                                       n = 190, r2 = 0.569
                                                                                                                                                               N = 2.08 x 10-3 exp-0.0227A
                                                                                                                            0.8%
                                                                                                                                                                    n = 69; r2 = 0.896
                             3           Mode = 1974
                                      n = 150, r2 = 0.992
                                                                                                                            0.6%

                             2
                                                                                                                            0.4%

                             1                                      B = 1.63 x 106 exp-0.0823A
                                                                         n = 12; r2 = 0.959                                 0.2%

                                      1880   1900     1920   1940     1960    1980    2000                                                1800       1840    1880     1920     1960          2000

  Figure 12. Rates of production of crude oil and naming of vertebrate organisms. A, US field production of crude oil from
  1858 to 2019. Skewed normal distribution to interval up to 2007 peaks at c. 1974; exponential distribution between 2007
  and 2019 reflects an increase of 8.2% per year. B, proportional naming of all new species of vertebrate organisms between
  1760 and 2020. Skewed normal distribution between 1760 and 1950 peaks at c. 1878; exponential distribution between 1950
  and 2020 defined by an annual increase of 2.23%. Abrupt changes in slope in both oil and taxonomic curves in part reflect
  changes in the nature of resource reservoirs being sampled.

                                                             © 2021 The Linnean Society of London, Biological Journal of the Linnean Society, 2021, XX, 1–21
You can also read