Computers in Biology

Trends in taxonomy revealed by

the published literature
rominent biologists have determining the numbers of speci- allows online searches of citations

P stressed t?e u.rgent need .for

systematIcs In conserVIng
biodiversity (e.g., Wilson 1989, May
mens in collections, the availability
of jobs, or the amount of funding for
museum research (e.g., Stuessy and
and abstracts to the biological litera-
ture in approximately 7500 journals
and monographs, covering the years

Downloaded from https://academic.oup.com/bioscience/article/48/2/125/243025 by guest on 04 January 2022

1990, Anonymous 1991), and wirh Thompson 1981, Clarke 1990). An- 1969 to the present. OUf searches
good reason. According to recent other possible approach, measuring ended with 1996 because indexing
estimates, between 69% and 96% of the numbers of currently active tax- of the 1997year was not complete as
extant species remain undescribed onomists, is complicated by the fact of our search date (21 October 1997).
(Lean et al. 1990, Systematics Agenda that many active taxonomists are On the DIALOG system, truncation
2000 1994). If knowledge of Earth's not employed in their specialty. They of search terms is indicated by a
biota is so incomplete, it follows that may be officially retired or employed question mark, and the symbol (w)
even the most basic kind of system- full- or part-time at other jobs (Win- indicates that two words must be
atics research, taxonomy-the inven- ston 1988). adjacent. Free text searching was used
tory, description, and classification In this article, we present a differ- because it provided the most accu-
of organisms-is also far from fin- ent way to analyze the health of the rate results, calling up all articles in
ished. However, although cladistics field. This approach uses the pub- which the search term appeared in
and molecular phylogenetic tech- lished scientific literature itself, re- the title, abstract, key words, or the
niques have attracted attention to trievable as citations, keywords, ab- "concept codes" tagged by the data-
systematics and awakened new in- stracts, and other forms from a base producer. By contrast, we found
terest among students, reports and variety of electronic databases (e.g., that limiting our search to the data-
surveys on the state of the discipline Aquatic Sciences and Fisheries Ab- base producer's assigned concept
have all cited problems such as dwin- stracts, BIOSIS Previews, Disserta- codes severely limited our data col-
dling resources, an aging population tion Abstracts Online, Enviroline, lection and missed pertinent cita-
of professional systematists, and low Environmental Bibliography, Oce- tions. BIOSIS differs from other da-
recruitment due to job scarcity (Steere anic Abstracts, Pollution Abstracts, tabases because it uses abstracts as
et a1. 1971, Stuessy and Thompson SciSearch, and Zoological Record printed in journal articles, creating
1981, Winston 1988, SA2000 1994). Online), whose records extend back its own abstracts only for book ar-
In Great Britain alone, 25-30% of over the last 20-30 years. Bio- ticles that lack them.
evolutionary biology positions are diversity-related information re- Figure 1a shows the growth of the
estimated to have disappeared be- trieved from such databases effec- biological literature between 1969
tween 1980 and 1990 (Clark 1990); tively supplements and complements and 1996. During that time, period
during the same period, the number the information provided by collec- yearly totals of biological articles
of theses in systematic biology de- tions databases. We used electronic indexed increased from 204,326 to
clined drastically relative to the num- databases to compile information 546,496. Most of this growth re-
ber in other fields of biology, such as that could help to answer the follow- flects an increase in the number of
biochemistry (Claridge and Ingrouille ing questions about the status of sys- biological journals, part of the "in-
1992). Thus, even if taxonomic data tematics: Is the field growing or de- formation explosion" predicted by
collection is still far from complete, clining? Is the basic job of taxonomy librarians in the early 1960s. The
it may be that science is finished with finished-that is, is the rate of descrip- growth rate appears to be leveling
taxonomy-that it is no longer a tion of new taxa slowing down, or is it off over the last few years, but
viable field. still increasing? Is species description, whether this pause is temporary or
in particular, still an important part of indicates a real change (e.g., as a
Assessing the state taxonomy? How much of taxonomists' result of factors such as negative
efforts is really spent on nomencla- global economic climate, declines in
of taxonomy tural matters? What trends in the field research funding, the breakup of the
How can the current state of tax- can be identified overthe last28 years? scientific apparatus of the former
onomy be assessed? One way is by Soviet Union, or even the beginning
Is taxonomy in decline? of a switch to electronic publication)
remains to be determined.
by Judith E. Winston We used the BIOSIS Previews data- Between 1969 and 1988, the num-
and Kristen L. Metzger base on the DIALOG system, which ber of articles published in taxonomy

February 1998 125

Figure 1. Biological a Are taxonomists too fnssy?
and taxonomic ar-
ticles, 1969-1996. (a) Electronic searches can provide use-
Growth of the hiologi- ful data on the nomenclatural as-
cal literature between pects of taxonomy. The last 25 years
1969 and 1996. (b) have also seen the prospect of in-
Growth of the taxo-
nomic literature No.
...... creasing support for revolutionary
01 changes in the present system of bio-
(search term: "taxo-

-
Artiekm JOOOOD
nom?") between 1969 logical nomenclature. Proponents of
and 1996. (c) Taxo- one change, name stabilization, wish
nomic literature as a to prevent confusing and expensive
percentage of the total changes in the names of common or
biological literature. • commercially used species by setting
up lists of accepted names that would
b remain immutable (Brandenburg
1991). Professional systematists, on

Downloaded from https://academic.oup.com/bioscience/article/48/2/125/243025 by guest on 04 January 2022

also grew (using the
search term "taXO- "" the other hand, defend their scien-
nom?"; Figure 1b), as tific freedom to incorporate new
did the percentage of knowledge into the discipline, even
the biological litera- if this approach mean:'> that species
No.
tUfe represented by of
.~
names may change over time (e.g., if
Artiel .... _
articles in taxonomy a species is moved to a new genus;
(Figure lc). Thenum-
ber of articles with
.... Anderson 1991). In february 1991,
an entire Svstematics Association
,~

"taxon om" in the symposium ~as devoted tothis topic,

title or abstract or c without achieving consensus on ei-
among the keywords YN' ther the extent of the problem or its
increased from 910 in solution (Hawksworth 1991).
1969 to a peak of Name changes fall under the search
7769 in 1988. Up term "nomenclature" in BIOSIS. As
until 1989 the field % Figure 3 shows, the perception that
was not only grow- taxonomists are spending an inordi-
ing numerically, but nate amount of time pursuing obscure
also increasing its details of nomenclature appears to be
share of the total false. The percentage of taxonomy
biological literature articles indexed under nomenclature
74 89
(from 0.45 % to
1..51 %). The num- " 84 94 (search term; "nomenclature? and
taxonom?") showed considerable year-
ber of taxonomy articles then de- or species(w)nov) and taxonom?") to-year variation but no clear trend.
clined sharply (to 5230 in 1992), increased from 7.1 % in 1969 to The lowest percentage was 7.2%, in
although they peaked again at 7882 47.4% in 1988, declined to 33.2% 1988, the highest was 17.7%, in 1993,
in 1994. Taxonomy has lost ground in 1992, returned to 43.9% i111995, and the average over the 28-year pe-
relative to biology as a whole since but dropped to 33.8% in 1996. Many riod was 13.3%.
1989; its share of the literature de- basic species descriptions are still However, the descriptor "nomen-
clined to 0.98% in 1992, although it being published. On average, at least clature" covered a variety of subtop~
rose to 1.46% in 1995. The effects of one new species is described in more ics, which might have obscured any
the decline in numbers of taxono- than one-third (35.7%) of the taxo- trend; consequently, we also ana-
mists (e.g., Holden 1989, Wilson nomic articles published since 1969. 1yzed one subtopic, designation of a
1989, Feldman and Manning 1991) Statistics for the BIOSIS database neotype, independently. in the
may only now be becoming apparent as a whole also indicate that descrip- present system of biological nomen-
due to the time lag between submis- tion of new taxa is a significant part clature, each scientific name is Ii nked
sion and publication of a paper. of taxonomists' work. New genera by its audlOr (the person who first
are reported in approximately 8%, publishes a description) to a type
Are new species still and new subspecies in approximately specimen. Although the International
being described? 3 %, of articles on taxonomy; higher- Zoological and International Botani-
level discoveries are rare. New fami- cal Codes of Nomenclature handle
The impression that the cataloging lies figure in fewer than one taxo- types somewhat differently, in both
stage of taxonomy is far from over is nomic article in 100, new orders in systems a type is essentially a sort of
upheld by yearly data for description fewerthan one in a 1000, a new class supervoucher specimen. In the course
of new species (Figure 2). The per- in one in 4000, and a new phylum or of revisionary work, a taxonomist
centage of articles describing new division in only approximately one may find that the type of a particular
species (search term: "(new(w)species in 10,000. taxon has been lost or misplaced, or

126 BioScience Vol. 48 No.2

 50,-________________________- ,
that it was never designated. est and most widely available
In such cases, the taxono- 45 method (Murphy et. al. 1990).
mist carrying out the revi- Articles based on electro-
40
sion may pick a new type phoretic methods are still
specimen, or neotype. This 35 being produced in slowly
procedure is recommended % 30 increasing numbers, but dur-
only as a last resort because of 25 ing the 1970s DNA tech-
mistakes can lead to real no- niques rapidly outdistanced
menclatural headaches, for Taxon. 20 other methods, with se-
example, if (as has hap- Lit. 15 quencing and restriction site
pened), the neotype mate- 10 analysis now surpassing
rial later turns our to belong 5L-__- L_ _ _ _L -_ _ ~ ____ ~ __ ~~
DNA-DNA hybridization in
to a completely different popularity. In the last few
taxon than was claimed or is 69 74 79 84 89 94 years, articles based on DNA
found to be a mixture of Year techniques have overtaken

Downloaded from https://academic.oup.com/bioscience/article/48/2/125/243025 by guest on 04 January 2022

taxa. It is reassuring, there- those based on protein elec-
Figure 2. Percentage of taxonomic articles in which at least
fore, that the percentage of one new species is described, for each year from 1969 through trophoresis (search term:
articles in which a neotype is 1996 \search term: "(new(w)species or species\w)nov) and "'taxonom? and (allozyme?
designated (search term: taxonom?"). Although there was no simple way to search for or isozyme? or electrophore-
"neotype? and taxonom?") total number of new species via this database) it was possible sis?)"}. Studies using se-
is very low (1.4% on aver- to search for articles in which at least one new species was quencing and restriction site
age) and has not increased described. On average, at least one new species was described analysis (search term:
over time. The infrequent use in 35.7% of all taxonomic articles over the last 28 years. "taxonom? and (mtDNA or
of this procedure indicates mt(w)DNA or rnitochon-
that taxonomists are not cre- drial(w)DNA) or sequence? or
Nomenclature
ating unnecessary work for 16 sequencing?") have pulled
themselves or others. 14 ahead of DNA-DNA hy-
Articles dealing with re- 11 bridization (search term: "tax-
description and revision 10 unum? andDNA(w)hybrid?")
(search term: "(revision or re- % 8 in popularity. However, ar-
description) and taxonom?") , Revision
ticles based on the less ex-
increased somewhat over the pen sive electrophoretic meth-
time period surveyed, from • ods available to almost any
approximately 2 % to 8 %. N90typa graduate student are still be-
Revision, with its goal of ing produced in slowly In-
14 19 84 89 94
increasing the phylogenetic Year
creasing numbers.
predictiveness of classifica-
tions, is often considered to Figure 3. Percentage of taxonomic articles dealing with nomen- Conclusions
be the most important ev- clature (search term: "nomenclature? and taxonom?"), rede~
eryday business of tax- scription or revision (search term: "(revision or redescription) Taxonomy, although ac-
onomy, but there is no sign and taxonom?"'), and the desi;gnation of a neotype (search knowledged as a necessity in
it is being carried to excess- term: "neotype? and taxonom?"). assessing Earth's biodiver-
in fact, it might be questioned sity, is often considered a
whether enough revision is being done. based on a phenetic approach (search dying field. A search of the BIOSIS
term: "'phenetic?") has not declined database for the past 28 years shows
Recent trends but, instead, is still slowly increas- that, in fact, the field was still grow-
ing. By examining sample abstracts, ing through 1988, but now it is de-
Literature analysis can also effec- we found that most of these articles clining or, at the very least, ceasing
tively document major trends in the deal with botanical systematics. For to grow. Taxonomic articles describ-
field. In 1969, phenetics, especially representation of the reticulate evo- ing new species also increased
the quantitative computer-based va- lutionary patterns produced by the through 1988; their production may
riety known as numerical taxonomy, genetic systems of some plants, phe- have slackened, but they still make
was in vogue. Since then, cladistics netics is apparently still a useful ap- up more than one-third of all work
has come to dominate much of the proach (Stace 1989). published over the time period.
field (Figure 4a). The number of ar- The last 28 years have also seen a Trends in the taxonomic literature
ticles using cladistic methods (search rapid increase in [he use of molecu- show that taxonomists have slightly
term: "cladistic?") has increased dra- lar techniques in taxonomy (Figure increased the production rate of their
matically, as has the percentage of 4b). In 1969, protein electrophoresis revisionary work, but not of their
taxonomy articles that mention cla- was the only molecular method avail- articles on nomenclature. There has
distics (i.e., from 0.7% in 1974, to able to most taxonomists; it is still been a large increase in publications
3.3% in 1984. to 6.3% in 19961. the method of choice for certain types based on cladistic methods (without,
However, the number of articles of studies and, in addition, is the cheap- however, a concomitant decline in

February 1998 127

 culture. Pages 23-3J in Improving the Sta-
Figure 4. Trends in taxo- a ". bility of Nallle~: Needs and Options,
nomic methodology over Hawksworth DL, ed. Taunus (Germany):
~~--------------L~~
thelast28vears. (aJ Num- no ~~- - - - - - - L Kocltz Scientific Books.
ber of tax~nomicartides No. of 300 f__----------,r-'-- Claridge M, Ingrouille FM. 1992. ~ystematic
biology and higher education in the U.K.
using cladistic (search ~.f__-------------,/~--- Pages 39-48 in An Appraisal of Taxonomy
term: "cladistic?") and Articles 200 f__--------------__;7'/~-----
in the 1990s.london: 'fhe lillllaean Society
phenetic (search term: ". f__~ -----------,f-/-------- and the Systematics Association.

~.,:.;:;~/~:/~==;""~
"phenetic?") methods. Clarke B. 1990. Systematics and evolutionary
(b) Number of articles
using different molecu-
'"r-
'00
o
biology in Britain. Association of System-
atic Collections Newsletter 19: 58-59.
lar phylogenetic methods 69 74 79 84 89 94 Fddman RM, Manning RB. 1991. Cri~is in
by year {search terms: Vea, systematic hi()logy in the "Age of Biodi-
"taxonom? and (allo- versity. "Journal of Paleontology 66: 157-
zyme? or isozyme? or 158.
electrophoresis?)" or Hawbworth 01., ed. 1991. Improving the Sta-
"taxonom: and (mtDNA billty of Names: Needs and Options. T aunLis
or mt(w)D~A or mito-
b 600 - (Germany): Koeltz Scientific Books.

Downloaded from https://academic.oup.com/bioscience/article/48/2/125/243025 by guest on 04 January 2022

S~quenclng
Holden C.19!!9 .Entomologists wane as insects
chondrial(w)DNAor se-
quence? or sequencing)" '" waX. Science 246: 754-756.
Lean G, Hinrichsen D, Markham A. 1990.
No. of 4OOf--------------li-
or "[axonom? and Atlas of the Environment. New York:
liybrldkalion
D::-JA(w)hybrid?)"). Articles lOCI - /1--
.-------- ----~------__; Prentice-Hall.
May R. 1990. Taxonomy as destiny. Nature
~ol__----- -- - ---=-;z./='::: -'" 347: 129-130.
those using phenetic
t:;;;~;;;;~~5~:L7"~~;.;~~
Electraphoresl8 Miller S. 1994. Biological collections database~
methods) and in ar- 10: available on Internet. Association of Sys-
ticles based on mo- tematics Collections Newsletter 22: 57.
69 74 79 84 89 94 Murphy RW, Sites JW, Buth DG, Haufter CH.
lecular methods, with Year 1990. Proteins J: Isozyme electrophoresis.
~equencing now the Pages 45-126 in Hillis DM, Moritz C, eds.
most popular technique. This analy- Although we looked at trends in Molecular Systematics. Sunderland (MA);
sis thus indicates that the "crisis" in taxonomy, similar electronic searches Sinauer.
taxonomy-the basic scientific under- would be useful in other areas. Litera- Stace C. 1989. plant Taxonomy and Biosys-
ternaties. 2nd ed.londoIl: Edward Arnold.
pinning of all the efforts of those inter- ture searches cannot replace surveys Steere WC,et. al. 1971. The Systematic Biology
ested in biodiversity and its preserva- and inventories, but much informa- Collections of the United States: An Essen-
tion-is not a figment of the tion useful in designing and focusing tial Resource, Part I. The Great Collections:
imaginations of taxonomists but can them is already publicly available Their :-.Iature, Imporrance, Condition, and
be objectively documented in the pub- and can be obtained at a modest future. The Bronx (NY): The New York
Botanical Gardens.
lished literature. cost. Although retrieving synonymic Stue~sy TF, Thompson KS, eds. 1981. Trends,
Systematists, accustomed to delv- information for a particular taxon Priorities, and Needs in Systematic Biology.
ing into 200 years of literature, have usually requires searching the taxo- Lawrence (KS): Association of Systematics
tended to overlook the potential of the nomic literature back much further Collecrions.
three decades of taxonomic literature than the last 20 or 30 years, being [SA2000J Systematics Agcnda2000. 1994. Sys-
tematics Agenda 2000: Charting the hio-
available online to contribute to taxo- able to retrieve the citations for that sphere. New York: SA 2000. Technical
nomic study, but we believe that period in minutes can cut weeks of Report. Available from: New York Botani-
these data can be of gteat practical preparation time from monographic cal Gardens, The Bronx, NY 10458.
use in focusing future taxonomic re- endea vors and rna y well uncover rel- Wilson EO. 1989. The coming pluralization of
biology and the stewardship of systematics.
search on the most important or ne- evant ecological or other biological BioScience 39: 242-245.
glected groups and geographic regions. citations that would otherwise have Wimton JE. 1988. The systematists' perspec-
Electronic searches of databases such been missed. Enlisting the aid of an tiye. Memoirs of the California Academy of
as BIOSIS or Zoological Record On- information specialist from the be- Sciences 13: 1-6.
line can provide a fast, relatively inex- ginning of a project could avoid the
pensive assessment of other aspects waste of considerable time and Judith E. Winston (e-mail: jwinston@/eo.
of taxonomy. For example, a 1991 money in recompiling or re-docu- vsla.edu), formerly curator and chair ofthe
search using Zoological Record Online, menting such material and would Department oflnvertehrates at the Ameri-
a database from which records ofthe jumpstart any biodiversity program. can Museum of Natural History in New
numbers of taxonomic surveys can York. is director of research at the Virginia
be easily compiled by geographic References cited Museum of Natural History. Martinsville,
area, showed that although marine [Anonymous]. 1991. Funding un sexy science. VA 24112. Her research interests include
bryozoan systematics and marine biodiver-
habitats cover 70% of the earth's Science 251: .177. sity. Kristen L. Metzger (e-mail: metzger@
su rface, only 13% of published taxo- Anderson WR. 1991. Shall we change the code? hhoi.edu) is librarian in the research library
nomic surveys were marine oriented. Concerns of a working taxonomist. Pages at the Harbor Branch Oceanographic In-
This information would certainly be 95-103 in Hawksworth DL, ed. Improving
the Stability of Names: Needs and OpTions. stitution, Fort Pierce, FL 44946. Her inter-
good ammunition in a request for Taunus (Germany): Koeltz Scientific Books. ests include marine science librarianship
support for funding of a marine tax- Brandenburg WA. 1991. The need for stabi- and information technology. © 1998 Ameri-
onomy project. lized plant names in agriculture and horti- can Institute of Biological Sciences.

128 BioScience Vol. 48 No.2