Journal of Threatened Taxa |
www.threatenedtaxa.org | 26 July 2020 | 12(10): 16204–16209
ISSN 0974-7907 (Online) | ISSN 0974-7893
(Print)
doi: https://doi.org/10.11609/jott.5517.12.10.16204-16209
#5517 | Received 31 October 2019 | Final
received 19 June 2020 | Finally accepted 21 June 2020
The worrisome conservation status
of ecosystems within the distribution range of the Spectacled Bear Tremarctos ornatus (Mammalia:
Carnivora: Ursidae) in Ecuador
José Guerrero-Casado 1 & Ramón H. Zambrano 2
1 Departamento de Medicina
Veterinaria, Facultad de Ciencias Veterinarias,
Universidad Técnica de Manabí, Avenida Urbina,
130105, Portoviejo, Ecuador.
2 Facultad de Ingeniería
Agropecuaria, Universidad Laica
Eloy Alfaro de Manabí, El Carmen, Manabí, 130401, Ecuador.
1 guerrero.casado@gmail.com
(corresponding author), 2 ramonhoracio@hotmail.com
Editor: David Mallon, Manchester
Metropolitan University, UK. Date of publication: 26
July 2020 (online & print)
Citation: Guerrero-Casado, J. & R.H.
Zambrano (2020). The worrisome conservation status
of ecosystems within the distribution range of the Spectacled Bear Tremarctos ornatus
(Mammalia: Carnivora: Ursidae) in Ecuador. Journal of Threatened Taxa 12(10): 16204–16209. https://doi.org/10.11609/jott.5517.12.10.16204-16209
Copyright: © Guerrero-Casado & Zambrano
2020. Creative Commons Attribution
4.0 International License. JoTT allows unrestricted use, reproduction, and
distribution of this article in any medium by providing adequate credit to the
author(s) and the source of publication.
Funding: The research received
no financial support.
Competing interests: The authors
declare no competing interests.
Author details: José
Guerrero-Casado, full professor at the Technical University of Manabí,
has worked in several research projects of wildlife management and conservation
in Spain and Ecuador. Ramón H. Zambrano, a biologist with a
master’s degree of biodiversity conservation at Università
della Calabria, is currently a PhD student at the
University of Córdoba, Spain.
Author contribution: Conceptualization JG-C; GIS data
management and analysis JG-C and RHZ;
manuscript preparation JG-C and RHZ.
Acknowledgements: This work was carried out using
freely available data of the IUCN Red List and the Ecuadorian Ministry of
Environment.
Abstract: The distribution range of
Spectacled Bear in Ecuador has been strongly fragmented owing to human
activities, thus studying the conservation status of the remaining patches in
which this species inhabits is essential to ensure its conservation. We
performed a descriptive analysis of the conservation status of the ecosystems
that form part of the Spectacled Bear distribution range in Ecuador, using
values of five indicators (fragmentation, connectivity, threat, vulnerability
and fragility) provided by the Ecuadorian Ministry of Environment. Moreover, we compared the conservation status
between scrubland & pastures and evergreen forest ecosystem according to
the five indicators. Overall, the Spectacled Bear distribution in Ecuador is
covered by a greater surface area of ecosystems classified with a medium level
of fragmentation (58.8%), low connectivity (45.1%), high fragility (45.6%),
high (38.2%) and medium threat (39.2%), as well as high (33.7%) and medium
(33.1 %) vulnerability. Scrubland and
pastures ecosystems had worse conservation status than evergreen forests, with
98% of their surface classified as having low or very low connectivity, 77.3%
as high fragility and 77.4% as high threat.
These results showed the worrisome conservation status of the ecosystems
shaping the distribution range of Spectacled Bear in Ecuador, particularly the
scrubland and pastures, and therefore, a wider national strategy (inside and
outside the limits of the bear distribution range) should be applied to ensure
the preservation of these ecosystems.
Keywords: Andean Bear, conservation,
ecosystem, endangered species, habitat loss, human activities, species
distribution.
Introduction
Habitat loss and degradation are
major threats for the world’s mammals (Brooks et al. 2002; Fahrig
2003; Dobson et al. 2006). Apart from
constraining the distribution range of mammal species, habitat loss entails the
fragmentation and isolation of remaining populations, increasing the risk of
extinction as a consequence of the loss of genetic flow and adaptive genetic
variation (Ernest et al. 2011), inbreeding (Keller & Waller, 2002),
demographic and environmental stochastic phenomena (Shaffer 1981; Crooks et al.
2011), and lower colonization rate (Fahrig
2003). Large mammals are most vulnerable
to habitat loss and fragmentation owing to their extensive home ranges, lower
densities, and lower population growth rates (Cardillo et al. 2005). Furthermore, large mammals are more prone to
poaching and direct persecution, and are often implicated in various kinds of
human-wildlife interactions such as damage to agriculture or attacks on
livestock (Woodroffe et al. 1998; Kansky & Knight
2014), which are more frequent in smaller populations surrounded by a
human-dominant matrix (Michalski & Peres 2005).
As a large mammal, the Andean
Bear or Spectacled Bear Tremarctos ornatus F.G. Cuvier, 1825 needs large areas to fulfil
its biological requirements (Cuesta et al. 2003; Castellanos 2011). This species is endemic to the Tropical Andes
and was formerly distributed widely across the mountain range (Peyton et al.
1998; Kattan et al. 2004; Velez–Liendo
& García-Rangel 2017), however, its current populations are distributed in
fragmented areas from Bolivia to Venezuela, which led to the species being
assessed as Vulnerable (VU) on the IUCN Red List (Velez–Liendo
& García-Rangel 2017). In Ecuador,
this species is listed as Endangered (EN) in the Red Book of Mammals of Ecuador
(Castellanos et al. 2011), owing to a reduction in both population size and
distribution as a consequence of habitat alteration (Peyton et al. 1998; Kattan et al. 2004; Peralvo et
al. 2005), thus affecting strongly the viability of the species (Kattan et al. 2004).
Therefore, it is important to assess the remaining patches in terms of
threat, vulnerability, and fragmentation, to establish measures favouring bear
conservation. Most studies in Ecuador
have been conducted to assess habitat preferences at small scales (Suarez 1988;
Cuesta et al. 2003; Peralvo et al. 2005; Clark 2008;
Castellanos 2011; Demay et al. 2014; Filipczyková et al. 2017), but none have characterized the
remaining fragments at national scale in which this species inhabits. The aim of this study was to characterize all
the remaining fragments of Andean Bears in Ecuador, taking into account five
indicators of the conservation status of the ecosystems (fragmentation,
connectivity, vulnerability, threat, and fragility) provided by the Ecuadorian
Ministry of Environment, in order to comprehensively assess the conservation
status of ecosystems in which the Spectacled Bear inhabits in Ecuador.
Material
and Methods
Species distribution range
The geo-referenced current range
of the bear within Ecuador was obtained from the IUCN Red List webpage (Velez–Liendo & García-Rangel 2017; Figure 1). An area is considered as extant if the
species is known or thought very likely to occur there presently, which
encompasses localities with current or recent (data updated in 2017) records
where suitable habitat at appropriate altitudes remain (IUCN 2018).
Ecosystem features
Geo-referenced shapefiles were
obtained from the Ecuador Ministry of Environment (available at http://ide.ambiente.gob.ec/mapainteractivo/)
wherein the ecosystems of
mainland Ecuador are classified into 91 types according to biotic and abiotic
factors (Ministerio del Ambiente
del Ecuador 2013). This shapefile also
contains information concerning fragmentation, connectivity, vulnerability,
threat, and fragility of each ecosystem.
The rate of fragmentation was calculated using three variables: the
number of patches, their mean size, and the coefficient of variation of the
size of the patch (Ministerio del Ambiente
del Ecuador 2015b). These variables were
used to rank the fragmentation of each ecosystem on four levels (very high,
high, medium, and low) according to the method of Jenks’ natural breaks (for
more details see the previous reference).
Connectivity was measured by the equivalent connected area index, which
is defined as the size of a single habitat patch (maximally connected) that
would provide the same value of the probability of connectivity than the actual
habitat pattern in the landscape (Saura et al.
2011). Using also Jenks’ natural breaks,
the connectivity rates of the ecosystems were classified into four categories:
high, medium, low, and very low (Ministerio del Ambiente del Ecuador 2017).
The vulnerability index was calculated at species level by using the
number of species listed in CITES, the number of endemic species, the number of
plants with a commercial value, and the number of endangered plant species
according to the IUCN Red List in Ecuador.
At the landscape level, vulnerability was calculated by the degree of
representativeness, fragmentation, and connectivity. The threat to ecosystems was assessed by five
variables: climate change, water resource use, forest exploitation, extraction
of natural resources, and the probability of land conversion. Vulnerability and threat were classified in
three categories (high, medium, and low) by using quantiles, and both
indicators were combined to get five levels of fragility (very high, high,
medium, low, and very low; Ministerio del Ambiente del Ecuador 2015a). No values were assigned to areas identified
as non-natural ecosystems, such as crops, urban areas, or planted forests. These indexes provided by the Environmental
Ministry have been used in previous scientific papers (e.g. Rivas et al. 2020).
Three different descriptive
analyses were performed. Firstly, the
proportions of each level (e.g., high, medium, low) of each of the five
indicators (fragmentation, connectivity, vulnerability, threat, and fragility)
were calculated in the whole bear distribution range, which was done by adding
the surface area of every ecosystem with the same level of each indicator. Secondly, these proportions were calculated
separately in each of 17 distribution patches, which allows us to estimate the
mean values and the variation among patches.
Thirdly, the ecosystems were grouped into the two main habitats of
Andean Bears in Ecuador (Peralvo et al. 2005):
forests and scrubland & pastures (including paramo), to compare the
conservation status between both habitats in accordance with the five
indicators.
Results
According to the IUCN data,
Spectacled Bears in Ecuador are located in 17 different patches (Figure 1)
covering an area of 19,940km2, with a mean patch size of 1,172km2
(± 890). This distribution range
includes 42 distinct ecosystems: 14 natural pastures and/or scrublands covering
35.7% of the range, 28 different forest ecosystems covering 57.4%, and the
remaining 6.9% classified as non-natural areas.
Overall, in the whole distribution range, the bear distribution contains
a higher proportion of ecosystems classified as medium fragmented (58.8%), with
low connectivity (45.1%), high fragility (45.6%), high (38.2%), and medium
threat (39.2%), as well as high (33.7%) and medium (33.1%) vulnerability
(Figure 2). The mean values obtained in
the 17 patches in which the Andean Bear inhabits showed greater average values
of medium level of fragmentation, low connectivity, high fragility, medium
threat, and high vulnerability (Table 1).
Finally, the comparison between
forests and scrublands & pastures, showed that the latter ecosystem type
has a poorer conservation status (Figure 3).
For instance, 98% of the area covered by scrublands & pastures are
classified as having low or very low connectivity, 77.3% with high fragility,
and 77.4% with high threat (Figure 3).
Discussion
Habitat fragmentation and its
consequent isolation of populations is one the main conservation problems for
the majority of large mammals (Crooks et al. 2011), including the Spectacled
Bear (Castellanos et al. 2005). The
development of the road network, land-use changes, and the establishment of
human settlements (Armenteras et al. 2003; Kattan et al. 2004; Peralvo et
al. 2005) have fragmented continuous populations into smaller and isolated
patches (Figure 1), which could jeopardize bears population viability. It is, therefore, important to understand the
features of these fragments, such as the conservation status of the ecosystems
shaping the distribution range of the Spectacled Bear, which could help to
establish remedial measures to conserve its habitats. In this work, however, we have shown that the
ecosystems occupied by this species are not always the best preserved, since a
great proportion of the potential distribution range of the Spectacled Bear in
Ecuador are ecosystems with some degree of threat.
Overall, a large proportion of
the distribution range of the Spectacled Bear in Ecuador is covered by
ecosystems catalogued as fragmented, poorly connected, fragile, threatened, and
vulnerable (Figure 2). This relatively
poor conservation status has important implications for the Spectacled Bear if
no measures are applied to conserve these ecosystems. A fragmented and poorly connected ecosystem
under high human pressure is vulnerable to reduction in surface area or even
complete disappearance. Therefore, the ecosystems in Andean Bear’s range could
be replaced by other land uses that are less suitable as Andean Bear habitats.
Several scientific works have
suggested that connectivity among patches still inhabited by Andean Bear is
essential to ensure the conservation of its populations (Kattan
et al. 2004; Velez–Liendo et al. 2014). According to our results, however, apart
from connecting the reaming patches, a wider strategy (inside and outside of
the distribution range) aimed at boosting the conservation status of the
important ecosystems for this species is also necessary. One strategy could be to use the Andean Bear
as an umbrella species (Crespo-Gascón &
Guerrero-Casado 2019) in order to conserve those ecosystems with a high degree
of threat that this species inhabits.
Furthermore, our results show
that scrublands and pastures (including paramo or moorland) which are an
important part of the Andean Bear’s distribution range are more threatened than
evergreen forests. The paramo in Ecuador has been extensively replaced by other
land uses (e.g., pine plantations, crops or artificial pastures) (Ross et al.
2017) and although it covers a large area in the Andes region, it is highly
threatened by anthropogenic activities (Ministerio
del Ambiente del Ecuador 2015b). Therefore, since the paramo is an important
ecosystem for the Andean Bear (Demay et al. 2014), it
covers an important proportion of its distribution, and is a fragile ecosystem,
conservation efforts should be also targeted at preserving this ecosystem at
the national scale, which in turn could help to conserve the Spectacled Bear
habitats.
Evergreen montane forests are the
main habitats for the Spectacled Bear in Ecuador (almost 60 % according to our
data), and as we have shown in this study, almost 50% of the distribution range
of the Spectacled Bear is covered by forests considered as high or very high
fragility (Figure 3). These ecosystems
have a high alpha diversity and they are well represented in protected areas
(Sierra et al. 2002), although they have been intensely deforested, thus, they
are very fragile ecosystems. Therefore,
conserving the larger blocks which this species inhabits (Kattan
et al. 2004) as well as smaller patches of evergreen forests to promote the
connectivity (Peralvo et al. 2005) should be
implemented to ensure the preservation of these ecosystems and the habitat of
the Andean Bear.
Conclusion
In summary, our results show that
the ecosystems in which the Spectacled Bear inhabits have a poor conservation
status, with an important proportion of the distribution range covered by
ecosystems classified as fragmented, poorly connected, vulnerable, threatened,
and fragile. Therefore, a national
conservation strategy should be developed to enhance the conservation status of
these ecosystems, which should include actions performed not only in areas
where bears are present, but also in areas where they are not. This would contribute to the conservation of
these ecosystems at a national scale, which would effectively preserve
Spectacled Bear habitats.
Table 1. Mean values ± standard
deviation (SD) of the percentage of surface area of the different levels of the
five indicators for the 17 distribution patches in which the Andean Bear
inhabits in Ecuador.
|
Indicator |
Mean |
±SD |
|
Fragmentation |
||
|
Very High |
7.94 |
6.77 |
|
High |
17.43 |
25.6 |
|
Medium |
55.28 |
27.28 |
|
Low |
9.35 |
15.46 |
|
Connectivity |
||
|
High |
10.83 |
17.47 |
|
Medium |
27.17 |
22.84 |
|
Low |
43.96 |
25.6 |
|
Very Low |
8.24 |
12.01 |
|
Fragility |
||
|
Very High |
15.28 |
29.13 |
|
High |
49.77 |
31.06 |
|
Medium |
6.21 |
8.47 |
|
Low |
10.81 |
15.6 |
|
Very Low |
8.14 |
12.54 |
|
Threat |
||
|
High |
32.11 |
25.1 |
|
Medium |
45.54 |
26.76 |
|
Low |
12.55 |
13.65 |
|
Vulnerability |
||
|
High |
43.2 |
32.3 |
|
Medium |
29 |
27.22 |
|
Low |
17.99 |
23.55 |
References
Armenteras, D., F. Gast & H. Villareal
(2003). Andean forest
fragmentation and the representativeness of protected natural areas in the
eastern Andes, Colombia. Biological Conservation 113(2): 245–256. https://doi.org/10.1016/S0006-3207(02)00359-2
Brooks, T.,
R.A. Mittermeier, C.G. Mittermeier, G.A.B. Da Fonseca, A.B. Rylands, W.R. Konstant, P. Flick, J. Pilgrim, S. Oldfield, G. Magin, C. Hilton-Taylor (2002). Habitat loss and extinction in
the hotspots of biodiversity. Conservation Biology 16(4): 909–923. https://doi.org/10.1046/j.1523-1739.2002.00530.x
Cardillo, M.,
G.M. Mace, K.E. Jones, J. Bielby, O.R.P. Bininda-Emonds, W. Sechrest,
C.D.L. Orme, A. Purvis (2005). Multiple causes of high extinction risk in large
mammal species. Science 309(5738): 1239–1241. https://doi.org/10.1126/science.1116030
Castellanos,
A. (2011). Andean Bear
home ranges in the Intag region, Ecuador. Ursus 22(1): 65–73. https://doi.org/10.2192/URSUS-D-10-00006.1
Castellanos,
A., M. Altamirano & G. Tapia (2005). Ecología y comportamiento
de osos andinos reintroducidos en la reserva biológica Maquipucuna, Ecuador: implicaciones
en la conservación. Revista Politécnica
26(1): 198–209.
Castellanos,
A., R. Cisneros, C.F. Cuesta, R. Narváez, L. Suárez
& D.G. Tirira (2011). Oso andino
(Tremarctos ornatus).
In: Libro Rojo de los mamíferos del Ecuador. Fundación Mamíferos
y Conservación, Pontificia
Universidad Católica del Ecuador & Ministerio del Ambiente del
Ecuador, Quito, 399pp. http://librorojo.mamiferosdelecuador.com/lista-de-especies-/carnivora/ursidae/597-tremarctos-ornatus-corto.html
Clark, M.R.
(2008). Using the
Spectacled Bear as a Conservation Tool in the Condor Bioreserve, Ecuador. Journal
of Sustainable Forestry 18(2–3): 223–236. https://doi.org/10.1300/J091v18n02_10
Crespo-Gascón, S. & J. Guerrero-Casado (2019). The role of the Spectacled Bear (Tremarctos ornatus) as an
umbrella species for Andean ecoregions. Wildlife Research 46(2):
176–183. https://doi.org/10.1071/WR18056
Crooks, K.R.,
C.L. Burdett, D.M. Theobald, C. Rondinini & L. Boitani (2011). Global patterns of fragmentation
and connectivity of mammalian carnivore habitat. Philosophical Transactions
of the Royal Society of London. Series B, Biological Sciences 366(1578):
2642–2651. https://doi.org/10.1098/rstb.2011.0120
Cuesta, F.,
M.F. Peralvo & F.T. Van Manen
(2003). Andean Bear
Habitat Use in the Oyacachi River Basin, Ecuador. Ursus 14(2): 198–209.
Demay, S.M., D.A. Roon,
J.L. Rachlow & R. Cisneros (2014). Selective foraging on bromeliads
by Andean bears in the Ecuadorian paamo. Ursus 25(2): 139–147. https://doi.org/10.2192/URSUS-D-14-00022.1
Dobson, A.,
D. Lodge, J. Alder, G. Cumming & J. Keymer
(2006). Habitat loss,
trophic collapse, and the decline of ecosystem services. Ecology 87(8):
1915–1924. https://doi.org/10.1890/0012-9658(2006)87[1915:HLTCAT]2.0.CO;2
Ernest, E.M.,
H. Haanes, S. Bitanyi, R.D.
Fyumagwa, P.L. Msoffe, G. Bjørnstad & K.H. Røed (2011).
Influence of
habitat fragmentation on the genetic structure of large mammals: evidence for
increased structuring of African buffalo (Syncerus caffer) within the Serengeti ecosystem. Conservation
Genetics 13(2): 381–391. https://doi.org/10.1007/s10592-011-0291-0
Fahrig, L. (2003). Effects of habitat fragmentation
on biodiversity. Annual Review of Ecology, Evolution, and Systematics
34: 487–515. https://doi.org/10.1146/annurev.ecolsys.34.011802.132419
Filipczyková, E., I.M.A. Heitkönig,
A. Castellanos, W. Hantson & S.M.J.G. Steyaert (2017). Marking behavior of Andean bears
in an Ecuadorian cloud forest: a pilot study. Ursus
27(2): 122–128. https://doi.org/10.2192/URSU-D-16-00002.1
IUCN (2018). Mapping Standards and Data
Quality for the IUCN Red List Categories and Criteria. Red List Technical
Working Group, 30pp.
Kansky, R. & A.T. Knight (2014). Key factors driving attitudes
towards large mammals in conflict with humans. Biological Conservation
179: 93–105. https://doi.org/10.1016/j.biocon.2014.09.008
Kattan, G., O.L. Hernández, I.
Goldstein, V. Rojas, O. Murillo, C. Gómez, Restrepo, H. & F. Cuesta (2004).
Range
fragmentation in the Spectacled Bear Tremarctos
ornatus in the northern Andes. Oryx 38(2):
155–163. https://doi.org/10.1017/S0030605304000298
Michalski, F.
& C.A. Peres (2005). Anthropogenic determinants of primate and carnivore local extinctions in
a fragmented forest landscape of southern Amazonia. Biological Conservation 124(3):
383–396. https://doi.org/10.1016/j.biocon.2005.01.045
Ministerio del Ambiente
del Ecuador (2013). Sistema de Clasificación de los Ecosistemas del Ecuador
Continental. Subsecretaría de Patrimonio
Natural, Quito, Ecuador, 232pp.
Ministerio del Ambiente
del Ecuador (2015a). Fragilidad de los
Ecosistemas del Ecuador Continental. Quito, Ecuador, 52pp.
Ministerio del Ambiente
del Ecuador (2015b). Fragmentación de los
ecosistemas del Ecuador Continental. Subsecretaría
de Patrimonio Natural, Quito, Ecuador, 22pp.
Ministerio del Ambiente
del Ecuador (2017). Conectividad de los
Ecosistemas del Ecuador Continental. Proyecto Sistema Único de Información Ambiental,
Sistema Nacional de Monitoreo de Bosques,
Quito, Ecuador, 26pp.
Peralvo, M.F., F. Cuesta & F. van Manen (2005). Delineating priority habitat areas for the
conservation of Andean Bears in northern Ecuador. Ursus
16(2): 222–233. https://doi.org/10.2192/1537-6176(2005)016[0222:DPHAFT]2.0.CO;2
Peyton, E.,
E. Yerena, D.I. Rumiz, J.
Jorgenson & J. Orejuela (1998). Status of wild Andean Bears and
policies for their management. Ursus 10:
87–100.
Rivas, C.A.,
R.M. Navarro-Cerillo & J. Guerrero-Casado (2020).
Dry forest is
more threatened but less protected than evergreen forest in Ecuador’s coastal
region. Environmental Conservation 47(2): 79–93. https://doi.org/10.1017/S0376892920000077
Ross, C., S. Fildes, A. Millington, C. Ross, S. Fildes
& A. Millington (2017). Land-use and land-cover change in the Páramo
of south-central Ecuador, 1979–2014. Land 6(3): 46. https://doi.org/10.3390/land6030046
Saura, S., C. Estreguil,
C. Mouton & M. Rodríguez-Freire (2011). Network analysis to assess
landscape connectivity trends: Application to European forests (1990–2000). Ecological
Indicators 11(2): 407–416. https://doi.org/10.1016/J.ECOLIND.2010.06.011
Shaffer, M.
(1981). Minimum
population sizes for species conservation. BioScience
31(2): 131–134.
Sierra, R.,
F. Campos & J. Chamberlin (2002). Assessing biodiversity
conservation priorities : ecosystem risk and representativeness in continental
Ecuador. Landscape and Urban Planning 59(2): 95–110. https://doi.org/10.1016/S0169-2046(02)00006-3
Suarez, L.
(1988). Seasonal
distribution and food habits of spectacled Bears Tremarctos
ornatus in the highlands of Ecuador. Studies
on Neotropical Fauna and Environment 23(3): 133–136. https://doi.org/10.1080/01650528809360755
Velez–Liendo, X., F. Adriaensen &
E. Matthysen (2014). Landscape assessment of habitat
suitability and connectivity for Andean bears in the Bolivian Tropical Andes. Ursus 25(2): 172–187. https://doi.org/10.2192/URSUS-D-14-00012.1
Velez–Liendo, X. & S. García-Rangel (2017). Tremarctos ornatus (errata version published in
2018). The IUCN Red List of Threatened Species 2017: e.T22066A123792952.
Downloaded on 23 June 2020. https://doi.org/10.2305/IUCN.UK.2017-3.RLTS.T22066A45034047.en
Woodroffe, R., J.R. Ginsberg, R. Lande, R.R. Knight, L.L. Eberhardt, J. Felsenstein
& B.N. McLellan (1998). Edge effects and the extinction of populations inside protected areas. Science
280(5372): 2126–2128. https://doi.org/10.1126/science.280.5372.2126