RES EARCH

BIODIVERSITY LOSS assessed species has an unknown conserva-

tion status (13). The diets of tropical snakes
Tropical snake diversity collapses after widespread include amphibians and their eggs, inverte-
brates (including oligochaetes and mollusks),
amphibian loss lizards, snakes, birds, and mammals, with most
species feeding on amphibians to some extent
Elise F. Zipkin1*, Graziella V. DiRenzo1,2, Julie M. Ray3, Sam Rossman1,4, Karen R. Lips5 (table S1). Although amphibian declines are
likely to negatively affect snakes through the
Biodiversity is declining at unprecedented rates worldwide. Yet cascading effects of biodiversity loss on loss of diet items, presumably many species
other taxa are largely unknown because baseline data are often unavailable. We document the collapse could persist by shifting to other prey.
of a Neotropical snake community after the invasive fungal pathogen Batrachochytrium dendrobatidis Our study occurred in Parque Nacional G. D.
caused a chytridiomycosis epizootic leading to the catastrophic loss of amphibians, a food source for Omar Torríjos Herrera, 8 km north of El Copé,
snakes. After mass mortality of amphibians, the snake community contained fewer species and was Panama. The amphibian community at the
more homogeneous across the study site, with several species in poorer body condition, despite no other study site (hereafter “El Copé”) contained
systematic changes in the environment. The demise of the snake community after amphibian loss >70 species pre-epizootic (11). Amphibian abun-
demonstrates the repercussive and often unnoticed consequences of the biodiversity crisis and calls dance declined by >75% immediately after
attention to the invisible declines of rare and data-deficient species. the Bd epizootic in late 2004, with extirpation
of at least 30 species (11, 12). The study site is

L
composed of mature secondary forest that re-
ong-term biodiversity trends indicate that Without a clear understanding of these cascad- mained undisturbed with no systematic changes

Downloaded from http://science.sciencemag.org/ on February 13, 2020

species extinction rates over the past two ing sequences, we risk undermining options documented within the abiotic environment
centuries are up to 100 times higher than available for effective conservation (5). (e.g., habitat, water quality, or contaminants;
throughout the rest of human history (1). Nowhere has biodiversity loss been more materials and methods). We conducted 594
Despite tremendous data collection ef- acute than in the tropics, which harbor two- surveys targeting all amphibians and reptiles
forts worldwide, empirical evidence of the eco- thirds of described species (6). Recent assess- on seven permanent transects during the 7 years
logical impacts of these losses is often lacking. ments suggest that nearly 12% of animal species pre-epizootic (December 1997 to December
Scientists rarely have the ability to predict in tropical countries are classified as endan- 2004) and 513 surveys on the same transects
impending change, precluding the opportunity gered, vulnerable, or near threatened, represent- during the 6 years post-epizootic (September
to collect adequate pre- and postdata to eva- ing 64% of all such classified species worldwide 2006 to July 2012).
luate ecosystem responses to species declines. (7). Amphibians, in particular, have suffered In El Copé, as with many tropical commun-
Yet biodiversity loss can cause cascading effects severe declines in the tropics from habitat loss, ities, a large fraction of species are rare and
within ecosystems, such as coextinction of mu- disease, and climate change (8, 9). Given that most are difficult to detect. For example, of the
tualist species, changes in energy flow and amphibians are important as both consumers 36 snake species ever observed on our stand-
primary production, and reduced resiliency and prey in aquatic and terrestrial habitats and ardized transect surveys during the 13-year
to climate and environmental change (2–4). that their abundance in the tropics can be quite study, 12 were detected only once. In an effort
high, the effects of amphibian losses likely per- to include the data from rarely observed spe-
meate to other taxa within ecosystems (10). cies while also accounting for imperfect detec-
1
Department of Integrative Biology; Ecology, Evolutionary We evaluated a Neotropical snake commu- tion and ecological variations among species,
Biology, and Behavior Program, Michigan State University, nity for changes in species richness, commu- we developed a hierarchical community model
East Lansing, MI 48824, USA. 2Ecology, Evolution, and
nity composition, occurrence rates, and body using a Bayesian approach for parameter es-
Marine Biology, University of California, Santa Barbara, CA
93101, USA. 3La MICA Biological Station, El Copé de La condition after the mass mortality of amphib- timation (14). Our model estimated occurrence
Pintada, Coclé, Republic of Panama. 4Hubbs-SeaWorld ians from chytridiomycosis caused by the rates, or the probability that both observed
Research Institute, Melbourne Beach, FL 32951, USA. invasive fungal pathogen Batrachochytrium and unobserved species used the survey tran-
5
Department of Biology, University of Maryland, College Park,
MD 20742, USA. dendrobatidis (Bd) (11, 12). Snakes are an un- sects, which we utilized to calculate species
*Corresponding author. Email: [email protected] derstudied taxon in which almost one in four richness pre- and post-epizootic (materials and

Fig. 1. Snake species richness and A B

composition before and after the 0.03 40.1
Mode 1.0 Pre
Mean Post
epizootic that led to amphibian loss. 52.7 Number of
(A) Observed (dashed lines) and esti- Npost 48.8 observed species
61.7 0.5
mated snake species richness (posterior
Posterior density

0.02 Npre
density plots with mean and mode) pre-
Axis 2

epizootic (Npre, blue) and post-epizootic 0.0

(Npost, orange). (B) Standard ellipses
representing observed snake composition 0.01 P[Npost < Npre] = 0.85
−0.5
pre-epizootic (blue) and post-epizootic
(orange). Points within the ellipses show
the dimensionless values of community 21 30
−1.0 P[Areapost < Areapre] = 0.99
0.00
composition for the seven transects pre-
20 40 60 80 100 120 −1.0 −0.5 0.0 0.5 1.0
and post-epizootic. The smaller area of the
Species richness (N) Axis 1
post-epizootic ellipse indicates a more
homogeneous snake community
compared with pre-epizootic.

Zipkin et al., Science 367, 814–816 (2020) 14 February 2020 1 of 3

RES EARCH | R E P O R T

methods). We focused on estimating proba- (61.7 versus 48.8), median (58 versus 45), and community has fewer species and is more homo-
bilities that species diversity and occurrence mode (52.7 versus 40.1) values of posterior geneous post-epizootic.
metrics changed from pre- to post-epizootic distributions all indicate that snake species Individual snake species responses to the
rather than reporting absolute values of these richness was higher pre-epizootic than post- loss of amphibians were variable, but most
metrics, which are inherently imprecise owing epizootic (Fig. 1A), although the 95% credible fared worse post-epizootic. Despite low detec-
to the many rare species within tropical snake intervals on richness estimates were wide both tion power for many species (figs. S1 and S2),
communities. pre-epizootic (38 to 105) and post-epizootic (28 we were able to confidently estimate the prob-
After the epizootic, the total number of to 89). Results of a nonmetric multidimen- ability that occurrence rates changed from pre-
observed snake species declined from 30 to sional scaling analysis show that the observed to post-epizootic for almost half of the observed
21, with an estimated 0.85 probability that snake community composition also changed snake species (tables S2 and S3). Of the 17
species richness was lower post-epizootic than from pre- to post-epizootic, as indicated by a species with at least five total observations,
pre-epizootic (Fig. 1A). Estimated species rich- shift of the centroid (0.93 probability of change) nine had occurrence rates that were lower post-
ness was considerably higher than the number and reduction in area (0.99 probability of epizootic (with ≥0.72 probability), four had
of observed snake species because of a high decrease) of standard ellipses comparing com- occurrence rates that were higher, and the
probability that many species were present and position across survey transects (Fig. 1B). remaining four species experienced no sub-
went undetected during sampling. The mean Collectively, these results reveal that the snake stantial change (Fig. 2). We compared body
condition (ratio of mass to snout-to-vent length
squared) for the six snake species with at least
Detections five samples both pre- and post-epizootic (table
Pre Post S4). Four of the six species had ≥0.97 prob-

Downloaded from http://science.sciencemag.org/ on February 13, 2020

13 0
ability of decreased body condition post-epizootic,
149 49
4 1
whereas two had body conditions that in-
4 1 creased (Fig. 3). Although there is no single life
7 0 history or diet attribute that provides a clear
7 0 explanation of the species results (table S1),
21 8 snakes that declined post-epizootic may have
5 0 had a difficult time switching their diets as am-
4 2 phibians declined and prey availability shifted.
69 69
For example, Sibon argus, which has been doc-
47 57
4 5
umented feeding on amphibian eggs at higher
3 5 levels than the three other Sibon species [primar-
19 59 ily molluscivores; (15)], experienced the most
13 38 severe declines of its genus despite otherwise
1 5 similar habitat requirements and behaviors.
3 32 Although most snake species were negatively
0.00 0.25 0.50 0.75 1.00 affected by the loss of amphibians, a few ex-
P[Occurrencepost < Occurrence pre] ploited this change, increasing in occurrence
and/or body condition. Thus, the Bd epizootic
Fig. 2. Changes in snake species occurrence rates after the epizootic that led to amphibian loss. indirectly produced a large number of “loser”
Probabilities (black circles) that occurrence rates were lower post-epizootic than pre-epizootic for snake species but also a few “winners,” an
the 17 snake species with at least five total detections across both time periods. High values (red-shaded ecological phenomenon frequently observed
zone) indicate that the occurrence rate decreased after the epizootic, whereas low values (blue-shaded after disturbance leading to biotic homoge-
zone) indicate that the occurrence rate increased. The gray zone represents no change. The number nization (16).
of detections pre- and post-epizootic on standardized survey transects is shown for each snake species to Our analyses demonstrate that widespread
the right of the figure. amphibian losses led to a smaller, less diverse

Sibon Oxybelis Sibon Imantodes Dipsas Leptodeira

annulatus brevirostris argus cenchoa sp. septentrionalis
P[Post< Pre] 1.00 1.00 0.98 0.97 0.19 0.11
0.75 1.30 0.75 0.43 0.70 1.10
Body Condition (g/cm )

0.55 0.85 0.65 0.37 0.55 0.93

0.35 0.40 0.55 0.31 0.40 0.75

Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post

Fig. 3. Average body condition for snake species before and after the epizootic in both time periods. Probabilities that body condition was lower post-epizootic
that led to amphibian loss. Body condition for the six snake species with at than pre-epizootic are shown for each species above the individual plots. High
least five samples available both pre-epizootic (blue) and post-epizootic (orange). probabilities (close to 1) indicate that body condition decreased after the epizootic,
Mean values (circles) and 95% credible intervals (lines) are plotted for each species whereas low probabilities (close to zero) indicate that body condition increased.

Zipkin et al., Science 367, 814–816 (2020) 14 February 2020 2 of 3

RES EARCH | R E P O R T

snake community, even with uncertainty in resulting in data deficiencies. However, data 14. E. F. Zipkin, J. A. Royle, D. K. Dawson, S. Bates,
the exact number of species that declined. Al- deficiencies can also arise because some species Biol. Conserv. 143, 479–484 (2010).
15. J. M. Ray, C. E. Montgomery, H. K. Mahon, A. H. Savitzky,
though there are no direct effects of the Bd are rare or have elusive behaviors and life his- K. R. Lips, Copeia 2012, 197–202 (2012).
pathogen on snakes, many of our focal species tory strategies, such that it can be difficult to 16. M. L. McKinney, J. L. Lockwood, Trends Ecol. Evol. 14, 450–453
(table S1) as well as others in Central America quantify species losses even with extensive (1999).
17. H. W. Greene, Mem. Cali. Acad. Sci. 12, 259–280 (1988).
(17) have been observed preying on amphibian sampling and advanced statistical models.
18. M. R. Whiles et al., Ecosystems (N. Y.) 16, 146–157 (2013).
adults and/or eggs. Our results suggest that the Despite a lack of data for many species, it is 19. C. N. Johnson et al., Science 356, 270–275 (2017).
snake community may be dependent on am- clear that biodiversity loss is a global problem 20. E. F. Zipkin, G. V. DiRenzo, J. M. Ray, S. Rossman, K. R. Lips,
phibians for a large portion of their diet and/or (1). Our results suggest that ecosystem struc- Code and data for Tropical snake diversity collapses after
widespread amphibian loss. Zenodo (2020); doi:10.5281/
the loss of amphibians disrupted the food tures could deteriorate faster than expected zenodo.3628038.
web to such an extent that other taxonomic from indirect and cascading effects generated
groups (e.g., lizards, another major food source) by disease, invasive species, habitat loss, and AC KNOWLED GME NTS
have also declined. The loss of amphibians and climate change. Fast-moving policies are essen- We thank the many people who contributed to data collection.
L. Brown, M. Farr, S. Saunders, A. Shade, A. Wright, and E. Zylstra
snakes might well cascade upward through tial for effective adaptation to ongoing species provided comments on the manuscript, and M. Newman helped
effects on higher-order predators, such as changes and to mitigate the impacts of the with figure design. Funding: E.F.Z. was funded by NSF EF-1702635
raptors and mammals (17), potentially causing world’s biodiversity crisis (19). during model development. G.V.D. was supported by NSF PRFB-
1611692. Field work was funded by NSF DEB-0717741 and
substantial changes to the food web structure.
DEB-0645875 to K.R.L. and IBN-0429223, IBN-0429223, and IOB-
Indeed, top-down effects from amphibian losses 0519458 to J.M.R. and A. Savitzky. The Smithsonian Tropical
RE FERENCES AND NOTES
on the food web are well documented, includ- Research Institute and Ministerio de Ambiente provided logistical
1. G. Ceballos et al., Sci. Adv. 1, e1400253 (2015). support in Panama. Author contributions: All authors conceived
ing changes to algae and detritus biomass,

Downloaded from http://science.sciencemag.org/ on February 13, 2020

2. R. K. Colwell, R. R. Dunn, N. C. Harris, Annu. Rev. Ecol. of the research. K.R.L. and J.M.R. led data collection. E.F.Z.,
reduced energy flow between streams and Evol. Syst. 43, 183–203 (2012). G.V.D., and S.R. built the models. All authors contributed to the
surrounding forested habitats, and lower rates 3. R. Dirzo et al., Science 345, 401–406 (2014). interpretation of results. E.F.Z., G.V.D., and K.R.L. wrote the paper.
4. F. Isbell et al., Nature 526, 574–577 (2015). All authors contributed edits. Competing interests: The authors
of nitrogen turnover (10, 18). Together, these 5. H. M. Pereira et al., Science 330, 1496–1501 (2010). declare no competing interests. Data and materials availability:
results demonstrate the indirect and cascading 6. C. J. Bradshaw, N. S. Sodhi, B. W. Brook, Front. Ecol. Environ. 7, All data and code are available at https://zipkinlab.github.io/
effects of the invasive Bd pathogen and high- 79–87 (2009). #community2020Z and are archived at Zenodo (20).
7. International Union for Conservation of Nature (IUCN), The
light the negative consequences of amphibian IUCN Red List of Threatened Species, version 2019-2 (IUCN,
losses on other taxonomic groups through both 2019); www.iucnredlist.org [accessed 4 November 2019]. SUPPLEMENTARY MATERIALS
top-down and bottom-up processes. 8. C. Hof, M. B. Araújo, W. Jetz, C. Rahbek, Nature 480, 516–519 science.sciencemag.org/content/367/6479/814/suppl/DC1
(2011).
The extent of global biodiversity loss is likely Materials and Methods
9. B. C. Scheele et al., Science 363, 1459–1463 (2019). Supplementary Text
underestimated because cascading effects of 10. M. R. Whiles et al., Front. Ecol. Environ. 4, 27–34 (2006). Figs. S1 and S2
disappearing species can lead to invisible de- 11. A. J. Crawford, K. R. Lips, E. Bermingham, Proc. Natl. Acad. Tables S1 to S4
clines of sympatric species. Tracking these pro- Sci. U.S.A. 107, 13777–13782 (2010). References (21–81)
12. K. R. Lips et al., Proc. Natl. Acad. Sci. U.S.A. 103, 3165–3170
cesses is particularly challenging because certain (2006). 2 July 2019; accepted 23 January 2020
taxa and geographic locations are understudied, 13. M. Böhm et al., Biol. Conserv. 157, 372–385 (2013). 10.1126/science.aay5733

Zipkin et al., Science 367, 814–816 (2020) 14 February 2020 3 of 3

Tropical snake diversity collapses after widespread amphibian loss
Elise F. Zipkin, Graziella V. DiRenzo, Julie M. Ray, Sam Rossman and Karen R. Lips

Science 367 (6479), 814-816.

DOI: 10.1126/science.aay5733

Cascading impacts of prey loss

The global pandemic caused by the amphibian fungal pathogen Batrachochytrium dendrobatidis has decimated
frog populations around the world. This decline has been called out as a potential catastrophe for amphibian species.
What has been less explored are the impacts of amphibian declines on other members of their ecological communities.
Using survey data collected over 13 years, Zipkin et al. looked at diversity and body condition of a tropical snake

Downloaded from http://science.sciencemag.org/ on February 13, 2020

community heavily affected by chytridiomycosis. They found that affected snake communities were less diverse and most
species were in decline, except for a few ''winning'' species.
Science, this issue p. 814

ARTICLE TOOLS http://science.sciencemag.org/content/367/6479/814

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2020/02/12/367.6479.814.DC1
MATERIALS

REFERENCES This article cites 65 articles, 8 of which you can access for free
http://science.sciencemag.org/content/367/6479/814#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. The title Science is a registered trademark of AAAS.
Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of
Science. No claim to original U.S. Government Works