A snapshot of online wildlife trade: Australian e-commerce trade of native and non-native pets

Abstract

The international trade of non-domesticated pets impacts both conservation and biosecurity via the harvest and release of live animals beyond their native distributions. The extent to which individual countries mitigate these impacts via regulation of trade is inconsistent, as is their capacity to monitor internet facilitated trade. We investigated the online trade of vertebrate pets within Australia, a country with a reputation for relatively stringent pet-importation regulations and world-class border biosecurity. Using semi-automated data mining (i. e., webscraping) techniques, we collected online pet trade data over the course of 14 weeks from 12 Australian e-commerce platforms selected using an a priori set of search terms. We analysed spatial, temporal and taxonomic biases in trade and identified instances of high rates of trade in: (i) threatened species, (ii) non-native species, (iii) and species not permissible for live import. We identified over 100,000 individual live animals across 1192 species, including: 667 non-native species for sale within Australia from 03/12/2019 to 20/03/2020 (mammals were excluded from our analysis). Our findings constitute a much greater scale (in terms of abundance and richness) of non-native species trade than previously recorded in Australia. Substantial changes to legislative control of domestically traded pets are needed at the national level to reduce the volume of non-native pets that may contribute to the establishment of invasive species in Australia. We suggest that contemporary examples of permit systems applied to native taxa may provide a valuable template for the implementation of such changes.

Full text

Biological Conservation 282 (2023) 110040
Available online 24 April 2023
0006-3207/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
A snapshot of online wildlife trade: Australian e-commerce trade of native
and non-native pets
Adam Toomes
a
,
*
, Stephanie Moncayo
a
, Oliver C. Stringham
a
,
b
,
c
, Charlotte Lassaline
a
,
Lisa Wood
a
, Mariah Millington
d
, Charlotte Drake
a
, Charlotte Jense
e
, Ashley Allen
f
,
Katherine G.W. Hill
a
, Pablo García-Díaz
g
, Lewis Mitchell
b
, Phillip Cassey
a
a
Invasion Science and Wildlife Ecology Group, The University of Adelaide, North Terrace, Adelaide, SA 5005, Australia
b
School of Mathematical Sciences, The University of Adelaide, North Terrace, Adelaide, SA 5005, Australia
c
Institute of Earth, Ocean, and Atmospheric Sciences, Rutgers, The State University of New Jersey, New Brunswick, NJ, USA
d
Australian Rivers Institute, Grifth University, Brisbane, QLD 4111, Australia
e
School of Natural Sciences, Department of Biological Sciences, University of Tasmania, Sandy Bay, TAS 7005, Australia
f
School of Social Sciences, The University of Adelaide, North Terrace, Adelaide, SA 5005, Australia
g
School of Biological Sciences, University of Aberdeen, Zoology Building, Aberdeen AB24 2TZ, UK
ARTICLE INFO
Keywords:
Biosecurity
E-commerce
Exotic pets
Invasive non-native species
Wildlife trade
ABSTRACT
The international trade of non-domesticated pets impacts both conservation and biosecurity via the harvest and
release of live animals beyond their native distributions. The extent to which individual countries mitigate these
impacts via regulation of trade is inconsistent, as is their capacity to monitor internet facilitated trade. We
investigated the online trade of vertebrate pets within Australia, a country with a reputation for relatively
stringent pet-importation regulations and world-class border biosecurity. Using semi-automated data mining (i.
e., webscraping) techniques, we collected online pet trade data over the course of 14 weeks from 12 Australian e-
commerce platforms selected using an a priori set of search terms. We analysed spatial, temporal and taxonomic
biases in trade and identied instances of high rates of trade in: (i) threatened species, (ii) non-native species, (iii)
and species not permissible for live import. We identied over 100,000 individual live animals across 1192
species, including: 667 non-native species for sale within Australia from 03/12/2019 to 20/03/2020 (mammals
were excluded from our analysis). Our ndings constitute a much greater scale (in terms of abundance and
richness) of non-native species trade than previously recorded in Australia. Substantial changes to legislative
control of domestically traded pets are needed at the national level to reduce the volume of non-native pets that
may contribute to the establishment of invasive species in Australia. We suggest that contemporary examples of
permit systems applied to native taxa may provide a valuable template for the implementation of such changes.
1. Introduction
The international wildlife trade, particularly the trade of live animals
as non-domesticated pets, has garnered growing research interest across
the last decade (e.g., Mohanty and Measey, 2019; Marshall et al., 2020);
primarily due to the conservation, criminological and biosecurity
threats posed by unsustainable trade practices (Warwick et al., 2018;
Lockwood et al., 2019). Contemporary investigation of wildlife trade has
largely focused on the cross-border movement and trade of species by
utilising import/export permit recording systems such as for CITES-
listed species or the US wildlife import-export recording system
(Harfoot et al., 2018; Watters et al., 2022). Documentation of illegal
components of the international pet trade have relied on seizure data
compiled by various border-security agencies of a wide variety of na-
tions (Ribeiro et al., 2019; Hitchens and Blakeslee, 2020), although this
data is rarely collected on a consistent basis subject to an international
standard (e.g., Nijman and Shepherd, 2021). Such sources of data have
nonetheless provided substantial improvements in our understanding of
pet trade trends and spatio-temporal dynamics (Harfoot et al., 2018;
Andersson et al., 2021). However, a considerable (yet not fully quanti-
ed) proportion of trade of internationally-sourced species takes place
within the domestic borders of individual nations (de Magalh˜
aes and
* Corresponding author at: Benham G18 North Terrace, Adelaide 5005, Australia.
E-mail address: adam.toomes@adelaide.edu.au (A. Toomes).
Contents lists available at ScienceDirect
Biological Conservation
journal homepage: www.elsevier.com/locate/biocon
https://doi.org/10.1016/j.biocon.2023.110040
Received 5 December 2022; Received in revised form 3 March 2023; Accepted 29 March 2023

Biological Conservation 282 (2023) 110040
2
S˜
ao-Pedro, 2012; Papavlasopoulou et al., 2014; Janssen and Leupen,
2019). Regulation and documentation of such domestic trade is con-
ducted on a case-by-case basis by individual nations (if at all) and is
often subject to taxonomic biases (as identied in Fukushima et al.,
2020).
Australia is a country widely regarded as having highly stringent
border security policies, which strictly controls the importation (and
exportation) of most live animals for commercial purposes (Whittington
and Chong, 2007; Schneider et al., 2018). These regulations, imple-
mented by the Commonwealth government, go far beyond Australia's
obligations as a signatory to CITES (UNEP-WCMC, 2022). However,
non-native species are nonetheless present in Australia, many of which
were imported prior to the implementation of such policies. There is also
a shortage of documentation for the domestic trade of both native and
non-native species taking place within Australia (Vall-llosera and Cas-
sey, 2017c; Woolnough et al., 2020; Millington et al., 2022a). Australia
is federated into six States and eight Territories (two mainland and six
external), and while Commonwealth-wide regulations are in place for
some taxa (e.g., the trade and private possession of non-native reptiles is
universally prohibited across Australia; see Toomes et al. (2019)), most
regulations pertaining to the pet trade are managed and enforced at the
individual State/Territory jurisdiction (see Toomes et al. (2022) and
Woolnough et al. (2020) for specic examples). This jurisdiction-specic
management ranges from simple prohibited lists to more complex
permit systems that would-be traders need to acquire before buying
specic taxa. As such, Australia does not consistently document the
trade of live pets across all taxa and jurisdictions, allowing an unknown
proportion of trade to occur without guarantee of sustainable or ethical
practice.
Such lack of oversight in wildlife trade is concerning for several
biosecurity and conservation-related reasons. From a biosecurity
perspective, non-native species, including species that are invasive
elsewhere in the world, are known to be illegally smuggled into
Australia, held in private captivity and escape into Australian ecosys-
tems (Toomes et al., 2019). There is also public desire to possess other
highly invasive species as non-domesticated pets in Australia (Toomes
et al., 2020), and non-native species that were brought into Australia
prior to importation bans are known to be widely (and legally) traded
and bred domestically (Woolnough et al., 2020). From a conservation
perspective, Australian native species are highly desirable and valuable
on the international pet market (Vall-llosera and Cassey, 2017a;
Marshall et al., 2020; Heinrich et al., 2021) and there is a known do-
mestic trade of threatened native species (Toomes et al., 2022). While
the trade of some Australian species can be supplied by captive breeding,
the slow life history traits and restricted distributions of many Australian
native (particularly endemic) taxa leave them vulnerable to trade-
incentivised harvesting of wild populations (e.g., Holocephalus bungar-
oides; Jolly et al. (2020)). When such biosecurity and conservation
concerns are considered alongside additional threats such as the trans-
mission of pathogens (Norval et al., 2020) and animal welfare concerns
associated with captive keeping/breeding (Wyatt et al., 2022), there is a
clear need to monitor and quantify the risk of domestic trade to ensure
that wildlife trade occurs sustainably and ethically, Yet, to date, no
systematic method of monitoring trade has been implemented by
Australian Commonwealth and State/Territory governments.
Throughout a complex legal landscape, the pet trade (and wildlife
trade more broadly) has undergone a rapid transition from traditional
brick-and-mortar marketplaces (e.g., pet stores) to online e-commerce
platforms over the last decade (Siriwat and Nijman, 2018, 2020; Fink
et al., 2021). Such online platforms include direct business-to-consumer
sites (e.g., online pet stores) as well as more centralised community-
based sites (e.g., large classieds) (Stringham et al., 2021). The ease-
of-access, potential anonymity and large consumer base afforded by e-
commerce has increased both the scale and diversity of pet trade (Paul
et al., 2020; Atoussi et al., 2022). Fortunately, this also provides re-
searchers with an opportunity for large-scale surveillance of trade
activity, assisted by the development of open-source data mining (a.k.a.
webscraping) resources. Such tools have recently been used to rapidly
collect large quantities of trade data beyond the capabilities of tradi-
tional manual surveillance (e.g., Marshall et al., 2020; Hughes et al.,
2021; Marshall et al., 2022) and can facilitate the analysis of taxonomic,
spatial and temporal wildlife trade dynamics in lieu of formal trade
monitoring and regulation.
Here, we took advantage of the increasing abundance of online data
to glean insights into the Australian vertebrate pet trade. We identied
Australia as a suitable candidate for the implementation of data mining-
based surveillance of the online pet trade due to the aforementioned lack
of consistent monitoring and the clear biosecurity and conservation
concerns. We developed t-for-purpose data mining tools to provide a
near-comprehensive snapshot of advertised pets for sale across major
Australian surface-web e-commerce platforms (see Stringham et al.
(2021) for descriptions of surface and deep web). Our objective was to
simultaneously use Australia as a case study to highlight domestic trade
as a crucial yet understudied facet of international pet trade, while also
assisting relevant Australian biosecurity and conservation stakeholders
by identifying trade of key species. Specically, we aimed to quantify
not only the diversity of pets traded in Australia but also the relative
quantity of individuals possessed, in order to examine the proportion of
trade that involves non-native and threatened taxa.
2. Methods
2.1. Surface web E-commerce
To identify relevant surface web e-commerce platforms (i.e., web-
sites) that trade live animals as pets, we followed the framework
developed in Stringham et al. (2021). Specically, we dened a series of
search phrases centred around our taxa of interest (freshwater aquarium
shes, marine aquarium shes, pet reptiles, pet amphibians, and pet
birds) and type of websites (pet stores, classieds or forums) within
Australia. We limited the taxonomic scope of our study to vertebrates as
they are the most commonly recorded taxa in trade, and because there
are (relatively) strongly resolved taxonomic databases that would
facilitate identication of advertised pets on a sufciently large scale for
the quantity of data collected. We did not search for mammalian pets
due to the very high quantity of e-commerce sites dedicated to the trade
of highly domesticated mammals (e.g., dogs, cats, rabbits, hamsters). In
total, we created 105 search phrases (see Appendix A for full list), which
we used to search for candidate websites using the Google search engine
during August 2019. For each search, we recorded the rst 50 results (i.
e., 5 pages of results with 10 URLs per page) and retrieved Alexa web
ranking, the number of page visits per month and the number of new
listings posted in August 2019 (if available; see Stringham et al. (2021)
for further details of web trafc statistics). In total this resulted in the
selection of 12 websites (eight pet stores, three classieds and one
forum).
2.2. Webscraping trade data
Once candidate websites were identied, we developed t-for-
purpose webscraping code in the Python programming language
(Sheridan, 2016) using the Selenium Webdriver, Beautiful Soup and
Requests modules (Patel, 2020), to acquire pet trade data (i.e., instances
of pets being advertised for sale online). Further details of this procedure
are provided in Appendix B. We recorded the following attributes, where
available, from each listing of all platforms (see Appendix C): scientic
name, common/trade names, quantity, price, location (at either State/
Territory or suburb level), listing date. We also collected image URLs to
assist with species identication in cases where scientic names were
not present and taxa could not be reasonably derived from free-form
listing text. We generated unique identication codes for each listing
based on a combination of the listing text and website-specic identier,
A. Toomes et al.

Biological Conservation 282 (2023) 110040
3
where available. If platforms did not provide a date of listing creation,
we assumed this to be the rst date that data was collected. Webscrapers
were constructed in a manner that did not unduly impact the selected
platforms and were compliant with the University of Adelaide HREC
approval (Projects H-2020-184 and H-2020-256). We determined the
frequency of sampling (daily, weekly or fortnightly) based on the fre-
quency of trade occurring on each individual platform to ensure we did
not miss new advertisements. Although our webscrapers also recorded
‘wanted ads’ i.e., listings where potential buyers express an interest in a
product, we limited our analysis to advertisements where pets were
being offered for sale. We identied wanted ads based on the presence of
the text strings ‘wanted’ or ‘wtb’ (meaning wanted to buy) in listing
descriptions, as most websites did not distinguish between wanted ads
and normal advertisements.
2.3. Generating a list of taxa names
We compiled a list of the scientic names of advertised pets and
manually standardised them to the Global Biodiversity Information
Facility (GBIF, 2021). Where a hybrid was advertised for sale, we
recorded the hybrid status and GBIF identication of both parent taxa, if
known. Additionally, we included as synonyms for each unique GBIF
record any terms frequently used by the community of online pet traders
and keepers that are context specic, including common names, incor-
rect/outdated scientic names and ‘trade names’. Outdated scientic
names were matched to current scientic names by manually cross
referencing advertised names against GBIF. Informal trade names were
matched to scientic names using hobby-specic knowledge from
naturalist and trade forums, as well as the authors own knowledge of
Australian trade. For example, ‘IRN’ is used in trade to refer to the Indian
ringneck parrot (Psittacula krameri).
Although we did not use data from ‘wanted ads’ in our analysis, we
did inspect the text of these listings in order to assist with the compi-
lation of standardised taxa names and synonyms used to search for taxa
that may be advertised for sale. In total we generated a library of 1583
scientic names, 1408 common names and 2743 trade names for a total
of 1381 species, 42 subspecies and 44 hybrids, with additional taxa only
identiable to genus (n =79), family (n =25) or higher (n =8) level.
While we have taken every effort to reduce the chances of non-target
character string matches occurring, we do acknowledge that this may
occur and lead to an overestimation of the frequency of trade in some
species. However, scientic, common and trade names were only
included in our library and used in string matching if they had been
encountered for sale or in wanted ads at least once during our pre-
liminary analysis. As such, we anticipate false matches to be infrequent.
2.4. Curation and analysis of advertised listings
All data curation and analyses were conducted in the R statistical
software version 4.0.3 (R Core Team, 2022), using base functions unless
otherwise specied. All data visualisation was generated using the
ggplot2 package (Wickham, 2016). We extracted webscraped data for a
14-week snapshot: 3rd December 2019–20th March 2020. This study
period was selected based on the date at which all our webscrapers
became operational until the date that Australia closed its borders to
non-resident human travel. Australia was not entirely unaffected by
COVID prior to 20th March 2020 (e.g., air trafc was reduced when
other nations closed their borders earlier in 2020) and therefore it is
impossible to capture circumstances that entirely represent pre-COVID
trade conditions. However, to the best of our knowledge, no other
research or government entity was systematically collecting online trade
data in Australia across this many platforms prior to Australia closing its
borders. Therefore, we believe our dataset to be the best available rep-
resentation of pre-COVID conditions and is referred to as a pre-COVID
snapshot hereafter.
We used literal character string (i.e., letter and number) matching
with the stringr package (Wickham, 2022) to identify listing titles or text
that contained scientic, common and trade names (in that respective
order of priority) from our reference library, at the taxonomic resolution
of species and subspecies. For the remaining unmatched listings, we
performed fuzzy string matching with the same list of names using a
Levenshtein edit distance of two (i.e., matches any string within any
combination of two-character additions, deletions or substitutions),
excluding names of six or fewer characters in length. We also manually
inspected cases where a fuzzy-string match yielded a notably higher
number of listings and excluded this string if matches did not contain the
target taxa. Finally, we repeated this process for unmatched listings
against names at the resolution of family and genera. For listings that
failed to match any literal or fuzzy string, we omitted them based on a
pre-dened list of exclusion terms (Appendix D) and manually inspected
the remaining unidentied listing text to determine if any pet was
advertised for sale. If one or more pets were advertised for sale, we
manually assigned them to the most specic taxonomic rank possible. In
some instances, a pet was advertised that had not yet been taxonomi-
cally described yet is present in trade and referred to using hobby-
specic terms/jargon (e.g., undescribed catsh). In such instances, we
recorded taxonomy at a coarser level (genus, family or order, where
possible).
For listings that matched multiple names, we manually inspected the
text and recorded each unique taxon that was advertised for sale,
ensuring that the unique listing identier was recorded for each taxon.
We omitted highly domesticated taxa from our analysis, namely pigeons
(Columba livia) and chickens (Gallus gallus). We generated species
accumulation curves by randomly sampling listings without replace-
ment and plotted the number of species detected against sampling effort.
For websites that provided a unique listing identier, we used this to
distinguish between unique listings, otherwise we used the unique
combination of listing title and text to distinguish between unique list-
ings. However, this does not account for the possibility that the same
product may be advertised multiple times in different listings that have
small differences in text description. Due to the considerable quantity of
listings selling pets (62,584, not including listings selling pet products),
we deemed it logistically infeasible to manually verify the uniqueness of
listings or to manually establish additional information such as the
quantity of pets for sale. If listings specied a ‘pair’ or ‘trio’ of animals,
quantity was assumed to be two or three respectively. Listings referring
to animals using a plural term (e.g., dragons, parrots) were assumed to
be advertising two individuals, noting that the actual number may be
higher. Listings that referred to a ‘colony’ or other collective terms were
conservatively assumed to be advertising ve individuals. We did not
determine listing quantity based on the presence of numerical character
strings (i.e., digits) due to the prevalence of information in free form text
that contained digits yet was unrelated to quantity (e.g., addresses,
phone numbers). Given the diversity of platforms, taxa and locations
covered by our online surveillance, as well as human ethical consider-
ations of contacting pet traders directly, we were unable to manually
verify the veracity of advertisements.
We collated International Union for Conservation of Nature (IUCN)
threat status of all traded species, and Global Invasive Species Database
(GISD) records of invasive species, to categorise advertised pets based on
their conservation status and history of invasions respectively. For birds
we also compared the species identied for sale with the ofine
aviculture records previously collated by Vall-llosera and Cassey
(2017c). We cross referenced scientic names and, where necessary,
upstream taxonomy against the Australian Commonwealth ‘List of
Specimens Taken to be Suitable for Live Import’ (Live Import List
hereafter). For the subset of listings that were identied to species level
and contained a specied location, we determined the rate of trade per
region (i.e., city, town or municipality). The native/non-native status of
reptile and bird species were determined by visually inspecting the
distribution records listed in GBIF (2021), excluding introduced pop-
ulations. Due to the large diversity of sh taxa detected, we cross-
A. Toomes et al.

Biological Conservation 282 (2023) 110040
4
referenced scientic names against the Australian Faunal Directory
(AFD) list of native species, including scientic name synonyms, in order
to determine native/non-native status (Australian Faunal Directory,
2021). Similarly, we also identied non-native species that are known to
be introduced using the AFD list.
3. Results
We have recorded a notable diversity of non-domesticated pets
traded online in Australia, with 1192 species detected, including 667
non-native species (56.0 %). Species accumulation curves reveal a
plateau in new bird species throughout our 14-week sampling period.
Notably, sh and reptile species continued to accumulate without
plateaux (Fig. 1). We detected a total of 62,584 listings advertising at
least 109,056 live animals (52,409 non-native; 47.6 %) at the species
level, including a minimum of 66,894 individual birds (24,899 non-
native; 37.2 %), 30,343 sh (27,455 non-native; 90.5 %), 11,603 rep-
tiles (all native), and 216 amphibians (55 non-native; 25.5 %). For
listings that contained location information, most trade occurred in
highly populous cities, namely Sydney (22,797 animals), Melbourne
(13,866 animals), Brisbane (10,424 animals) and Perth (9854 animals).
The highest volume of trade was concentrated in the most populous
Australian States, namely New South Wales (35,181 animals), Queens-
land (26,781 animals), and Victoria (17,188 animals) (see Appendix E
for summaries of trade frequency per region). The vast majority of trade
took place on classieds sites (60,306 listings; 96.4 %), followed by pet
stores (2089 listings; 3.34 %) and forums (189 listings; 0.302 %). There
was a high diversity of species that were not found on more than one
website (600 species, 50.3 %), implying a high level of e-commerce
specialisation catering to specic hobbies or consumer types.
Fish were the most species-rich taxon traded with 885 distinct taxa
— 805 species, one subspecies and eight hybrids, including taxa that
could only be identied at the level of genus (n =53), family (n =15),
and order (n =3). 553 of identied species are non-native (62.5 %;
constituting 18,850 listings). A total of 279 non-native sh species are
illegal to import into Australia based on the Live Import List yet were
detected in our trade snapshot. Perciformes were the most species-rich
order of sh in trade (perch and relatives, 483 species), followed by
Siluriformes (catshes, 88 species), Characiformes (characins, 57 spe-
cies) and Cypriniformes (carp and relatives, 56 species), which collec-
tively account for 85.0 % of identied sh species richness (Fig. 2).
We detected 228 distinct taxa of birds — 184 species, 11 subspecies,
nine hybrids and two domesticated breeds, including taxa that could
only be identied at the level of genus (n =18) and family (n =4). 113
of identied species are non-native species (61.4 %; constituting 16,345
listings). The most species-rich bird order in trade was Psittaciformes
(parrots, 99 species), followed by Passeriformes (passerines, 48 species)
and Galliformes (fowl and relatives, 16 species). The native red-collared
lorikeet (Trichoglossus rubritorquis) and four species of non-native birds
were not already listed on the 2007 inventory of known bird species
traded in Australia, implying that they have been newly introduced into
the trade since this inventory was created (DAWE, 2021). While the
updated classication of T. rubritorquis (previously the rainbow lorikeet
(Trichoglossus moluccanus)), may have obscured their trade in this earlier
inventory, there is no such explanation for the non-native Pacic par-
rotlet (Forpus coelestis), olive-headed lorikeet (Trichoglossus euteles),
yellow-fronted canary (Crithagra mozambica) or orange-breasted waxbill
(Amandava subava). Of the 197 non-native bird species previously
identied by Vall-llosera and Cassey (2017c), 91 species were not
detected in our online surveillance.
We detected 237 distinct taxa of reptiles — 186 species, 25 subspe-
cies and 14 hybrids, including taxa that could only be identied at the
level of genus (n =7), family (n =3), suborder (n =1), and order (n =
1). All detected species were native, although we did detect two ex-
pressions of interest (i.e., ‘wanted’ advertisements) for the prohibited
non-native corn snake (Pantherophis guttatus). Lizards (122 species) were
Fig. 1. Species accumulation curve for reptile, bird and sh taxa detected in
Australian e-commerce trade. Raw data is displayed after randomly sampling
species without replacement from all listings.
A. Toomes et al.

Biological Conservation 282 (2023) 110040
5
(caption on next page)
A. Toomes et al.

Biological Conservation 282 (2023) 110040
6
the most species-rich reptile taxa in trade, followed by Serpentes
(snakes, 44 species), Testudines (turtles, 18 species) and Crocodilians
(crocodiles, 2 species).
Amphibian trade was relatively sparse, with 18 distinct taxa detec-
ted, including 17 species, one of which is non-native (5.88 %; consti-
tuting 55 listings). Frogs (Anura) were most species-rich taxa in trade,
with 16 species. The only other amphibian species was the axolotl
(Ambystoma mexicanum), the sole non-native amphibian. There was a
low diversity and abundance of native amphibians relative to reptiles in
Australia, with the magnitude of the disparity between taxa not repre-
sented in other studies (Hughes et al., 2021). This may be due to the low
diversity of Australian amphibian fauna (247 species of anurans
compared to 1034 species of reptile; AmphibiaWeb, 2023; Melville
et al., 2021).
Twenty of the traded non-native pet species identied here are
invasive elsewhere in the world, according to GISD (Appendix F). In
addition, a total of 22 traded non-native sh species have introduced
populations in Australia, including species that are invasive elsewhere
such as jaguar cichlids (Parachromis managuensis) (Holmes et al., 2020)
and species whose invasion potential has yet to be realised, such as Si-
amese ghting sh (Betta splendens) (Hammer et al., 2019). Of the 1192
species identied in our trade snapshot, 81 were classied by the IUCN
as threatened (12 Critically Endangered, 35 Endangered, 34 Vulner-
able), and 35 classied as Near Threatened. Most taxa were classied as
Least Concern (797), with the remaining taxa classied as Data Decient
(38) or simply Not Listed (241). Many examples of species not listed,
such as Peckoltia compta and Symphysodon discus, have highly restricted
known range sizes and it is possible that their eventual assessment will
categorise them as Threatened.
4. Discussion
4.1. Scale of the non-native pet trade
Our online surveillance has captured a considerable richness of
traded non-native pets (667 species) and, to the best of our knowledge,
provided the only contemporary and systematic survey of online pet
trade frequency in Australia. While there are existing audits of non-
native species such as compiled avicultural records (197 bird species;
Vall-llosera and Cassey, 2017c) and a species inventory compiled by the
Australian government in collaboration with the ornamental sh in-
dustry (447 sh species; Millington et al., 2022b), our online surveil-
lance reveals that contemporary understanding of the domestic non-
native pet trade is far from comprehensive. The lack of saturation in
the accumulation of new species (for sh and reptiles) despite extensive
sampling of tens of thousands of advertisements suggests that the true
diversity of non-native taxa traded in Australia has yet to be determined
and implies that the biosecurity threat posed by the pet-release pathway
continues to be underestimated. This is further evidenced by our sur-
veillance failing to detect 91 species identied from ofine aviculture
records (Vall-llosera and Cassey, 2017c). Additional trade may be taking
place across the deep web, namely social media platforms (see Appendix
G for considerations of Deep Web surveillance).
Further temporal sampling is underway to facilitate analysis of
greater quantities of data taking place across multiple years. However,
the immediate and long-term effects of COVID-19 on the Australian pet
trade have yet to be investigated, which may frustrate efforts to
exhaustively quantify the full suite of traded taxa if online trade is
occurring less frequently than previously. Most e-commerce platforms
provide user feedback metrics as a proxy for online reputation, meaning
there is incentive for traders to advertise pets accurately (Bojang et al.,
2017). Nonetheless, we acknowledge that the advertised information
does not necessary accurately reect the attributes of the pet for sale,
and that some fake/misleading advertisements may be present within
our dataset.
Although our research focused on the trade and regulation of non-
native species nationally in Australia, we also note that the majority
of the 667 traded non-native species are not regulated at a State/Ter-
ritory level. Even high-risk species that are regulated or prohibited are
not done so uniformly across jurisdictions. For example, P. krameri is
prohibited in Tasmania and Western Australia yet can be traded without
regulation or permits in other States (Woolnough et al., 2020). Such
inconsistent regulation is rarely successful; rather creating opportunities
for subversion of trade via other jurisdictions (e.g., Raghavan et al.,
2013). We recommend that State/Territory governments use our
collected data to cross-reference against their jurisdiction-specic reg-
ulations and identify non-compliant trade. Alternatively, we recom-
mend that research and government authorities work collaboratively to
collate all legislation pertaining to the domestic keeping and trading of
pets across all Australian jurisdictions, in order to provide a resource
that can be readily cross-examined against trade data analogous to the
data collected in our research.
The lack of regulation not only hinders the ability of Australian
biosecurity authorities to control the trade of high-risk species, such as
well-known invasive species listed in GISD, but it also deprives those
authorities of a systematic means of recording data pertaining to trade
and escapes. For example, South Australia's permit system for the
keeping of native species obligates permit holders to keep a record of the
number of individuals that have been sold, bred and escaped over a
given reporting period, yet no equivalent system is in place for non-
native species. As such, the trade-related propagule pressure remains
unquantied for hundreds of non-native species. The ndings of Toomes
et al. (2022) suggest that, for native pets, propagule pressure is pro-
portional to the quantity of possession. Assuming this pattern extends to
non-native species, our surveillance data provides a proxy measure of
relative propagule pressure and may assist with the creation of priority
lists for future management strategies/interventions.
4.2. Comparison with illegal seizures
The 111 species of non-native reptile detected during smuggling
attempts or from illegal captivity in Australia (Toomes et al., 2019) were
not detected in our surface web surveillance. Recent investigation of
illicit e-commerce suggest that illegal pet trade is similarly rare on dark
web platforms (Harrison et al., 2016; Stringham et al., 2022), though
deep web (i.e., social media) trade warrants further investigation (see
Section 4.3).
In contrast to the paucity of nationally prohibited species recorded
here, non-uniformly prohibited species (e.g., P. krameri in Western
Australia and Tasmania) were routinely recorded in prohibited juris-
dictions, albeit in lower abundances than permitted jurisdictions. While
part of this trade may be due to a lack of awareness surrounding the
specic and varying trade regulations in different jurisdictions, their
availability may instead illustrate the blatant disregard for trade regu-
lations. Future communication with the traders responsible for in-
fringements may reveal the extent to which taxa are traded knowingly.
Regardless, our results show a clear parallel between Australia's policy
regarding domestic trade of non-native species and both the quantity
and diversity of contemporary trade. Non-native sh and birds, while
mostly illegal to import, are legal to trade without quota or
Fig. 2. Total number of listings (A and C) and species richness (B and D) of e-commerce trade by taxonomic order for native and non-native species (A and B), and for
threatened and non-threatened species (C and D), displayed on a square-root scale. Threat status was determined based on the IUCN Red List, with the Endangered,
Critically Endangered and Vulnerable categories being classed as threatened. (For interpretation of the references to colour in this gure legend, the reader is referred
to the web version of this article.)
A. Toomes et al.

Biological Conservation 282 (2023) 110040
7
documentation unless specically declared as prohibited (usually via
the Biosecurity Act 2015 (DAWR, 2019)) by a State or Territory. In
contrast, all non-native reptiles are prohibited except for non-
commercial purposes. This inconsistency in policy is worthy of further
interrogation because there is no evidence that biosecurity threat posed
by reptile and non-reptile taxa are fundamentally different, as evidenced
by the number of introduced and known invasive vertebrates currently
present in Australia (Vall-llosera and Cassey, 2017b). Additionally,
educating the public and the pet supply chain on trade regulations
specic to each State and Territory may aid in reducing the incidence of
non-uniformly prohibited species advertisements in prohibited
jurisdictions.
4.3. Trade of threatened taxa
The impacts of wildlife trade, be they biosecurity, animal welfare or
conservation related, are often difcult to identify (Morton et al., 2021).
Many threatened taxa are traded globally, yet trade is not a threatening
process if conducted sustainably (i.e., via captive breeding (Tensen,
2016)). We found examples of both native and non-native species in our
analysis that are known to be threatened by wild harvest, including the
broad-headed snake (Hoplocephalus bungaroides; Jolly et al., 2020) and
Lake Malawi cichlids (Cichlidae; Msukwa et al., 2021). However, we
cannot estimate the proportion of trade recorded in our analysis that was
captive-bred versus wild-caught, as most traders did not provide this
information. Indeed, there is no onus to provide traded pet species origin
information in Australia despite calls for green certication (Millington
et al., 2022a), which would simultaneously educate the general public
and allow potential consumers to make an informed decision to pur-
chase pets based on sustainability. One measure to ensure that the pet
trade is not a driver of unsustainable trade is the use of a permit system
to regulate the trade of threatened taxa (e.g., by issuing permit quotas or
by requiring proof of captive-bred provenance). Currently, permit sys-
tems only exist in some Australian jurisdictions for certain taxa, such as
in South Australia (Toomes et al., 2022). Various State and Territory
departments tasked with wildlife management could use South Aus-
tralia's system as a template, with the decision to control or reduce trade
based on species' life history traits and rate-of-trade data.
4.4. Taxonomy and trade
Pet traders are often abreast of contemporary taxonomy, however
there are inevitably instances whereby outdated taxonomy is used when
advertising pets for sale. There are also instances where a trade/hobby
community acknowledge a taxonomic revision yet continue to use a
longstanding yet outdated scientic synonyms, for example ‘Nephrurus
milii’ is often used to refer to barking geckos (Underwoodisaurus milii).
Many hybrids are also commonly traded, yet the origin species that
constitute the hybrid are not always conclusively known. This is
exemplied by the popular owerhorn cichlid (see Fig. 3), which is
believed to originate from a multi-generation hybrid of several Cichla-
soma species with Vieja synspila (Nico et al., 2007). Other examples
include red Texas cichlids (Cichlidae sp.), lemon bristlenose catsh
(Ancistrus sp.) and pigeon blood discus (Symphysodon sp.). Such in-
stances need to be considered during future efforts to monitor online
trade, and synonyms should be considered wherever possible when
querying character strings against large volumes of trade data.
There were many ornamental sh that have not been formerly
described and yet are nonetheless widely known and traded both in
previous research and during our surveillance (Tan and Armbruster,
2016). This lack of taxonomic resolution sties efforts to evaluate both
the biosecurity threat of traded sh, as well as the risk trade poses to
their conservation. For example, there are several undescribed cichlid
sh from Lake Malawi that are known only as captive-bred colour
morphs (Msukwa et al., 2022). Similarly, there are a diversity of catsh
that can only be identied to genus level yet are partitioned into
‘pseudo’ taxonomic units by traders using so-called ‘L numbers’ (Glaser
and Glaser, 1995), representing as-yet undescribed taxa within the
family Loricariidae that do not necessarily map to distinct species
(Cardoso et al., 2016).
Fig. 3. Examples of traded pet sh that are
difcult to taxonomically identify yet are none-
theless referred to by traders using pseudo-
taxonomic units. Clockwise from top-left: ow-
erhorn cichlid (multi-species hybrid of Cichla-
soma species with Vieja synspila); hongi
(undescribed Labidochromis sp. erroneously
referred to as Labidochromis hongi); pigeon blood
discus (captive-bred colour morph of unknown
Symphysodon sp.); gold nugget pleco (Bar-
yancistrus xanthellus, previously referred to as
L018 and L085 before being formerly described
in 2011 (Py-Daniel et al., 2011)). Image credit,
clockwise from top-left: patanasak (Getty Im-
ages); ArtEvent ET (Getty Images); vojce (Getty
Images); Mirko_Rosenau (Getty Images). (For
interpretation of the references to colour in this
gure legend, the reader is referred to the web
version of this article.)
A. Toomes et al.

Biological Conservation 282 (2023) 110040
8
Undescribed and/or hybrid sh are nonetheless known to be intro-
duced (Maciaszek et al., 2019) or invasive (Herder et al., 2012) else-
where in the world. Similarly, undescribed species can still face
conservation threats: approximately 28,000 individual sh are har-
vested from Lake Malawi each year to supply the ornamental trade, the
majority of which are undescribed, which limits capacity to understand
whether overharvesting is occurring (Msukwa et al., 2021). Consider-
able effort is therefore required to keep abreast of hobbyist naming
conventions, particularly if future taxonomic resolution occurs (e.g.,
recent scientic description of Geophagus sp. “Tapajos Red head” as
Geophagus pyrocephalus (Chuctaya et al., 2022)). To this end, the work
conducted by Nov´
ak et al. (2022) provides a useful template of how
hobbyist pseudo-taxonomic units such as L numbers can be matched (in
some cases) to current taxonomy.
5. Conclusion
Australia's biosecurity priorities are commendable, yet its manage-
ment of non-native pets falls short of a system that comprehensively
reduces known and/or identiable risks. We have provided the rst
instance of a systematic survey identifying a large diversity of non-
native taxa including the rst known systematic record of the fre-
quency of online trade in Australia. Our results include undescribed taxa
as well as hybrids with poorly documented provenance. A high diversity
of threatened taxa are also traded, though the sustainability of trade is
difcult to verify considering the paucity of information regarding
captive-bred status. We recommend continued online surveillance in
lieu of the lack of the saturation in species accumulation, as well as an
expansion of this methodology to deep web platforms, as we likely did
not detect all species in the trade. Ultimately such surveillance can
support evidence-informed policy changes to more closely align the
trade of non-native pets with a nation's biosecurity priorities.
Funding
This project was funded by the Centre for Invasive Species Solutions
(Project PO1-I-001). Adam Toomes was additionally supported by the FJ
Sandoz PhD Scholarship. Pablo García-Díaz was funded by NERC grants
NE/S011641/1 (Newton LATAM programme) and
2022GCBCCONTAIN.
CRediT authorship contribution statement
Adam Toomes: Conceptualization, Methodology, Software, Formal
analysis, Investigation, Data curation, Writing – original draft. Stepha-
nie Moncayo: Methodology, Validation, Data curation, Writing – re-
view & editing. Oliver C. Stringham: Conceptualization, Methodology,
Software, Data curation, Writing – review & editing. Charlotte Lassa-
line: Validation, Data curation, Writing – review & editing. Lisa Wood:
Data curation, Writing – review & editing. Mariah Millington: Data
curation, Writing – review & editing. Charlotte Drake: Data curation,
Writing – review & editing. Charlotte Jense: Data curation, Writing –
review & editing. Ashley Allen: Data curation, Writing – review &
editing. Katherine G.W. Hill: Validation, Data curation, Writing – re-
view & editing. Pablo García-Díaz: Conceptualization, Writing – review
& editing, Supervision. Lewis Mitchell: Conceptualization, Writing –
review & editing, Supervision. Phillip Cassey: Conceptualization, Re-
sources, Funding acquisition, Writing – review & editing, Supervision.
Declaration of competing interest
We declare no conicts of interest.
Data availability
As our data contains potentially identiable or re-identiable
information, we have chosen not to publish it in any publicly available
archive. However, we have published a dataset summarising the rate of
trade for native and non-native species within Australia, which can be
found at: https://doi.org/10.6084/m9.gshare.20956339.v1.
Supplementary Data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.biocon.2023.110040.
References
AmphibiaWeb, 2023. List of amphibians in Australia. (University of California, Berkeley,
CA, USA). Available at: https://amphibiaweb.org/cgi/amphib_query?rel-isocc=like
&orderbyaw=Order&where-isocc=Australia [Accessed 14 February 2023].
Andersson, A.A., Lau, W., Tilley, H.B., Dudgeon, D., Bonebrake, T.C., Dingle, C., 2021.
CITES and beyond: illuminating 20-years of global, legal wildlife trade. Global
Ecology and Consservation 26 (e01455).
Atoussi, S., Razkallah, I., Ameziane, I.N., Boudebbouz, A., Bara, M., Bouslama, Z.,
Houhamdi, M., 2022. Illegal wildlife trade in Algeria, insight via online selling
platforms. Afr. J. Ecol. 60, 175–181.
Australian Faunal Directory, 2021. ‘Pisces. Australian Biological Resources Study’
(Canberra, Australia). Available at: https://biodiversity.org.au/afd/taxa/PISCES
[Viewed 01 April 2021].
Bojang, I., Medvedev, M.A., Spasov, K.B., Matvevnina, A.I., 2017. Determinants of trust
in B2C e-commerce and their relationship with consumer online trust. In: AIP
Conference Proceedings. AIP Publishing LLC.
Cardoso, A.L., Carvalho, H.L.S., Benathar, T.C.M., Serrao, S.M.G., Nagamachi, C.Y.,
Pieczarka, J.C., Sousa, L.Md., Ready, J.S., Noronha, R.C.R., 2016. Integrated
cytogenetic and mitochondrial DNA analyses indicate that two different phenotypes
of hypancistrus (L066 and L333) belong to the same species. Zebrash 13, 209–216.
Chuctaya, J., Nitschke, P., Andrade, M.C., Wingert, J., Malabarba, L.R., 2022. A new
species of geophagus (Teleostei: Cichlidae): naming a cichlid species widely known
in the aquarium hobby as “Geophagus sp. Tapajos red head”. J. Fish Biol. https://
doi.org/10.1111/jfb.15207.
de Magalh˜
aes, A.L.B., S˜
ao-Pedro, V.A., 2012. Illegal trade on non-native amphibians and
reptiles in Southeast Brazil: the status of e-commerce. PhyllomedusaJ. Herpetol. 11,
155–160.
Department of Agriculature and Water Resources, 2019. Biosecurity Act 2015. Available
at. https://www.legislation.gov.au/Details/C2019C00097 [Accessed 01 July].
Department of Agriculture, Water and the Environment, 2021. 2007 Inventory of Exotic
(non-native) Bird Species known to be in Australia. Available at: https://www.agric
ulture.gov.au/sites/default/les/documents/inventory-exotic-common.pdf
[Accessed 01 February 2021].
Fink, C., Toivonen, T., Correia, R.A., Di Minin, E., 2021. Mapping the online songbird
trade in Indonesia. Appl. Geogr. 134, 102505.
Fukushima, C.S., Mammola, S., Cardoso, P., 2020. Global wildlife trade permeates the
tree of life. Biol. Conserv. 247, 108503.
Glaser, U., Glaser, W., 1995. Loricariidae all L-numbers. Aqualog Verlag A.C.S. GmbH,
Germany.
Global Biodiversity Information Facility, 2021. GBIF Home Page. Available at. https://
www.gbif.org/en/citation-guidelines [Accessed 01 December].
Hammer, M.P., Skarlatos Simoes, M.N., Needham, E.W., Wilson, D.N., Barton, M.A.,
Lonza, D., 2019. Establishment of Siamese ghting sh on the Adelaide River
oodplain: the rst serious invasive sh in the Northern Territory, Australia. Biol.
Invasions 21, 2269–2279.
Harfoot, M., Glaser, S.A., Tittensor, D.P., Britten, G.L., McLardy, C., Malsch, K.,
Burgess, N.D., 2018. Unveiling the patterns and trends in 40 years of global trade in
CITES-listed wildlife. Biol. Conserv. 223, 47–57.
Harrison, J.R., Roberts, D.L., Hernandez-Castro, J., 2016. Assessing the extent and nature
of wildlife trade on the dark web. Conserv. Biol. 30, 900–904.
Heinrich, S., Toomes, A., Janssen, J., 2021. Legal or unenforceable? Violations of trade
regulations and the case of the Philippine sailn lizard Hydrosaurus pustulatus
(Reptilia: squamata: Agamidae). J. Threat. Taxa 13, 18532–18543.
Herder, F., Schliewen, U.K., Geiger, M.F., Hadiaty, R.K., Gray, S.M., McKinnon, J.S.,
Walter, R.P., Pfaender, J., 2012. Alien invasion in Wallace's dreamponds: records of
the hybridogenic" owerhorn" cichlid in Lake Matano, with an annotated checklist of
sh species introduced to the Malili Lakes system in Sulawesi. Aquat. Invasions 7,
521–535.
Hitchens, R.T., Blakeslee, A.M., 2020. Trends in illegal wildlife trade: analyzing personal
baggage seizure data in the Pacic northwest. PLoS One 15, e0234197.
Holmes, B.J., Williams, S.M., Power, T.N., 2020. Evidence of naturalisation of the
invasive jaguar cichlid Parachromis managuensis (Günther, 1867), in Queensland,
Australia. BioInvas. Records 9, 146–157.
Hughes, A.C., Marshall, B.M., Strine, C., 2021. Gaps in global wildlife trade monitoring
leave amphobians vulnerable. elife 10, e70086.
Janssen, J., Leupen, B.T., 2019. Traded under the radar: poor documentation of trade in
nationally-protected non-CITES species can cause fraudulent trade to go undetected.
Biodivers. Conserv. 28, 2797–2804.
Jolly, C., von Takach, B., Webb, J., 2020. Slow life history leaves endangered snake
vulnerable to illegal poaching. Sci. Rep. 11, 5380.
A. Toomes et al.

Biological Conservation 282 (2023) 110040
9
Lockwood, J.L., Welbourne, D.J., Romagosa, C.M., Cassey, P., Mandrak, N.E.,
Strecker, A., Leung, B., Stringham, O.C., Udell, B., Episcopio-Sturgeon, D.J., 2019.
When pets become pests: the role of the exotic pet trade in producing invasive
vertebrate animals. Front. Ecol. Environ. 17, 323–330.
Maciaszek, R., Marcinek, D., Eberhardt, M., Wilk, S., 2019. Alien freshwater sh,
Xiphophorus interspecies hybrid (Poeciliidae) found in articial lake in Warsaw,
Central Poland. World Sci. News 132, 291–299.
Marshall, B.M., Strine, C., Hughes, A.C., 2020. Thousands of reptile species threatened by
under-regulated global trade. Nat. Commun. 11, 1–12.
Marshall, B.M., Strine, C.T., Fukushima, C.S., Cardoso, P., Orr, M.C., Hughes, A.C., 2022.
Searching the web builds fuller picture of arachnid trade. Commun. Biol. 5, 1–13.
Melville, J., Chapple, D.G., Keogh, J.S., Sumner, J., Amey, A., Bowles, P., Brennan, I.G.,
Couper, P., Donnellan, S.C., Doughty, P., 2021. A return-on-investment approach for
prioritization of rigorous taxonomic research needed to inform responses to the
biodiversity crisis. PLoS Biol. 19, e3001210.
Millington, M.D., Holmes, B.J., Balcombe, S.R., 2022a. Systematic review of the
Australian freshwater ornamental sh industry: the need for direct industry
monitoring. Manag. Biol. Invas. 13, 406–434.
Millington, M., Sierp, M., Gaylard, S., 2022. Assessing the Invasiveness Risk of Non-
indigenous Fish in the Australian Ornamental Trade. Department of Agriculture and
Fisheries, Brisbane, Australia.
Mohanty, N.P., Measey, J., 2019. The global pet trade in amphibians: species traits,
taxonomic bias, and future directions. Biodivers. Conserv. 28, 3915–3923.
Morton, O., Scheffers, B.R., Haugaasen, T., Edwards, D.P., 2021. Impacts of wildlife trade
on terrestrial biodiversity. Nature Ecology & Evolution 5, 540–548.
Msukwa, A.V., Cowx, I.G., Harvey, J.P., 2021. Vulnerability assessment of Lake Malawi's
ornamental sh resources to export ornamental trade. Fish. Res. 238, 105869.
Msukwa, A.V., Cowx, I.G., Harvey, J.P., 2022. Ornamental sh export trade in Malawi.
J. Fish Biol. 100, 300–314.
Nico, L.G., Beamish, W.H., Musikasinthorn, P., 2007. Discovery of the invasive Mayan
cichlid sh “Cichlasoma” urophthalmus (Günther 1862) in Thailand, with comments
on other introductions and potential impacts. Aquat. Invas. 2, 197–214.
Nijman, V., Shepherd, C.R., 2021. Underestimating the illegal wildlife trade: a ton or a
tonne of pangolins? Biol. Conserv. 253, 108887.
Norval, G., Halliday, B., Sih, A., Sharrad, R.D., Gardner, M.G., 2020. Occurrence of the
introduced snake mite, Ophionyssus natricis (Gervais, 1844), in the wild in
Australia. Acarologia 60, 559–565.
Nov´
ak, J., Hofmann, J., Hohl, D., Magalh˜
aes, A.L.B., Patoka, J., 2022. Enigmatic
armoured catshes (Siluriformes: callichthyidae and Loricariidae) in ornamental
aquaculture: a new insight into neotropical sh diversity. Aquaculture 547, 737460.
Papavlasopoulou, I., Vardakas, L., Perdikaris, C., Kommatas, D., Paschos, I., 2014.
Ornamental sh in pet stores in Greece: a threat to biodiversity? Mediterr. Mar. Sci.
15, 126–134.
Patel, J.M., 2020. Web scraping in Python using beautiful soup library. In: Getting
Structured Data from the Internet. Springer, New York, USA, pp. 31–84.
Paul, K., Miles, K., Huffer, D., 2020. Two Clicks Away: Wildlife Sales on Facebook.
Alliance to Counter Crime Online.
Py-Daniel, L.R., Zuanon, J., Oliveira, R.Rd., 2011. Two new ornamental loricariid
catshes of Baryancistrus from Rio Xingu drainage (Siluriformes: Hypostominae).
Neotrop. Ichthyol. 9, 241–252.
Raghavan, R., Dahanukar, N., Tlusty, M.F., Rhyne, A.L., Kumar, K.K., Molur, S.,
Rosser, A.M., 2013. Uncovering an obscure trade: threatened freshwater shes and
the aquarium pet markets. Biol. Conserv. 164, 158–169.
R Core Team, 2022. R: A Language and Environment for Statistical Computing. R
Foundation for Statistical Computing, Vienna, Austria.
Ribeiro, J., Reino, L., Schindler, S., Strubbe, D., Vall-llosera, M., Araújo, M.B.,
Capinha, C., Carrete, M., Mazzoni, S., Monteiro, M., 2019. Trends in legal and illegal
trade of wild birds: a global assessment based on expert knowledge. Biodivers.
Conserv. 28, 3343–3369.
Schneider, K., Fraser, H., Dodd, A.J., Robinson, A., Arndt, E., 2018. Evaluating the Health
of Australia's Biosecurity System. Centre for Excellence of Biosecurity Risk Analysis:
Melbourne, Australia.
Sheridan, C., 2016. The Python Language Reference Manual. North Carolina, USA, Lulu
Press Inc.
Siriwat, P., Nijman, V., 2018. Illegal pet trade on social media as an emerging
impediment to the conservation of Asian otters species. J. Asia-Pac. Biodivers. 11,
469–475.
Siriwat, P., Nijman, V., 2020. Wildlife trade shifts from brick-and-mortar markets to
virtual marketplaces: a case study of birds of prey trade in Thailand. J. Asia-Pac.
Biodivers. 13, 454–461.
Stringham, O.C., Maher, J., Lassaline, C., Wood, L., Moncayo, S., Toomes, A.,
Heinrich, S., Drake, C., Chekunov, S., Hill, K.G., 2022. The Dark Web Trades
Wildlife, but Mostly as Drugs. EcoEvoRxiv. https://doi.org/10.32942/osf.io/yk7tg.
Stringham, O.C., Toomes, A., Kanishka, A.M., Mitchell, L., Heinrich, S., Ross, J.V.,
Cassey, P., 2021. A guide to using the internet to monitor and quantify the wildlife
trade. Conserv. Biol. 35, 1130–1139.
Tan, M., Armbruster, J.W., 2016. Two new species of spotted hypancistrus from the Rio
Negro drainage (Loricariidae, Hypostominae). ZooKeys 552, 123–135.
Tensen, L., 2016. Under what circumstances can wildlife farming benet species
conservation? Glob. Ecol. Conserv. 6, 286–298.
Toomes, A., García-Díaz, P., Stringham, O.C., Ross, J.V., Mitchell, L., Cassey, P., 2022.
Drivers of the Australian native pet trade: the role of species traits, socioeconomic
attributes and regulatory systems. J. Appl. Ecol. 59, 1268–1278.
Toomes, A., García-Díaz, P., Wittmann, T.A., Virtue, J., Cassey, P., 2019. New aliens in
Australia: 18 years of vertebrate interceptions. Wildl. Res. 47, 55–67.
Toomes, A., Stringham, O.C., Mitchell, L., Ross, J.V., Cassey, P., 2020. Australia’s wish
list of exotic pets: biosecurity and conservation implications of desired alien and
illegal pet species. NeoBiota 60, 43–59.
UNEP-WCMC, 2022. Checklist of CITES species. CITES Secretariat, Geneva, Switzerland
and UNEP-WCMC (Cambridge, United Kingdom) [Accessed 01 March 2022].
Vall-llosera, M., Cassey, P., 2017a. ‘Do you come from a land down under?’
Characteristics of the international trade in Australian endemic parrots. Biol.
Conserv. 207, 38–46.
Vall-llosera, M., Cassey, P., 2017b. Leaky doors: private captivity as a prominent source
of bird introductions in Australia. PLoS One 12, e0172851.
Vall-llosera, M., Cassey, P., 2017c. Physical attractiveness, constraints to the trade and
handling requirements drive the variation in species availability in the Australian
cagebird trade. Ecol. Econ. 131, 407–413.
Warwick, C., Steedman, C., Jessop, M., Arena, P., Pilny, A., Nicholas, E., 2018. Exotic pet
suitability: understanding some problems and utilizing a labeling system to aid
animal welfare, environment, and consumer protection. J. Vet. Behav. 42, 53–63.
Watters, F., Stringham, O., Shepherd, C.R., Cassey, P., 2022. Demand for wildlife not
listed in the CITES multilateral treaty. Conserv. Biol. https://doi.org/10.1111/
cobi.13978.
Whittington, R., Chong, R., 2007. Global trade in ornamental sh from an Australian
perspective: the case for revised import risk analysis and management strategies.
Prev. Vet. Med. 81, 92–116.
Wickham, H., 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag, New
York, USA.
Wickham, H., 2022. Stringr: simple, consistent wrappers for common string operations.
Available at: https://stringr.tidyverse.org. Accessed 01 March 2022.
Woolnough, A.P., de Milliano, J.W., Petel, TvP, Cassey, P., 2020. A policy approach to
non-indigenous bird management in Victoria: managing potential threats to
biodiversity, social amenity and economic values. Victorian Nat. 137, 203–209.
Wyatt, T., Maher, J., Allen, D., Clarke, N., Rook, D., 2022. The welfare of wildlife: an
interdisciplinary analysis of harm in the legal and illegal wildlife trades and possible
ways forward. Crime Law Soc. Chang. 77, 69–89.
A. Toomes et al.