research paper summary

R E G U L A R P A P E R

The endangered White’s seahorse Hippocampus whitei chooses artificial over natural habitats

Michael Simpson1 | Rebecca L. Morris2 | David Harasti3 | Ross A. Coleman1

1Coastal and Marine Ecosystems, School of

Life and Environmental Sciences, University of

Sydney, Sydney, New South Wales, Australia

2National Centre for Coasts and Climate,

School of Biosciences, The University of

Melbourne, Melbourne, Victoria, Australia

3Fisheries Research, NSW Department of

Primary Industries, Taylors Beach, New South

Wales, Australia

Correspondence

Michael Simpson, Marine Ecology Laboratories

(A11), The University of Sydney, NSW 2006,

Australia.

Email: michael.simpson@sydney.edu.au

Abstract

To explore whether the endangered White’s seahorse Hippocampus whitei would

choose to inhabit artificial over natural habitats, 10 adult H. whitei individuals were

put through a series of binary choice trials in aquaria, during which they were offered

different paired combinations of natural (different types of macroalga and seagrass)

and artificial habitat (panels of swimming-net material). It was found that H. whitei

displayed a significant choice for swimming-net material over all other available natu-

ral habitats and chose habitats according to the following rankings: (a) Net;

(b) Sargassum sp.; (c) Posidonia australis; (d) Zostera muelleri. Hippocampus whitei’s

choice of swimming net material over natural habitat suggests that these artificial

structures could be a useful conservation measure for seahorses in areas where natu-

ral habitat is becoming less favourable due to declines in abundance or quality.

K E Y W O R D S

conservation, habitat selection, Hippocampus whitei, seagrass, Syngnathidae, urbanised

habitats

1 | INTRODUCTION

Increased urbanisation has resulted in a significant decline in natural

habitat availability for marine animals (Airoldi & Beck, 2007; Lai et al.,

2015). As a result of decreased natural habitat availability, which often

affects species in combination with other negative influences such as

disturbance and overfishing, marine animals have experienced signifi-

cant population declines, with many species being threatened with

extinction (Airoldi et al., 2008; Mace & Lande, 1991; Stuart et al.,

2004; Waycott et al., 2009). Large areas of natural shorelines have

been replaced by artificial structures in what is referred to as ocean

sprawl, a phenomenon that is recognised as a significant threat to

marine ecosystems (Airoldi & Beck, 2007; Dafforn et al., 2015; Firth

et al., 2013). These artificial structures are deployed for a number of

purposes, such as coastal defence (e.g., seawalls, breakwaters) and

commercial (e.g., ports) or recreational use (e.g., swimming nets,

marinas; Dafforn et al., 2015).

Artificial structures are considered to have a number of ecological

implications for marine communities, including loss of biodiversity

(Bulleri & Chapman, 2010; Gittman et al., 2016) and spread of invasive

species (Dafforn et al., 2009; Simkanin et al., 2012). While the effect

of artificial structures is considered mostly negative, there are cases

of rare or threatened species seeking refuge on artificial structures

where natural habitat may be absent or sparse (Claassens et al., 2018;

García-Gómez et al., 2011). Purpose-built structures in the form of

artificial reefs have been commonly built to replace degraded habitat

(Ambrose, 1994; Feary et al., 2011). There is also an increasing inter-

est in building microhabitats in marine artificial structures to enhance

native species colonisation through ecological engineering (i.e., the

combination of ecological and engineering principles to design struc-

tures that benefit both humans and nature; Chapman & Underwood,

2011). Eco-engineering has been targeted at threatened species in

some cases; e.g., drilling of pits into seawalls significantly increased

the density of an over-exploited limpet in Portugal due to their associ-

ation with this microhabitat on natural rocky shores (Martins et al.,

2010). There exists the possibility that artificial structures, either pre-

existing or purposefully created could be used as a tool in the conser-

vation of threatened marine fishes.

Received: 6 November 2018 Accepted: 9 May 2019

DOI: 10.1111/jfb.14002

FISH

J Fish Biol. 2019;95:555–561. wileyonlinelibrary.com/journal/jfb © 2019 The Fisheries Society of the British Isles 555

The Syngnathidae are a family of marine fish that are especially

susceptible to population collapse following habitat loss or overfishing

due to low population densities, weak swimming abilities and small

home range (Foster & Vincent, 2004; Vincent et al., 2011). A number

of seahorse species (Hippocampus spp.) are considered threatened and

are listed on the IUCN Red List as Vulnerable (n = 12) or Endangered

(n = 2; IUCN, 2018). Natural seahorse habitats, such as macroalgae,

sponges, soft coral and seagrass provides shelter from predators and

an attachment holdfast for seahorses to grip with their prehensile tail

while feeding or resting (Quin et al., 2014; Woods, 2002). Seahorses

may be habitat specific (Lourie & Randall, 2003) and display strong

habitat preferences depending on the species and life stage of the

seahorse (Gristina et al., 2014; Harasti et al., 2014). In Brazil, the alga

Caulerpa racemosa was the most frequently inhabited holdfast in situ

for the slender seahorse Hippocampus reidi Ginsburg 1933 (Dias &

Rosa, 2003). The long snouted seahorse Hippocampus guttulatus

Cuvier 1829 and the short snouted seahorse Hippocampus hippocam-

pus (L. 1758) have inhabited artificial habitat structures in a Ria For-

mosa lagoon (Portugal; 37� 010 N, 08� 000 W) that had suffered

extensive natural habitat loss and showed strong selections for spe-

cific design materials and holdfast densities (Correia et al., 2013,

2015). The endangered seahorse Hippocampus capensis Boulenger

1900 were more likely to inhabit artificial structures over natural habi-

tat in in situ choice chambers (Claassens et al., 2018). Use of various

habitats by seahorses may vary according to scale, location and spe-

cies, but a pattern of strong affiliation with artificial habitat, even in

the presence of natural habitats, is apparent.

In Australia, White’s seahorse Hippocampus whitei Bleeker 1855

has undergone significant declines in abundance, most likely as a

result of habitat loss (Harasti, 2016). In Port Stephens (NSW; 32�

420 S, 152� 060 E), H. whitei display ontogenetic differences in habitat

use; juveniles tend to be found on gorgonian fan habitats whilst adults

prefer soft coral and sponge habitats (Harasti et al., 2014). In Sydney

Harbour, these preferred natural habitats (Gorgonian fans, soft coral

and sponge) are largely absent. However, H. whitei has been observed

inhabiting protective swimming nets, as well as natural habitats that

are present in the harbour, such as the seagrasses Posidonia australis

and Zostera muelleri and the brown alga Sargassum sp. (Vincent et al.,

2005). Hippocampus whitei has been shown to inhabit swimming nets

in relatively high numbers compared with patches of natural habitats

around the harbour (Harasti et al., 2010). Most swimming nets are

deployed for the entire year in Sydney Harbour, but some may be

removed during the winter months or for cleaning (Harasti et al.,

2010). When a new net is constructed, seahorses can take up to

4 months to colonise the new netting and numbers of seahorses on

permanent netting were 10 to 100 times greater compared with those

that were removed seasonally (Clynick, 2008). Further research found

that the addition of further material to a new net to increase struc-

tural complexity resulted in increased seahorse abundance on experi-

mental nets in the field, with habitat preference assays in the

labratory demonstrating that seahorses displayed a significant prefer-

ence for swimming net material of higher complexity (Hellyer

et al., 2011).

While we know about seahorse preferences of natural habitat in

areas absent of swimming nets (Harasti et al., 2014) and seahorse

preference of swimming-net material of varying complexity (Hellyer

et al., 2011), it is not currently understood whether seahorses choose

to inhabit artificial structures such as swimming nets over natural hab-

itat when the two are in close proximity. This information is important

because it may help explain the greater seahorse densities found on

swimming nets than on natural habitat within Sydney Harbour. It is

currently unknown whether seahorses inhabit swimming nets in high

densities through choice or availability of suitable habitat. A series of

experiments were held in aquaria to assess whether H. whitei choose

swimming-net material over the natural habitats found in close prox-

imity to nets in Sydney Harbour, while also exploring which of these

natural habitats were more frequently chosen by H. whitei. The

research tested the hypothesis that, in each binary habitat choice trial,

the time spent on either habitat would differ significantly from 50%,

indicating a choice has been made. From the results of these experi-

ments, a choice ranking of H. whitei habitats is constructed.

2 | MATERIALS AND METHODS

This research was conducted under Department of Primary Industries

scientific collections permit No: P17/0028-1.1 and with approval from

the Sydney University Ethics Committee, project No. 2016/1066.

2.1 | Seahorse collection and housing

Ten H. whitei (five male and five female) were collected from the

swimming net located at Chowder Bay, Sydney Harbour by scuba div-

ing (11 May 2017). All individuals collected were at least 7 cm total

length from the top of the coronet to the end of the tail as this size

means they considered adults. This measurement was used as a mini-

mum length for H. whitei collection as it is at this size that the brood

pouch becomes well defined, allowing for visual determination of sex

(Harasti et al., 2012). Visual implant fluorescent elastomer (VIFE;

Northwest Marine Technologies; www.nmt.us) was used as per

Woods and Martin-Smith (2004) to give each H. whitei a unique tag

for identification during and after the experiment. VIFE tagging is a

harmless and effective method of applying individual identifiable tags

to seahorses and has been successfully used for identification of

experimental animals (Harasti et al., 2010). Tagging was performed to

track individual H. whitei as they were moved among different tanks.

The provision of a unique tag also allowed for some degree of moni-

toring of fish health after experiment completion, as all animals were

released back onto the net at Chowder Bay after the aquaria habitat-

choice experiments were completed.

Hippocampus whitei were kept in aquaria for a total of 3–4 weeks.

They were given a week-long acclimation period, in which they were

all kept in a 100 l tub with flow-through system of seawater from

Chowder Bay. The water pumped through the system was at ambient

sea temperature of 19.5–20.5�C, as measured at Chowder Bay during

the month of May 2017 and was pumped into each tub at a mean rate

556 SIMPSON ET AL.FISH

of 0.5 l min−1. Holdfasts consisting of nylon rope, dowel wood poles

and kelp were provided during this period to reduce stress. None of

these nominal habitats were used in the choice experiments to pre-

vent any initial bias. Hippocampus whitei were fed daily on a diet of

enriched brine shrimp, as well as caprellid amphipods cultured on-site

at Sydney Institute of Marine Science (SIMS).

2.2 | Choice experiment habitats and design

Two of the habitats used in this experiment, the alga Sargassum

sp. and seagrass Posidonia australis, are native taxa found in Sydney

Harbour, each known to be used as habitat by H. whitei (Harasti et al.,

2014). The seagrass Zostera muelleri is also commonly found in Sydney

Harbour and was included in this experiment to establish whether H.

whitei would choose to inhabit P. australis over Z. muelleri when these

seagrasses are equally available, as observed in previous field studies

(Harasti et al., 2014; Manning et al., 2018).

Panels of netting were cut in situ from remnant netting material at

Clifton Gardens. Only net panels with sufficient levels of epibiotic

growth were used in this experiment, as previous research has shown

that H. whitei prefers fouled to clean netting (Harasti et al., 2010) and

bio-fouled nets are most frequently found in the wild; the exception

being the immediate months after net cleaning occurs, or the installa-

tion of a new net. All habitats were placed in the flow-through system

for 1 week prior to the experiment beginning to ensure associated

food sources for H. whitei would be flushed, as to not introduce bias

in their habitat selection.

It is recognised that while H. whitei were collected from one of the

habitats that would be featured in the choice experiments (swimming

net), any bias in seahorse habitat choice as a result of this is likely to

be negligible. All natural habitats used in the choice experiments are

present in close proximity to the Chowder Bay swimming net, so it is

reasonable to expect that seahorses have encountered all of these

habitats during their lives. Previous observations of tagged seahorses

moving between habitats (Harasti et al., 2010), as well as current

observations (M. Simpson, unpubl. data), indicate that H. whitei move

between the swimming net and surrounding natural habitats at Chow-

der Bay, thus demonstrate that they can and do change habitats. The

1 week acclimation period that H. whitei experienced prior to the

experiment also existed to minimise any pre-existing bias towards

habitats featured in the choice experiments.

Hippocampus whitei were placed into 75 l habitat-choice chambers

in which they were able to choose to inhabit one of two available hab-

itats (Figure 1). Enough habitat was placed into each choice chamber

to cover each respective tub wall (c. 30 × 30 cm). Surface area, as

opposed to frond or holdfast density, was used in this experiment due

to the wide variation in holdfast–frond size between habitats. All habi-

tats were weighted to the tank floor with steel nuts, with nets being

attached to the top of the tub with cable ties to ensure the material

would stay upright. Fibreglass window mesh was needed to keep

seagrass rhizomes intact, so mesh was placed underneath all habitats

to remove any confounding effect of H. whitei habitat selection.

At the start of each experimental run, H. whitei were placed into

the centre of the choice chamber in small plastic transport container,

through which H. whitei could assess each habitat. Hippocmpus whitei

were released after 5 min and their selection of habitat was recorded

every 0.5 h for 12 h to give 24 data points in total using a Go-Pro

camera (www.gopro.com) located 30 cm directly above each tank.

After each 12 h period was completed, the proportion of time (i.e,

data points) that H. whitei spent attached to each habitat in the cham-

ber was calculated. A space of 24 h was held between experiments.

Six binary choice trial experiments were run, with each combina-

tion of habitats: (a) net v. Z. muelleri; (b) Z. muelleri v. Sargassum sp.;

(c) net v. Sargassum sp.; (d) P. australis v. Z. muelleri; (e) P. australis v.

Sargassum sp.; (f) P. australis v. net.

All H. whitei were tested in each habitat choice experiment and as

they were being reused, a cross-over design was employed in order to

dictate the order through which each H. whitei undertook subsequent

experiments, as per Ratkowsky et al. (1992). Such a design allows for

each fish to undertake the habitat-choice experiments in a different

sequence, removing possible bias that may carry over from any previ-

ous experiment. To account for any possible left–right bias of

seahorses, five seahorses were used in each experiment in an XY (i.e.,

habitat X on the left and habitat Y on the right) arrangement, while

the other five experienced choices in a YX (i.e., habitat Y on the left,

habitat X on the right) arrangement. Different patches of habitat were

used in each choice chamber experiment as to prevent any potential

effect of co-specific odours in habitat selection. To maximise effi-

ciency, 10 choice chamber tanks were used at a time in each experi-

mental run so that each H. whitei would be put through a choice

experiment on a daily basis.

If a H. whitei was to spend 100% of time on the first habitat in the

experiment title (for example, net in the net v. Z. muelleri experiment,

then that data point will read 12 for that replicate. If the opposite

occurs, the data point will read −12. Thus, if a H. whitei was to display

zero choice and spend 50:50 time on either habitat in the experiment,

FIGURE 1 A 75 l tub 30 × 65 × 40 cm (base) choice chamber where Hippocampus whitei has the choice to inhabit either Zostera muelleri or Sargassum sp., each of c. 30 × 30 cm. Cameras were placed 30 cm above the surface of the water

SIMPSON ET AL. 557FISH

a score of 0 will be recorded for that fish. Within experiments, one

sample t-tests were used to compare the mean score of the

10 H. whitei to a hypothetical test value of zero.

A contingency table of rank frequencies (1st or 2nd) for each habi-

tat was constructed in order to determine a ranking of habitats. The

transitivity of H. whitei habitat selections will be explored when

observing habitat selection rankings, meaning that if habitat A is

selected over habitat B and habitat B is selected of habitat C, then

habitat A will be also be selected over C with no contradictions in

choice (Regenwetter et al., 2011).

3 | RESULTS

3.1 | Habitat choice experiments

The mean choice score across all H. whitei differed significantly from

the test-score of zero, indicating that choices of habitat had been

made in all experiments (Table 1). Swimming net was chosen over Z.

muelleri, Sargassum sp. and P. australis in experiments 1, 3 and

6, respectively. Sargassum sp. was chosen over Z. muelleri and P. aus-

tralis in experiments 2 and 5, respectively. In experiment 4, P. australis

was chosen over Z. muelleri (Figure 2).

3.2 | Ranking of habitats

Net was chosen the greatest number of times throughout the experi-

ments, with no other habitat being chosen over net throughout

(Table 2 and Figure 2). Thus, net is ranked first out of the four habitats

tested. Sargassum sp. was ranked second, as it was chosen twice over

both species of seagrass. Of the two seagrass species, P. australis

(ranked third) was chosen only over Z. muelleri, which was ranked

fourth and chosen over no other habitats (Table 2 and Figure 2). This

pattern is reflected in Figure 3, which shows that across all experi-

ments, net had the greatest mean number of selections, followed by

Sargassum, P. australis and Z. muelleri. As these data were not indepen-

dent (time spent on one habitat is dependent on the time spent on

the other), ANOVA tests were not possible. Hippocampus whitei habi-

tat selections in this experiment were transitive, with the results of

each binary trial supporting a clear ranking of habitats; i.e., habi-

tat A > B > C > D.

TABLE 1 Results of Student’s t-tests run for each experiment, which tested the null hypotheses that the mean choice scores for Hippocampus whitei (within each experiment) were equal to zero. In a significant A v. B test, a positive t test indicates that habitat A has been chosen and vice versa

Comparison t df P

Net v. Z. muelleri 4.918 9 <0.001

Zostera. muelleri v. Sargassum sp. −9.291 9 <0.001

Net v. Sargassum sp. 2.965 9 <0.05

Posidonia australis v. Z. muelleri 6.527 9 <0.001

P. australis v. Sargassum sp. −3.039 9 <0.05

P. australis v. Net −3.262 9 <0.05

0

5

10

15

–5

–10

Ti m

e on

h ab

ita t

(n 3

0 m

in –1

)

–15 Net v.

z. muelleri z. muelleri

v. sargassum sp.

Net v. sargassum sp.

P. australis v.

z. muelleri

P. australis v.

Sargassum sp.

P. australis v. net

FIGURE 2 Mean (±SE) time that Hippocampus whitei spent on the first-choice habitat A in each pair of habitats presented as an A v. B test (n = 10; Table 1). Habitats: Net, Swimming-net material; Zostera muelleri; Posidonia australis; Sargassum sp

TABLE 2 Contingency table of rank frequencies for each habitat. This table displays the frequency each habitat was ranked 1st or 2nd in a habitat choice experiment

Ranks

Habitats 1st 2nd

Net 3 0

Sargassum sp. 2 1

Posidonia australis 1 2

Zostera. muelleri 0 3

40

60

80

20N o.

o f �

m es

h ab

ita t

ch os

en ac

ro ss

e xp

er im

en ts

0 Net Sargassum sp. Posidonia

australis Zostera muelleri

FIGURE 3 Mean (± SE) number of times each habitat was chosen by Hippocampus whitei across all experiments (n = 10)

558 SIMPSON ET AL.FISH

4 | DISCUSSION

As predicted, H. whitei displayed a significant choice of habitat during

every choice trial, with the resulting ranking of chosen habitats show-

ing that the swimming-net material is actively chosen over other

equally available natural habitats. Of the natural habitats used in this

experiment, Sargassum sp. was the highest ranked, with P. australis

and Z. muelleri in 3rd and 4th respectively.

When dealing with experiments that involve threatened animals,

researchers often have to compromise by using low animal numbers

to decrease any possible threat or damage to the populations of these

species that are already in decline. An obvious disadvantage is that a

low sample size may impede robust statistical results. While the num-

ber of H. whitei used in this research (n = 10) was adequate to test for

habitat choice among H. whitei as a whole, we were unable to test dif-

ferences in habitat choice between sexes (n = 5).We feel this limita-

tion is minor, however, as Harasti et al. (2014) showed that there

were no differences in habitat preference of H. whitei. There are

methods by which low numbers of study animals can still be used suc-

cessfully in experiments to gain robust results, such as the re-use of

animals for subsequent experiments through the implementation of a

cross-over design. This was the design used here for the habitat

choice experiment. However, when interpreting results from an

experiment with low numbers such as this, it is advisable that the

inevitable limitations be kept in mind.

Similar patterns to those observed here were obtained from work

done on Hippocampus capensis in the Knysna Estuary (South Africa;

Claassens et al., 2018). Much like H. whitei in Sydney Harbour, this

seahorse inhabits artificial structures (Reno mattresses filled with

rocks) in high densities despite the availability of natural habitat

(Claassens et al., 2018). Claassens et al. (2018) showed that, when

placed in a choice chamber and given the option to inhabit both, H.

capensis would choose to inhabit artificial structures over equally pre-

sent Zostera capensis. With habitat availability the suspected cause of

Hippocampus spp. population declines in Ria Formosa lagoon, artificial

holdfast units were deployed as a potential conservation tool for the

species H. guttulatus and H. hippocampus (Correia et al., 2015). These

artificial habitats were inhabited by large densities of Hippocampus

spp., though results suggested that these units may have limited

effect when placed in close proximity to undamaged, high complexity

natural habitats (Correia et al., 2015). Although it is known that artifi-

cial habitats may be selected by seahorses such as H. whitei, it is not

known how these choices work at scale, nor where large amounts of

preferred habitat are present; i.e., soft coral and sponge habitats.

A number of factors dictate seahorse habitat choice and many of

these may explain why H. whitei may choose swimming net over natu-

ral habitat. It is important to address a number of these factors that

can be ruled out as a result of the design of this experiment. Often

habitats in the wild will be selected because they are simply available

in a higher abundance, offer a greater or more readily available food

source for the animal (Hellyer et al., 2011), or because one habitat

offers greater protection from predation pressure (Chick & Mlvor,

1997). As habitats were offered to H. whitei at equal availability,

rinsed of natural food sources and in isolation from potential preda-

tors, these reasons can be discarded as an explanation for habitat

choice in this experiment. It is reasonable to presume, therefore, that

H. whitei may be relying on other cues when choosing a habitat, such

as visual cues (colour, shape, size) or the quality and suitability of the

habitat as a holdfast. It is possible that H. whitei choose to inhabit

swimming nets as these structures share some similarities with the

natural branching habitats that H. whitei have inhabited in the wild.

As seahorses grasp onto holdfasts for shelter from water cur-

rents during feeding (Quin et al., 2014), they prefer solid, tubular

holdfasts with greater complexity (Harasti et al., 2014; Perante

et al., 1998). Swimming-net material would seem to fit these criteria

more than the other habitats used in this experiment, despite not

being natural. Seagrasses, while often a successful habitat for

seahorses in the wild, may be less favourable of a holdfast because

it possesses blade-like holdfast structures rather than solid tube-like

holdfasts. While it is possible for H. whitei to grasp these structures

and inhabit them if necessary, they may not be as suitable or stable

as other habitat types. Among the two seagrass species used in this

experiment, it was not surprising that seahorses chose P. australis,

the species with the larger, more stable holdfast structure, over Z.

muelleri. This was supported by studies in Nelson Bay, which

showed that H. whitei are found in higher densities in P. australis

than Z. muelleri due to its greater blade length and density (Manning

et al., 2018).

Juvenile seahorses may display a preference for Sargassum sp. in

the wild as the smaller fronds provide a more suitable holdfast for the

juveniles’ smaller tails to grasp, while adult seahorses preferred bulkier

holdfasts (Harasti et al., 2014; Perante et al., 1998). Thus, it is easier

to understand why H. whitei may choose a sturdier, bulkier, similarly

complex habitat such as net over Sargassum sp. These results suggest

that holdfast stability and thickness is equally or more important to

adult H. whitei when choosing habitat than complexity. Future

research can implement studies on juvenile H. whitei habitat choice

that include both artificial and natural habitat to further explore these

differences in habitat choice. From the experiments in aquaria, we

would expect to find H. whitei in higher densities on Sargassum

sp. than on seagrass in the wild, but anecdotal evidence from surveys

in Sydney Harbour suggest that this is not the case. It is possible that,

while Sargassum sp. is the preferred habitat of H. whitei in a laboratory

environment where both are of equal availability, Sargassum sp. is sim-

ply too patchy and seasonal in the wild, whereas P. australis tends to

grow in larger beds and is year round. It may also be the case that,

once food availability and protection from predation are taken into

account, P. australis may be a preferable habitat for seahorses in the

wild, as recently indicated by Manning et al., 2018. Studies have

shown that fragmented islands of habitat result in less successful ani-

mal populations than larger, less fragmented blocks (Andren, 1994).

While not explored in the habitat choice experiment, as the net panels

used had only a small amount of epibiotic algal growth, a reason for H.

whiteis greater affiliation with swimming net as a habitat may be that

the net has the capability to act as a base structure for other habitat-

SIMPSON ET AL. 559FISH

forming species to grow upon. Soft coral and sponge species can be

found growing on some of the lesser maintained swimming nets

around Sydney Harbour, as well as Sargassum sp. and the kelp Ecklonia

radiate, which may further attract seahorses.

The results of the choice experiment emphasise the potential of

artificial habitats as a valuable tool in the conservation of H. whitei in

Australia and potentially worldwide. The observation that H. whitei

will choose to inhabit swimming-net material in the presence of

other natural habitats that may be in decline in the wild indicates

that swimming nets could potentially be used to support H. whitei

populations when natural habitat is dwindling, becoming patchier

and as a result, a less viable habitat. In the future, priority should be

placed on seahorse-friendly swimming-net management (Harasti

et al., 2010), cleaning and construction materials. A key consider-

ation when managing artificial habitats for H. whitei is that the artifi-

cial habitat presence results in increased H. whitei production at a

site, rather than simply an attraction of H. whitei away from their

natural habitat (Brickhill et al., 2005; Chapman et al., 2018). Similarly,

it is important to explore whether H. whitei population fitness is

equal on both artificial and natural habitats to ensure that H. whitei

are not choosing a lesser quality habitat as a result of instinctual

cues. This phenomenon is observed in the case of ecological traps

(Battin, 2004; Schlaepfer et al., 2002; Sherley et al., 2017) and can

be an unintended consequence of restoration projects (Hale &

Swearer, 2017). As H. whitei have been shown to not discriminate

against viable habitats due to their artificiality, this research suggests

that H. whitei will choose to inhabit well designed artificial habitat

that may be used to bolster H. whitei populations where natural hab-

itat is absent.

ACKNOWLEDGEMENTS

We would like to thank C. McArthur for her help with experimental

design and analyses. We would also like to thank the staff at the

SIMS aquarium for their help housing the collected seahorses, as

well as helping set up the flow through system used in the experi-

ment. We acknowledge the contribution of our field volunteers

W. Dakin, J. Haberstroh and C. Dittes for their help with seahorse

collection. This is SIMS publication 245 (number to be provided on

publication).

CONTRIBUTIONS

M.S., R..LM., D.H. and R.A.C. conceived the ideas and designed the

methodology; M.S. collected and analysed the data; M.S., R.L.M.,

D.H. and R.A.C. interpreted the data and wrote the manuscript. All

authors contributed critically to the drafts and gave final approval for

publication.

ORCID

Michael Simpson https://orcid.org/0000-0002-2978-7607

REFERENCES

Airoldi, L., Balata, D., & Beck, M. W. (2008). The gray zone: Relationships

between habitat loss and marine diversity and their applications in

conservation. Journal of Experimental Marine Biology and Ecology, 366

(1–2), 8–15. Airoldi, L., & Beck, M. W. (2007). Loss, status and trends for coastal marine

habitats of Europe. Oceanography and marine biology, 45, 357–417. Ambrose, R. F. (1994). Mitigating the effects of a coastal power plant on a

kelp forest community: Rationale and requirements for an artificial

reef. Bulletin of Marine Science, 55(2–3), 694–708. Andren, H. (1994). Effects of habitat fragmentation on birds and mammals

in landscapes with different proportions of suitable habitat: A review.

Oikos, 71, 355–366. Battin, J. (2004). When good animals love bad habitats: Ecological traps

and the conservation of animal populations. Conservation Biology, 18,

1482–1491. Brickhill, M. J., Lee, S. Y., & Connolly, R. M. (2005). Fishes associated with

artificial reefs: Attributing changes to attraction or production using

novel approaches. Journal of Fish Biology, 67, 53–71. Bulleri, F., & Chapman, M. G. (2010). The introduction of coastal infra-

structure as a driver of change in marine environments. Journal of

Applied Ecology, 47(1), 26–35. Chapman, M. G., & Underwood, A. J. (2011). Evaluation of ecological engi-

neering of “armoured” shorelines to improve their value as habitat. Journal of Experimental Marine Biology and Ecology, 400(1–2), 302–313.

Chapman, M. G., Underwood, A. J., & Browne, M. A. (2018). An assess-

ment of the current usage of ecological engineering and reconciliation

ecology in managing alterations to habitats in urban estuaries. Ecologi-

cal Engineering, 120, 560–573. Chick, J. H., & Mlvor, C. C. (1997). Habitat selection by three littoral zone

fishes: Effects of predation pressure, plant density and macrophyte

type. Ecology of Freshwater Fish, 6(1), 27–35. Claassens, L., Booth, A. J., & Hodgson, A. N. (2018). An endangered sea-

horse selectively chooses an artificial structure. Environmental Biology

of Fishes, 101, 1–11. Clynick, B. G. (2008). Harbour swimming nets: A novel habitat for

seahorses. Aquatic Conservation: Marine and Freshwater Ecosystems,

18, 483–492. Correia, M., Koldewey, H., Andrade, J. P., & Palma, J. (2015). Effects of

artificial holdfast units on seahorse density in the Ria Formosa

lagoon, Portugal. Journal of Experimental Marine Biology and Ecology,

471, 1–7. Correia, M., Palma, J., Koldewey, H., & Andrade, J. P. (2013). Can artificial

holdfast units work as a habitat restoration tool for long-snouted sea-

horse (Hippocampus guttulatus Cuvier)? Journal of Experimental Marine

Biology and Ecology, 448, 258–264. Dafforn, K. A., Glasby, T. M., Airoldi, L., Rivero, N. K., Mayer-Pinto, M., &

Johnston, E. L. (2015). Marine urbanization: An ecological framework

for designing multifunctional artificial structures. Frontiers in Ecology

and the Environment, 13, 82–90. Dafforn, K. A., Johnston, E. L., & Glasby, T. M. (2009). Shallow moving

structures promote marine invader dominance. Biofouling, 25(3),

277–287. Dias, T. L. P., & Rosa, I. L. (2003). Habitat preferences of a seahorse spe-

cies, Hippocampus reidi (Teleostei: Syngnathidae) in Brazil, AQUA.

International Journal of Ichthyology, 6, 165–176. Feary, D. A., Burt, J. A., & Bartholomew, A. (2011). Artificial marine habi-

tats in the Arabian Gulf: review of current use, benefits and manage-

ment implications. Ocean & Coastal Management, 54(10), 742–749. Firth, L. B., Mieszkowska, N., Thompson, R. C., & Hawkins, S. J. (2013). Cli-

mate change and adaptational impacts in coastal systems: The case of

sea defences. Environmental Science: Processes and Impacts, 15,

1665–1670.

560 SIMPSON ET AL.FISH

Foster, S. A., & Vincent, A. C. J. (2004). Life history and ecology of

seahorses: Implications for conservation and management. Journal of

Fish Biology, 65(1), 1–61. García-Gómez, J. C., López-Fé, C. M., Espinosa, F., Guerra-García, J. M., &

Rivera-Ingraham, G. A. (2011). Marine artificial micro-reserves: a possi-

bility for the conservation of endangered species living on artificial

substrata. Marine Ecology, 32(1), 6–14. Gittman, R. K., Scyphers, S. B., Smith, C. S., Neylan, I. P., & Grabowski, J. H.

(2016). Ecological consequences of shoreline hardening: A meta-analy-

sis. Bioscience, 66(9), 763–773. Gristina, M., Cardone, F., Carlucci, R., Castellano, L., Passarelli, S., &

Corriero, G. (2014). Abundance, distribution and habitat preference of

Hippocampus guttulatus and Hippocampus hippocampus in a semi-

enclosed Central Mediterranean marine area. Marine Ecology, 36,

57–66. https://doi.org/10.1111/maec.12116. Hale, R., & Swearer, S. E. (2017). When good animals love bad restored

habitats: How maladaptive habitat selection can constrain restoration.

Journal of Applied Ecology, 54(5), 1478–1486. Harasti, D. (2016). Declining seahorse populations linked to loss of essen-

tial marine habitats. Marine Ecology Progress Series, 546, 173–181. Harasti, D., Glasby, T. M., & Martin-Smith, K. M. (2010). Striking a balance

between retaining populations of protected seahorses and maintaining

swimming nets. Aquatic Conservation: Marine and Freshwater Ecosys-

tems, 20, 159–166. Harasti, D., Martin-Smith, K., & Gladstone, W. (2012). Population dynamics

and life history of a geographically restricted seahorse, Hippocampus

whitei. Journal of Fish Biology, 81, 1297–1314. Harasti, D., Martin-Smith, K., & Gladstone, W. (2014). Ontogenetic and

sex-based differences in habitat preferences and site fidelity of

White's seahorse Hippocampus whitei. Journal of Fish Biology, 85,

1413–1428. Hellyer, C. B., Harasti, D., & Poore, A. G. B. (2011). Manipulating artificial

habitats to benefit seahorses in Sydney Harbour, Australia. Aquatic

Conservation: Marine and Freshwater Ecosystems, 21, 582–589. Lai, S., Loke, L. H., Hilton, M. J., Bouma, T. J., & Todd, P. A. (2015). The

effects of urbanisation on coastal habitats and the potential for eco-

logical engineering: A Singapore case study. Ocean & Coastal Manage-

ment, 103, 78–85. Lourie, S. A., & Randall, J. E. (2003). A new pygmy seahorse, Hippocampus

denise (Teleostei Syngnathidae) from the Indo-Pacific. Zoological

Studies-Taipei, 42(2), 284–291. Mace, G. M., & Lande, R. (1991). Assessing extinction threats: Toward a

reevaluation of IUCN threatened species categories. Conservation Biol-

ogy, 5, 148–157. Martins, G. M., Thompson, R. C., Neto, A. I., Hawkins, S. J., & Jenkins, S. R.

(2010). Enhancing stocks of the exploited limpet Patella candei d’Orb- igny via modifications in coastal engineering. Biological Conservation,

143(1), 203–211. Manning, C., Forster, S., Harasti, D., & Vincent, A. (2018). A holistic investi-

gation of the ecological correlates of abundance and body size for the

endangered White's seahorse Hippocampus whitei. Journal of Fish Biol-

ogy, 93, 649–663. https://doi.org/10.1111/jfb.13745. Quin, G., Zhang, Y., Huang, L., & Lin, Q. (2014). Effects of water current on

swimming performance, ventilation frequency and feeding behavior of

young seahorses (Hippocampus erectus). Journal of Experimental Marine

Biology and Ecology, 461, 337–343. Perante, N. C., Vincent, A. C. J., & Pajaro, M. G. (1998). Demographics of

the seahorse Hippocampus comes in the central Philippines. Marine

Biology of the South China Sea, 3, 439–448. Ratkowsky, D., Alldredge, R., & Evans, M. A. (1992). Cross-over experiments:

Design, analysis and application (Vol. 135). Florida, United States: CRC

Press.

Regenwetter, M., Dana, J., & Davis-Stober, C. P. (2011). Transitivity of

preferences. Psychological Review, 118, 42–56. Schlaepfer, M. A., Runge, M. C., & Sherman, P. W. (2002). Ecological and

evolutionary traps. Trends in Ecology & Evolution, 17, 474–480. Sherley, R. B., Ludynia, K., Dyer, B. M., Lamont, T., Makhado, A. B.,

Roux, J.-P., … Votier, S. C. (2017). Metapopulation tracking juvenile penguins reveals an ecosystem-wide ecological trap. Current Biology,

27, 563–568. Simkanin, C., Davidson, I. C., Dower, J. F., Jamieson, G., & Therriault, T. W.

(2012). Anthropogenic structures and the infiltration of natural ben-

thos by invasive ascidians. Marine Ecology, 33(4), 499–511. Stuart, S. N., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues, A. S. L.,

Fischman, D. L., & Waller, R. W. (2004). Status and trends of amphib-

ian declines and extinctions worldwide. Science, 306, 1783–1786. Vincent, A. C., Evans, K. L., & Marsden, A. D. (2005). Home range behav-

iour of the monogamous Australian seahorse, Hippocampus whitei.

Environmental Biology of Fishes, 72(1), 1–12. Vincent, A. C., Foster, S. J., & Koldewey, H. J. (2011). Conservation and

management of seahorses and other Syngnathidae. Journal of Fish Biol-

ogy, 78(6), 1681–1724. Waycott, M., Duarte, C. M., Carruthers, T. J., Orth, R. J., Dennison, W. C.,

Olyarnik, S., … Kendrick, G. A. (2009). Accelerating loss of seagrasses across the globe threatens coastal ecosystems. Proceedings of the

National Academy of Sciences., 106(30), 12377–12381. Woods, C. M. C. (2002). Natural diet of the seahorse Hippocampus

abdominalis. New Zealand Journal of Marine and Freshwater Research,

36, 655–660. Woods, C. M., & Martin-Smith, K. M. (2004). Visible implant fluorescent

elastomer tagging of the big-bellied seahorse, Hippocampus

abdominalis. Fisheries research, 66(2-3), 363–371.

How to cite this article: Simpson M, Morris RL, Harasti D,

Coleman RA. The endangered White’s seahorse Hippocampus

whitei chooses artificial over natural habitats. J Fish Biol. 2019;

95:555–561. https://doi.org/10.1111/jfb.14002

SIMPSON ET AL. 561FISH