Lizard tail-loss rates on islands are not governed by longer life spans

Abstract

We recently studied whether, on islands, predation or intraspecific aggression is the main driver of tail-loss, a common defense mechanism among lizards. We concluded the latter was the stronger driver (Itescu et al. 2017). Werner (2017) suggested that we failed to falsify an alternative hypothesis. He claims that on low-predation islands lizards live longer. Thus while tail loss is caused by predators, it accumulates over longer periods, resulting in overall higher tail-loss rates in populations experiencing weak predation. Here we test this hypothesis and three other arguments he presented, and fail to support them. We therefore adhere to our original conclusion that intraspecific aggression is the main driver of lizard tail loss on islands.
Keywords: conspecific aggression; geckos; islands; longevity; predation; tail autotomy

Full text

<UN>
Israel Journal of Ecology & Evolution, 2017
http://dx.doi.org/10.1163/22244662-06301011
Introduction
Tail loss (autotomy) is a common defense mechanism
among many lizard groups (Bateman & Fleming 2009).
While traditionally it is considered an anti-predatory
mechanism (Arnold 1988; Chapple & Swain, 2002; Pafilis
et al. 2009a), accumulating evidence suggest it functions
also as defense against intraspecific aggression (Pafilis
etal. 2009b; Brock et al. 2015; Donihue et al. 2016).
Recently we studied whether tail loss of two species
of geckos (Mediodactylus kotschyi and Hemidactylus
turcicus) on islands is mainly driven by predation or by
intraspecific aggression (Itescu et al. 2017). We provided
several lines of evidence supporting the latter possibility,
and could find no data to suggest the former was impor-
tant in this system. Werner (2017) argues that we failed
to falsify an alternative hypothesis: that the high tail-loss
frequencies on predator-poor islands do result from preda-
tion, rather than from conspecific aggression. He claims
that on such islands lizards live longer, due to lower preda-
tion, and thus even if tail loss rates due to predation inci-
dents are low, they accumulate over longer periods. This
results in greater overall population-level regenerated-tail
frequencies (Werner 2017). His only reservation concerns
predator-free islands: he concurs with us that the high
tail-loss rates we have found on such islands are driven
by intraspecific aggression. We acknowledge that his
hypothesis is feasible in principle. We disagree, howev-
er, to the notion that it could aptly explain tail-loss fre-
quencies in our study system. Werner (2017) presented
no new observational or experimental data or analyses.
He provided only circumstantial evidence to support his
speculations. Therefore we tested the following hypotheses
he presented:
1. Tail-loss is expected to increase with body size, es-
pecially when sexes are examined separately, since
body size reflects lizard age.
2. Mediodactylus kotschyi inhabits many more islands
than H. turcicus because their regenerated tails keep
their cryptic form and function for avoiding preda-
tion, unlike the relatively conspicuous regenerated
tails of H. turcicus.
3. Mediodactylus kotschyi has higher autotomy rates
than H. turcicus among the populations we stud-
ied (Itescu et al. 2017) on both the islands and the
mainland.
4. Mediodactylus kotschyi is more exposed to predation
(i.e. encounters more predator species) than H. turci-
cus because it is cathemeral (active both at day and at
night), and this explains their higher tail-loss rates.
Methods
To test prediction #1 we used the sex-specific tail-loss
rates reported in Table S1 in Itescu et al. (2017) and sex-
specific body-size means for each studied population
(not presented in Itescu et al. 2017) that we report here
(Appendix1).
To test prediction #2 we used data from a literature
survey conducted by Itescu (2017) to gather all published
locality records of reptile species from Aegean and Ionian
Sea islands (our study system). We counted the numbers
of reptilian predator species of the two focal gecko species
on all the islands which only one of them inhabits, and
tested whether islands inhabited only by M. kotschyi have
more predators on average than those inhabited only by
H. turcicus.
Lizard tail-loss rates on islands are not governed by longer life spans
Yuval Itescua,*, Rachel Schwarza, Shai Meiria and Panayiotis Pafilisb
aSchool of Zoology, Tel Aviv University, Tel Aviv 6997801, Israel;
bSection of Zoology and Marine Biology, Department of Biology, National and Kapodistrian
University of Athens, Panepistimioupolis, Ilissia, Athens 157-84, Greece
Abstract We recently studied whether, on islands, predation or intraspecific aggression is the main driver of tail-loss, a
common defense mechanism among lizards. We concluded the latter was the stronger driver (Itescu et al. 2017). Werner
(2017) suggested that we failed to falsify an alternative hypothesis. He claims that on low-predation islands lizards live
longer. Thus while tail loss is caused by predators, it accumulates over longer periods, resulting in overall higher tail-loss
rates in populations experiencing weak predation. Here we test this hypothesis and three other arguments he presented, and
fail to support them. We therefore adhere to our original conclusion that intraspecific aggression is the main driver of lizard
tail loss on islands.
Keywords: conspecic aggression; geckos; islands; longevity; predation; tail autotomy
*Corresponding author. E-mail: yuvitescu@gmail.com
© Koninklijke Brill NV, Leiden, 2017

<UN>
2Y. Itescu et al.
We used the data we reported in the text and Table1
of Itescu et al. (2017) to test predictions #3 (both for the
islands and for the mainland) and #4. To test prediction
#4 we also counted diurnal predator richness (cathemeral
species were counted as nocturnal) on each island that both
geckos were sampled from. The nocturnal and cathemeral
predators of geckos that were excluded from this count
were Martes foina, Meles meles, Felis (sylvestris) catus,
Erinacerus concolor, Athene noctua, Eryx jaculus, Tele-
scopus fallax, and Vipera ammodytes (see Appendix S1 in
Itescu et al. 2017 for all potential predators on the islands).
We then divided autotomy rates of M. kotschyi by those of
H. turcicus on each such island and regressed that against
diurnal-predator richness on the island. We expected a pos-
itive relationship if indeed M. kotschyi are subject to more
predatory attacks from diurnal species (and thus more au-
totomy) due to their less-nocturnal life style.
Results
Regressing tail-loss rates on body size for each sex sepa-
rately, showed that the pattern is similar to what we reported
previously (Itescu et al. 2017) regardless of sex: mean body
size (which, to Werner, is equivalent to mean age) is not
related to tail-loss rates across insular populations in either
M. kotschyi (males: slope=0.009±0.008, n=41 islands,
R2=0.035, p= 0.236; females: slope = −0.003 ± 0.008,
n = 41, R2 = 0.005, p = 0.649) or H. turcicus (males:
slope=0.028±0.017, n=17, R2=0.159, p=0.118; females:
slope=−0.004±0.030, n=17, R2=0.001, p=0.898). This
refutes hypothesis 1.
The data we gathered from the literature on 427 islands
in the Aegean and Ionian Seas shows that M. kotschyi was
recorded on 289 islands, while H. turcicus is only known
from 121 islands. In contrast to expectations from hypothe-
sis #2, mean predator richness was higher on islands inhab-
ited only by H. turcicus, than on those inhabited only by
M. kotschyi (mean predator richness on 203 islands with
only M. kotschyi: 0.24± 0.04 [range: 0–4, median = 0],
mean predator richness on 35 islands with only H. turcicus:
2.94±0.75 [range: 0–14, median=1], t=3.59, p=0.001).
Thus a cryptic regenerated tail does not seem to be an ad-
aptation to increased predation, refuting hypothesis #2.
Mediodactylus kotschyi had higher autotomy rates
on 10 of the 17 islands from which we sampled both spe-
cies (Itescu et al. 2017), while H. turcicus had higher rates
on 7 islands. The dierence is not statistically significant
(χ2= 1.059, p=0.303). Furthermore, a paired t-test com-
paring tail-loss rates across these 17 islands shows no sig-
nificant dierence in tail-loss rate between these species
(M. kotschyi mean rate = 0.72 ± 0.02, H. turcicus =
0.66 ± 0.04, t = 1.257, p= 0.226). A χ2 test showed no
dierence on the mainland as well (χ2=10.23, p=0.332).
Thus prediction #3 is refuted for both the mainland and
the islands.
Regressing the M. kotschyi/H. turcicus tail-loss ratio
against diurnal-predator richness across islands showed
no significant relationship (slope = 0.050 ± 0.029, n =
17 islands, R2 = 0.166, p = 0.104). This demonstrates
M. kotschyi tail-loss rates do not increase due to their
less-nocturnal activity habits and hence their exposure to
diurnal predators in addition to nocturnal ones, and refutes
prediction #4.
Discussion
We have found that on predator-free islands tail loss rates
(self-evidently caused by conspecifics) are the highest
(Itescu et al. 2017). Thus, if conspecifics can drive tail loss
on these islands, we see no reason to infer that this does not
happen elsewhere, where predators exist, but simultane-
ously intraspecific competition is high. Furthermore, even
if on low-predation islands lizards live longer, tail-loss is
unlikely to be solely caused by predation, and not at all by
conspecific aggression.
Werner’s assumption requires an unsubstantiated fac-
tor: that the reduction in predation rates on islands is lower
than the relative increase in longevity (so that overall is-
land individuals will encounter more predation attempts
over their lifetimes). We are aware of no data that substan-
tiates either the notion that longevity is indeed higher on
islands in general, or that overall predation rates decrease
less than longevity increases. Another potential explana-
tion could be that insular predators are less ecient than
mainland predators, and hence more of their predation
attempts end in a lizard losing its tail, rather than being
preyed upon. Again we are unaware of any data that sup-
port this notion.
We view these explanations, without supporting evi-
dence, as non-parsimonious: we have shown that increased
gecko density and decreased number of predators result in
higher rates of tail loss. If on islands with few predators
lizards live longer, then the probability of tail breakage re-
sulting from intraspecific aggression will also increase (we
know of no record of intraspecific killing of adults in either
H. turcicus or M. kotschyi). To assume that reduced preda-
tion results in more predator-induced tail losses requires
more unsubstantiated assumptions as to the relative rate of
predation decrease vs. longevity increase, or unsubstanti-
ated assumptions regarding predation eciency.
Werner (2017) implies (without data), that mean adult
body size on islands reflects mean adult age in reptiles. We
suspect this is untrue when multiple insular populations
of a single species are considered: in our study system
on the Aegean islands, lizards such as Podarcis spp. and
M. kotschyi have inter-island variability in size that can
amount to almost double in SVL and threefold dierence
in body weight (Pafilis etal. 2009b; Meiri et al. 2014; Itescu
et al. 2018). In fact, adults in some populations are small-
er than sub-adults on other islands (Fig.1). Furthermore,
across its large range we have shown that size maxima for
H. turcicus occur on islands that harbor predators (Itescu
et al. 2016). These were not expected to be particularly
long-living, according to Werner, and hence would be
predicted to be small. Furthermore, inter-island size dif-
ferences are already observed in hatchlings (e.g., Pafilis et
al. 2009b), and are maintained through growth to adult-
hood (Schwarz 2016). The age argument therefore does not
seem to hold. Without data on actual age structure of each
population (which we currently do not possess), regressing

Israel Journal of Ecology & Evolution 3
<UN>
body size against autotomy rates across populations is the
best test we can provide for this hypothesis. We found and
reported no correlation between body size and autotomy
rates before (Itescu et al. 2017) and here we showed that
analyzing males and females separately does not change
the picture. Thus, neither our previous nor our new results
provide any support for this hypothesis.
Werner (2017) further argues that the ability for ef-
fective crypsis for M. kotschyi is kept even after tail loss
and regeneration, while that of H. turcicus does not. He
postulates that this allows M. kotschyi to better cope with
predators and possibly populate more islands. Our results
imply the exact opposite: H. turcicus inhabits islands with
more predators, suggesting that if anything, H. turcicus
can survive better on predator-rich islands. Mediodacty-
lus kotschyi on the other hand abounds on small, predator
free islands as well as on large ones. We suspect this may
reflect colonization or survival ability rather than an anti-
predatory mechanism.
The statement that our data show M. kotschyi has
higher autotomy rates than H. turcicus is simply errone-
ous. Table1 of Itescu et al. (2017) and associated analyses
here show this is wrong. We also suspect that the argument
regarding the reduced exposure of H. turcicus to preda-
tors because of its more nocturnal habits is neither correct
nor relevant. Although M. kotschyi is indeed active both
at day and at night, while H. turcicus is mainly nocturnal,
the main predators of both these geckos on Aegean islands
(i.e., Vipera ammodytes and Telescopus fallax, Cattaneo
2010) are mainly nocturnal, and all islands from which
we sampled H. turcicus harbor nocturnal predators. Our
results here show that the diurnal predators are not driv-
ing increased tail loss in M. kotschyi. That said, activity
habit dierence producing more autotomy in M. kotschyi
is probably irrelevant since there is no dierence between
tail-loss rates in these species.
We therefore conclude that Werner’s hypotheses, al-
though feasible in principle, are neither the most parsimo-
nious nor are they supported by the data that were available
to him or by the new data presented herein. Consequently,
we adhere to our original conclusion (Itescu et al. 2017):
intraspecific aggression has a stronger impact than preda-
tion on tail-loss rates across the insular lizard populations
we studied. Our findings come to corroborate previous
studies that clearly demonstrated the predominant role of
intraspecific competition in caudal autotomy (Brock et al.
2015; Cooper et al. 2015; Donihue et al. 2016).
Acknowledgments
The study was funded by an Israel Science Foundation (ISF)
Grant #1005/12. We thank Herve Seligmann and an anonymous
referee for comments on an earlier draft of this manuscript.
References
Arnold EN. (1988). Caudal autotomy as a defense. In: Biology of
the Reptilia. New York: Alan R. Liss. pp. 235–273.
Bateman PW, Fleming PA. (2009). To cut a long tail short: a re-
view of lizard caudal autotomy studies carried out over the
last 20 years. J Zool. 277, pp. 1–14.
Brock KM, Bedneko PA, Pafilis P, Foufopoulos J. (2015). Evo-
lution of antipredator behavior in an island lizard species,
Podarcis erhardii (Reptilia: Lacertidae): The sum of all
fears? Evolution. 69, pp. 216–231.
Cattaneo A. (2010). Note eco-morfologiche su alcune specie
ofidiche Egee, con particolare riferimento alle popolazioni
delle Cicladi centro-orientali (Reptilia). Natural. Sicil. 34,
pp. 319–350.
Chapple DG, Swain R. (2002). Eect of caudal autotomy on lo-
comotor performance in a viviparous skink, Niveoscincus
metallicus. Funct Ecol. 16, pp. 817–825.
Cooper WE, Dimopoulos I, Pafilis P. (2015). Sex, age, and popu-
lation density aect aggressive behaviors in island lizards
promoting cannibalism. Ethology 121, pp. 260–269.
Donihue CM, Brock KM, Foufopoulos J, Herrel A. (2016). Feed
or fight: testing the impact of food availability and intraspe-
cific aggression on the functional ecology of an island lizard.
Funct Ecol. 30, pp. 566–575.
Itescu Y, Schwarz R, Donihue C, Slavenko A, Roussos SA, Sago-
nas K,Valakos ED, Foufopoulos J, Pafilis P, Meiri S. (2018).
Inconsistent patterns of body size evolution in co-occurring
island reptiles. Global Ecol Biogeogr. In press.
Figure1. Mediodactylus kotschyi specimens demonstrating body size dierences across island populations: left – a young individual from
Amorgos Island; middle – an adult female from Amorgos Island; right – an adult female from Gavdos Island. Specimens are from Natural
History Museum Crete. Photo taken by Yuval Itescu.

Y. Itescu et al.
<UN>
4
Itescu Y, Schwarz R, Meiri S, Pafilis P. (2017). Intra-specific com-
petition, not predation, drives lizard tail loss on islands. J
Anim Ecol 86, pp. 66–74.
Itescu Y, Schwarz R, Moses M, Pafilis P, Meiri S. (2016). Record
sizes for the Turkish house gecko, Hemidactylus turcicus,
from Aegean islands, Greece. Herpetol. Bull. 138, pp. 24–26.
Meiri S, Kadison AE, Novosolov M, Pafilis P, Foufopoulos J,
Itescu Y, Raia P, Pincheira-Donoso D. (2014). The number of
competitor species is unlinked to sexual dimorphism. J Anim
Ecol 83, pp. 1302–1312.
Pafilis P, Foufopoulos J, Poulakakis N, Lymberakis P, Valakos
ED. (2009a). Tail shedding in island lizards [Lacertidae,
Reptilia]: decline of antipredator defenses in relaxed preda-
tion environments. Evolution 63, pp. 1262–1278.
Pafilis P, Meiri S, Foufopoulos J, Valakos E. (2009b). Intraspe-
cific competition and high food availability are associated
with insular gigantism in a lizard. Naturwissenschaften 96,
pp. 1107–1113.
Schwarz R. (2016). Evolution of life history traits in lizards: A
local and global approach. M.Sc. thesis, School of Zoology,
Tel-Aviv University.
Werner YL. (2017). Commentary on the factors governing the
rate of tail loss in island lizards. Isr J Ecol Evol. In press.
Appendix 1. Data for sex-specific mean body size (in mm) in insular populations.
Island Mediodactylus kotschyi Hemidactylus turcicus
Females Males Females Males
Agios Eustathios 44.06 43.10
Amorgos 47.27 46.75 52.49 49.83
Anafi 46.05 41.66 52.09 50.41
Andreas 47.84 44.27
Andros 45.40 41.49 53.37 51.78
Antiparos 41.68 40.59 45.64 49.36
Apano Kufonisi 46.29 46.72 48.81 52.06
Aspronisi 46.16 43.61
Despotiko 42.43 39.81 49.55 48.60
Folegandros 49.32 47.25
Glaronisi 47.33 44.76
Ios 43.81 42.66 49.36 49.58
Iraklia 45.89 43.09 46.91 44.24
Karpathos 38.97 35.02
Kasos 37.81 34.96
Kato Fira 42.00 41.18
Kato Kufonisi 44.15 45.29
Kimolos 44.75 42.56 55.31 54.08
Kitriani 39.10 38.33
Kopria 45.21 47.01
Kythnos 43.44 39.62 49.22 49.23
Megali Fteno 44.35 47.21
Mikri Fteno 45.85 46.54
Milos 45.45 44.36
Mykonos 45.94 42.72 49.84 50.54
Naxos 42.99 42.69 48.42 50.87
Nikouria 46.82 45.15
Pachia 47.07 44.86
Palakida 34.57 34.04
Panteronisi 43.61 42.73
Paros 43.59 41.75 49.80 45.51
Polyaigos 44.56 44.38
Saria 38.32 35.97
Schinoussa 49.18 46.98 49.64 46.47
Serifos 45.71 42.49 46.92 46.59
Sifnos 43.08 41.63
Sikinos 43.69 42.72 44.93 46.28
Syrna 48.01 45.52
Syros 42.40 42.32 46.10 47.89
Tinos 42.50 40.71
Tsimintiri 41.80 40.38
Venetiko 43.74 41.61