Translocation of cave fish (Poecilia mexicana) within and between natural habitats along a toxicity gradient

Page created by Paula Foster
 
CONTINUE READING
Ecology of Freshwater Fish 2012                                                                                Ó 2012 John Wiley & Sons A/S
Printed in Malaysia  All rights reserved
                                                                                                               ECOLOGY OF
                                                                                                             FRESHWATER FISH

Translocation of cave fish (Poecilia mexicana)
within and between natural habitats along a
toxicity gradient
Ingo Schlupp1, Timothy J. Colston2, Brandon L. Joachim1, Rüdiger Riesch3
1
 Department of Biology, University of Oklahoma, 730 Van Vleet Oval, Norman, OK 73019, USA
2
 Biology Department, University of Mississippi, Box 1848, University, MS 38677, USA
3
 Department of Biology, North Carolina State University, 127 David Clark Labs, Raleigh, NC 27695-7617, USA

Accepted for publication October 16, 2012

Abstract – During ecological speciation, diverging taxa have the potential to remain in close spatial proximity
(i.e., sympatry or micro-allopatry) theoretically allowing for continued contact and gene flow. In a system where
incipient speciation of populations of the Atlantic molly (Poecilia mexicana) appears to be driven by abiotic
factors, we investigated whether one of these factors, hydrogen sulphide (H2S) toxicity, also constitutes an
effective barrier to slow migration within and between habitats. We addressed this experimentally by
translocating individuals from high toxicity to lower toxicity within a toxic cave and by translocating individuals
from the toxic cave to a nontoxic surface habitat. Using a stepwise-backwards Cox regression, we found that
overall mortality was low, but statistically significant mortality occurred when individuals were transferred from
higher toxicity to lower toxicity. In addition, only males were negatively affected by being transferred from low
levels of toxicity to nontoxic, surface waters. This indicates that in addition to abiotic factors, other mechanisms,
like predation and sexual selection, must be important in maintaining population separation.

Key words: ecological speciation; divergence; habituation; acclimation; troglobite

                                                                                 mechanisms in detail likely also advances our under-
Introduction
                                                                                 standing of how sympatric speciation occurs.
While it has long been accepted that geographic                                     Abiotic factors often play an important role in
isolation can lead to speciation (i.e., allopatric specia-                       maintaining population separation and divergence
tion; Mayr 1963; Coyne & Orr 2004), a number of                                  (Tobler et al. 2006). However, these factors often
mechanisms, including sexual (Maan & Seehausen                                   represent soft gradients, rather than sharp clines
2011) and natural selection (Schluter 1993; Riesch                               (i.e., abrupt demarcation with presence of a factor on
et al. 2011b), can lead to population divergence and                             one side and absolute absence of that factor on the
eventually to ecological speciation even in sympatry                             other), along which populations are positioned. It is
(Rundle & Nosil 2005). Once populations show signs                               important to understand to which degree these differ-
of phenotypic and genetic divergence, and are on an                              ent gradients contribute to divergence, and a sharper
evolutionary trajectory that could lead to speciation,                           cline is thought to lead to conditions under which
mechanisms must be in place to reduce the probabil-                              divergence can evolve easier (Nosil et al. 2009a,b).
ity of reticulation, or the collapse of the nascent spe-                         Classical work in botany, for example, has estab-
cies (Woehrer et al. 2012). Mechanisms responsible                               lished how important sharp clines in soil composition
for maintaining species separation were potentially                              can be in population divergence (Dickinson &
also important in generating divergence (but see                                 Antonovics 1973; Antonovics & Thrall 1994). A
Coyne & Orr 2004), and thus studying these                                       combination of sharp and soft gradients, on the other

Correspondence: I. Schlupp, Department of Biology, University of Oklahoma, 730 Van Vleet Oval, Norman, OK 73019, USA. E-mail: schlupp@ou.edu

doi: 10.1111/eff.12017                                                                                                                         1
Schlupp et al.
hand, along which organisms can potentially move          selection, which would enable us to study the under-
up and down, occur in a cluster of populations of the     lying genomic architecture of divergence (Nosil
Atlantic molly (Poecilia mexicana) from Tabasco,          2012).
Mexico (Tobler & Plath 2011). Here, a population of          Previous work on mechanisms isolating the cave
Atlantic mollies has colonised a toxic cave, the          and the surface populations has documented that sud-
Cueva del Azufre (or Cueva de Villa Luz), that has        den translocation of surface fish into nonfamiliar hab-
been under study since its scientific discovery in the     itat (i.e., toxic to nontoxic and vice versa) leads to
1960s (Gordon & Rosen 1962; Parzefall 2001). The          high mortality (Tobler et al. 2009; Plath et al. 2010),
Cueva del Azufre is divided into 13 different cham-       while sudden translocation within toxic habitats (i.e.,
bers, with chamber XIII being the innermost chamber       toxic surface to toxic cave and vice versa) yielded no
(Gordon & Rosen 1962). Several springs in the cave        measurable mortality (Tobler et al. 2009). Thus,
(mainly in chamber X) release water that is extremely     toxicity limits successful migration and promotes iso-
rich in hydrogen sulphide, thus forming a toxicity        lation between fish from nontoxic and toxic habitats.
gradient within the cave that varies in levels of         The proposed mechanism behind the high mortality
hydrogen sulphide depending on the relative position      is physiological incompatibility with the new condi-
to the springs (Tobler et al. 2006). As the toxic water   tions for either form (Tobler et al. 2009). Nonethe-
flows downstream in the cave, it is oxygenated and,        less, this result comes as somewhat of a surprise,
as a result, becomes less toxic with greater distance     because over the last couple of decades, experimental
from the springs. The creek that flows through the         populations of P. mexicana have successfully and
cave eventually leaves the cave and becomes the           repeatedly been established from toxic (surface and
toxic El Azufre, a surface creek also containing          cave) habitats in benign (i.e., nontoxic) laboratory
hydrogen sulphide. While toxicity represents a some-      environments after acclimation to nontoxic water.
what soft gradient, the change from darkness within       Why then was mortality in the field experiments so
the cave to light at the mouth of the cave is abrupt.     high? This could be due to the sudden exposure to
   The community of organisms inhabiting the waters       either relatively high levels of hydrogen sulphide (for
of the Cueva del Azufre and El Azufre is extremely        fish from nontoxic waters) or relatively high levels of
depauperate given its geographic location in the tro-     oxygen (for fish from toxic waters) following translo-
pics (Gordon & Rosen 1962). The Atlantic molly is         cation (Tobler et al. 2009).
the only fish species that has permanently colonised          Thus, while previous translocation experiments
the spring habitats within the cave (Tobler et al.        have addressed how well individuals would fair after
2006; Riesch et al. 2009). Populations of P. mexicana     rapid transport into a foreign environment, the more
inhabiting the toxic cave are fundamentally different     plausible scenario of slow and gradual acclimation
from surface-living populations in almost every           of individuals to a new habitat has so far not been
aspect of their natural history (e.g., behaviour: Plath   considered experimentally. As toxicity in the cave
et al. 2004; morphology: Tobler et al. 2008a; life his-   molly system represents a gradient, one can hypoth-
tory: Riesch et al. 2010b), and using common-garden       esise that slower, more gradual movement along the
experiments, these differences have repeatedly been       gradient would allow individuals a higher probability
shown to have a strong heritable component (e.g.,         of survival when migrating from more toxic to less
behaviour: Parzefall 2001; morphology: Tobler et al.      toxic conditions. This question is also important
2008a; life history: Riesch et al. 2011a; Physiology:     relative to the surprisingly strong genetic structuring
Plath & Tobler 2010). Thus, the combined evidence         within the toxic cave. While there is strong genetic
indicates a system in which ongoing, incipient specia-    structure even between chambers within the Cueva
tion is occurring (Plath et al. 2007a; Tobler & Plath     del Azfure, there is, nonetheless, evidence for low
2011). Clearly such early splitting events are labile     levels of gene flow between different cave chambers.
and vulnerable to collapse if conditions change (Wu       On the other hand, we have so far found no evi-
& Ting 2004; Hendry 2009; Nosil et al. 2009a,b;           dence for any gene flow from migrants between the
Bolnick 2011; Yeaman & Whitlock 2011; Woehrer             toxic surface and toxic cave on one side and the
et al. 2012), although so far we have no evidence for     nontoxic surface habitats on the other side, further
this happening in P. mexicana. In fact, we have           corroborated by FST values that are usually above
strong evidence indicating that population divergence     0.2 for any possible pairwise comparison (Plath
is relatively stable because phenotypic and genetic       et al. 2007a, 2010; Tobler et al. 2008a, 2009). For
divergence has remained strong even if faced by           example, Tobler et al. 2009 were able to demon-
strong temporary disturbance due to natural catastro-     strate that, based on microsatellite DNA, less than
phes (like extreme flooding events; Plath et al. 2010).    10% first-generation migrants move from the inside
This said, there are other outcomes possible, includ-     of the Cueva del Azufre into the adjacent toxic
ing stable divergence with limited gene flow and           surface stream, while first-generation migrants within

2
Translocation of cave fish
these habitats (e.g., between different cave cham-        and because seining is impossible due to the topogra-
bers) ranged anywhere from 5 to 35%. Interestingly,       phy of the cave. Fishes were subsequently moved to
a study investigating the population genetic struc-       individual containers (30.48 9 8.89 9 17.78 cm) (L
ture, following a millennium flood in 2007 (Plath          9 H 9 W) and placed in cave chambers that featured
et al. 2010), found slightly higher numbers of            lower concentrations of H2S than the local area of
migrants within the cave immediately after the            collection (i.e., chamber X fish were moved to cham-
event, pointing to a potential role of rare but cata-     ber V; chamber V fish were moved to chamber II and
strophic events. What maintains population structur-      chamber V fish moved to the surface habitat). As a
ing between (and within) the Cueva del Azufre and         control, fishes from chamber V were captured and
the nontoxic surface habitats (Plath et al. 2007a,        placed in individual containers within the same
2010; Tobler et al. 2008a)? Are individuals so well       chamber. We aimed for an even sex ratio in the
adapted to the local conditions that dispersal even       experimental fish, but were somewhat limited by the
within the cave is limited?                               fish available in the respective chambers (Table 1).
   In this study, we addressed the following ques-           For the experiment in which chamber V fish were
tions: how do fish fair if they are translocated along     translocated to the surface habitat, we used a modi-
the toxicity gradient within the cave, where the other    fied protocol. Fishes were collected as described,
environmental      gradient     (darkness)     remains    kept in two 19-litre coolers in their habitat-specific
unchanged? Furthermore, does acclimation to the           water. This water was aerated for 4 h using portable
local conditions play a role when fish are transferred     air pumps, which slowly removed hydrogen sul-
from the toxic cave to nontoxic surface waters?           phide and oxygenized the water. After 4 h, we grad-
Finally, do the sexes respond differently to transloca-   ually replaced the cave water with river water from
tions? To tackle these questions, we conducted three      the nearby Río Oxolotán, a freshwater surface habi-
separate translocation experiments, all of which were     tat. The next morning fish were then placed in indi-
designed to take individuals from higher toxicity to      vidual containers in a surface habitat, Arroyo
lower toxicity. In the first experiment, we translocat-    Bonita, a small tributary to the Río Oxolotán (both
ed fish from the very toxic chamber X to the less          the Río Oxolotán and the Arroyo Bonita are natu-
toxic chamber V. In our second experiment, we             rally populated by the surface form of P. mexicana).
moved individuals from chamber V to the even less         Overnight mortality was low; during our overnight
toxic chamber II, and finally, in a third experiment,      acclimation period of chamber V fish, only one male
we translocated fishes from chamber V to a nontoxic        died.
surface habitat after acclimating them to nontoxic           Previous studies have utilised large containers
water overnight.                                          (buckets with holes drilled into them) with multiple
                                                          fish in the same container (Tobler et al. 2009). How-
                                                          ever, we chose to use smaller individual containers to
Materials and methods
                                                          minimise any stress that may be associated with
                                                          crowding. Additionally, small holes were drilled in
Study system
                                                          each container to allow for the circulation of water.
All experiments were conducted in January 2010            Each container contained one fish and was weighed
near Tapijulapa, Tabasco, Mexico. We selected fish         down with small rocks that were gathered locally.
from habitats containing high hydrogen sulphide           Containers were placed in a shallow area with low
concentrations (chamber X: 91.5 ± 32.3 lM), low           flow to avoid having the containers washed away.
concentrations of hydrogen sulphide (chamber V:           Due to the low oxygen concentrations in the Cueva
28.4 ± 15.3 lM) and very low concentrations of            del Azufre, cavernicolous P. mexicana spend the
hydrogen sulphide (chamber II: 24.0 ± 21.6 lM)            majority of their time at the water’s surface perform-
(Tobler et al. 2006). Additionally, fish were translo-     ing aquatic surface respiration (ASR) to exploit
cated to a nontoxic surface stream, the Arroyo            increased oxygen levels (Tobler et al. 2006). There-
Bonita (17o25′37.42″N, 92o45′6.98″W). The Arroyo          fore, we did not fully submerge our containers to
Bonita is approximately 2 km upstream from where          avoid mortality that could be attributed to behaviour-
the outflow of the Cueva del Azufre connects to the        al limitations. Once all fish were placed in individual
Río Oxolotán, and does not contain hydrogen sul-          containers, we allowed 1 h for acclimatisation and
phide (Tobler et al. 2008a).                              subsequently checked the condition of the fish in
                                                          each habitat (see Table 1). At the termination of each
                                                          experiment (after 24 h), fish were removed from
Local adaptation and natural selection
                                                          containers, sacrificed and preserved in 5% formalde-
Fish were collected by hand with small dip nets in        hyde to avoid inadvertent mixing of populations. All
each cave chamber to minimise stress from handling        individuals used in the experiment were adult. For

                                                                                                              3
Schlupp et al.
Table 1. Detail of the translocation experiments.
                                                                                             (a)
                                         Sample size       Mortality
Experiment    Date       From → to       (males/females)   (males/females)

1a            090111     X   →V          7/13              2/2
1b            100111     X   →V          9/11              2/1
2a            090111     V   → II        9/11              0/0
2b            100111     V   → II        10/10             0/0
3             120111     V   →V          10/10             0/0
4             130111     V   → Bonita    21/19             6/0

logistical reasons, we were unable to conduct a study
that would have measured mortality over a longer
term. Similarly, we were unable to conduct a control
experiment within chamber X.                                                                 (b)

Statistical analysis
We used a stepwise-backwards Cox regression based
on likelihood ratios to analyse individual survival of
P. mexicana as implemented in IBM® SPSS® Statis-
tics 19.0.0 (IBM Corporation). In the original model,
‘treatment’ (PSX-to-PSV, PSV-to-PSV, PSV-to-PSII
and PSV-to-AB), ‘sex’ (male vs. female) and the
interaction of ‘treatment-by-sex’ were included as
covariates.

Results                                                                      Fig. 1. Survival plot indicating cumulative survival (%) over time
When analysing the survival of cave-dwelling                                 for (a) fish translocated between from chamber X to chamber V
                                                                             within the toxic Cueva del Azufre and (b) fish translocated from
P. mexicana in different habitats, the covariate ‘sex’                       the inside of the toxic Cueva del Azufre (chamber V) to a
(P = 0.868) was removed from the final Cox regres-                            nontoxic surface stream (Arroyo Bonita).
sion model during the stepwise-backwards approach.
Only the covariates ‘treatment’ (B = 0.84,
SE = 0.39, Wald = 4.69, d.f. = 1, P = 0.03) and the                          gence and ultimately ecological speciation. Sharp
interaction term ‘treatment-by-sex’ (B = 0.58,                               clines, like the difference between dark and light in
SE = 0.29, Wald = 3.91, d.f. = 1, P = 0.048) were                            the cave molly system, are much more likely to aid
retained as statistically significant in the final model                       population separation than soft gradients (Nosil et al.
( 2 log likelihood = 127.42).                                                2009a,b), as exemplified by the hydrogen sulphide
   Overall we detected very low mortality across                             gradient known for the cave molly within the Cueva
treatments. No mortality was uncovered in two treat-                         del Azufre. In our particular system, we examined
ments (PSV-to-PSV and PSV-to-PSII), while some                               how differences in hydrogen sulphide toxicity affect
mortality occurred in the other two treatments (PSX-                         the survival of immigrants, and therefore population
to-PSV and PSV-to-AB, Fig. 1; significant ‘treat-                             structuring, between populations within the Cueva
ment’ effect). Furthermore, while males and females                          del Azufre. Previous research indicated that abrupt
suffered roughly similar mortality rates when trans-                         changes in water conditions (i.e., toxic to nontoxic
ferred from highly toxic waters (PSX) to low levels                          and vice versa) lead to very high mortality in translo-
of toxicity (PSV), only males were negatively                                cated fishes, but that translocation between chamber
affected by being transferred from low levels of                             II and the El Azufre (i.e., toxic cave and toxic sur-
toxicity (PSV) to nontoxic waters (AB, Fig. 1;                               face) resulted in no measurable mortality (Tobler
significant ‘treatment-by-sex’ effect).                                       et al. 2009). In the experiment presented here, in
                                                                             which we tested for survivorship during gradual
                                                                             migration from highly toxic to lesser toxic microhabi-
Discussion
                                                                             tats (i.e., cave chambers) within the cave, we found
How animals can move along relevant environmental                            low levels of mortality in two of four treatments. This
gradients is important for our understanding of diver-                       indicates that differences in toxicity found within the

4
Translocation of cave fish
cave do not fully explain the genetic structure uncov-       induced translocation during a religious ceremony, in
ered by previous studies (Plath et al. 2007a, 2010;          which cave fishes are poisoned and immobilized
Tobler et al. 2008a). Clearly, factors other than the        resulting in passive downstream drifting (Tobler et al.
toxicity gradient must play a role here in limiting          2011a), is not likely to lead to successful gene flow
between-chamber movement, and we can only specu-             due to the high mortality induced by rapid transloca-
late that they could include sexual selection, preda-        tion between extreme and benign environments (To-
tion and age-specific mortality. Furthermore, we              bler et al. 2009). Under the scenario that we
found that translocation from a toxic cave habitat to a      investigated here that envisions gradual movement
nontoxic surface habitat results in relatively low mor-      along a gradient allowing for potential acclimatisation,
tality if the water is slowly detoxified and fish have         previously described additional selective forces acting
time to adjust to the new conditions.                        against migrants (e.g., predation: Tobler 2009; Riesch
   Interestingly, males turned out to be significantly        et al. 2010a; sexual selection: Tobler et al. 2008b) are
more likely to die than females, when translocated           likely to contribute more to reproductive isolation than
from toxic waters within the cave to a nontoxic              previously assumed (Tobler 2009). Finally, the higher
surface water habitat. Sex-specific responses to physi-       survival of females, nonetheless, raises the interesting
ological stress are widespread among animal taxa (e.         possibility of sex-specific dispersal that should be
g., insects: Freitak et al. 2012; rats: Sterrenburg et al.   investigated further in future studies.
2011; humans: Verma et al. 2011), and even though
this depends on the taxa under investigation and the
stressors involved, males often seem to be more sen-         Acknowledgements
sitive to stressful conditions (Freitak et al. 2012; but     We are very grateful to the people and authorities of Tapijula-
see Riesch et al. 2011b). In our system, it is not clear     pa and the Federal authorities of Mexico for granting permits.
what exactly is causing this. Nonetheless, our finding        Francisco García de León and Lenin Arias-Rodriguez helped
is congruent with data from previous field experi-            logistically and with fieldwork. Funding for this project came
ments in this system in which males usually also per-        from NSF (DEB-0743406). This work was conducted under
formed worse than females (Plath et al. 2007a,b).            University of Oklahoma IACUC R09-030. We are also thank-
                                                             ful to the reviewers who helped greatly to improve the manu-
Likely males are physiologically less capable of buf-
                                                             script. IS gratefully acknowledges support from the Alexander
fering adverse conditions than females, potentially
                                                             von Humboldt Foundation.
due to the lower body condition of cave molly males
exhibited within the Cueva del Azufre, indicating the
males might be more energy limited than females              References
(Riesch et al. 2010b, 2011a,b). Additional indirect
                                                             Antonovics, J. & Thrall, P.H. 1994. The cost of resistance and
evidence for this comes from studies investigating
                                                               maintenance of genetic polymorphism in host-pathogen sys-
male–male aggression and the propensity of males to            tems. Proceedings of the Royal Society of London. Series
engage in sexual harassment of females (e.g., Parze-           B: Biological Sciences (London) B257: 105–110.
fall 2001; Bierbach et al. 2012). While these behav-         Bierbach, D., Klein, M., Sassmannshausen, V., Schlupp, I.,
iours are widespread in P. mexicana from benign                Riesch, R., Parzefall, J. & Plath, M. 2012. Divergent evolu-
habitats, such costly behaviours are highly reduced in         tion of male aggressive behaviour: another reproductive iso-
the Cueva del Azufre, likely as an adaptation to pre-          lation barrier in extremophile Poeciliid fishes? International
serve much needed energy for reproduction (Parzefall           Journal of Evolutionary Biology. DOI:10.1155/2012/148745.
2001; Bierbach et al. 2012). Finally, while our results      Bolnick, D.I. 2011. Sympatric speciation in Threespine Stick-
could also indicate that buffering ability might poten-        leback: why not? International Journal of Ecology.
tially increase with size (and P. mexicana males are           DOI:10.1155/2011/942847.
                                                             Coyne, J.A. & Orr, H.A. 2004. Speciation. Sunderland: Sina-
on average smaller than females; Riesch et al. 2010b,
                                                               uer Associates.
2011a,b), this interpretation is unlikely because a pre-     Dickinson, H. & Antonovics, J. 1973. Theoretical consider-
vious study has shown that smaller adults actually             ations of sympatric divergence. The American Naturalist
have a higher sulphide tolerance than larger individu-         107: 256–274.
als (Tobler et al. 2011b).                                   Freitak, D., Knorr, E., Vogel, H. & Vilcinskas, A. 2012.
   In summary, our experiments indicate that under             Gender- and stressor-specific microRNA expression in
certain conditions (i.e., gradual transfer along the           Tribolium castaneum. Biology Letters 8: 860–863.
gradient from toxic to nontoxic), factors other than         Gordon, M.S. & Rosen, D.E. 1962. A cavernicolous form of
toxicity must also play an important role in maintain-         the poeciliid fish Poecilia sphenops from Tabasco, Mexico.
ing reproductive isolation reported on by previous             Copeia 1962: 360–368.
studies (Plath et al. 2007a, 2010; Tobler et al. 2008a).     Hendry, A.P. 2009. Ecological speciation! Or the lack thereof?
                                                               Canadian Journal of Fisheries and Aquatic Sciences 66:
Clearly, sudden translocation as it is possible during
                                                               1383–1398.
extreme flooding events (Plath et al. 2010) or human-

                                                                                                                          5
Schlupp et al.
Maan, M.E. & Seehausen, O. 2011. Ecology, sexual selection          Riesch, R., Plath, M. & Schlupp, I. 2011b. Speciation in
  and speciation. Ecology Letters 14: 591–602.                        caves: experimental evidence that permanent darkness pro-
Mayr, E. 1963. Animal species and evolution. Harvard: Har-            motes reproductive isolation. Biology Letters 7: 909–912.
  vard University Press.                                            Rundle, H.D. & Nosil, P. 2005. Ecological speciation. Ecol-
Nosil, P. 2012. Ecological speciation. Oxford: Oxford Univer-         ogy Letters 8: 336–352.
  sity Press.                                                       Schluter, D. 1993. Adaptive radiation in sticklebacks: size,
Nosil, P., Harmon, L.J. & Seehausen, O. 2009a. Ecological             shape, and habitat use efficiency. Ecology 74: 699–709.
  explanations for (incomplete) speciation. Trends in Ecology       Sterrenburg, L., Gaszner, B., Boerrigter, J., Santbergen, L.,
  & Evolution 24: 145–156.                                            Bramini, M., Elliott, E., Chen, A., Peeters, B.W.M.M., Rou-
Nosil, P., Funk, D.J. & Ortiz-Barrientos, D. 2009b. Divergent         bos, E.W. & Kozicz, T. 2011. Chronic stress induces sex-
  selection and heterogeneous genomic divergence. Molecular           specific alterations in methylation and expression of cortico-
  Ecology 18: 375–402.                                                tropin-releasing factor gene in the rat. PLoS ONE 6:
Parzefall, J. 2001. A review of morphological and behaviour-          e28128.
  al changes in the cave molly, Poecilia mexicana, from             Tobler, M. 2009. Does a predatory insect contribute to the
  Tabasco, Mexico. Environmental Biology of Fishes 62:                divergence between cave- and surface-adapted fish popula-
  263–275.                                                            tions? Biology Letters 5: 506–509.
Plath, M. & Tobler, M. 2010. Subterranean Fishes of Mexico          Tobler, M. & Plath, M. 2011. Living in extreme environ-
  (Poecilia mexicana, Poeciliidae). In: Trajano, E., Bichuette,       ments. In: Evans, J.P., Pilastro, A. & Schlupp, I., eds. Ecol-
  M.E. & Kapoor, B.G., ed. Biology of subterranean fishes.             ogy and evolution of poeciliid fishes. Chicago: University of
  pp. 281–330.                                                        Chicago Press, pp. 120–127.
Plath, M., Parzefall, J., Körner, K.E. & Schlupp, I. 2004. Sex-     Tobler, M., Schlupp, I., Heubel, K.U., Riesch, R., Garcia de
  ual selection in darkness? Female mating preferences in sur-        Leon, F.J., Giere, O. & Plath, M. 2006. Life on the edge:
  face- and cave-dwelling Atlantic mollies, Poecilia mexicana         hydrogen sulfide and the fish communities of a Mexican
  (Poeciliidae, Teleostei) Behavioural Ecology and Sociobiol-         cave and surrounding waters. Extremophiles 10: 577–585.
  ogy 55: 596–601.                                                  Tobler, M., DeWitt, T.J., Schlupp, I., García de León, F.J.,
Plath, M., Hauswaldt, J.S., Moll, K., Tobler, M., Garcia De           Herrmann, R., Feulner, P.G.D., Tiedemann, R. & Plath, M.
  Leon, F.J., Schlupp, I. & Tiedemann, R. 2007a. Local adapta-        2008a. Toxic hydrogen sulfide and dark caves: phenotypic
  tion and pronounced genetic differentiation in an extremo-          and genetic divergence across two abiotic environmental
  phile fish, Poecilia mexicana, inhabiting a Mexican cave with        gradients in Poecilia mexicana. Evolution 62: 2643–2659.
  toxic hydrogen sulphide. Molecular Ecology 16: 967–976.           Tobler, M., Schlupp, I. & Plath, M. 2008b. Does divergence
Plath, M., Tobler, M., Riesch, R. & García de León, F.J.              in female mate choice affect male size distributions in two
  2007b. Survival in an extreme habitat: the roles of behaviour       cave fish populations? Biology Letters 4: 452–454.
  and energy limitation. Naturwissenschaften 94: 991–996.           Tobler, M., Riesch, R., Tobler, C.M., Schulz-Mirbach, T. &
Plath, M., Hermann, B., Schröder, C., Riesch, R., Tobler, M.,         Plath, M. 2009. Natural and sexual selection against immi-
  de Leon, F.J.G., Schlupp, I. & Tiedemann, R. 2010. Locally          grants maintains differentiation among micro-allopatric pop-
  adapted fish populations maintain small-scale genetic differ-        ulations. Journal of Evolutionary Biology 22: 2298–2304.
  entiation despite perturbation by a catastrophic flood event.      Tobler, M., Culumber, Z.W., Plath, M., Winemiller, K.O. &
  BMC Evolutionary Biology 10: 256.                                   Rosenthal, G.G. 2011a. An indigenous religious ritual
Riesch, R., Duwe, V., Herrmann, N., Padur, L., Ramm, A.,              selects for resistance to a toxicant in a livebearing fish. Biol-
  Scharnweber, K., Schulte, M., Schulz-Mirbach, T., Ziege,            ogy Letters 7: 229–232.
  M. & Plath, M. 2009. Variation along the shy-bold contin-         Tobler, M., Palacios, M., Chapman, L.J., Mitrofanov, I.,
  uum in extremophile fishes (Poecilia mexicana, Poecilia              Bierbach, D., Plath, M., Arias-Rodriguez, L., Garcia de
  sulphuraria). Behavioral Ecology and Sociobiology 63:               Leon, F.J. & Mateos, M. 2011b. Evolution in extreme envi-
  1515–1526.                                                          ronments: replicated phenotypic differentiation in livebear-
Riesch, R., Oranth, A., Dzienko, J., Karau, N., Schiessl, A.,         ing fish inhabiting sulfidic springs. Evolution 65: 2213–
  Stadler, S., Wigh, A., Zimmer, C., Arias-Rodriguez, L.,             2228.
  Schlupp, I. & Plath, M. 2010a. Extreme habitats are not           Verma, R., Balhara, Y.P. & Gupta, C.S. 2011. Gender
  refuges: poeciliids suffer from increased aerial predation risk     differences in stress response: role of developmental and
  in sulphidic southern Mexican habitats. Biological Journal          biological determinants. Industrial Psychiatry Journal 20:
  of the Linnean Society 101: 417–426.                                4–10.
Riesch, R., Plath, M. & Schlupp, I. 2010b. Toxic hydrogen           Woehrer, M., Hougen, D. & Schlupp, I. 2012. Sexual Selec-
  sulfide and dark caves: life-history adaptations in a livebear-      tion, Resource Distribution, and Population Size in Synthetic
  ing fish (Poecilia mexicana, Poeciliidae). Ecology 91: 1494–         Sympatric Speciation. Artificial Life 13: 137–144.
  1505.                                                             Wu, C.I. & Ting, C.T. 2004. Genes and speciation. Nature
Riesch, R., Plath, M. & Schlupp, I. 2011a. Toxic hydrogen             Reviews Genetics 5: 114–122.
  sulphide and dark caves: pronounced male life-history             Yeaman, S. & Whitlock, M.C. 2011. The genetic architecture
  divergence among locally adapted Poecilia mexicana                  of adaptation under migration-selection balance. Evolution
  (Poeciliidae). Journal of Evolutionary Biology 24: 596–606.         65: 1897–1911.

6
You can also read