Content uploaded by Kenneth Chapin
Author content
All content in this area was uploaded by Kenneth Chapin on Nov 12, 2019
Content may be subject to copyright.
Content uploaded by Ingi Agnarsson
Author content
All content in this area was uploaded by Ingi Agnarsson on Feb 06, 2019
Content may be subject to copyright.
Content uploaded by Kenneth Chapin
Author content
All content in this area was uploaded by Kenneth Chapin on Jul 30, 2018
Content may be subject to copyright.
Ecology and Evolution . 201 8 ;1–1 3 .
|
1
www.ecolevol.org
Received:22January2018
|
Revised:14Ju ne2018
|
Accepted:17June2018
DOI:10.1002/ece 3.4333
ORIGINAL RESEARCH
Island biogeography and ecological modeling of the amblypygid
Phrynus marginemaculatus in the Florida Keys archipelago
Kenneth J. Chapin1 | Daniel E. Winkler2 | Patrick Wiencek3 | Ingi Agnarsson3
Thisisanop enaccessarticleundert hetermsoft heCreat iveCommonsAttr ibutio nLicense ,whichpe rmitsu se,dist ributi onandrep roduc tioninanym edium,
provide dtheorig inalworkisproper lycited.
©2018TheAut hors.Ecology and Evolutionpu blishedbyJohnWiley&SonsLtd .
1Depar tmentofEcology&Evolutiona ry
Biolog y,Universi tyofCalifornia,Los
Angeles,LosAngeles,C alifornia
2Depar tmentofEcology&Evolutiona ry
Biolog y,Universi tyofCalifornia,Irvi ne,
Irvine,California
3Depar tmentofBiology,Universit yof
Vermont,Burling ton,Verm ont
Correspondence
KennethJ.Chapin,Depar tmentofEco logy
&EvolutionaryBiology,Universit yof
Califo rnia,LosAngeles,LosAngeles,C A.
Email:chapinkj@gmail.com
Funding information
TheodoreRooseveltMemorialFundoft he
AmericanMuse umofNatur alHistory;Edwin
W.PauleyFell owship,UCLA;Department
ofEcolog y&EvolutionaryBiology,UCL A;
Depar tmentofEco logy&Evol utionary
Biolog y,UCI;NationalGeo graphi cSociet y
(WW-203R-17)toIA
Abstract
Aim:ThebiogeographyofterrestrialorganismsacrosstheFloridaKeysarchipelagois
poorly understood.Weusedpopulationgenetics andspatioecological modeling of
the Ambly pygi Phrynus marginemaculatusto un derstand t he genetic stru cture and
metapopulationdynamicsofKeyspopulationsthatareotherwiseisolatedbyhuman
developmentandocean.
Location:TheFloridaKeysarchipelagoandmainlandFlorida.
Methods: We sequenced a 1,238bp fragment of mtDNA for 103 individuals of
P. marginemaculatusfrom13sitesintheFloridaKeysandSouthFlorida,binnedinto
four regions. We used population genetic analy ses to understan d the population
structureofthespecies throughout its US range.Furthermore,weusedecological
modelingwithclimate,habitat,andhumandevelopmentdatatodevelophabitatsuit-
abilityestimatesforthespecies.
Results: We found clear genet ic structure b etween local ities. The Lower Keys, in
particular, support populations separate from those in other regions studied.
Ecologicalmodelingandgeneticanalysesshowedthehighesthabitatsuitabilityand
geneticisolationintheLowerKeys,buturbandevelopmentacrossthespeciesrange
hasresultedinthelossofmosthistoricalhabitat.
Main conclusions:A mainland-metapopulation model best fits P. marginemaculatus
geneflowpatternsintheFloridaKeysandmainland.Oceancurrentslikelyplayarole
inmetapopulationdynamicsandgeneflowforterrestrialKeysspecieslikeP. margin-
emaculatus,andgeneticpatternsalsomatchedpatternsconsistentwithgeologichis-
tory.Suitablehabitat,however,islimitedandunderthreatofhumandestruction.The
fewremainingpockets ofthemostsuitable habitattendtooccurinparksandpro-
tected areas.Weargue that conser vation efforts for this speciesandothers inthe
terrestrialFloridaKeyswouldbenefitfromadeeperunderstandingofthepopulation
geneticstructureandecologyofthearchipelago.
KEY WORDS
Amblypygi,ecologicalmodel,FloridaKeys,habitatsuitability,islandbiogeography,MaxEnt,
metapopulation,pinerockland,populationgenetics,urbandevelopment
Present address
KennethJ.Chapin,UniversityofArizona,
Depar tmentofEcologyandEvolutionary
Biolog y,Tucson,Arizona.
2
|
CHAPIN et Al.
1 | INTRODUCTION
Acentralgoalofbiologyistounderstandhowtimeandspaceshaped
the evolut ionary his tory of life (Ma cArth ur & Wilson, 1967; Podas,
Crisci,&Katinas,2006;Warrenetal.,2015).TheFloridaKeysisabio-
diversearchipelagowithhighendemism(Forys&Allen,2005;Kautz&
Cox,2001),butthebiogeographyofterrestrialpopulationsalongthe
islandchainandnearbymainlandFloridaremainspoorlyunderstood.
This is par ticularly su rprising conside ring the Florid a Keys was the
studysiteusedtodevelopsomeofthemosticonicbiogeographythe-
ory (e.g., Mac Arthur&Wilson, 1967;Simberloff, 1976;Simberloff &
Wilson,1969,Simberloff&Wilson1970;Wilson&Simberlof f,1969).
The Florida Keys may support metapopulation spatial struc ture
for some species or genetic divergence and speciation in others
(Hanski, 1998; S hrestha, W irshing, & Har asewych, 2015 ). In either
case,thedistributionofspeciesacrosstheKeysisimport antforun-
derstanding both specieslong-termsur vivaland biodiversity.Thisis
especiallytrue considering the precariousness ofFlorida Keys habi-
tat sinthef aceofhumandisturbance,includinghumandevelopment,
deforestation,nonnativespeciesintroductions, andhuman-induced
climate change impacts like sea level rise, increased catastrophic
storms, and altered fire regimes (Bancroft, Strong, & Carrington,
1995;Forys,2005;Maschinskietal.,2011;Ross,O’Brien,&daSilveira
LoboSternberg,1994;Ross, O’Brien,Ford,Zhang, &Morkill,20 09).
Fur the rm or e, th eFlo ri daKey sr em ai namaj or to ur is td es tina ti on wi th
over4.5milliontouristsvisitingannually(McClenachan,2013).
Our understanding of the genetic structure of Florida Keys or-
ganismslargelycomesfromresearchofmarinespecies,whereocean
current s play a major role in gene flow and migration (Apodaca,
Trexler,Jue,Schrader,&Travis,2013;DeBiasse,2010;Kirk,Andras,
Harvell,Santos,&Coffroth,2009;Lacson&Morizot,1991).Genetic
patter ns of terrest rial specie s are expec ted to differ co nsiderabl y,
as the mari ne ecosystem a cts as an uninh abitable mat rix and the
formation of terrestrial habitat occurred on different timescales
(Hoffmeister&Multer,1968;Shresthaetal.,2015).
Geneti c research using n ative terrest rial species i s scarce, but
not absent . The mosqu ito Aedes aegyptishowednogeneticstruc-
ture along the Florida Keys,likely because they disperse via flight
(Brown, O bas, Morley, & Powell , 2013). The invasive br own anole
(Anolis sagrei) and greenhouse frog (Eleutherodactylus planirostris)
showed introductions to the Florida Keys from Cuba but cannot
inform patterns for nativespeciesacross the Keys (Heinicke, Diaz,
& Hedges, 2 011; Kolbe etal. , 2004). An alloz yme elect rophoretic
study ontheFloridaTreeSnail (Liguus fasciatus)revealedlowlevels
ofgenetic diversity, likely duetoa recent introductionfrom Cuba,
orthelowresolutionofallozymeapproaches(Hillis,Dixon,&Jones,
1991).ThelandsnailCerion incanumshowedsomehaplotypestruc-
turebetweenUpperandLowerKeys,likelycausedbydifferencesin
thetimingofformationoftheKeys.Shresthaetal.(2015)proposed
that the C. incanumspread southwesterly to colonizenew Keys as
theyformed,withLowerKeypopulationsbeingtheyoungest.Lastly,
antgutmicrobiotashowedgeneticstructurebetweentheupperand
lowerkeys(Huetal.2013).Thatsaidmostpastbiogeographicstudies
havebeenlimitedtononnativespeciesorexcludedFloridamainland
populations.Nostudieshaveinvestigatedthebiogeographyofana-
tivespeciesoccupyingtheentirearchipelagoandmainland,orhave
anyinvestigatedhumanimpact sonstructureandconnectivity.
The human population of the Florida Keys has drastically im-
pacted ecosystems therein. The human population of Monroe
County, which includes the Florida Keys and a portion ofrural land
westofEvergladesNationalPark,hasmorethandoubledsince1950
(US Census B ureau). While po pulation sizes of re sidents may have
stabilized over the last 25years (currently ca. 77,000 residents),
the numb er of tourists vis iting the Keys is enor mous. In 2014, an
estimated 4.516million tourists visited the Florida Keys (Key West
Ch am berofCo mm erc e, 20 17 ). KeyWe st, thema jo rc ity of th eF lor id a
Keys,locatedonanisland ofonly 19km2,hasover52,000housing
units. T he natural are as that remain are m ostly protec ted as state
parks or n ational wildlife r efuges, but are co ntinually impa cted by
nearb yhuma nact ivit y(P eterson,L opez,Frank,P or te r,&Silvy,20 04).
Weused fieldobservationscoupledwithlanduseandclimate
data to model Phrynus marginemaculatus distribution in South
Florida a nd the Florida Keys . Species dis tribution mo dels (SDM)
have been employed in many conservation, evolutionary, and
ecologic al applications (Elith & Leathwick, 20 09). These include
studiesofspatial patternsof diversity(Hoagstrom,Ung,&Taylor,
2014;Peterson, 2011;Walt ari&Guralnick,2009) ,genetic struc-
ture (Gotelli & Stanton-Geddes,2015),and the historicspread of
invasivespecies(Li,Dlugosch,&Enquist,2015;Václavík,Kupfer,&
Meentemeyer,2012). Additionally, SDM have identified suitable
habitatforspeciesofconcern(Tittensoretal.,2009)andidentified
climaticfactorsdrivingspeciesdistributions, including responses
toclimatechange(Feng&Pap eş,2015;Ficeto la,T huiller,&Miaud,
2007). Ad ditionally, SDM tech niques have advanc ed in the past
decade (Guisan & Thuiller, 2005) to allow for predictive power
with presence-only data (Bradley, 2015; Bradie & Leung, 2016;
Elithetal.,2006; Jiménez-Valverde,Decae, &Arnedo, 2011)and
smallsamples(Pearson,Raxworthy,Nakamura,&Peterson,2007;
Proosdij,Sosef,Wieringa,&Raes,2016;Wiszetal.,2008).
Weaimedtoquantifythegeneticandecologicalcharacteristicsof
P. marginemaculatuspopulationsintheUS.In particular,weaimed to
uncover the genetic structureofthe speciesacross theFlorida Keys
archipelago,understandtheevolutionaryhistoryofKeyspopulations
inrelationtothe speciesrangevia phylogeneticanalysis,and identify
suitablehabitatandlocationsofputativepopulationsthroughoutthe
species’ potentialrange.Together,theseresultswillbe thefirst toex-
amineP. marginemaculatusinthewild,andwillprovidecritic alinforma-
ti o napp l i c abl e tom a nyt e r res t r i ali s lan d s peci e san d t hei r con s e r v ati o n .
2 | MATERIALS AND METHODS
2.1 | Study site
The FloridaKeysisaca250-km-longarchipelagoamountingtoca
350km2ofdryland,extendingfromKeyLargoborderingmainland
|
3
CHAPIN e t Al.
Florida s outhwest to Key West 140km f rom Cuba. The Keys ar e
made from t wo geologic form ations that both for med during the
Tarantian Pleistocene (0.126–0.0117mya) and raising above sea
level during the Wisconsin glaciation (ca. 100,000 before present;
Hoffmeister&Multer,1968;Shresthaetal.,2015).TheLowerKeys
(BigPineKeytoKeyWest)formedfromcementedsandbars,result-
inginoöliticlimestone(termedMiamiLimestone).TheseKeyscurve
westandorientlaterallyduetogulfstreamcurrents.TheUpperKeys
(Bahia HondaKey to Key Largo), however,are constituted of fossil
coralreefs(termed KeyLargoLimestone)withoutstrong lateraliza-
tion.SomespeciesshowendemismtotheLowerorUpperKeys,but
notbothbecauseofthesedifferences(e.g.,Peck&Howden,1985).
Florida m ixed hardwoo d forests h ave persiste d since the bir th
ofterrestrialFlorida25mya(Webb,1990).Sincethen,specieshave
hadtworoutestocolonizingFloridahabitats:alandmigrationdown
theFloridapeninsulafromNorthAmerica,orwatermigrationnorth
orwestfromtheC aribbeanandBahamas,asFloridawasnevercon-
nected totheCaribbeanislands (Snyder,1990).Thus,we might ex-
pectthediversityofFloridatobeshapedbytropicalspeciesableto
dispers e over water, and temper ate species o nly able to est ablish
vialand.ThishasresultedinadominanceofvertebratesfromNorth
AmericabutflorafromtheCaribbean(Snyder,1990).Exceptionsin-
cludeninebirdspecies,andtwospecieseachofbat,frogandlizard,
allofwhichhave Caribbeanorigins. That being said,natural migra-
tionsareonlyclearforafewofthesespecies;manymighthavebeen
introduced via humans, and still others have gone extinct(atleast
locall y) since their discover y (Snyder, 1990). Inver tebrate biogeo-
graphicpatternsare more mixed, but still fit themodel of tropical
watermigratorsversustemperatelandmigrators.Forexample,most
Florida ant species have Nor th American origins,while Butterflies
arelargelyCaribbean(Lenczewski,1980).
Land to support the growing humanpopulations of Miami and
theFloridaKeyshavebeenlargelyobtainedbyclearingpinelandand
hammock(Snyder,Herndon,&Robertson,1990),amajorconserva-
tionissuethatresearchershavebeenbringingattentiontofornearly
acentury (Small, 1929).The firstsettlers in Southern Floridawere
concentrated in the Florida Keys. Early settlers exploited pineand
hardwoo d trees (espe cially Mahog any) for lumber, fue l, and slash-
and-burn agriculturalpractices(Small, 1917;Browder,Littlejohn,&
Young,1976;Wilson&Porras, 1983).Asa result,very few stands
of rocklan d include origi nal forest. In dustrial log ging was enabl ed
by the Flor ida East Coa st Railroa d, which reac hed Miami in 1896.
Rocklan d habitat was subje ct to clear-cutting fo r timber and fuel
but made pooragricultural land due to an abundance of limestone
rocksthatmadesoilunworkable.Greatlyexpandingagricultureof-
tentimessparedrocklandhabitatinfavorofdraininggladesforcrops
untiltheinventionoftherockplowinthe1950s.Thisenabledlime-
stonerockstobecollectedandseparatedfromsoil,therebyenabling
agricultural access. Lime stone rocks co llected via h and, plow, and
mine, were used as buildingmaterials andcanbeseeninhistorical
building s, walls, an d gardens of th e Florida Keys to day.Ro cklands
arenutrientandwaterdepauperate,andthusrequiretheheavyuse
ofirrigationandchemicalfertilizerstobeagriculturallyusable.Asa
result,abandonedrocklandsshowlittleresemblancetotheoriginal
ecosystem, and are often dominated byinvasive species (Loope &
Dunevit z,1981).Theabilityto turn rocklandintospaceforagricul-
ture and housing has led to the steep decline in rockland habit ats
that conti nue to the pres ent, with p ractic ally no hope of r eestab-
lishment without human inter vention (Dorn,1956;Meyers& Ewel,
1990;Possley,Maschinski,Maguire,&Guerra,2014;Snyder,1990).
2.2 | Study species
We used the amblypygid species P. marginemaculatusC.L.Koch,
1840 as amodel to understandthegeneral biogeographicpattern
ofterrest rialFloridaKeysspecies.Amblypygidsareasmallar achnid
order(ca.220spp.)oflarge nocturnalpredators(Chapin&Hebets,
2016).Phrynus marginemaculatusistheonlyamblypygidspeciesin
theUSeastoftheMississippiriverandthemostcommonlystudied
speciesofamblypygid(Chapin&Hebets,2016;Figure1).Laborator y
research has shown that P. marginemaculatus exhibit ritualized
agonisti c displays (Fowle r-Finn& H ebets, 2 006), and ca n learn to
navigatemazesusingtactilecues(Santer&Hebets,2009a, 2009b)
via exceptional brain structures and sensory systems (Chapin &
Hebets,2016;Santer&Hebets,2011).Whilef ascinatinglaboratory
researchhasbeen conduc ted on the species, no researchontheir
habitat requirements, distribution, population ecology, or popu-
lation gen etics has ever be en published (C hapin & Hebet s, 2016;
Weygoldt,20 00).
Historical records indicate that the species was found as far
northasMartinCounty,FL,butrecentrecordsareabsent.Thespe-
cies is also foundon severalBahamian islands, Cuba,Jamaica,and
Hispaniola(Muma,1967;Quintero,1981).Research onthespecies,
FIGURE1 PhotographsoftheamblypygidPhrynus
marginemaculatusintheFloridaKeys.(a)Notetheelongatedfirst
pairoflegsadaptedassensorystructures.(b)Closeupofthebody
highlight scolorationandpat terning
5 cm
(a)
(b)
4
|
CHAPIN et Al.
however,hasonlyoccurredincaptivity,withanimalscollectedfrom
asingleisland(BigPineKey;Hebets&Chapman,2000;Fowler-Finn
&Hebets, 2006;Spence &Hebets20 06; Santer & Hebet s,2009a,
2009b),thepettrade(Rayor&Taylor,2006),orboth(Graving,2015).
Interestingly, P. marginemaculatus has evolved a plastron to
breatheunderwater,whichtheycandoforupwardsof24hr(Hebets
& Chapman , 2000 ). While the fu nction of th e plastron i s not well
unders tood, it likely i ncreases chan ces of surviv al during floo ding
in their terrestrial retreats. This may be particularly impor tant in
the Flor ida Keys, where annua l hurricanes c an result in floodi ng.
Furthermore, hurricanes, alongwith ocean currents, likely promote
oceanicdispersal(Fleming&Murray,2009;Gillespieetal.,2011).A
plastronallowingunderwaterbreathinglikelyextends thedispersal
propens ities across b odies of water, and th e likelihood of su rvival
duringoceanmigration,inP. marginemaculatus.
Twomolecular phylogenetic studies have focused on amblypy-
gids. First, a phylogeny of the Damon variergatusgroup delineated
two crypticspecieswithinDamon,an African genus of amblypygid
(Prendini,Weygoldt,&Wheeler,2005).Second,phylogeneticanaly-
sesofPhrynusspeciesinPuertoRicorevealedhiddendimensionsof
diversit yacross cavepopulations(Espositoetal.,2015).In par ticu-
lar,Espositoetal.(2015)notedhigh levels ofdiversityacrossmito-
chondrialbutnotnuclear,genomes.Herewefocusonmitochondrial
sequencesto examineifsimilargenetic structure acrosssmallgeo-
graphicalscalesisevidentintheFloridaKeys.
2.3 | Specimen collection
Wecollected P. marginemaculatusspecimensfrom13locationsin
southernmainlandFloridaandtheFloridaKeysarchipelagousing
anonrandomsamplingmethod( Table1;Figure2).Welimitedour
survey to upland habitat types, as we assumed that P. margin-
emaculatuswouldnotbefoundininter tidalhabitat slikemangrove
swamps an d floodplains. Additional survey areas were selected
fromhistoricrecords,whichwereallassociatedwiththeseupland
habitat types(Table1).P. marginemaculatushideunderdebris,es-
peciallylimestone rocks, during theday (Chapin&Hebets,2016;
Hebets&Chapman,2000).Oursamplingregimeincludedwalking
trails an d looking for P. marginemaculatusunderrocks,logs,and
other larger debris. When found, we collected genetic samples
andstored them in 95%ethanol ondry ice. Wedissectedmuscle
tissuefromoneorafewappendages,dependingonthesizeofthe
specimen.
2.4 | Extraction, amplification, and sequencing
WeextractedgenomicDNAusingQIAGENDNeasyBlood&Tissue
Kits.Wefollowedthestandardkitprotocolbutusedchilledethanol
(−20°C) anda50-mLfinalelution. Weamplified a1, 238nucleotide
sequenceofthemitochondrialgenecy tochromecoxidasesubunit1
(COI)byperforming34iterationsofthefollowingcycleonathermal
cycler:30sat94°C,35sat48°C,and90sat65°C,beginningwithan
initialcycleof 2minat 94°Cand ending with10minat72°C.Using
illustraPuReTaqReady-To-GoPCRbeadsand40 0-nMforwardand
reverse primers,the longor shor t read of COI was sequenced for
eachsample.LCOI1490wasuse dast heforwardprimerforbotht he
longandshortreads,whileHCOI2198wasusedasthereverseprimer
fortheshort readsandC1-N-2776wasusedasthereverse primer
fo r t he lo ngre a ds (LCO I14 90 GG TCA AC A A ATC ATA A AG ATAT TG G ,
HCOI2198TAAACTTCAGGGTGACCAAAAAATCA,andC1-N-2776
GGATAATCAGAATATCGTCGAGG; Folmer, Black, Hoeh, Lutz, &
Vrijenhoek,1994).WechoseamtDNAsequenceasmtDNAismuch
more informative than nuclear DNA among Amblypygi (Esposito
etal.,2015).
2.5 | Cleanup and alignment
AmplifiedfragmentsweresenttotheUniversityofArizonaGenetic
Core and Genewizfor sequencing. Subsequently sequences were
assembled using the Chromaseq module (Maddison & Maddison,
2016a) in Mesqu ite 3.02 (Maddiso n & Maddison, 2016b) th rough
Phred and P hrap (Ewing & Green , 1998; Ewing, Hilli er, Wendl, &
Green,1998; Green,1999;Green&Ewing,2002), and thenproof-
read in Me squite. Seq uences were a ligned with C lustal W2 (Larkin
etal.,2007)inMesquite.
2.6 | Genetic analysis
Wegroupedlocalitiesintofourregionsbymajorgeologic features:
the mainland, Key Largo, Upper Keys (excluding Key Largo), and
LowerKeys.Weproducedgeneticdiversit yindicesforeachregion
and used a hierarchical analysis of molecular variance (AMOVA;
pop N H G λ λc Hexp π
LeyLargo 18 2.44 8.53 0.883 0.935 0.032 0.003
Mainland 28 3.12 20.63 0.952 0.987 0 .105 0.010
UpperKeys 13 2.56 13.0 0 0.923 1.000 0.150 0.013
LowerKeys 44 3.54 28.47 0.965 0.987 0.115 0.010
Tot al 103 4.39 6 7. 57 0 .985 0.995 0.131 0.013
Note. GisSto dda rtan dTa yl or ’s in de xo fM LGd ivers it y;His th eS ha nno n–W ie ne rI nd exof mu lt il ocus
genotype(MLG)diversity;Hexpis Nei’sunbiased genediversity; πisnucleotidediversity; Nist he
numberofindividualssequenced;λisSimpson’sIndex;λcisSimpson’sindexcorrectedforvariation
insamplesize.
TABLE1 Geneticdiversityindicesof
COIsequencesformajorregionsinwhich
Phrynus marginemaculatus occur
|
5
CHAPIN e t Al.
Excoffier,Smouse,&Quattro,1992)toestimatethevariancewithin
and betweenlocalities and regions. Wetested for isolation by dis-
tance (IBD) by testing for a correlationbetween Nei’s genetic dis-
tance and maximum geographic distance among samples with a
mantel tes t (Mantel, 1967). We used disc riminant ana lysis of prin-
cipal components (DAPC)with cross-validation to examinegenetic
divergencebetweenregionsandcalculatedpairwise GSTasanes-
timate ofmigrationbetween regions. WeusedtheR2.3.2 (R Core
Team)packages“ade4”(Dray&Dufour,2007),“adegenet”(Jombart,
2008; Jombar t & Ahmed, 2011), “mmod” (W inter, 2012), “pegas”
(Paradi s, 2010), and “poppr” ( Kamvar, Brooks, & Gr ünwald, 2015;
Kamvar,Tabima,&Grünwald,2014)forgeneticanalyses.
2.7 | Ecological modeling
Weusedlocalitydatafromthe103field-collectedspecimenstoes-
timatethegeographicrangeofP. marginemaculatususingtheniche
modeling software Maxent 3.3.3 (Phillips, Anderson, & Schapire,
2006;Phillips &Dudík,2008).Maxent usesamaximum-entropyal-
gorithm to predict species geographic ranges using presence-onl y
dataandenvironmentalGISlayers.Weevaluated19BioClimclimate
variables(BIO1–19;Hijmans,Cameron,Parra,Jones,&Jarvis,2005)
ata30-arc-secondresolution(ca.1km2)forinclusioninourmodels.
Weincludedelevationandageologicmapinourfirstsetofmodels
totesttheircontributiontoinformingmodelpredictions.Bothlay-
erswereobtainedfromtheUnitedStatesGeologicalSurvey(http://
www.usgs.gov). We ran a second set of models thatincludedland
usedatabased onimagerymadepublicallyavailable bytheFlorida
DepartmentofEnvironmentalProtection(http://geodata.dep.state.
fl.us/). This dataset included 195categories ofland use based pri-
marilyonhumanuse(e.g.,agriculture,urbandevelopment,transpor-
tation corridors) but alsoincluded subcategoriesofvegetation and
other ecologically relevanthabitattypes (e.g., pinelands, mangrove
swamps, cabbage palm hammock). This included all land cover
categories used in the Florida L and Cover Classification Systems
(Kawula,2009).
All data we re clipped to a regi onal extent of so uthern Flori da
and the Fl orida Keys at app roximately la titude 28° N using ArcG IS
v10.2.2 . This nor thern latit ude is locate d approximate ly along the
freezelineinFlorida(Miller&Glantz,1988).Thereisnoevidenceto
indicatethatP. marginemaculatusoccurs beyond thisline(Quintero,
1981).Wetestedalllayersforpairwisecorrelationacrossthestudy
area using t he package ‘Raster’ in R 3.3 .2 (Hijmans & van Ette n,
2012).Weretained12ofthe19BioClimlayersthathadcorrelation
coefficients under |0.75|. Theseclimate variables represent annual
andseasonaltrends,aswellasextremesintemperatureandprecip-
itation.Temperaturevariables included annual mean temperature,
mean diurnalrange, isothermality,and mean temperaturesofboth
the wettest and driest quar ters. Precipitation variables included
annual temperature, precipitation during the wettest and driest
months, precipitationseasonality,andprecipitation ofthewarmest
andcoldestquarters.
We ran 100 model replic ates using a randoml y selected 75%
of the occur rence record s to calibrate t he model an d 25% to test
it(Phillips etal., 2006),wellbeyond the ideal minimumsample size
toobt ainreliable results(Proosdijetal.,2016).Each model was as-
sessed with the area under the receiver operating characteristic
curve (AUC;Hanley &McNeil,1982).AUCvaluesrepresent amea-
sure of the MaxEnt model’s ability to discriminate betwe en suit-
able and unsuitable areas in the modeled distribution (Anderson
& Gonzal ez, 2011). AUC values ra nge from zero to one, w ith one
indicating a perfect differentiation of suitable and unsuitable hab-
itat.Wecompared predicted models against distribution literature
for P. marginemaculatus (Quintero, 1981). Models that performed
poorly (AUCscores<0.75)orthatvaried substantiallyfromhistor-
ical recordswere discarded. Jackknife tests were usedtoevaluate
FIGURE2 MapofSouthernFlorida
withgeographicregions(colors),ocean
currents(bluearrows;adaptedfromLee
&Smith,2002),localitieswherePhrynus
marginemaculatussampleswerecollected
(opencircles),andpairwiseGST(thicker
linesindicatelowerGSTt;Gst range:
0.174–0.620)indicatingmigration
Ma
i
nlan
d
U
pper
ke
y
s
L
L
o
o
w
w
er
e
ke
ys
k
e
y
lar
g
o
6
|
CHAPIN et Al.
the impo rtance of eac h environment al and abiotic var iable to ex-
plain the range ofP. marginemaculatus.Last,we calculated areas of
overlap between human land use features (i.e., urban and rural
developments, transportation, communication, and utilities, and
agricultural lands) and model outputs of suitable habitat using the
image-processing software ImageJ (Abr amoff,Magalhaes ,& Ram,
2004).Wecalculated declinesin suitablehabitat atfourthresholds
ofmodeledsuitablehabitat (>0.1, 0.1, 0.3,0.5, and 0.9)caused by
humandevelopment.
3 | RESULTS
3.1 | Population genetics
Wesequenceda1, 238bp region ofthemitochondrialCOIgeneof
103individuals.Sequenceshadanoverallbasecompositionof24.6%
adenosine,25.2%cytosine,16.5%guanine,and33.7%thymine.Key
Largoexhibitedthelowestgeneticdiversityamongregions,withthe
LowerKeysandmainlandregionsexhibitingconsiderablyhigherge-
netic dive rsity ( Table1).A dditional ly,a r ange-wide mante l test for
IBDofgeographiccoordinatesandNei’sdistancewasnonsignificant
(Mantelr =−0.07,P = 0.637).AhierarchicalAMOVArevealedpopu-
lation st ructure a mong popula tions, but not r egions (Table2). The
AMOVAindicated significant genetic structure between localities,
suggesting that oceans limit dispersal. Stratified cross-validation
of DAPC resul ted in a mean success ful assignment of 0.92% a nd
88.2%conservedvariancewith20principalcomponents.Allregions
separatedintodistinctclusters,furtherevidencingregionalgenetic
structure (Figure3). Furthermore, pairwise GST showed relatively
high divergence of KeyLargo localitiesfrom the rest ofthe range,
and low divergence between mainland and island sites, indicating
ongoinggeneflow(Figure2).
3.2 | Ecological modeling
Specie s distribut ion modelin g using MaxEnt fou nd good mode l fit
for climate-only models (mean AUC=0.978±0.02, n = 100 m od-
els;Figure4a–c).Habitatsuitabilitywas highestin the FloridaKeys
(Figure4), bu t also exten ded to the so utheaste rn end of mainl and
Florida. Areas of predicted suitable habitat on the southeastern
mainland corresponded to the geologic features of the peninsula,
which had a p ermutation importance of 5.8%. H owever, suitabl e
habitat was identified primarily by environmental variables that
contribu ted most to the mo del: precipit ation of the colde st quar-
ter(74.8%permutationimportance)andmeandiurnal temperature
range (14.7%). Altit ude also had predi ctive power with 2 .4% per-
mutatio n import ance in the fir st set of models . Jackknife te sts of
variablesinisolationfromallothersrevealedthatannualmeantem-
peraturehadthehighesttraininggainformodels,followedbymean
temperatureofthedriestquarter,meandiurnaltemperaturerange,
andmeantemperatureofwarmestquarter.
Models including land use categoriesperformed slightly worse
(mean AUC=0.873±0.06, n=100 models) than those without
landusebutappeartohaverefinedthehabitatsuitabilityofP. mar-
ginemaculatus (Figure4d–f). L and use categories had the highest
permutationimportance(50.5%)followedbymeantemperatureof
the drie st quarte r (18%),m ean diurnal t emperatu re range (12.7%),
TABLE2 Hierarchicalanalysisofmolecularvariance(AMOVA)
usingcytochromecoxidaseI(COI)sequencesofPhrynus
marginemaculatusinFloridaindicatinggeneticstructure
σ% variance ϕp
Withinlocalities 1.377 38.37 0.616 <0.001
Betweenlocalities 1.9 97 55.65 0.592 <0.001
Betweenregions 0. 214 5.97 0.060 0 .160
FIGURE3 Discriminantanalysisof
principalcomponents(DAPC)ofPhrynus
marginemaculatuspopulationregions
PCAandDAEigenvaluesispresented
asinsets.Dashedlineistheminimum
spanningtreeofregions.Stratifiedcross-
validationofDAPCresultedinamean
successfulassignmentof0.75%and93.9%
conservedvariancewith10principal
componentsandthreediscriminant
functions.Allregionsseparatedinto
distinctclusters
K
e
y
l
ar
g
o
Mainland
Upp
er
k
e
y
s
L
o
w
er
k
e
s
s
ys
s
s
DA eigenvaluesPCA eigenvalues
|
7
CHAPIN e t Al.
geology(7%),andaltitude (3%). Similarto our modelswithout land
use catego ries, annu al mean tempe rature had t he highest t raining
gainformodelsbutvariedinsubsequentvariableimportance.Mean
temperature during the driest quarter, mean diurnal temperature
range, an d mean temper ature of the war mest quar ter followed in
ter msofvaria bleim port anceinisolatio n.Mod elswithlanduseiden-
tified regions of the Lower andUpper Keys, pockets in Everglades
National Park,andseveral coastal areas ofMiami asthe most suit-
ablehabitat.Additionalsuitableareas wereidentifiedon KeyLargo
and in Big CypressNational Preserve.Smaller pockets ofpotential
habitatrangeduptheeastandwestcoast s.
We saw an alarm ing 22%–34% human-ind uced declin e in suit-
able habitat under the best fit model (Figure5; Tables3 and 4).
Modelsthatincludedlanduseshowedsteeperdeclinesof29%–48%
ofsuitablehabitatduetohumandevelopment(Tables3and4).This
indicates that, while climatically identified habitat shows consid-
erable decline, humanimpac ts particularly target landuse habitat
typesimportantforthespecies.
FIGURE4 MaxEntsuitabilitymapforP. marginemaculatusinsouthernFlorida.Colorscaleindicatesprobabilityofoccurrencebasedon
presence-onlydata.Minimum,mean,andmaximumsuitabilitiesusingonlyclimatedatasets(a–c;meanAUC=0.92±0.02);minimum,mean,
andmaximumsuitabilitiesusingclimateandvegetationcommunitiesdatasets(d–f;meanAUC=0.87±0.6)
(a)
(f)(e)(d)
(c)(b)
8
|
CHAPIN et Al.
4 | DISCUSSION
Pine rocklands succeedtotropicalhardwoodhammocksafter two
or three decades of fire suppression (Loope & Dunevitz, 1981;
Robertson, 1981), but it remains unclearif P. marginemaculatus oc-
cursinbothhabitats.Bothperiodicseawaterfloodingandfireoccur
naturally in rockland habitats (Snyder etal., 1990) and P. margin-
emaculatus,likemuchofthefaunaandfloraspeciesinthesehabitats
(Hofstet ter,1975;Robertson,1953,1962),have evolvedto survive
thesestochasticevents.
The mainland and Lower Keys show the greatest genetic di-
versity.This is likely because these tworegionsincludethelargest
expansesofarea,andbotharerepresentedbysmallpatchesofsuit-
able habitatwithin an inhospitablematrix. Key Largoshowed rela-
tively low diversity,despite the island’ssize.However,onlya small
portionof Key Largo’slandareaissuitable habitat—mostoftheis-
landishighlydeveloped.
Genetic variation was significant between localities but not
larger regions (Table2), suggesting a more complicated genetic
st ru cturethanouraprioriregionaldeline at io ns .Pop ul at io n- leve lg e-
neticstruc tureisalignedwithintuition,consideringallpopulationsin
ourstudyoccuronKeysorislandsofhabitatsurrounded byhuman
disturbance (Figure2). Regional structure, however, is somewhat
surprising.Weselectedregionsa prioribasedongeologicfeatures:
theKey Largo regionforitssizeand proximityto themainland;the
Upper an d Lower Keys for the ir geologic va riation in fo rmation, if
nottimingthereof;andthemainland,asanobviousdelineatorfrom
island populations.This generally alignswith the population struc-
ture of nati ve Cerionland snails, which showed isolation between
the Uppe r and Lower Keys (Shre stha etal., 2015 ). Species on th e
LowerKeyslikelyhaveauniqueevolutionar yhistoryseparatefrom
therestoftheKeys.
OurDAPCanalysisshowscleargeneticstructureacrossregions,
withkeypopulationsallbeingclosestrelatedtomainland localities.
This pat tern suggest s that a mainland–met apopulation m odel de-
scribedthelandscapegenetics oftheKeys.Asmentioned,genetic
analysesoflandsnailsintheFloridaKeysalsoshowanUpper–Lower
Keydivision(Shresthaetal.,2015). But researchon marinespecies
like bicolor damselfish (Eupomacentrus partitus) and common reef
sponge(Callyspongia vaginalis)foundmuchlowerlevelsofdivergence
betweenregions than we found for P. marginemaculatus(DeBiasse,
Richards,& Shivji,2010; Lacsonetal.,1989).Thisis likely because
P. marginemaculatus is much more disper sal limited tha n a marine
fish.Interestingly,theleastdivergentlocalitypairsofthesesmarine
speciesincludedKeyLargo,whichisthesamepatternwefindinour
study.Thismatching patternsupport stheideathatcurrents playa
majorroleinP. marginemaculatusgeneticstructure.
PairwiseGSTshowedthat thegreatest divergencewasbetween
Key Largo an d the other region s (Figure2). This coul d be caused
bymigrants from the Bahamas, the closestislands ofwhichareca.
100km from Key Largo.Surprisingly,the mainlandhad low Gstes-
timates with the Upper and LowerKeys. We posit that this is due
toongoinggeneflow betweentheseregions. Ocean currentslikely
push raftingP. marginemaculatustotheLowerKeysasmightmajor
weather events including hurricanes (Fleming & Murray, 2009).
Experim entswit hGPS-equi pp edbuoyss howthatt hisisamajorcur-
rentpathway(Lee&Smith,2002)andP. marginemaculatus,withthe
abilitytobreatheunderwater,areaptlysuitedtosurvivethevoyage
(Hebets&Chapman,2000).Otherresearchhasemphasizedtheim-
portanceofoceancurrentsintheFloridaKeysingeneticstructure,
butthishasbeenlimitedtomarinespecies.Forexample,astudyof
threemarineinver tebratesshowedhighgeneflowandconnectivit y
acrosstheKeys,withapatternofsouthernmigration(Richardsetal.,
2007).Thesaltmarshsnake(Nerodia clarkii),whichissomewhatre-
strictedtoshallowwaters,showedgeneticstructurebetweenUpper
FIGURE5 (a)MaxEntsuitabilitymapofthebestfitmodel
(climate-onlymean)withhumandevelopmentoverlayforthe
amblypygidPhrynus marginemaculatusintheFloridaKeysandSouth
Florida.Whiteareasindicatelandconvertedforhumanuse.Panels
b–dare10×magnificationofareasindicatedinpanela,which
include(b)KeyLargo,(c)theotherUpperKeys,and(d)theLower
Keys.Arrowsindicatenorth.TheKeysincludeparticularlyhigh
suitabilityhabitat,especiallytheLowerKeys.46
0.00 0.92 100 km
10 km
(d)
(a)
N
(c)
N
(b)
N
|
9
CHAPIN e t Al.
andLowerKeysthatwasexplainedby IBD (Jansenetal., 2007).In
general,ourSDMpredictedhabitatsuitabilityforP. marginemacula-
tusinpartsofMonroe,Miami-Dade,Collier,Lee,andHendrycoun-
ties. This limited range generally matches museum records, with
occasionalrecords fromcountiesasfar NorthasCharlot te county
(Quintero,1981).Thesenor therlyreco rdscouldbef romsparsep op-
ulations,wheredetectingthespeciesisdifficult,orcollectionscould
be made by va grant alloa nthropic ind ividuals as sociated onl y with
human st ructures , and not viabl e populations . Additional ly,mo del
predic tionso fhab it at su it abilitywith inth eurb an ar ea sofsouth ea st-
ernFloridamainlandshouldbecautiouslyinterpretedasrealizedvia-
blehabitatismuchlessbecauseofhumandevelopment .
Few studies of Florida biogeography have been conducted,
but generally align withourresults.Ant gut microbiota show simi-
largenetic structurebetweentheUpper andLowerKeys,but also
shoe divergenceamongthe Lower Keys, which might be indicative
of Carib bean migration (Hu e tal., 2014). The mosquito A. aegypti
showed practically no genetic stru cture among the Florida Keys,
likelybecausetheydisperse viaflight(Brownetal.,2013).Shrestha
etal. (2015) proposed that the C. incanum spread southwesterly
tocolonize new Keys as theyformed, withLower Key populations
beingtheyo ungest.L astly,antg utmicr obi otashowedgeneticst ruc-
turebetweentheUpperandLowerKeys(Huetal.,2013).
Ourmodelsidentifiedareasoftropicalhardwoodhammocksand
pinerocklandsasthemostsuitablehabitattypesforP. marginemac-
ulatus.Indeed,thisiswherealmostallofourobservationsoccurred
andiscorroboratedbypublished collection sites of the speciesfor
laborat ory resea rch (Hebets & C hapman, 20 00; Weygoldt , 2000).
The two ha bitat types , being at relativel y high elevation, ar e the
primar ytargetsforhuman development inFlorida (Noss,LaRoe,&
Scott , 1995;Snyde r etal., 1990). Thu s, our mode ling result s show
that P. marginemaculatus suitable habitats are also areas where
human disturbance has been, and continues to be, an imminent
threattothehabitatandspecies.
Pinerockland forests,oncecommon throughout southeastern
Florida,arenowoneofthe mostthreatenedhabitats globally,with
atleast98%globalloss(Nossetal.,1995).Thisincludesca.8,0 00ha
in Everglades National Park and a mere 920ha outside the park’s
bounda ry (Bradl ey,20 05). Pine roc kland habit at was identif ied as
oneofthemostsuitablehabitatsbasednotonlyonourmodelsthat
includedlandusecategoriesbutalsothose with onlyclimatevari-
ables and geolog y.Pine rockland habitats occur on exposed lime-
stonesubstrates where limestone rock outcroppings are common
andprovideimportantmicrohabitatforP. marginemaculatus(Chapin
&Hebets, 2016;Hebets& Chapman,2000).Humandevelopment,
fire supp ression, and climate change have altered or ent irely re-
movedmanyareasoncedominatedbythiscommunity(Kautz&Cox,
2001;Possley,Woodmansee,&Maschinski,2008;Rossetal.,1994).
This habitatlosshasresulted in fivefederallylistedanimal and 21
rare, endemicplantspeciessympatricwithP. marginemaculatus, all
ofwhicharedependent on remaining fragments of pine rockland
habitat (FloridaNaturalAreasInventory,2010).Furthermore,lime-
stonerocksthatmakeupcriticalhabitatforP. longipesareoftencol-
lectedandusedforconstructionandlandscaping.
Some are as of pine rockland a nd upland hardwo od forest are
protectedtoday.Theseinclude patcheswithin EvergladesNational
Park, Big Cypress National Preserve, Key Deer National Wildlife
Refuge,andst ate-m an agedlands .M os th abita to ut sid eofthe se pr o-
tectedareashavealreadybeendestroyed,andmanysur vivingfrag-
mentsremainthreatenedandatriskofextirpationwithintheareas
ofMiami,surroundingsuburbanareas,andinthetourist-dominated
Florida Keys.Thishasdramaticimplications forP. marginemaculatus
andtheendemic,endangeredcommunit yinwhichtheyoccur.
Human development is not the only threat to pine rockland
and uplan d hardwood for est habitat s; sea level ris e brought on by
human-induced climate c hanges is also threatening these habitats
TABLE3 AreaofsuitablehabitatoftheamblypygidPhrynus marginemaculatusinsouthernFloridaandtheFloridaKeys.Areaofhabitat
withsuitabilitythresholdsof0.1,0.5,and0.9underaclimate-onlyandclimate-pluslanduseMaxEntmodels.Lossindicatespercentlossof
habitatatagiventhresholdbyhumandevelopment.ModelfitisindicatedbytheAreaUndertheCur veandst andarddeviation(AUC±SD)
Model
0.1 threshold 0.5 threshold 0.9 threshold
AUC ± SDkm2% loss km2% loss km2% loss
Climateonly 3,545.88 27. 5 8 179.95 3 4.91 4.64 22.44 0.978±0.02
Climate+landuse 2 , 42 7.96 48.68 3 00 .97 43. 81 12 .74 28.65 0.873±0.6 0
TABLE4 Reductioninsuitablehabitatfortheamblypygid
Phrynus marginemaculatuspredictedbyMaxEntmodelingcausedby
humandevelopmentinSouthFloridaandtheFloridaKeys.
Thresholdisthelowerlimitfortheindexofhabitatsuitability;
Habitatisthetotalareainkm2withouthumandevelopment
Threshold Habitat DevelopmentaPercent loss
Climate-only(AUC=0.98±0.02)
0.1 3,545.88 2, 5 67. 7 9 27. 5 8
0.3 532.32 400.15 24.8 3
0.5 17 9.9 5 11 7.1 2 34 .91
0.9 4.64 3.60 22.44
Climate+Landuse(AUC=0.873±0.6)
0.1 4, 247.9 6 2, 17 9.85 48.68
0.3 958. 57 558.18 41.7 7
0.5 30 0.97 1 69.11 43.81
0.9 12.74 9.0 9 28.65
aDevelopmentisthereducedareaafterconsideringhabitatdegradedby
humanuse.
10
|
CHAPIN et Al.
(Maschinskietal.,2011;Rossetal., 2009).Inthissense,increases in
thefrequencyandintensit yof hurricane stormsurges reshapepine
rockland and similar vegetatio n communities (Ross etal., 1994). It
remains unknown how climate-induced changes in storm systems
may have alrea dy impacted the co nsiderably fr agile extant P. ma r-
ginemaculatuspopulations.Wecan,however,gleaninsightfrom
studie s of other species . For example, Hu rricane An drew dramati-
cally al tered pine roc kland commu nities when i t struck po rtions of
Everglade s National Par k and Big Cypres s National Pre serve, lea d-
ingto ca.90%mortalityofmaturepinetrees(Maguire,1995)which
negative ly impacted p lant and animal co mmunities (L loyd & Slater,
2012; Orr & Ogden, 1992; Williams, Wang, Borchetta, & Gaines,
2007). Majorstorms and humandisturbance not only alterhabitats
butcanalsoleadto isolationof populationsandreshape population
geneticsofspeciesatrisk(e.g.,Villanova,Hughes,&Hof fman,2017).
Ourresearchwason specimenscollectedin2015, priortothe 2016
Hu rri can eM att h ew and20 17Hu rr i can eI rma ev ent s.Fu t ur ere se a rc h
willbenefitfromexaminingtheimpactsoftheseandothers tormson
populationstructureofP. marginemaculatusinsouthernFlorida.
Both our genetic and ecologicalresultsare limited byour data-
set, whichis constrained in both time andspace. Spatially, we only
sampledP. marginemaculatusinso uther nFl oridaandtheFloridaKeys
archi pe la go,butt hespeciesoccu rs asfarso ut hwes tasHi sp aniola,in -
clud in gp op ulation si nt he Ba ha mas ,Cub a, Ja maica ,a nd th eTu rksan d
Caicos (Quintero, 1981). In par ticular,gene flow fromthe Bahamas
and Cuba to Key L argo and th e Lower Keys could be o ngoing, but
thisremainsunexamined.Temporally,weexcludedhistoricalsamples
withtheconcernthattheywouldnotinformmodernbiogeographic
patterns. While historicalmaterial wouldincreasesample sizes, our
moderncollections spanthespatialrangeof Floridahistorical sam-
ples bar northern limits, where populations may no longer occur.
Sampl ingwa sunev enacr osssi tes,w hi chcouldbiasre su lt s.Lastly,w e
usedonlyonesequence inouranalysis. Wechosea mtDNA marker
because nDNA appears highly conser ved in Amblypygi (Esposito
etal.,2015),andmtDNAisthus,moreinformative.Cautionshouldbe
taken,however,ininterpretingresultsfromonlyonemtDNAmarker,
andfutureresearchusingmorethoroughgenomicsequencingcould
revealmorehigh-resolutionbiogeographicpatterns.
4.1 | Conservation
Our resul ts point to a prima ry threat to the p opulation hea lth of
P. marginemaculatusinthewild:habitatlossbyhumandevelopment.
Approximately 77,000 people permanently reside in the Florida
Keys (US Censu s Bureau). Given th e small land area of t he archi-
pelago, thisaccountsfor an average density of 205.54peopleper
squarekilometerand leavesapproximately150km2uninhabitedby
humans, muchof which isseasonallyorpermanentlyflooded habi-
tat unsui table for ma ny terrest rial specie s. Furthe rmore, th is does
notinclude the impactfromcommercial,transportation,andutili-
ties developments. Fortunately, muchof the remaining habitat for
P. marginemaculatus is within protect ed areas managed by f ederal
andstateagencies.This,however,doesnotprotecthabit atsfromall
threats,includingtheimpact sofincreasedhurricanes,sea-levelrise,
poaching,andmicrohabitatalterations.
Secondarily,P. marginemaculatus is collec ted from the wild f or
saleinthepettrade. While we donothavedataon the numberof
individualscollectedforsaleinthepetindustr y,personal obser va-
tionsleadustobelievethatitmustbeinthehundredstothousands.
WildpopulationsofP. marginemaculatuswouldbenefitfromcaptive
breedin g that allows thes e fascinating an imals to be kept as pet s
without reducingnumbersin the wild.Moreresearch on wild pop-
ulations is n eeded to asse ss populati on health, a nd we encoura ge
researchers to conduct both fieldand laboratorystudieson these
fascinatingorganisms.
ACKNOWLEDGMENTS
Major funding was provided by theTheodoreRoosevelt Memorial
FundoftheAmericanMuseumofNaturalHistory.Fundingwasalso
provided b y the Edwin W. Pauley Fello wship, UCLA , Departm ent
ofEcology & EvolutionaryBiology,UCLA ,Department ofEcology
& Evolutiona ry Biolog y,UCI a nd the Nationa l Geograph ic Society
(WW-203R-17) to IA. Thanks also to laboratory collections man-
ager Emily Chen, Sarah Hsu, and other undergraduate assistants
for maintaining samples, dissecting tissue for DNA extraction, and
databasingliteraturesearches.Thisresearchwasconducted under
EvergladesNationalPark Scientific ResearchandCollectingPermit
EVER–2014–SCI–0057,Nation alW ildlife RefugeSystem Rese arch
andMonitoringSpecial Use PermitFFO4RFKD–2015–001,Florida
Park Ser vice Scientific Collecting Permit 08051410, and Miami-
DadeCountyParks&RecreationResearchPermit241.
CONFLICT OF INTEREST
Nonedeclared.
AUTHOR CONTRIBUTIONS
K.J.C .conceived thestudy; K.J.C and D.E.W collected and extracted
sa mp le s;I.A. an dP.W.se qu en ce d,clean ed ,a ndalignedDN A; K.J .C .a n-
aly zedgenet icdat a;D.E.W.analy zedecologicaldata;K. J.C.a ndD.E .W.
draftedthemanuscript;Allauthorscontributedtowritingandediting.
DATA ACCESSIBILITY
DNA sequences have been under accession number
MH478912- MH479012.
ORCID
Kenneth J. Chapin http://orcid.org/0000-0002-8382-4050
REFERENCES
Abramoff,M. D.,Mag alhaes,P.J.,&R am,S.J.(2004). Imageprocessing
withImageJ.Biophotonics International,11,36–42.
|
11
CHAPIN e t Al.
Ander son, R. P., & Gonzale z, J. R. I. (2011). Scpe cies-s pecific tuni ng in-
creasesrobustnesstosamplingbiasinmodelsofspeciesdistributions:
AnimplementationwithMaxent.Ecological Modelling,222,27 96 –2811 .
Apodaca,J.J., Trexler,J. C., Jue, N.K., Schrader,M., &Travis,J. (2013).
Large-scale natural disturbancealters geneticpopulationstructure
of the sail fin molly, Poecilia latipinna. The American Naturalist, 181,
254–263.https://doi.org/10.1086/668831
Bancrof t,G.T.,Strong,A.M.,&Carring ton,M.(1995).Defores tationandits
effec tsonforest-nestingbirdsintheFloridaKeys.Conservation Biology,
9,835–844.https://doi.org/10.1046/j.1523-1739.1995.09040 835.x
Bradie,J.,&Leung,B.(2016).Aquantitativesynthesisoftheimportance
ofvariables used in MaxEnt species distribution models. Journal of
Biogeography,44,1344–1361.https://doi.org /10.1111/jbi.12894
Br ad ley,K . A.(2 0 05 ). De li ne ationan do rg an izati onofnatura lf or es tcom-
munitiesofMiami-DadeCounty.DatabasesubmittedtoMiami-Dade
County Department of Environmental Resources Management.
Miami,FL:TheInstituteforRegionalConservation.
Bradley,B.A.(2016).Predictingabundancewithpresence-onlymodels.
Landscape Ecology,31(1),19–30.
Browder, J., Lit tlejoh n, C., & Young, D. (1976). The South Florida study:
South Florida, seeking a balance of man and nature. Ga inesville, FL :
UniversityofFlorida.
Brown,J.E.,Obas,V.,Morley,V.,&Powell,J.R .(2013).Phylogeography
and spatio-temporal genetic variation of Aedes aegypti (Diptera:
Culicidae) populations in the Florida Keys. Journal of Medical
Entomology,50,294–299.https://doi.org/10.1603/ME12173
Chapin,K.J.,& Hebets, E. A. (2016). Behavioral ecology ofAmblypygi.
Journal of Arachnology,42,1–14.https://doi.org/10.1636/V15- 62.1
DeBiasse, M. B., Richards,V.P.,& Shivji, M.S .(2009).Genetic assess-
mentof connectivit yin thecommonreefsponge,C allyspongiavag-
inalis(Demospongiae:Haplosclerida)revealshighpopulationstruc-
turealongtheFloridareeftract.Coral Reefs,27,47–55.
DeBiasse, M. B., Richards,V.P.,&Shivji,M. S. (2010). Genetic assess-
ment ofconnectivit yinthecommon reef sponge, Callyspongia vag-
inalis(Demospongiae:Haplosclerida)revealshighpopulation struc-
ture along the Florida reef tract. Coral Reefs,29,47–55.ht tps://doi.
org/10.1007/s00338-009-0554-0
Dorn, H . W.(1956). Mango gr owing aroun d early Mia mi. Teq ue s ta, 16,
37–5 3 .
Dray,S .,& Du four,A .B .(2007) .Thea de 4p ackage:Imp le me ntingth edu -
alitydiagramforecologis ts.Journal of Statistical Sof tware,22,1–20.
Elith,J.,Graham,C.H.,Anderson,R.P.,Dudík,M.,Ferrier,S.,G uisan,A.,
…Zimmermann,N.E. (2006).Novelmet hods improve prediction of
species’distributionsfromoccurrencedata.Ecography,29,129–151.
https://doi.org/10.1111/j.2006.0906-7590.04596.x
Elith,J.,&Leathwick,J.R.(2009).Speciesdistributionmodels:Ecological
explanation and predict ion across space and time. Annual Review
of Ecology, Evolution, and Systematics, 40, 677–697. https://doi.
org /10.1146/annurev.ecols ys.110 30 8.120159
Esposito,L. A., Bloom, T.,Caicedo,L., Alicia-Serrano,A., Sánchez-Ruíz,
J., May-Co llado Lau ra, J., … Ag narsson , I. (2015). Isl ands withi n is-
lands:Diversificationoftaillesswhipspiders(Amblypygi,Phrynus)in
Caribbeancaves.Molecular Phylogenetics and Evolution,93,107–117.
https://doi.org/10.1016/j.ympev.2015.07.005
Ewing,B.,&Green,P.(1998).Basecallingofautomatedsequencertraces
using phred. II. Error probabilitie s. Genome Research, 8, 186–194.
htt ps://doi.org/10.1101 /gr.8. 3.18 6
Ewing, B. , Hillier, L., Wend l, M., & Gree n, P.(1998). Base calling of au-
tomated sequencer traces using phred. I. Accur acy assessment.
Genome Research,8,175–185.https://doi.org/10.1101/gr.8.3.175
Excoff ier,L .,Smo use ,P.E.,&Quattro,J.M.(1992).Analysisofm olecularv ari-
anceinferredfrommetricdistancesamongDNAhaplotypes:Application
tohumanmitochondrialDNAre stric tiondat a.Genetics,131,479–491.
Feng, X ., & Papeş, M . (2015). Ecologi cal niche mo delling conf irms po-
tential north-east range expansion of the nine-banded armadillo
(Dasypus novemcinctus)intheUSA.Journal of Biogeography,42,803–
807.htt ps: //doi.o rg /10.1111 /jb i.12427
Ficetola , G. F., Thuille r,W., & Miau d, C. (20 07). Predic tion and v alida-
tion ofthe potentialglobal distributionofa problematicalieninva-
sivespecies—TheAmericanbullfrog.Diversity and Distributions, 13,
476–4 85.ht tps://doi.or g/10 .1111/j.1472- 46 42.20 07.0 0377.x
Fleming,T.H.,&Murr ay,K .L.(2009).Populationandgeneticconsequences
of hurric anes for thre e species of West In dian phyllost omid bats.
Biotropica,41,250–256.https://doi.org/10.1111/j.1744-7429.2008.
00466.x
FloridaNatural Areas Inventory (2010). Pine rockland. In (Ed.), Guide to
the natural communities of Florida (2010 Editio n, pp. 1–8). Flori da:
FloridaNaturalA reasInventory.
Folmer,O.,Black, M., Hoeh,W.,Lutz, R., & Vrijenhoek, R. (1994).DNA
primer s for amplification of mitochondrial cytochrome c oxidase
subunit I from diverse metazoan invertebrates. Molecular Marine
Biology and Biotechnology,3,2 94 –299.
Forys,E.A.,&Allen,C.R.(2005).Theimpac tsofsprawlonbiodiversity:
The ant fa una of the lowe r Florida Key s. Ecology & Society, 10, 25 .
htt ps://doi.org /10.5751/ES-013 07-100125
Fowler-Finn, K. D., & Hebets, E. A . (2006). An examination of ago-
nistic interactions in the whip spider Phrynus marginemaculatus
(Arachnida, Amblypygi). Journal of Arachnology, 34,62–76. https://
doi.org /10.1636/S04-104.1
Gillespie, R. G., Baldwin, B. G., Water s,J.M .,Fraser,C. I., Nikula,R., &
Roderick ,G .K .( 2011). Lo ng-dis ta nc ed ispe rs al :Afr am ewor ks forhy -
pothesistesting.Trends in Ecolog y & Evolution,27,47–56.
Gotelli,N.J.,&Sta nto n-G ed des ,J.(2015).Climatec han ge,geneticmar k-
ersandspecies distribution modelling. Journal of Biogeography, 42,
1577–1585.https://doi.org/10.1111/jbi.12562
Graving,J.M.(2015).Nocturnal homing in the amblypygidPhrynusmar-
ginemaculatus. T hesis, Bowling Green State University, Bowling
Green,Ohio,p.44.
Green,P.(1999).PHRAP.Retrievedfromhttp://phrap.org/
Green,P.,&Ewing,B.(2002).PHRED.Retrievedfromht tp://phrap.org/
Guisan,A.,&Thuiller,W.(2005).Predictingspeciesdistribution:Offering
more than simple habitat models. Ecology Letters, 8, 993–1009.
https://doi.org/10.1111/j.1461-0248.2005.00792.x
Hanski,I.(1998).Metapopulationdynamics.Nature,396,41–49.https://
doi.org /10.103 8/23876
Hanley, J. A., & McNeil, B. J. (1982). The meaning and use of thearea
underareceiveroperatingcharacteristic(ROC)curve.Radiology,143,
29–36 .
Hebet s, E. A., & Cha pman, F. R. (200 0). Survi ving the floo d: Plastron
respirationinthenon-tracheatearthropodPhrynus marginemacula-
tus (Ambl ypygi: Arach nida). Journal of Insect Physiology, 46, 13–19.
htt ps://doi.org /10.1016/S0022-1910(99)0 00 96-7
Heinicke , M. P.,Dia z, L. M., & H edges, S. B . (2011). Origin of inv asive
FloridafrogstracedtoCuba.Biology Letters,7,407–410.https://doi.
org/10.1098/rsbl.2010.1131
Hijmans,R.J.,Cameron,S.E.,Parra,J.L.,Jones,P.G.,&Jarvis,A.(2005).
Very high re solution inter polated climate s urfaces for gl obal land
areas.International Journal of Climatology,25,1965–1978.https://doi.
org /10.1002/(I SSN)1097-0 08 8
Hijmans, R. J., & van Et ten, J. (2012). raster: Geographic analysis and
modeling with raster data.R package version 2.0-12.Retrieved from
http://CRAN.R-project.org/package=raster
Hillis, D. M., Dixon, M. T.,& Jones, A. L .(1991).Minimal geneticvaria-
tion in a morphologically diverse species (Florida Tree Snail, Liguus
fasciatus).Journal of Heredity,82,282–286.https://doi.org/10.1093/
oxfordjournals.jhered.a111087
Hu, Y.,Ukasik, P.Ł., More au, C. S., & Russel,J. A .(2014).Correlatesof
gutcommunitycompositionacrossanantspecies(Crphalotesvari-
ans) elucidates causes and consequences of symbiotic variabilit y.
Molecular Ecology,23,1284–1300.
12
|
CHAPIN et Al.
Hoagstrom,C.W.,Ung,V.,&Taylor,K.(2014).Mioceneriversandtaxon
cycles clarify the comparative biogeography of North American
highland fishes. Journal of Biogeography, 41, 644–65 8. https://doi.
org /10.1111/jbi.1 224 4
Hoffmeister, J. E., & Multer, H. G. (1968). Geology and origin of
the Florida Keys. The Geological Societ y of America Bulletin, 79,
1487–1502. http s://doi.org /10.1130/0016-7606(1968)79[1487:G
AOOTF]2.0.CO;2
Hofstetter,R.H.(1975).Theeffectoffireonthepinelandandsawgrass
communitiesofSouthernFlorida.InP.J.Gleason(Ed.),Environements
of Southern Florida: Present and past(pp.201–212).Miami,FL:Miami
GeologicalSociety.
Jansen,K.P.,Mushinsky,H.R.,&Karl,S.A. (2007).Populationgenetics
ofthemangrovesaltmarshsnake,Nerodiaclarkiicompressicauda,in
alinear,fragmentedhabitat.Conservation Genetics,9,401–410.
Jiménez-Valverde,A.,Decae,A.E.,&Arnedo,M.A.(2011).Environmental
suitability of new reported localities of the funnelweb spider
Macrothele calpeiana: An assessment using potential distribution
modelling with presence-onl ytechnique s.Journal of Biogeography,
38,1213–1223.https://doi.org/10.1111/j.1365-2699.2010.02465.x
Jombar t, T. (2008). a degenet: A R pa ckage for the m ultivaria te analy-
sis of genet ic markers. Bioinformatics, 24, 1403–1405. https://doi.
org/10.1093/bioinformatics/btn129
Jombart ,T.,&Ah med,I.(2011).ade genet1.3-1:Newtoo lsfortheanaly-
sisofgenome-wideSNPdata.Bioinformatics,27,307 0–3071 .h t tps ://
doi.org/10.1093/bioinformatics/btr521
Kamvar,Z.N.,Brooks,J.C.,&Grünwald,N .J. (2015).NovelRtoolsfor
analysisof genome-wide population geneticdatawith emphasison
clonality.Frontiers in Genetics,6,208.
Kamvar,Z.N.,Tabima,J.F.,&Grünwald,N.J.(2014).Poppr:AnRpackage
forgeneticanalysis ofpopulationswith clonal,partiallyclonal,and/or
sexualreproduction.Pee rJ,2,e281.https://doi.org/10.7717/peerj.281
Kautz,R.S., &Cox,J.A.(20 01).Strategic habit atsforbiodiversity con-
servation in Florida. Conservation Biology, 15, 55–77. https://doi.
org /10.1111/j.1523 -1739.2001 .98355. x
Kawula,R.(2009).Florida land cover classification system.Tallahassee,FL:
Center for Spatial Analysis, Fish and Wildlife Research Institute,
FloridaFishandWildlifeConservationCommission.
KeyWestChamberofCommerce(2017).KeyWestandMonroeCount y
demographics & economy. Retrieved from http://www.keywest-
chamber.org/demographics-economy.html.
Kirk, N. L .,Andras,J.P., Harvell, C. D., Santos,S .R., &Coffroth, M. A.
(2009).PopulationstructureofSymbiodiniumsp.associatedwiththe
commonseafan,Gorgonia ventalina,in theFlorida Keys acrossdis-
tance,depth,andtime.Marine Biology,156, 160 9–1623.ht tps://doi.
org/10.1007/s00227-009-1196-z
Kolbe,J.J.,Glor,R.E., Schettino,L.R.,Lara,A.C.,Larson,A.,&Losos,J.B.
(2004).GeneticvariationincreasesduringbiologicalinvasionbyaCuban
lizard.Nature,431,177–181.https://doi.org/10.1038/nature02807
Lacs on, J. M., R iccardi, V. M., C alhoun , S. W., & Morizot, D. C. (1989).
Geneticdifferentiationofbicolordamselfish(Eupomacentrusparti-
tus)populationsintheFloridaKeys.Marine Biology,103,4 45–451.
Lacson,J.M.,&Morizot,D.C .(1991).Temporalgeneticvariationinsub-
populationsofbicolordamselfish(Stegastes partitus)inhabitingcoral
reefs in the FloridaKeys. Marine Biology,110,353–357.ht tps://doi.
org /10.1007/BF013 44 354
Larkin,M.A.,Blackshields,G.,Brown, N.P.,Chenna, R., McGettigan,P.
A., McW illiam, H. , … Higgins, D. G . (2007). C lustalW a nd Clust alX
version 2 . Bioinformatics, 23, 2947–2948. https ://doi.or g/10.1093 /
bioinformatics/btm404
Lee, T.N., &Smith, N. (2002).Volumetranspor tvariabilitythroughthe
Florida Keys tidal channels. Continental Shelf Research, 22, 1 361–
1377.https://doi.org/10.1016/S0278-4343(02)00003-1
Lenczewski,B.(1980).ButterfliesofEvergladesNationalPark.Everglades
NationalPark,SouthFloridaResearchCenterReportNo.T-588.
Li,Y.M.,Dlugosch,K .M.,& Enquis t,B.J. (2015). Novel spatialanalysis
methods reveal scale-dependent spread and inferlimiting factors
ofinvasion byS ahara mustard. Ecography,38,311–320.https://doi.
org /10.1111/eco g.007 22
Lloyd, J. D., & Slater, G. L. (2012). Fire his tory and the structure of
pine-rocklandbirdassemblages.Natural Areas Journal, 32, 2 47–25 9.
https://doi.org/10.3375/043.032.0303
Loope, L. L., &Dunevitz,V.L. (1981).Investigations of early plant succes-
sion on abandoned farmland in Everglades National Park. Everglade s
Nat.ParkSouthFloridaRs.Cent.Rep.T-645
MacArthur,R.H.,&Wilson,E.O.(1967).The theory of isla nd biogeography
(Vo l. 1).Princeton,NJ:PrincetonUniversityPress.
Maddison,D.R.,&Maddison,W.P.(2016a).Chromaseq: A Mesquite pack-
age for analyzing sequence chromatograms , version 1.2.Retrievedfrom
https://www.mesquiteproject.org/
Maddis on, W. P., & Maddis on, D. R. (2016b). Mesquite: A modular sys-
tem for evolutionary analysis, version 3.10.Retrievedfromhttp://mes-
quiteproject.org
Maguire,J.(1995).Restoration pl an for Dade County ’s pine Rockland for ests
following Hurricane Andrew.Miami,FL:DadeCountyDepar tment of
EnvironmentalResourcesManagement.
Mantel, N.(1967).The detec tionof diseaseclusteringandageneralized
regressionapproach.Cancer Research,27,209–220.
Maschin ski, J., Ross, M . S., Liu, H., O’ Brien, J., vo n Wettberg, E . J., &
Haskins , K. E. (2011). Sinkin g ships: Conser vation optio ns for en-
demic taxa threatened by sea levelrise. Climatic Change, 107, 14 7–
167.https://doi.org /10.1007/s10584- 011-0 083-z
McClen achan, L . (2013). Recreati on and the “rig ht to fish” movem ent:
Angler sandecologicaldegradationintheFloridaKeys.Environmental
Histor y,18,76–87.https://doi.org/10.1093/envhis/ems110
Meyers,R . L., & Ewel,J.J.(1990).Ecosystems of Florida(pp. 230–280).
Orlando,FL,Universit yofCentralFloridaPress.
Miller, K. A ., & Glantz, M . H. (1988). Climate and e conomic compet i-
tivene ss: Florida f reezes and t he global ci trus proce ssing indus try.
Climatic Change,12,135–164.https://doi.org/10.1007/BF00138936
Muma,M.H.(1967).Scorpions,WhipScorpionsand Wind Scorpionsof
Florida.InArthropods of Florida and Neighbouring land areas(Volume
4,pp.1–28).FL:FloridaDepartmentofAgriculture.
Muma,M.H.(1967).Scorpions,WhipScorpionsand Wind Scorpionsof
Florida.p.1-28InArthropodsofFloridaandNeighboringLandAreas,
Volume4.FloridaDepartmnetofAgriculture,Florida.
Noss, R . F.,L aRoe, E. T., & Scott , J. M. (1995). Endangered ecosystems
of the United States: A preliminary assessment of loss and degradation
(Vo l. 28). Washing ton, DC: USDepar tment of the Interior,National
BiologicalSer vice.
Orr,D.W., & Ogden ,J.C. (1992).Theimpactof HurricaneAndrew on
theecosystemsofSouthFlorida.Conservation Biology,6,488–490.
Paradis,E.(2010). pegas:An Rpackage forpopulationgeneticswithan
integrated-modular approach. Bioinformatics, 26 , 419–420. https://
doi.org/10.1093/bioinformatics/btp696
Pearson,R.G.,Raxworthy,C.J.,Nakamura,M.,&Peterson,A.T.(2007).
Predic ting spec ies distri butions fr om small num bers of occu rrence
records: A test case using cryptic geckos inMadagascar.Journal of
Biogeography,34,102–117.
Peck, S.B.,&Howden, H.F.(1985).Biogeographyof scavengingscarab
beetles in the Florida Keys: Post-Pleistocene land-bridge islands.
Canadian Journal of Zoology,63,2730 –2737.htt ps://doi .o rg/1 0.1139/
z85-4 07
Peterson,A.T.(2011).Ecologicalnicheconservatism:Atime-structured
reviewofevidence.Journal of Biogeography,38,817–827.https://doi.
org /10.1111/j.1365 -2699.2010 .02456.x
Peterso n, M. N., Lo pez, R. R ., Frank, P. A., Por ter, B. A., & Sil vy, N. J.
(200 4) .Ke yd ee rf aw nres po ns etour banizati on :I ssusta in ab le de vel-
opmentpossible?Wildlife Society Bulletin,32,493–499.htt ps: //doi. or
g/10.2193/0091-7648(2004)32[493:KDFRTU]2.0.CO;2
|
13
CHAPIN e t Al.
Phillips ,S.J.,Anderson,R.P.,&Schapire,R.E.(20 06).Maximumentropy
modeli ng of species geogr aphic distri butions. Ecological Modelling,
190,231–259.https://doi.org/10.1016/j.ecolmodel.2005.03.026
Phillips,S.J.,&Dudík,M.(2008).Modelingofspeciesdistributionswith
Maxent:Newextensionsandacomprehensiveevaluation.Ecography,
31,161–175.https://doi.org/10.1111/j.0906-7590.2008.5203.x
Podas, P., Crisci, J. V., & Katinas, L. (2006). Historical biogeography:
A review of it s basic concepts and critical issues. Journal of Arid
Environments,66,389–403.
Possley, J. E., Maschinski, J. M., Maguire, J., & G uerra, C. (2014).
Vegetation m onitoring to gui de management de cisions in Miami’s
urbanpine rocklandpreserves.Natural Areas Journal,34, 154–165.
https://doi.org/10.3375/043.034.0205
Possley,J.,Woodmansee,S.W.,&Maschinski,J.(2008).Patternsofplant
compositioninfragmentsofgloballyimperiledpinerocklandforest:
Effectsofsoilt ype,recentfirefrequency,andfragmentsize.Natural
Areas Journal,28,379–394.https://doi.org/10.3375/0885-8608(20
08)28[379:POPCIF]2.0.CO;2
Prendini, L., Weygoldt, P., & Wheeler, W. C. (2005). Systematics of
the Damon variegatus group of African whip spiders (Chelicerata:
Amblypygi): Evidence from behaviour, morphology and DNA.
Organisms Diversity & Evolution, 5, 203–236. https://doi.
org/10.1016/j.ode.2004.12.004
Pr oos di j ,A .S. ,So se f, M.S ., Wie rin ga, J. J. ,&R ae s ,N .( 2 016).Min imu mr e-
quirednumberofspecimenrecordstodevelopaccuratespeciesdis-
tributionmodels. Ecography, 39, 542–552. http s://doi. org /10.1111/
ecog.01509
Quintero, D. (1981). The amblypygid genus Phrynus in the Americas
(Amblypygi,Phrynidae).Journal of Arachnology,9,117–166.
Rayor,L.S.,&Taylor,L.A.(2006).Socialbehaviorin amblypygids,anda
reassessmentofarachnidsocialpat terns.Journal of Arachnology,34,
399–421.
RCoreTeam(2013).R: A language and environment for statistical comput-
ing.Vienna,Austria:RFoundationforStatisticalComputing.
Richar ds, V.P., Thom as, J. D., Stanhop e, M. J., & Shivj i, M. S. (2006).
Geneticconnectivityin the Florida reefsystem:Comparative phy-
logeographyofcommensalinvertebrateswithcontrastingreproduc-
tivestrategies.Molecular Ecology,16,139–157.
Robertson, W.B.Jr(1953).A survey of the effects of fire in the everglades
national park.Homestead,FL:NationalParkSer vice.
Rober tson, W. B. Jr (1962). Fire an d vegetatio n in the Evergla des. Tall
TimbersFireEcologyConferenceProceedings,1,67–80.
Ross,M.S.,O’Brien,J.J.,&daSilveiraLoboSternberg,L.(1994).Sea-level
riseandthereductioninpineforests intheFloridaKeys. Ecological
Applications,4,144–156.https://doi.org/10.23 07/1942124
Ross, M. S ., O’Brie n, J. J., Ford, R . G., Zhang , K., & Mor kill, A. (2 009).
Disturbance and the risingtide: The challenge of biodiversityman-
agement on low-island e cosystems. Frontiers in Ecolog y and the
Environment,7,471–478.https://doi.org/10.1890/070221
Santer,R .D., &Hebets,E.A.(20 09a).Tactilelearningbyawhipspider,
Phrynus marginemaculatusC.L.Koch(Arachnida,Amblypygi).Journal
of Comparative Physiolog y A,195,393–399.https://doi.org/10.1007/
s00 359-0 09-0 417-8
Santer, R. D., & H ebets, E . A. (20 09b). Prey capt rure by the w hip spi-
der Phrynus marginemaculatusC .L. Koch.Journal of Arachnology,37,
109–112.https://doi.org/10.1636/ST07-139.1
Santer,R.D.,&Hebets,E.A.(2011).Thesensotyandbehaviouralbiolog y
ofship spiders(A rachnida,Amblypygi).Advances in insect physiology
41,1– 6 4.
Simberloff,D.S., &Wilson, E. O.(1969).Experimentalzoogeographyof
islands:Thecolonizationofemptyislands.Ecology,50,278–296.
Simberloff,D.,&Wilson,E.O.(1970).Experimentalzoogeography ofis-
lands.at wo-yearrecordofcolonization.Ecology,51,934–937.
Simberloff, D.(1976).E xperimental zoogeographyofislands: effectsof
islandsize.Ecology,57,827–848.
Shres tha, Y., Wirshi ng, H. H., & H arasew ych, M. G. (2 015). The genu s
Cerion (Gastro poda: Ceri onidae) in the F lorida Keys. PLoS One, 10,
e0137325.https://doi.org/10.1371/journal.pone.0137325
Small, J. K. (1929).From eden to Sahara, Florida’s tragedy.Lancaster,PA:
SciencePress.
Small, J.K.(1917).The tree cactioftheFloridaKeys.Journal of the New
York Botanical Garden,18,199–203.
Snyder, J.R., Herndon, A ., &Robertson, W. B. Jr (1990). Sout hFlorida
rocklands.InR.L.Myers,&J.J.Ewel(Eds.),Ecosystems of Florida(pp.
230–277).Orlando,FL:UniversityofCentralFloridaPress.
Spence,A .J.,& Hebets, E. A. (20 06). Anatomy andphysiologyof giant
neuronsintheantenniformlegoftheamblypygidPhrynusmarginac-
ulatus.Journal of Arachnolog y,34,566–577.
Titte nsor, D. P.,Baco , A. R., B rewin, P. E., Clark , M. R., Co nsalvey, M.,
Hall-Spen cer, J., … Rogers, A. D. (2 009). Predict ing global habit at
suitabilityforstonycoralsonseamounts.Journal of Biogeography,36,
1111–1128.https: //doi.o rg /10.1111 /j.1365 -2699.2 00 8.02062 .x
USCensusBureau,U.S.DecennialCensus,UnitedStatesCensusBureau.
Retrievedfromhttp://www.census.gov/prod/www/decennial.html
Václavík,T.,Kupfer,J.A .,& Meentemeyer, R. K. (2012). Accountingfor
multi-sc alespatialautocorrelationimprovesper formanceofinvasive
species distribution modelling(iSDM). Journal of Biogeography, 39,
42–55.h tt ps://doi.org /10.1111/j.1365 -2699.2011.02 589.x
Villanova, V.L ., Hughes, P.T., & Hoffman, E. A . (2017). Combining ge-
neticstructureanddemographicanalysestoestimatepersistencein
endangered Key deer (Odocoileus virginianus clavium).Conservation
Genetics,5,1–16.
Waltari,E.,&Gura lnick ,R .P.(2 00 9).Ecolog ic aln ichem odell ingofm ontane
mammalsintheGre atBasin ,NorthAmerica:Exami ningpas tandpresent
connectivityofspeciesacrossbasinsandranges.Journal of Biogeography,
36,148–161.https://doi.org/10.1111/j.1365-2699.2008.01959.x
Warren,B.H.,Simb erloff,D.,Ricklefs,R.E.,Aguilée,R.,Condamine,F.L.,
Gravel,D.,…Thébaud,C.(2015).Islandsasmodelsystemsinecology
andevolution:ProspectsfiftyyearsafterMacA rthur-Wilson.Ecology
Letter s,18,20 0–217.https://doi.org/10.1111/ele.12398
Webb,D. S.(1990).Historicalbiogeography.InR .L.Myer s,& J.J.Ewel
(Eds.), Ecosystems of Florida(pp.70–100).Orlando,FL:Universit yof
CentralFloridaPress.
Weygoldt, P.(2000).Whip spiders (Chelicerata: Amblypygi): Their biology,
morphology and systematics.Stenstrup,Denmark:ApolloBooks.
Williams,D.A.,Wang,Y.,Borchetta ,M.,&Gaines,M .S.(20 07).Genetic
diversity and spatialstruc ture of a keystonespecies in fragmented
pinerocklandhabitat.Biological Conservation,138,256–268.https://
doi.org/10.1016/j.biocon.2007.04.024
Wilson,L .D.,& Porras, L.(1983). The ecologicalimpact of man on the
South Florida herpetofauna. Univ. Kansas Mus. Nat. Hidt. Spec.
Publ.,9,1–89.
Winter, D. J. (2012). mmo d: An R library for thecalculation of popula-
tiondifferentiationstatistics.Molecular Ecology Resources,12,1158–
1160.ht tp s://doi.org/10.1111/j.1755-0998.2012.0 3174. x
Wisz,M.S.,Hijmans,R.J.,Li,J.,Peterson,A .T.,Graham,C .H.,&Guisan,
A.(2008). Effec tsof samplesizeontheperformance ofspeciesdis-
tributionmodels.Diversity and Distributions,14,76 3–773.htt ps ://doi .
org /10.1111/j.1472-4 642. 20 08 .0 04 82 .x
How to cite this article:ChapinKJ,WinklerDE,WiencekP,
AgnarssonI.Islandbiogeographyandecologicalmodelingof
theamblypygidPhrynus marginemaculatusintheFloridaKeys
archipelago.Ecol Evol. 2018;00:1–13. h t tp s : //d o i .
org /10.1002/ece3.433 3
