ArticlePDF Available

Island biogeography and ecological modeling of the amblypygid Phrynus marginemaculatus in the Florida Keys archipelago

Authors:

Abstract

Aim The biogeography of terrestrial organisms across the Florida Keys archipelago is poorly understood. We used population genetics and spatioecological modeling of the Amblypygi Phrynus marginemaculatus to understand the genetic structure and metapopulation dynamics of Keys populations that are otherwise isolated by human development and ocean. Location The Florida Keys archipelago and mainland Florida. Methods We sequenced a 1,238 bp fragment of mtDNA for 103 individuals of P. marginemaculatus from 13 sites in the Florida Keys and South Florida, binned into four regions. We used population genetic analyses to understand the population structure of the species throughout its US range. Furthermore, we used ecological modeling with climate, habitat, and human development data to develop habitat suitability estimates for the species. Results We found clear genetic structure between localities. The Lower Keys, in particular, support populations separate from those in other regions studied. Ecological modeling and genetic analyses showed the highest habitat suitability and genetic isolation in the Lower Keys, but urban development across the species range has resulted in the loss of most historical habitat. Main conclusions A mainland‐metapopulation model best fits P. marginemaculatus gene flow patterns in the Florida Keys and mainland. Ocean currents likely play a role in metapopulation dynamics and gene flow for terrestrial Keys species like P. marginemaculatus, and genetic patterns also matched patterns consistent with geologic history. Suitable habitat, however, is limited and under threat of human destruction. The few remaining pockets of the most suitable habitat tend to occur in parks and protected areas. We argue that conservation efforts for this species and others in the terrestrial Florida Keys would benefit from a deeper understanding of the population genetic structure and ecology of the archipelago.
Ecology and Evolution . 201 8 ;1–1 3 .    
|
 1
www.ecolevol.org
Received:22January2018 
|
  Revised:14Ju ne2018 
|
  Accepted:17June2018
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
Thisisanop enaccessarticleundert hetermsoft heCreat iveCommonsAttr ibutio nLicense ,whichpe rmitsu se,dist ributi onandrep roduc tioninanym edium,
provide dtheorig inalworkisproper lycited.
©2018TheAut hors.Ecology and Evolutionpu blishedbyJohnWiley&SonsLtd .
1Depar tmentofEcology&Evolutiona ry
Biolog y,Universi tyofCalifornia,Los
Angeles,LosAngeles,C alifornia
2Depar tmentofEcology&Evolutiona ry
Biolog y,Universi tyofCalifornia,Irvi ne,
Irvine,California
3Depar tmentofBiology,Universit yof
Vermont,Burling ton,Verm ont
Correspondence
KennethJ.Chapin,Depar tmentofEco logy
&EvolutionaryBiology,Universit yof
Califo rnia,LosAngeles,LosAngeles,C A.
Email:chapinkj@gmail.com
Funding information
TheodoreRooseveltMemorialFundoft he
AmericanMuse umofNatur alHistory;Edwin
W.PauleyFell owship,UCLA;Department
ofEcolog y&EvolutionaryBiology,UCL A;
Depar tmentofEco logy&Evol utionary
Biolog y,UCI;NationalGeo graphi cSociet y
(WW-203R-17)toIA
Abstract
Aim:ThebiogeographyofterrestrialorganismsacrosstheFloridaKeysarchipelagois
poorly understood.Weusedpopulationgenetics andspatioecological modeling of
the Ambly pygi Phrynus marginemaculatusto un derstand t he genetic stru cture and
metapopulationdynamicsofKeyspopulationsthatareotherwiseisolatedbyhuman
developmentandocean.
Location:TheFloridaKeysarchipelagoandmainlandFlorida.
Methods: We sequenced a 1,238bp fragment of mtDNA for 103 individuals of
P. marginemaculatusfrom13sitesintheFloridaKeysandSouthFlorida,binnedinto
four regions. We used population genetic analy ses to understan d the population
structureofthespecies throughout its US range.Furthermore,weusedecological
modelingwithclimate,habitat,andhumandevelopmentdatatodevelophabitatsuit-
abilityestimatesforthespecies.
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.
Ecologicalmodelingandgeneticanalysesshowedthehighesthabitatsuitabilityand
geneticisolationintheLowerKeys,buturbandevelopmentacrossthespeciesrange
hasresultedinthelossofmosthistoricalhabitat.
Main conclusions:A mainland-metapopulation model best fits P. marginemaculatus
geneflowpatternsintheFloridaKeysandmainland.Oceancurrentslikelyplayarole
inmetapopulationdynamicsandgeneflowforterrestrialKeysspecieslikeP. margin-
emaculatus,andgeneticpatternsalsomatchedpatternsconsistentwithgeologichis-
tory.Suitablehabitat,however,islimitedandunderthreatofhumandestruction.The
fewremainingpockets ofthemostsuitable habitattendtooccurinparksandpro-
tected areas.Weargue that conser vation efforts for this speciesandothers inthe
terrestrialFloridaKeyswouldbenefitfromadeeperunderstandingofthepopulation
geneticstructureandecologyofthearchipelago.
KEY WORDS
Amblypygi,ecologicalmodel,FloridaKeys,habitatsuitability,islandbiogeography,MaxEnt,
metapopulation,pinerockland,populationgenetics,urbandevelopment
Present address
KennethJ.Chapin,UniversityofArizona,
Depar tmentofEcologyandEvolutionary
Biolog y,Tucson,Arizona.
2 
|
   CHAPIN et Al.
1 | INTRODUCTION
Acentralgoalofbiologyistounderstandhowtimeandspaceshaped
the evolut ionary his tory of life (Ma cArth ur & Wilson, 1967; Podas,
Crisci,&Katinas,2006;Warrenetal.,2015).TheFloridaKeysisabio-
diversearchipelagowithhighendemism(Forys&Allen,2005;Kautz&
Cox,2001),butthebiogeographyofterrestrialpopulationsalongthe
islandchainandnearbymainlandFloridaremainspoorlyunderstood.
This is par ticularly su rprising conside ring the Florid a Keys was the
studysiteusedtodevelopsomeofthemosticonicbiogeographythe-
ory (e.g., Mac Arthur&Wilson, 1967;Simberloff, 1976;Simberloff &
Wilson,1969,Simberloff&Wilson1970;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,thedistributionofspeciesacrosstheKeysisimport antforun-
derstanding both specieslong-termsur vivaland biodiversity.Thisis
especiallytrue considering the precariousness ofFlorida Keys habi-
tat sinthef aceofhumandisturbance,includinghumandevelopment,
deforestation,nonnativespeciesintroductions, andhuman-induced
climate change impacts like sea level rise, increased catastrophic
storms, and altered fire regimes (Bancroft, Strong, & Carrington,
1995;Forys,2005;Maschinskietal.,2011;Ross,O’Brien,&daSilveira
LoboSternberg,1994;Ross, O’Brien,Ford,Zhang, &Morkill,20 09).
Fur the rm or e, th eFlo ri daKey sr em ai namaj or to ur is td es tina ti on wi th
over4.5milliontouristsvisitingannually(McClenachan,2013).
Our understanding of the genetic structure of Florida Keys or-
ganismslargelycomesfromresearchofmarinespecies,whereocean
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;Shresthaetal.,2015).
Geneti c research using n ative terrest rial species i s scarce, but
not absent . The mosqu ito Aedes aegyptishowednogeneticstruc-
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 nativespeciesacross the Keys (Heinicke, Diaz,
& Hedges, 2 011; Kolbe etal. , 2004). An alloz yme elect rophoretic
study ontheFloridaTreeSnail (Liguus fasciatus)revealedlowlevels
ofgenetic diversity, likely duetoa recent introductionfrom Cuba,
orthelowresolutionofallozymeapproaches(Hillis,Dixon,&Jones,
1991).ThelandsnailCerion incanumshowedsomehaplotypestruc-
turebetweenUpperandLowerKeys,likelycausedbydifferencesin
thetimingofformationoftheKeys.Shresthaetal.(2015)proposed
that the C. incanumspread southwesterly to colonizenew Keys as
theyformed,withLowerKeypopulationsbeingtheyoungest.Lastly,
antgutmicrobiotashowedgeneticstructurebetweentheupperand
lowerkeys(Huetal.2013).Thatsaidmostpastbiogeographicstudies
havebeenlimitedtononnativespeciesorexcludedFloridamainland
populations.Nostudieshaveinvestigatedthebiogeographyofana-
tivespeciesoccupyingtheentirearchipelagoandmainland,orhave
anyinvestigatedhumanimpact sonstructureandconnectivity.
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 ofrural land
westofEvergladesNationalPark,hasmorethandoubledsince1950
(US Census B ureau). While po pulation sizes of re sidents may have
stabilized over the last 25years (currently ca. 77,000 residents),
the numb er of tourists vis iting the Keys is enor mous. In 2014, an
estimated 4.516million tourists visited the Florida Keys (Key West
Ch am berofCo mm erc e, 20 17 ). KeyWe st, thema jo rc ity of th eF lor id a
Keys,locatedonanisland ofonly 19km2,hasover52,000housing
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 yhuma nact ivit y(P eterson,L opez,Frank,P or te r,&Silvy,20 04).
Weused fieldobservationscoupledwithlanduseandclimate
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
studiesofspatial patternsof diversity(Hoagstrom,Ung,&Taylor,
2014;Peterson, 2011;Walt ari&Guralnick,2009) ,genetic struc-
ture (Gotelli & Stanton-Geddes,2015),and the historicspread of
invasivespecies(Li,Dlugosch,&Enquist,2015;Václavík,Kupfer,&
Meentemeyer,2012). Additionally, SDM have identified suitable
habitatforspeciesofconcern(Tittensoretal.,2009)andidentified
climaticfactorsdrivingspeciesdistributions, including responses
toclimatechange(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;
Elithetal.,2006; Jiménez-Valverde,Decae, &Arnedo, 2011)and
smallsamples(Pearson,Raxworthy,Nakamura,&Peterson,2007;
Proosdij,Sosef,Wieringa,&Raes,2016;Wiszetal.,2008).
Weaimedtoquantifythegeneticandecologicalcharacteristicsof
P. marginemaculatuspopulationsintheUS.In particular,weaimed to
uncover the genetic structureofthe speciesacross theFlorida Keys
archipelago,understandtheevolutionaryhistoryofKeyspopulations
inrelationtothe speciesrangevia phylogeneticanalysis,and identify
suitablehabitatandlocationsofputativepopulationsthroughoutthe
species’ potentialrange.Together,theseresultswillbe thefirst toex-
amineP. marginemaculatusinthewild,andwillprovidecritic alinforma-
ti o napp l i c abl e tom a nyt e r res t r i ali s lan d s peci e san d t hei r con s e r v ati o n .
2 | MATERIALS AND METHODS
2.1 | Study site
The FloridaKeysisaca250-km-longarchipelagoamountingtoca
350km2ofdryland,extendingfromKeyLargoborderingmainland
    
|
 3
CHAPIN e t Al.
Florida s outhwest to Key West 140km f rom Cuba. The Keys ar e
made from t wo geologic form ations that both for med during the
Tarantian Pleistocene (0.126–0.0117mya) and raising above sea
level during the Wisconsin glaciation (ca. 100,000 before present;
Hoffmeister&Multer,1968;Shresthaetal.,2015).TheLowerKeys
(BigPineKeytoKeyWest)formedfromcementedsandbars,result-
inginoöliticlimestone(termedMiamiLimestone).TheseKeyscurve
westandorientlaterallyduetogulfstreamcurrents.TheUpperKeys
(Bahia HondaKey to Key Largo), however,are constituted of fossil
coralreefs(termed KeyLargoLimestone)withoutstrong lateraliza-
tion.SomespeciesshowendemismtotheLowerorUpperKeys,but
notbothbecauseofthesedifferences(e.g.,Peck&Howden,1985).
Florida m ixed hardwoo d forests h ave persiste d since the bir th
ofterrestrialFlorida25mya(Webb,1990).Sincethen,specieshave
hadtworoutestocolonizingFloridahabitats:alandmigrationdown
theFloridapeninsulafromNorthAmerica,orwatermigrationnorth
orwestfromtheC aribbeanandBahamas,asFloridawasnevercon-
nected totheCaribbeanislands (Snyder,1990).Thus,we might ex-
pectthediversityofFloridatobeshapedbytropicalspeciesableto
dispers e over water, and temper ate species o nly able to est ablish
vialand.ThishasresultedinadominanceofvertebratesfromNorth
AmericabutflorafromtheCaribbean(Snyder,1990).Exceptionsin-
cludeninebirdspecies,andtwospecieseachofbat,frogandlizard,
allofwhichhave Caribbeanorigins. That being said,natural migra-
tionsareonlyclearforafewofthesespecies;manymighthavebeen
introduced via humans, and still others have gone extinct(atleast
locall y) since their discover y (Snyder, 1990). Inver tebrate biogeo-
graphicpatternsare more mixed, but still fit themodel of tropical
watermigratorsversustemperatelandmigrators.Forexample,most
Florida ant species have Nor th American origins,while Butterflies
arelargelyCaribbean(Lenczewski,1980).
Land to support the growing humanpopulations of Miami and
theFloridaKeyshavebeenlargelyobtainedbyclearingpinelandand
hammock(Snyder,Herndon,&Robertson,1990),amajorconserva-
tionissuethatresearchershavebeenbringingattentiontofornearly
acentury (Small, 1929).The firstsettlers in Southern Floridawere
concentrated in the Florida Keys. Early settlers exploited pineand
hardwoo d trees (espe cially Mahog any) for lumber, fue l, and slash-
and-burn agriculturalpractices(Small, 1917;Browder,Littlejohn,&
Young,1976;Wilson&Porras, 1983).Asa 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 pooragricultural land due to an abundance of limestone
rocksthatmadesoilunworkable.Greatlyexpandingagricultureof-
tentimessparedrocklandhabitatinfavorofdraininggladesforcrops
untiltheinventionoftherockplowinthe1950s.Thisenabledlime-
stonerockstobecollectedandseparatedfromsoil,therebyenabling
agricultural access. Lime stone rocks co llected via h and, plow, and
mine, were used as buildingmaterials andcanbeseeninhistorical
building s, walls, an d gardens of th e Florida Keys to day.Ro cklands
arenutrientandwaterdepauperate,andthusrequiretheheavyuse
ofirrigationandchemicalfertilizerstobeagriculturallyusable.Asa
result,abandonedrocklandsshowlittleresemblancetotheoriginal
ecosystem, and are often dominated byinvasive species (Loope &
Dunevit z,1981).Theabilityto turn rocklandintospaceforagricul-
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. marginemaculatusC.L.Koch,
1840 as amodel to understandthegeneral biogeographicpattern
ofterrest rialFloridaKeysspecies.Amblypygidsareasmallar achnid
order(ca.220spp.)oflarge nocturnalpredators(Chapin&Hebets,
2016).Phrynus marginemaculatusistheonlyamblypygidspeciesin
theUSeastoftheMississippiriverandthemostcommonlystudied
speciesofamblypygid(Chapin&Hebets,2016;Figure1).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
navigatemazesusingtactilecues(Santer&Hebets,2009a, 2009b)
via exceptional brain structures and sensory systems (Chapin &
Hebets,2016;Santer&Hebets,2011).Whilef ascinatinglaboratory
researchhasbeen conduc ted on the species, no researchontheir
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
northasMartinCounty,FL,butrecentrecordsareabsent.Thespe-
cies is also foundon severalBahamian islands, Cuba,Jamaica,and
Hispaniola(Muma,1967;Quintero,1981).Research onthespecies,
FIGURE1 PhotographsoftheamblypygidPhrynus
marginemaculatusintheFloridaKeys.(a)Notetheelongatedfirst
pairoflegsadaptedassensorystructures.(b)Closeupofthebody
highlight scolorationandpat terning
5 cm
(a)
(b)
4 
|
   CHAPIN et Al.
however,hasonlyoccurredincaptivity,withanimalscollectedfrom
asingleisland(BigPineKey;Hebets&Chapman,2000;Fowler-Finn
&Hebets, 2006;Spence &Hebets20 06; Santer & Hebet s,2009a,
2009b),thepettrade(Rayor&Taylor,2006),orboth(Graving,2015).
Interestingly, P. marginemaculatus has evolved a plastron to
breatheunderwater,whichtheycandoforupwardsof24hr(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, alongwith ocean currents, likely promote
oceanicdispersal(Fleming&Murray,2009;Gillespieetal.,2011).A
plastronallowingunderwaterbreathinglikelyextends thedispersal
propens ities across b odies of water, and th e likelihood of su rvival
duringoceanmigration,inP. marginemaculatus.
Twomolecular phylogenetic studies have focused on amblypy-
gids. First, a phylogeny of the Damon variergatusgroup delineated
two crypticspecieswithinDamon,an African genus of amblypygid
(Prendini,Weygoldt,&Wheeler,2005).Second,phylogeneticanaly-
sesofPhrynusspeciesinPuertoRicorevealedhiddendimensionsof
diversit yacross cavepopulations(Espositoetal.,2015).In par ticu-
lar,Espositoetal.(2015)notedhigh levels ofdiversityacrossmito-
chondrialbutnotnuclear,genomes.Herewefocusonmitochondrial
sequencesto examineifsimilargenetic structure acrosssmallgeo-
graphicalscalesisevidentintheFloridaKeys.
2.3 | Specimen collection
Wecollected P. marginemaculatusspecimensfrom13locationsin
southernmainlandFloridaandtheFloridaKeysarchipelagousing
anonrandomsamplingmethod( Table1;Figure2).Welimitedour
survey to upland habitat types, as we assumed that P. margin-
emaculatuswouldnotbefoundininter tidalhabitat slikemangrove
swamps an d floodplains. Additional survey areas were selected
fromhistoricrecords,whichwereallassociatedwiththeseupland
habitat types(Table1).P. marginemaculatushideunderdebris,es-
peciallylimestone rocks, during theday (Chapin&Hebets,2016;
Hebets&Chapman,2000).Oursamplingregimeincludedwalking
trails an d looking for P. marginemaculatusunderrocks,logs,and
other larger debris. When found, we collected genetic samples
andstored them in 95%ethanol ondry ice. Wedissectedmuscle
tissuefromoneorafewappendages,dependingonthesizeofthe
specimen.
2.4 | Extraction, amplification, and sequencing
WeextractedgenomicDNAusingQIAGENDNeasyBlood&Tissue
Kits.Wefollowedthestandardkitprotocolbutusedchilledethanol
(−20°C) anda50-mLfinalelution. Weamplified a1, 238nucleotide
sequenceofthemitochondrialgenecy tochromecoxidasesubunit1
(COI)byperforming34iterationsofthefollowingcycleonathermal
cycler:30sat94°C,35sat48°C,and90sat65°C,beginningwithan
initialcycleof 2minat 94°Cand ending with10minat72°C.Using
illustraPuReTaqReady-To-GoPCRbeadsand40 0-nMforwardand
reverse primers,the longor shor t read of COI was sequenced for
eachsample.LCOI1490wasuse dast heforwardprimerforbotht he
longandshortreads,whileHCOI2198wasusedasthereverseprimer
fortheshort readsandC1-N-2776wasusedasthereverse primer
fo r t he lo ngre a ds (LCO I14 90 GG TCA AC A A ATC ATA A AG ATAT TG G ,
HCOI2198TAAACTTCAGGGTGACCAAAAAATCA,andC1-N-2776
GGATAATCAGAATATCGTCGAGG; Folmer, Black, Hoeh, Lutz, &
Vrijenhoek,1994).WechoseamtDNAsequenceasmtDNAismuch
more informative than nuclear DNA among Amblypygi (Esposito
etal.,2015).
2.5 | Cleanup and alignment
AmplifiedfragmentsweresenttotheUniversityofArizonaGenetic
Core and Genewizfor 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 thenproof-
read in Me squite. Seq uences were a ligned with C lustal W2 (Larkin
etal.,2007)inMesquite.
2.6 | Genetic analysis
Wegroupedlocalitiesintofourregionsbymajorgeologic features:
the mainland, Key Largo, Upper Keys (excluding Key Largo), and
LowerKeys.Weproducedgeneticdiversit yindicesforeachregion
and used a hierarchical analysis of molecular variance (AMOVA;
pop N H G λ λc Hexp π
LeyLargo 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
UpperKeys 13 2.56 13.0 0 0.923 1.000 0.150 0.013
LowerKeys 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. GisSto dda rtan dTa yl or ’s in de xo fM LGd ivers it y;His th eS ha nno n–W ie ne rI nd exof mu lt il ocus
genotype(MLG)diversity;Hexpis Nei’sunbiased genediversity; πisnucleotidediversity; Nist he
numberofindividualssequenced;λisSimpson’sIndex;λcisSimpson’sindexcorrectedforvariation
insamplesize.
TABLE1 Geneticdiversityindicesof
COIsequencesformajorregionsinwhich
Phrynus marginemaculatus occur
    
|
 5
CHAPIN e t Al.
Excoffier,Smouse,&Quattro,1992)toestimatethevariancewithin
and betweenlocalities and regions. Wetested for isolation by dis-
tance (IBD) by testing for a correlationbetween 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 examinegenetic
divergencebetweenregionsandcalculatedpairwise GSTasanes-
timate ofmigrationbetween regions. WeusedtheR2.3.2 (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)forgeneticanalyses.
2.7 | Ecological modeling
Weusedlocalitydatafromthe103field-collectedspecimenstoes-
timatethegeographicrangeofP. marginemaculatususingtheniche
modeling software Maxent 3.3.3 (Phillips, Anderson, & Schapire,
2006;Phillips &Dudík,2008).Maxent usesamaximum-entropyal-
gorithm to predict species geographic ranges using presence-onl y
dataandenvironmentalGISlayers.Weevaluated19BioClimclimate
variables(BIO1–19;Hijmans,Cameron,Parra,Jones,&Jarvis,2005)
ata30-arc-secondresolution(ca.1km2)forinclusioninourmodels.
Weincludedelevationandageologicmapinourfirstsetofmodels
totesttheircontributiontoinformingmodelpredictions.Bothlay-
erswereobtainedfromtheUnitedStatesGeologicalSurvey(http://
www.usgs.gov). We ran a second set of models thatincludedland
usedatabased onimagerymadepublicallyavailable bytheFlorida
DepartmentofEnvironmentalProtection(http://geodata.dep.state.
fl.us/). This dataset included 195categories ofland use based pri-
marilyonhumanuse(e.g.,agriculture,urbandevelopment,transpor-
tation corridors) but alsoincluded subcategoriesofvegetation and
other ecologically relevanthabitattypes (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
freezelineinFlorida(Miller&Glantz,1988).Thereisnoevidenceto
indicatethatP. marginemaculatusoccurs beyond thisline(Quintero,
1981).Wetestedalllayersforpairwisecorrelationacrossthestudy
area using t he package ‘Raster’ in R 3.3 .2 (Hijmans & van Ette n,
2012).Weretained12ofthe19BioClimlayersthathadcorrelation
coefficients under |0.75|. Theseclimate variables represent annual
andseasonaltrends,aswellasextremesintemperatureandprecip-
itation.Temperaturevariables included annual mean temperature,
mean diurnalrange, isothermality,and mean temperaturesofboth
the wettest and driest quar ters. Precipitation variables included
annual temperature, precipitation during the wettest and driest
months, precipitationseasonality,andprecipitation ofthewarmest
andcoldestquarters.
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 etal., 2006),wellbeyond the ideal minimumsample size
toobt ainreliable results(Proosdijetal.,2016).Each model was as-
sessed with the area under the receiver operating characteristic
curve (AUC;Hanley &McNeil,1982).AUCvaluesrepresent amea-
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.Wecompared predicted models against distribution literature
for P. marginemaculatus (Quintero, 1981). Models that performed
poorly (AUCscores<0.75)orthatvaried substantiallyfromhistor-
ical recordswere discarded. Jackknife tests were usedtoevaluate
FIGURE2 MapofSouthernFlorida
withgeographicregions(colors),ocean
currents(bluearrows;adaptedfromLee
&Smith,2002),localitieswherePhrynus
marginemaculatussampleswerecollected
(opencircles),andpairwiseGST(thicker
linesindicatelowerGSTt;Gst range:
0.174–0.620)indicatingmigration
Ma
i
nlan
d
U
pper
ke
y
s
L
L
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 ofP. 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).Wecalculated declinesin suitablehabitat atfourthresholds
ofmodeledsuitablehabitat (>0.1, 0.1, 0.3,0.5, and 0.9)caused by
humandevelopment.
3 | RESULTS
3.1 | Population genetics
Wesequenceda1, 238bp region ofthemitochondrialCOIgeneof
103individuals.Sequenceshadanoverallbasecompositionof24.6%
adenosine,25.2%cytosine,16.5%guanine,and33.7%thymine.Key
Largoexhibitedthelowestgeneticdiversityamongregions,withthe
LowerKeysandmainlandregionsexhibitingconsiderablyhigherge-
netic dive rsity ( Table1).A dditional ly,a r ange-wide mante l test for
IBDofgeographiccoordinatesandNei’sdistancewasnonsignificant
(Mantelr =−0.07,P = 0.637).AhierarchicalAMOVArevealedpopu-
lation st ructure a mong popula tions, but not r egions (Table2). The
AMOVAindicated 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%conservedvariancewith20principalcomponents.Allregions
separatedintodistinctclusters,furtherevidencingregionalgenetic
structure (Figure3). Furthermore, pairwise GST showed relatively
high divergence of KeyLargo localitiesfrom the rest ofthe range,
and low divergence between mainland and island sites, indicating
ongoinggeneflow(Figure2).
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;Figure4a–c).Habitatsuitabilitywas highestin the FloridaKeys
(Figure4), 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%permutationimportance)andmeandiurnal 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
variablesinisolationfromallothersrevealedthatannualmeantem-
peraturehadthehighesttraininggainformodels,followedbymean
temperatureofthedriestquarter,meandiurnaltemperaturerange,
andmeantemperatureofwarmestquarter.
Models including land use categoriesperformed slightly worse
(mean AUC=0.873±0.06, n=100 models) than those without
landusebutappeartohaverefinedthehabitatsuitabilityofP. mar-
ginemaculatus (Figure4d–f). L and use categories had the highest
permutationimportance(50.5%)followedbymeantemperatureof
the drie st quarte r (18%),m ean diurnal t emperatu re range (12.7%),
TABLE2 Hierarchicalanalysisofmolecularvariance(AMOVA)
usingcytochromecoxidaseI(COI)sequencesofPhrynus
marginemaculatusinFloridaindicatinggeneticstructure
σ% variance ϕp
Withinlocalities 1.377 38.37 0.616 <0.001
Betweenlocalities 1.9 97 55.65 0.592 <0.001
Betweenregions 0. 214 5.97 0.060 0 .160
FIGURE3 Discriminantanalysisof
principalcomponents(DAPC)ofPhrynus
marginemaculatuspopulationregions
PCAandDAEigenvaluesispresented
asinsets.Dashedlineistheminimum
spanningtreeofregions.Stratifiedcross-
validationofDAPCresultedinamean
successfulassignmentof0.75%and93.9%
conservedvariancewith10principal
componentsandthreediscriminant
functions.Allregionsseparatedinto
distinctclusters
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%),andaltitude (3%). Similarto our modelswithout land
use catego ries, annu al mean tempe rature had t he highest t raining
gainformodelsbutvariedinsubsequentvariableimportance.Mean
temperature during the driest quarter, mean diurnal temperature
range, an d mean temper ature of the war mest quar ter followed in
ter msofvaria bleim port anceinisolatio n.Mod elswithlanduseiden-
tified regions of the Lower andUpper Keys, pockets in Everglades
National Park,andseveral coastal areas ofMiami asthe most suit-
ablehabitat.Additionalsuitableareas wereidentifiedon KeyLargo
and in Big CypressNational Preserve.Smaller pockets ofpotential
habitatrangeduptheeastandwestcoast s.
We saw an alarm ing 22%–34% human-ind uced declin e in suit-
able habitat under the best fit model (Figure5; Tables3 and 4).
Modelsthatincludedlanduseshowedsteeperdeclinesof29%–48%
ofsuitablehabitatduetohumandevelopment(Tables3and4).This
indicates that, while climatically identified habitat shows consid-
erable decline, humanimpac ts particularly target landuse habitat
typesimportantforthespecies.
FIGURE4 MaxEntsuitabilitymapforP. marginemaculatusinsouthernFlorida.Colorscaleindicatesprobabilityofoccurrencebasedon
presence-onlydata.Minimum,mean,andmaximumsuitabilitiesusingonlyclimatedatasets(a–c;meanAUC=0.92±0.02);minimum,mean,
andmaximumsuitabilitiesusingclimateandvegetationcommunitiesdatasets(d–f;meanAUC=0.87±0.6)
(a)
(f)(e)(d)
(c)(b)
8 
|
   CHAPIN et Al.
4 | DISCUSSION
Pine rocklands succeedtotropicalhardwoodhammocksafter two
or three decades of fire suppression (Loope & Dunevitz, 1981;
Robertson, 1981), but it remains unclearif P. marginemaculatus oc-
cursinbothhabitats.Bothperiodicseawaterfloodingandfireoccur
naturally in rockland habitats (Snyder etal., 1990) and P. margin-
emaculatus,likemuchofthefaunaandfloraspeciesinthesehabitats
(Hofstet ter,1975;Robertson,1953,1962),have evolvedto survive
thesestochasticevents.
The mainland and Lower Keys show the greatest genetic di-
versity.This is likely because these tworegionsincludethelargest
expansesofarea,andbotharerepresentedbysmallpatchesofsuit-
able habitatwithin an inhospitablematrix. Key Largoshowed rela-
tively low diversity,despite the island’ssize.However,onlya small
portionof Key Largo’slandareaissuitable habitat—mostoftheis-
landishighlydeveloped.
Genetic variation was significant between localities but not
larger regions (Table2), suggesting a more complicated genetic
st ru cturethanouraprioriregionaldeline at io ns .Pop ul at io n- leve lg e-
neticstruc tureisalignedwithintuition,consideringallpopulationsin
ourstudyoccuronKeysorislandsofhabitatsurrounded byhuman
disturbance (Figure2). Regional structure, however, is somewhat
surprising.Weselectedregionsa prioribasedongeologicfeatures:
theKey Largo regionforitssizeand proximityto themainland;the
Upper an d Lower Keys for the ir geologic va riation in fo rmation, if
nottimingthereof;andthemainland,asanobviousdelineatorfrom
island populations.This generally alignswith the population struc-
ture of nati ve Cerionland snails, which showed isolation between
the Uppe r and Lower Keys (Shre stha etal., 2015 ). Species on th e
LowerKeyslikelyhaveauniqueevolutionar yhistoryseparatefrom
therestoftheKeys.
OurDAPCanalysisshowscleargeneticstructureacrossregions,
withkeypopulationsallbeingclosestrelatedtomainland localities.
This pat tern suggest s that a mainland–met apopulation m odel de-
scribedthelandscapegenetics oftheKeys.Asmentioned,genetic
analysesoflandsnailsintheFloridaKeysalsoshowanUpper–Lower
Keydivision(Shresthaetal.,2015). But researchon marinespecies
like bicolor damselfish (Eupomacentrus partitus) and common reef
sponge(Callyspongia vaginalis)foundmuchlowerlevelsofdivergence
betweenregions than we found for P. marginemaculatus(DeBiasse,
Richards,& Shivji,2010; Lacsonetal.,1989).Thisis likely because
P. marginemaculatus is much more disper sal limited tha n a marine
fish.Interestingly,theleastdivergentlocalitypairsofthesesmarine
speciesincludedKeyLargo,whichisthesamepatternwefindinour
study.Thismatching patternsupport stheideathatcurrents playa
majorroleinP. marginemaculatusgeneticstructure.
PairwiseGSTshowedthat thegreatest divergencewasbetween
Key Largo an d the other region s (Figure2). This coul d be caused
bymigrants from the Bahamas, the closestislands ofwhichareca.
100km from Key Largo.Surprisingly,the mainlandhad low Gstes-
timates with the Upper and LowerKeys. We posit that this is due
toongoinggeneflow betweentheseregions. Ocean currentslikely
push raftingP. marginemaculatustotheLowerKeysasmightmajor
weather events including hurricanes (Fleming & Murray, 2009).
Experim entswit hGPS-equi pp edbuoyss howthatt hisisamajorcur-
rentpathway(Lee&Smith,2002)andP. marginemaculatus,withthe
abilitytobreatheunderwater,areaptlysuitedtosurvivethevoyage
(Hebets&Chapman,2000).Otherresearchhasemphasizedtheim-
portanceofoceancurrentsintheFloridaKeysingeneticstructure,
butthishasbeenlimitedtomarinespecies.Forexample,astudyof
threemarineinver tebratesshowedhighgeneflowandconnectivit y
acrosstheKeys,withapatternofsouthernmigration(Richardsetal.,
2007).Thesaltmarshsnake(Nerodia clarkii),whichissomewhatre-
strictedtoshallowwaters,showedgeneticstructurebetweenUpper
FIGURE5 (a)MaxEntsuitabilitymapofthebestfitmodel
(climate-onlymean)withhumandevelopmentoverlayforthe
amblypygidPhrynus marginemaculatusintheFloridaKeysandSouth
Florida.Whiteareasindicatelandconvertedforhumanuse.Panels
b–dare10×magnificationofareasindicatedinpanela,which
include(b)KeyLargo,(c)theotherUpperKeys,and(d)theLower
Keys.Arrowsindicatenorth.TheKeysincludeparticularlyhigh
suitabilityhabitat,especiallytheLowerKeys.46
0.00 0.92 100 km
10 km
(d)
(a)
N
(c)
N
(b)
N
    
|
 9
CHAPIN e t Al.
andLowerKeysthatwasexplainedby IBD (Jansenetal., 2007).In
general,ourSDMpredictedhabitatsuitabilityforP. marginemacula-
tusinpartsofMonroe,Miami-Dade,Collier,Lee,andHendrycoun-
ties. This limited range generally matches museum records, with
occasionalrecords fromcountiesasfar NorthasCharlot te county
(Quintero,1981).Thesenor therlyreco rdscouldbef romsparsep op-
ulations,wheredetectingthespeciesisdifficult,orcollectionscould
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 tionso fhab it at su it abilitywith inth eurb an ar ea sofsouth ea st-
ernFloridamainlandshouldbecautiouslyinterpretedasrealizedvia-
blehabitatismuchlessbecauseofhumandevelopment .
Few studies of Florida biogeography have been conducted,
but generally align withourresults.Ant gut microbiota show simi-
largenetic structurebetweentheUpper andLowerKeys,but also
shoe divergenceamongthe Lower Keys, which might be indicative
of Carib bean migration (Hu e tal., 2014). The mosquito A. aegypti
showed practically no genetic stru cture among the Florida Keys,
likelybecausetheydisperse viaflight(Brownetal.,2013).Shrestha
etal. (2015) proposed that the C. incanum spread southwesterly
tocolonize new Keys as theyformed, withLower Key populations
beingtheyo ungest.L astly,antg utmicr obi otashowedgeneticst ruc-
turebetweentheUpperandLowerKeys(Huetal.,2013).
Ourmodelsidentifiedareasoftropicalhardwoodhammocksand
pinerocklandsasthemostsuitablehabitattypesforP. marginemac-
ulatus.Indeed,thisiswherealmostallofourobservationsoccurred
andiscorroboratedbypublished collection sites of the speciesfor
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 ytargetsforhuman development inFlorida (Noss,LaRoe,&
Scott , 1995;Snyde r etal., 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
threattothehabitatandspecies.
Pinerockland forests,oncecommon throughout southeastern
Florida,arenowoneofthe mostthreatenedhabitats globally,with
atleast98%globalloss(Nossetal.,1995).Thisincludesca.8,0 00ha
in Everglades National Park and a mere 920ha outside the park’s
bounda ry (Bradl ey,20 05). Pine roc kland habit at was identif ied as
oneofthemostsuitablehabitatsbasednotonlyonourmodelsthat
includedlandusecategoriesbutalsothose with onlyclimatevari-
ables and geolog y.Pine rockland habitats occur on exposed lime-
stonesubstrates where limestone rock outcroppings are common
andprovideimportantmicrohabitatforP. marginemaculatus(Chapin
&Hebets, 2016;Hebets& Chapman,2000).Humandevelopment,
fire supp ression, and climate change have altered or ent irely re-
movedmanyareasoncedominatedbythiscommunity(Kautz&Cox,
2001;Possley,Woodmansee,&Maschinski,2008;Rossetal.,1994).
This habitatlosshasresulted in fivefederallylistedanimal and 21
rare, endemicplantspeciessympatricwithP. marginemaculatus, all
ofwhicharedependent on remaining fragments of pine rockland
habitat (FloridaNaturalAreasInventory,2010).Furthermore,lime-
stonerocksthatmakeupcriticalhabitatforP. longipesareoftencol-
lectedandusedforconstructionandlandscaping.
Some are as of pine rockland a nd upland hardwo od forest are
protectedtoday.Theseinclude patcheswithin EvergladesNational
Park, Big Cypress National Preserve, Key Deer National Wildlife
Refuge,andst ate-m an agedlands .M os th abita to ut sid eofthe se pr o-
tectedareashavealreadybeendestroyed,andmanysur vivingfrag-
mentsremainthreatenedandatriskofextirpationwithintheareas
ofMiami,surroundingsuburbanareas,andinthetourist-dominated
Florida Keys.Thishasdramaticimplications forP. marginemaculatus
andtheendemic,endangeredcommunit yinwhichtheyoccur.
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
TABLE3 AreaofsuitablehabitatoftheamblypygidPhrynus marginemaculatusinsouthernFloridaandtheFloridaKeys.Areaofhabitat
withsuitabilitythresholdsof0.1,0.5,and0.9underaclimate-onlyandclimate-pluslanduseMaxEntmodels.Lossindicatespercentlossof
habitatatagiventhresholdbyhumandevelopment.ModelfitisindicatedbytheAreaUndertheCur veandst andarddeviation(AUC±SD)
Model
0.1 threshold 0.5 threshold 0.9 threshold
AUC ± SDkm2% loss km2% loss km2% loss
Climateonly 3,545.88 27. 5 8 179.95 3 4.91 4.64 22.44 0.978±0.02
Climate+landuse 2 , 42 7.96 48.68 3 00 .97 43. 81 12 .74 28.65 0.873±0.6 0
TABLE4 Reductioninsuitablehabitatfortheamblypygid
Phrynus marginemaculatuspredictedbyMaxEntmodelingcausedby
humandevelopmentinSouthFloridaandtheFloridaKeys.
Thresholdisthelowerlimitfortheindexofhabitatsuitability;
Habitatisthetotalareainkm2withouthumandevelopment
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+Landuse(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
aDevelopmentisthereducedareaafterconsideringhabitatdegradedby
humanuse.
10 
|
   CHAPIN et Al.
(Maschinskietal.,2011;Rossetal., 2009).Inthissense,increases in
thefrequencyandintensit yof hurricane stormsurges reshapepine
rockland and similar vegetatio n communities (Ross etal., 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-
ginemaculatuspopulations.Wecan,however,gleaninsightfrom
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-
ingto ca.90%mortalityofmaturepinetrees(Maguire,1995)which
negative ly impacted p lant and animal co mmunities (L loyd & Slater,
2012; Orr & Ogden, 1992; Williams, Wang, Borchetta, & Gaines,
2007). Majorstorms and humandisturbance not only alterhabitats
butcanalsoleadto isolationof populationsandreshape population
geneticsofspeciesatrisk(e.g.,Villanova,Hughes,&Hof fman,2017).
Ourresearchwason specimenscollectedin2015, priortothe 2016
Hu rri can eM att h ew and20 17Hu rr i can eI rma ev ent s.Fu t ur ere se a rc h
willbenefitfromexaminingtheimpactsoftheseandothers tormson
populationstructureofP. marginemaculatusinsouthernFlorida.
Both our genetic and ecologicalresultsare limited byour data-
set, whichis constrained in both time andspace. Spatially, we only
sampledP. marginemaculatusinso uther nFl oridaandtheFloridaKeys
archi pe la go,butt hespeciesoccu rs asfarso ut hwes tasHi sp aniola,in -
clud in gp op ulation si nt he Ba ha mas ,Cub a, Ja maica ,a nd th eTu rksan d
Caicos (Quintero, 1981). In par ticular,gene flow fromthe Bahamas
and Cuba to Key L argo and th e Lower Keys could be o ngoing, but
thisremainsunexamined.Temporally,weexcludedhistoricalsamples
withtheconcernthattheywouldnotinformmodernbiogeographic
patterns. While historicalmaterial wouldincreasesample sizes, our
moderncollections spanthespatialrangeof Floridahistorical sam-
ples bar northern limits, where populations may no longer occur.
Sampl ingwa sunev enacr osssi tes,w hi chcouldbiasre su lt s.Lastly,w e
usedonlyonesequence inouranalysis. Wechosea mtDNA marker
because nDNA appears highly conser ved in Amblypygi (Esposito
etal.,2015),andmtDNAisthus,moreinformative.Cautionshouldbe
taken,however,ininterpretingresultsfromonlyonemtDNAmarker,
andfutureresearchusingmorethoroughgenomicsequencingcould
revealmorehigh-resolutionbiogeographicpatterns.
4.1 | Conservation
Our resul ts point to a prima ry threat to the p opulation hea lth of
P. marginemaculatusinthewild:habitatlossbyhumandevelopment.
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, thisaccountsfor an average density of 205.54peopleper
squarekilometerand leavesapproximately150km2uninhabitedby
humans, muchof which isseasonallyorpermanentlyflooded habi-
tat unsui table for ma ny terrest rial specie s. Furthe rmore, th is does
notinclude the impactfromcommercial,transportation,andutili-
ties developments. Fortunately, muchof the remaining habitat for
P. marginemaculatus is within protect ed areas managed by f ederal
andstateagencies.This,however,doesnotprotecthabit atsfromall
threats,includingtheimpact sofincreasedhurricanes,sea-levelrise,
poaching,andmicrohabitatalterations.
Secondarily,P. marginemaculatus is collec ted from the wild f or
saleinthepettrade. While we donothavedataon the numberof
individualscollectedforsaleinthepetindustr y,personal obser va-
tionsleadustobelievethatitmustbeinthehundredstothousands.
WildpopulationsofP. marginemaculatuswouldbenefitfromcaptive
breedin g that allows thes e fascinating an imals to be kept as pet s
without reducingnumbersin the wild.Moreresearch on wild pop-
ulations is n eeded to asse ss populati on health, a nd we encoura ge
researchers to conduct both fieldand laboratorystudieson these
fascinatingorganisms.
ACKNOWLEDGMENTS
Major funding was provided by theTheodoreRoosevelt Memorial
FundoftheAmericanMuseumofNaturalHistory.Fundingwasalso
provided b y the Edwin W. Pauley Fello wship, UCLA , Departm ent
ofEcology & EvolutionaryBiology,UCLA ,Department ofEcology
& 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
databasingliteraturesearches.Thisresearchwasconducted under
EvergladesNationalPark Scientific ResearchandCollectingPermit
EVER–2014–SCI–0057,Nation alW ildlife RefugeSystem Rese arch
andMonitoringSpecial Use PermitFFO4RFKD–2015–001,Florida
Park Ser vice Scientific Collecting Permit 08051410, and Miami-
DadeCountyParks&RecreationResearchPermit241.
CONFLICT OF INTEREST
Nonedeclared.
AUTHOR CONTRIBUTIONS
K.J.C .conceived thestudy; K.J.C and D.E.W collected and extracted
sa mp le s;I.A. an dP.W.se qu en ce d,clean ed ,a ndalignedDN A; K.J .C .a n-
aly zedgenet icdat a;D.E.W.analy zedecologicaldata;K. J.C.a ndD.E .W.
draftedthemanuscript;Allauthorscontributedtowritingandediting.
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). Imageprocessing
withImageJ.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-
creasesrobustnesstosamplingbiasinmodelsofspeciesdistributions:
AnimplementationwithMaxent.Ecological Modelling,222,27 96 –2811 .
Apodaca,J.J., Trexler,J. C., Jue, N.K., Schrader,M., &Travis,J. (2013).
Large-scale natural disturbancealters geneticpopulationstructure
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 tationandits
effec tsonforest-nestingbirdsintheFloridaKeys.Conservation Biology,
9,835–844.https://doi.org/10.1046/j.1523-1739.1995.09040 835.x
Bradie,J.,&Leung,B.(2016).Aquantitativesynthesisoftheimportance
ofvariables 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 ationan do rg an izati onofnatura lf or es tcom-
munitiesofMiami-DadeCounty.DatabasesubmittedtoMiami-Dade
County Department of Environmental Resources Management.
Miami,FL:TheInstituteforRegionalConservation.
Bradley,B.A.(2016).Predictingabundancewithpresence-onlymodels.
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 :
UniversityofFlorida.
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 ofAmblypygi.
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-
mentof connectivit yin thecommonreefsponge,C allyspongiavag-
inalis(Demospongiae:Haplosclerida)revealshighpopulationstruc-
turealongtheFloridareeftract.Coral Reefs,27,47–55.
DeBiasse, M. B., Richards,V.P.,&Shivji,M. S. (2010). Genetic assess-
ment ofconnectivit yinthecommon reef sponge, Callyspongia vag-
inalis(Demospongiae:Haplosclerida)revealshighpopulation 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) .Thea de 4p ackage:Imp le me ntingth edu -
alitydiagramforecologis 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).Novelmet hods improve prediction of
species’distributionsfromoccurrencedata.Ecography,29,129–151.
https://doi.org/10.1111/j.2006.0906-7590.04596.x
Elith,J.,&Leathwick,J.R.(2009).Speciesdistributionmodels: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:Diversificationoftaillesswhipspiders(Amblypygi,Phrynus)in
Caribbeancaves.Molecular Phylogenetics and Evolution,93,107–117.
https://doi.org/10.1016/j.ympev.2015.07.005
Ewing,B.,&Green,P.(1998).Basecallingofautomatedsequencertraces
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).Analysisofm olecularv ari-
anceinferredfrommetricdistancesamongDNAhaplotypes:Application
tohumanmitochondrialDNAre stric tiondat 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)intheUSA.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 ofthe potentialglobal distributionofa problematicalieninva-
sivespecies—TheAmericanbullfrog.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).Populationandgeneticconsequences
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
FloridaNatural Areas Inventory (2010). Pine rockland. In (Ed.), Guide to
the natural communities of Florida (2010 Editio n, pp. 1–8). Flori da:
FloridaNaturalA reasInventory.
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).Theimpac tsofsprawlonbiodiversity:
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 ed ispe rs al :Afr am ewor ks forhy -
pothesistesting.Trends in Ecolog y & Evolution,27,47–56.
Gotelli,N.J.,&Sta nto n-G ed des ,J.(2015).Climatec han ge,geneticmar k-
ersandspecies distribution modelling. Journal of Biogeography, 42,
1577–1585.https://doi.org/10.1111/jbi.12562
Graving,J.M.(2015).Nocturnal homing in the amblypygidPhrynusmar-
ginemaculatus. T hesis, Bowling Green State University, Bowling
Green,Ohio,p.44.
Green,P.(1999).PHRAP.Retrievedfromhttp://phrap.org/
Green,P.,&Ewing,B.(2002).PHRED.Retrievedfromht tp://phrap.org/
Guisan,A.,&Thuiller,W.(2005).Predictingspeciesdistribution: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).Metapopulationdynamics.Nature,396,41–49.https://
doi.org /10.103 8/23876
Hanley, J. A., & McNeil, B. J. (1982). The meaning and use of thearea
underareceiveroperatingcharacteristic(ROC)curve.Radiology,143,
29–36 .
Hebet s, E. A., & Cha pman, F. R. (200 0). Survi ving the floo d: Plastron
respirationinthenon-tracheatearthropodPhrynus 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
FloridafrogstracedtoCuba.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 geneticvaria-
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).Correlatesof
gutcommunitycompositionacrossanantspecies(Crphalotesvari-
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).Mioceneriversandtaxon
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).Theeffectoffireonthepinelandandsawgrass
communitiesofSouthernFlorida.InP.J.Gleason(Ed.),Environements
of Southern Florida: Present and past(pp.201–212).Miami,FL:Miami
GeologicalSociety.
Jansen,K.P.,Mushinsky,H.R.,&Karl,S.A. (2007).Populationgenetics
ofthemangrovesaltmarshsnake,Nerodiaclarkiicompressicauda,in
alinear,fragmentedhabitat.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 ytechnique 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 genet1.3-1:Newtoo lsfortheanaly-
sisofgenome-wideSNPdata.Bioinformatics,27,307 0–3071 .h t tps ://
doi.org/10.1093/bioinformatics/btr521
Kamvar,Z.N.,Brooks,J.C.,&Grünwald,N .J. (2015).NovelRtoolsfor
analysisof genome-wide population geneticdatawith emphasison
clonality.Frontiers in Genetics,6,208.
Kamvar,Z.N.,Tabima,J.F.,&Grünwald,N.J.(2014).Poppr:AnRpackage
forgeneticanalysis ofpopulationswith clonal,partiallyclonal,and/or
sexualreproduction.Pee rJ,2,e281.https://doi.org/10.7717/peerj.281
Kautz,R.S., &Cox,J.A.(20 01).Strategic habit atsforbiodiversity 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,
FloridaFishandWildlifeConservationCommission.
KeyWestChamberofCommerce(2017).KeyWestandMonroeCount 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).PopulationstructureofSymbiodiniumsp.associatedwiththe
commonseafan,Gorgonia ventalina,in theFlorida Keys acrossdis-
tance,depth,andtime.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).GeneticvariationincreasesduringbiologicalinvasionbyaCuban
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).
Geneticdifferentiationofbicolordamselfish(Eupomacentrusparti-
tus)populationsintheFloridaKeys.Marine Biology,103,4 45–451.
Lacson,J.M.,&Morizot,D.C .(1991).Temporalgeneticvariationinsub-
populationsofbicolordamselfish(Stegastes partitus)inhabitingcoral
reefs in the FloridaKeys. 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).Volumetranspor tvariabilitythroughthe
Florida Keys tidal channels. Continental Shelf Research, 22, 1 361–
1377.https://doi.org/10.1016/S0278-4343(02)00003-1
Lenczewski,B.(1980).ButterfliesofEvergladesNationalPark.Everglades
NationalPark,SouthFloridaResearchCenterReportNo.T-588.
Li,Y.M.,Dlugosch,K .M.,& Enquis t,B.J. (2015). Novel spatialanalysis
methods reveal scale-dependent spread and inferlimiting factors
ofinvasion byS 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-rocklandbirdassemblages.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.ParkSouthFloridaRs.Cent.Rep.T-645
MacArthur,R.H.,&Wilson,E.O.(1967).The theory of isla nd biogeography
(Vo l. 1).Princeton,NJ:PrincetonUniversityPress.
Maddison,D.R.,&Maddison,W.P.(2016a).Chromaseq: A Mesquite pack-
age for analyzing sequence chromatograms , version 1.2.Retrievedfrom
https://www.mesquiteproject.org/
Maddis on, W. P., & Maddis on, D. R. (2016b). Mesquite: A modular sys-
tem for evolutionary analysis, version 3.10.Retrievedfromhttp://mes-
quiteproject.org
Maguire,J.(1995).Restoration pl an for Dade County ’s pine Rockland for ests
following Hurricane Andrew.Miami,FL:DadeCountyDepar tment of
EnvironmentalResourcesManagement.
Mantel, N.(1967).The detec tionof diseaseclusteringandageneralized
regressionapproach.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 levelrise. 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 sandecologicaldegradationintheFloridaKeys.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 yofCentralFloridaPress.
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,WhipScorpionsand Wind Scorpionsof
Florida.InArthropods of Florida and Neighbouring land areas(Volume
4,pp.1–28).FL:FloridaDepartmentofAgriculture.
Muma,M.H.(1967).Scorpions,WhipScorpionsand Wind Scorpionsof
Florida.p.1-28InArthropodsofFloridaandNeighboringLandAreas,
Volume4.FloridaDepartmnetofAgriculture,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: USDepar tment of the Interior,National
BiologicalSer vice.
Orr,D.W., & Ogden ,J.C. (1992).Theimpactof HurricaneAndrew on
theecosystemsofSouthFlorida.Conservation Biology,6,488–490.
Paradis,E.(2010). pegas:An Rpackage forpopulationgeneticswithan
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 inMadagascar.Journal of
Biogeography,34,102–117.
Peck, S.B.,&Howden, H.F.(1985).Biogeographyof scavengingscarab
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).Ecologicalnicheconservatism:Atime-structured
reviewofevidence.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 yd ee rf aw nres po ns etour banizati on :I ssusta in ab le de vel-
opmentpossible?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).Maximumentropy
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).Modelingofspeciesdistributionswith
Maxent:Newextensionsandacomprehensiveevaluation.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
urbanpine rocklandpreserves.Natural Areas Journal,34, 154–165.
https://doi.org/10.3375/043.034.0205
Possley,J.,Woodmansee,S.W.,&Maschinski,J.(2008).Patternsofplant
compositioninfragmentsofgloballyimperiledpinerocklandforest:
Effectsofsoilt ype,recentfirefrequency,andfragmentsize.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 mr e-
quirednumberofspecimenrecordstodevelopaccuratespeciesdis-
tributionmodels. 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,117166.
Rayor,L.S.,&Taylor,L.A.(2006).Socialbehaviorin amblypygids,anda
reassessmentofarachnidsocialpat terns.Journal of Arachnology,34,
399–421.
RCoreTeam(2013).R: A language and environment for statistical comput-
ing.Vienna,Austria:RFoundationforStatisticalComputing.
Richar ds, V.P., Thom as, J. D., Stanhop e, M. J., & Shivj i, M. S. (2006).
Geneticconnectivityin the Florida reefsystem:Comparative phy-
logeographyofcommensalinvertebrateswithcontrastingreproduc-
tivestrategies.Molecular Ecology,16,139–157.
Robertson, W.B.Jr(1953).A survey of the effects of fire in the everglades
national park.Homestead,FL:NationalParkSer vice.
Rober tson, W. B. Jr (1962). Fire an d vegetatio n in the Evergla des. Tall
TimbersFireEcologyConferenceProceedings,1,67–80.
Ross,M.S.,O’Brien,J.J.,&daSilveiraLoboSternberg,L.(1994).Sea-level
riseandthereductioninpineforests intheFloridaKeys. 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 risingtide: The challenge of biodiversityman-
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).Tactilelearningbyawhipspider,
Phrynus marginemaculatusC.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 marginemaculatusC .L. Koch.Journal of Arachnology,37,
109–112.https://doi.org/10.1636/ST07-139.1
Santer,R.D.,&Hebets,E.A.(2011).Thesensotyandbehaviouralbiolog y
ofship spiders(A rachnida,Amblypygi).Advances in insect physiology
41,1– 6 4.
Simberloff,D.S., &Wilson, E. O.(1969).Experimentalzoogeographyof
islands:Thecolonizationofemptyislands.Ecology,50,278–296.
Simberloff,D.,&Wilson,E.O.(1970).Experimentalzoogeography ofis-
lands.at wo-yearrecordofcolonization.Ecology,51,934–937.
Simberloff, D.(1976).E xperimental zoogeographyofislands: effectsof
islandsize.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:
SciencePress.
Small, J.K.(1917).The tree cactioftheFloridaKeys.Journal of the New
York Botanical Garden,18,199–203.
Snyder, J.R., Herndon, A ., &Robertson, W. B. Jr (1990). Sout hFlorida
rocklands.InR.L.Myers,&J.J.Ewel(Eds.),Ecosystems of Florida(pp.
230–277).Orlando,FL:UniversityofCentralFloridaPress.
Spence,A .J.,& Hebets, E. A. (20 06). Anatomy andphysiologyof giant
neuronsintheantenniformlegoftheamblypygidPhrynusmarginac-
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
suitabilityforstonycoralsonseamounts.Journal of Biogeography,36,
1111–1128.https: //doi.o rg /10.1111 /j.1365 -2699.2 00 8.02062 .x
USCensusBureau,U.S.DecennialCensus,UnitedStatesCensusBureau.
Retrievedfromhttp://www.census.gov/prod/www/decennial.html
Václavík,T.,Kupfer,J.A .,& Meentemeyer, R. K. (2012). Accountingfor
multi-sc alespatialautocorrelationimprovesper formanceofinvasive
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-
neticstructureanddemographicanalysestoestimatepersistencein
endangered Key deer (Odocoileus virginianus clavium).Conservation
Genetics,5,1–16.
Waltari,E.,&Gura lnick ,R .P.(2 00 9).Ecolog ic aln ichem odell ingofm ontane
mammalsintheGre atBasin ,NorthAmerica:Exami ningpas tandpresent
connectivityofspeciesacrossbasinsandranges.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).Islandsasmodelsystemsinecology
andevolution:ProspectsfiftyyearsafterMacA rthur-Wilson.Ecology
Letter s,18,20 0–217.https://doi.org/10.1111/ele.12398
Webb,D. S.(1990).Historicalbiogeography.InR .L.Myer s,& J.J.Ewel
(Eds.), Ecosystems of Florida(pp.70–100).Orlando,FL:Universit yof
CentralFloridaPress.
Weygoldt, P.(2000).Whip spiders (Chelicerata: Amblypygi): Their biology,
morphology and systematics.Stenstrup,Denmark:ApolloBooks.
Williams,D.A.,Wang,Y.,Borchetta ,M.,&Gaines,M .S.(20 07).Genetic
diversity and spatialstruc ture of a keystonespecies in fragmented
pinerocklandhabitat.Biological Conservation,138,256–268.https://
doi.org/10.1016/j.biocon.2007.04.024
Wilson,L .D.,& Porras, L.(1983). The ecologicalimpact 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 thecalculation of popula-
tiondifferentiationstatistics.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 tsof samplesizeontheperformance ofspeciesdis-
tributionmodels.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:ChapinKJ,WinklerDE,WiencekP,
AgnarssonI.Islandbiogeographyandecologicalmodelingof
theamblypygidPhrynus marginemaculatusintheFloridaKeys
archipelago.Ecol Evol. 2018;00:1–13. h t tp s : //d o i .
org /10.1002/ece3.433 3
... Whip spiders (Amblypygi Thorell, 1883), an order of tropical arachnids related to spiders (Weygoldt, 2000), which reach their greatest diversity in the Mexican Neotropics (Armas, 2006;Harvey, 2003), offer a model system to address the drivers of tropical diversification. Unlike spiders, most of which are capable of dispersal on air currents (Weyman, 1993), whip spiders are limited to walking and possibly rafting (Chapin et al., 2018). Their vagility is further constrained by an ecological requirement for humidity, which restricts them to forests, caves or the vicinity of permanent water sources in arid areas (Weygoldt, 2000). ...
... Their vagility is further constrained by an ecological requirement for humidity, which restricts them to forests, caves or the vicinity of permanent water sources in arid areas (Weygoldt, 2000). Although phylogeographical studies of whip spiders remain scarce, the few published to date confirm that their diversity is underestimated and informative for understanding diversification in the tropics (Chapin et al., 2018;Esposito et al., 2015;Prendini et al., 2005;Reveillion et al., 2020). ...
... The ingroup comprised 26 specimens of Acanthophrynus from 16 localities (a proxy for populations), covering most of the distribution ( (Weygoldt, 2000). cytochrome c oxidase subunit I (COI), were selected for efficacy in reconstructing the phylogeny of whip spiders (Chapin et al., 2018;Esposito et al., 2015;Harrison et al., 2017;Prendini et al., 2005). Loci were amplified using universal primers and/or, in the case of COI, 28S ...
Article
The tropics contain many of the most biodiverse regions on Earth but the processes responsible for generating this diversity remain poorly understood. This study investigated the drivers of diversification in arthropods with stenotopic ecological requirements and limited dispersal capability using as model the monotypic whip spider (Amblypygi) genus Acanthophrynus, widespread in the tropical deciduous forests of Mexico. We hypothesized that for these organisms, the tropical deciduous forests serve as a conduit for dispersal, with their disappearance imposing barriers. Given that these forests are located in a region of complex geological history and fluctuated in extent during the Pliocene-Pleistocene glacial/interglacial cycles we couple molecular clock dating, paleoclimatic niche modeling and ancestral area reconstruction to test if and how habitat fragmentation promoted diversification in Acanthophrynus. Concomitant with the expected role of landscape change, we demonstrate that orogeny of the Trans-Mexican Volcanic Belt, in the Late Miocene/Early Pliocene (6.95-5.21 mya), drove the earliest divergence of Acanthophrynus through vicariance. Similarly, as expected, the later onset of glaciations strongly impacted diversification. Whereas a more stable climate in the southern part of the distribution enabled further diversification, a marked loss of suitable habitat during the glaciations only allowed dispersal and diversification in the north to occur later, resulting in a lower overall diversity in this region. Importantly, barriers and diversification patterns identified in Acanthophrynus are reflected in the phylogeography of codistributed vertebrates and arthropods, emphasizing the profound impact of Trans-Mexican Volcanic Belt orogeny and glacial/interglacial cycles as drivers of diversification in the Mexican Neotropics.
... This can be exemplified in the hardwood hammock habitats of P. marginemaculatus. Odor plumes present within the Florida Keys and the Everglades environments, which experience occasional strong winds (Chapin et al., 2018;Greller, 1980), should be more continuous with lower peak concentrations because of lower overall turbulence levels (Justus et al., 2002;Moore and Crimaldi, 2004;Murlis, 1997). Conversely, more complex rainforests, such as the habitats of P. laevifrons in Costa Rica, will have reduced wind velocity and a greater number of elements producing flow eddies resulting in odor stimuli that are more heterogeneous (Chapin and Hebets, 2016;Chazdon et al., 2010;Lawton and Putz, 1988;Moore and Crimaldi, 2004;Wiegmann et al., 2016;Wilson et al., 1985). ...
Article
Full-text available
Recent improvements in genetic analyses have paved the way in using molecular data to answer questions regarding evolutionary history, genetic structure, and demography. Key deer are a federally endangered subspecies assumed to be genetically unique, homogeneous, and have a female-biased population of approximately 900 deer. We used 985 bp of the mitochondrial cytochrome b gene and 12 microsatellite loci to test two hypotheses: (1) that Key deer are isolated and have reduced diversity compared to mainland deer and (2) that isolation of the Florida Keys has led to a small population size and a high risk of extinction. Our results indicate that Key deer are indeed genetically isolated from mainland white-tailed deer and that there is a lack of genetic substructure between islands. While Key deer exhibit reduced levels of genetic diversity compared to their mainland counterparts, they contain enough diversity to uniquely identify individual deer. Based on genetic identification, we estimated a census size of around 1000 individuals with a heavily skewed female-biased adult sex ratio. Furthermore, our genetic and contemporary demographic data were used to generate a species persistence model of the Key deer. Sensitivity tests within the population viability analysis brought to light the importance of fetal sex ratio and female survival as the primary factors at risk of driving the subspecies to extinction. This study serves as a prime example of how persistence models can be used to evaluate population viability in natural populations of endangered organisms.
Article
Full-text available
Arachnologists have uncovered tantalizing details about amblypygid behavioral ecology—the study of the fitness consequences of their behavior. Thus, it is the aim of this review to position Amblypygi as a useful system in which to investigate the principles of animal behavioral ecology. We synthesize amblypygid habitat preference and navigation modalities; predator, prey, parasite, parasitoid, cannibal, and commensal interactions; resource contests and territoriality; mating systems and mate choice; parental investment and sociality; and genetics and genomics as they relate to behavioral ecology. We present ideas for future research in each of these areas and discuss future directions for Amblypygi behavioral ecology research as they relate to four areas of behavioral ecology: adaptation, evolutionary history, mechano-sensory control of behavior, and behavioral development. We conclude by identifying several avenues of Amblypygi behavioral ecology that we think have the highest potential for transformative discoveries.
Article
Aim To synthesize the species distribution modelling (SDM) literature to inform which variables have been used in MaxEnt models for different taxa and to quantify how frequently they have been important for species’ distributions. Location Global. Methods We conducted a quantitative synthesis analysing the contribution of over 400 distinct environmental variables to 2040 MaxEnt SDMs for nearly 1900 species representing over 300 families. Environmental variables were grouped into 24 related factors and results were analysed by examining the frequency with which variables were found to be most important, the mean contribution of each variable (at various taxonomic levels), and using TrueSkill™, a Bayesian skill rating system. Results Precipitation, temperature, bathymetry, distance to water and habitat patch characteristics were the most important variables overall. Precipitation and temperature were analysed most frequently and one of these variables was often the most important predictor in the model (nearly 80% of models, when tested). Notably, distance to water was the most important variable in the highest proportion of models in which it was tested (42% of 225 models). For terrestrial species, precipitation, temperature and distance to water had the highest overall contributions, whereas for aquatic species, bathymetry, precipitation and temperature were most important. Main conclusions Over all MaxEnt models published, the ability to discriminate occurrence from reference sites was high (average AUC = 0.92). Much of this discriminatory ability was due to temperature and precipitation variables. Further, variability (temperature) and extremes (minimum precipitation) were the most predictive. More generally, the most commonly tested variables were not always the most predictive, with, for instance, ‘distance to water’ infrequently tested, but found to be very important when it was. Thus, the results from this study summarize the MaxEnt SDM literature, and can aid in variable selection by identifying underutilized, but potentially important variables, which could be incorporated in future modelling efforts.
Article
This book had its origin when, about five years ago, an ecologist (MacArthur) and a taxonomist and zoogeographer (Wilson) began a dialogue about common interests in biogeography. The ideas and the language of the two specialties seemed initially so different as to cast doubt on the usefulness of the endeavor. But we had faith in the ultimate unity of population biology, and this book is the result. Now we both call ourselves biogeographers and are unable to see any real distinction between biogeography and ecology.
Article
Shipboard acoustic Doppler current profiler (ADCP) data from six synoptic surveys are combined with current meter, bottom pressure and wind data to describe tidal and low-frequency exchanges through four tidal channels in the Middle Florida Keys. ADCP crossings provide transport rates for Channel 2, Channel 5, Long Key Channel and the Seven Mile Bridge channels. Predictions of tidal current speed and water level provide mid-channel, vertically-integrated transport for times corresponding to total channel transport given by the ADCP data. For Channel 2, Channel 5 and Long Key Channel, flood and ebb tide data are separated, and results from the surveys are pooled. Total channel transport is regressed against mid-channel transport, and the slope of the regression equation serves as a conversion factor to translate time series of mid-channel transport into time series of total channel transport. Tidal transport is estimated for the principal tidal constituents in each channel. M2 total flood and ebb discharges vary from 13.13 x 106 m3 for Channel 2 to 122.60 x 106 m3 for the Seven Mile Bridge channels. Time series of mid-channel current speed and bottom pressure from Long Key Channel are used to investigate the response to wind forcing. The mean mid-channel current speed is an inflow of 2.4 cm s-1, and the mean transport is an inflow of 32 m3 s-1. Both are the result of forcing by strong east-northeasterly winds typical of the fall season.