phylogeny of the aphnaeinae: myrmecophilous african butterflies … · 2015. 12. 1. · ah-95-z951...

Systematic Entomology (2015), 40, 169–182 DOI: 10.1111/syen.12098

Phylogeny of the Aphnaeinae: myrmecophilous Africanbutterflies with carnivorous and herbivorous lifehistories

J O H N H . B O Y L E 1,2, Z O F I A A . K A L I S Z E W S K A 1,2, M A R I A N N EE S P E L A N D 1,2,3, T A M A R A R . S U D E R M A N 1,2, J A K E F L E M I N G 2,4,A L A N H E A T H 5 and N A O M I E . P I E R C E 1,2

1Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, U.S.A., 2Museum of ComparativeZoology, Harvard University, Cambridge, MA, U.S.A., 3Museum of Natural History and Archaeology, Norwegian University ofScience and Technology, Trondheim, Norway, 4Department of Geography, University of Wisconsin, Madison, WI, U.S.A. and 5IzikoSouth African Museum, Cape Town, South Africa

Abstract. The Aphnaeinae (Lepidoptera: Lycaenidae) are a largely African subfamilyof 278 described species that exhibit extraordinary life-history variation. The larvae ofthese butterflies typically form mutualistic associations with ants, and feed on a widevariety of plants, including 23 families in 19 orders. However, at least one species ineach of 9 of the 17 genera is aphytophagous, parasitically feeding on the eggs, brood orregurgitations of ants. This diversity in diet and type of symbiotic association makesthe phylogenetic relations of the Aphnaeinae of particular interest. A phylogenetichypothesis for the Aphnaeinae was inferred from 4.4 kb covering the mitochondrialmarker COI and five nuclear markers (wg, H3, CAD, GAPDH and EF1𝛼) for each of 79ingroup taxa representing 15 of the 17 currently recognized genera, as well as threeoutgroup taxa. Maximum Parsimony, Maximum Likelihood and Bayesian Inferenceanalyses all support Heath’s systematic revision of the clade based on morphologicalcharacters. Ancestral range inference suggests an African origin for the subfamily witha single dispersal into Asia. The common ancestor of the aphnaeines likely associatedwith myrmicine ants in the genus Crematogaster and plants of the order Fabales.

Introduction

The subfamily Aphnaeinae (Lepidoptera: Lycaenidae) consistsof 278 species of butterflies with an unusual diversity of lifehistories. Several of the 17 genera are endemic to southernAfrica, but many of the genera are distributed throughoutAfrica. The majority (88%) of the described species are foundin Africa and/or Arabia, with the exception of 33 of the 71species of Cigaritis Donzel, which are found in Asia as fareast as Japan. Aphnaeines occur in many habitats and consumeand/or lay eggs on a wide range of host plants (Table 1).Like most known lycaenid species, aphnaeine larvae formassociations with ants (Pierce et al., 2002). These associationsare typically mutualistic: ants defend the larvae from predators

Correspondence: Naomi Pierce, Museum of Comparative Zoology,Harvard University, 26 Oxford St., Cambridge, MA 02138, U.S.A.E-mail: [email protected]

and parasitoids, and larvae produce nutritious secretions forthe ants. In some cases, however, larvae parasitize ants, eitherby inducing trophallaxis from ant workers or by consumingant brood. Interactions can range from facultative associationsin which lycaenid larvae intermittently associate with manyspecies of ants, to specialized, obligate symbioses in whichlarvae are never found without ants, often associating with onlyone or a few closely related species (Pierce et al., 2002).

The Aphnaeinae are notable among the Lycaenidae forexhibiting considerable variability in feeding strategies within asingle subfamily, especially considering the relatively small sizeof the group. Although 8 of the 17 genera consist of species thatare phytophagous and mutualistically associated with ants, theremaining 9 genera contain at least one species that is ‘aphy-tophagous’ (i.e. feeding obligately on substances other thanplants during at least some portion of the lifetime) and parasiti-cally associated with ants (Pierce et al., 2002, A. Heath, personal

© 2014 The Royal Entomological Society 169

170 J. H. Boyle et al.

Tab

le1.

Taxa

incl

uded

inth

isst

udy,

with

vouc

her

iden

tifica

tion,

and

info

rmat

ion

abou

tdie

t,ra

nge,

larv

alan

tass

ocia

tion

and

gene

regi

ons

sequ

ence

d.

Gen

bank

acce

ssio

nnu

mbe

r

Taxo

nV

ouch

erL

ocal

ityA

phyt

o-ph

agya

Ran

geb

Ant

asso

ciat

ecH

ostp

lant

dC

OI

CA

DE

F1𝛼

GA

PD

HH

3W

g

Pse

udal

etis

:24

spec

ies

wor

ldw

ide

W.A

fric

ato

Uga

nda

Cre

mat

ogas

ter

(Myr

mic

inae

)A

lgae

P.ag

ripp

ina

Dru

ceSC

-99-

T50

0E

bogo

,Cam

eroo

nC

ongo

lian

Cre

mat

ogas

ter

(M)

Alg

aeK

F787

284

KF7

8755

4K

F787

472

eK

F787

202

KF7

8740

0

P.cl

ymen

usD

ruce

RD

-98-

U08

7B

eni,

D.R

.Con

goC

ongo

lian

–A

lgae

KF7

8728

5K

F787

555

KF7

8747

3e

KF7

8720

3K

F787

401

Lip

aphn

aeus

:4sp

ecie

sw

orld

wid

eA

fric

aC

rem

atog

aste

r(M

)M

yrsi

nace

ae

L.a

dern

aPl

ötz

AJG

-07-

E10

5M

undw

iji,Z

ambi

aZ

ambe

zian

,Con

golia

nC

rem

atog

aste

r(M

)E

rica

les

KF7

8727

8K

F787

548

––

KF7

8719

6K

F787

394

L.l

eoni

nabi

tje

Dru

ceA

JG-0

7-N

567

Nch

ilaW

ildlif

eR

eser

ve,Z

ambi

aC

ongo

lian

Cre

mat

ogas

ter

(M)

–K

F787

279

KF7

8754

9K

F787

467

–K

F787

197

–

L.l

oxur

aR

ebel

RD

-98-

U15

2B

eni,

D.R

.Con

goZ

ambe

zian

,Con

golia

nC

rem

atog

aste

r(M

)–

KF7

8728

0K

F787

550

KF7

8746

8K

F787

333

KF7

8719

8K

F787

395

Chl

oros

elas

:13

spec

ies

wor

ldw

ide

Yes

E.A

fric

a,Sa

udiA

rabi

aC

rem

atog

aste

r(M

)Fa

bace

ae

C.a

zure

aB

utle

rSC

-99-

T19

6Pa

ngan

i,Ta

nzan

iaZ

ambe

zian

,Som

alia

n–

–K

F787

253

KF7

8752

3K

F787

442

KF7

8732

2K

F787

180

KF7

8736

6C

.maz

oens

isT

rim

enA

JG-0

7-D

625

Mos

aH

ill,Z

ambi

aS.

Afr

ican

,Zam

bezi

an–

Faba

les

KF7

8725

4K

F787

524

KF7

8744

3K

F787

323

KF7

8718

1K

F787

367

C.o

verl

aeti

Stem

pffe

rA

JG-0

7-D

616

Mos

aH

ill,Z

ambi

aZ

ambe

zian

––

KF7

8725

5K

F787

525

KF7

8744

4K

F787

324

KF7

8718

2K

F787

368

C.p

seud

ozer

itis

Tri

men

DJM

-07-

A07

5L

aiki

pia,

Ken

yaY

es,A

S.A

fric

an,Z

ambe

zian

,C

ongo

lian,

Som

alia

nC

rem

atog

aste

r(M

)Fa

bale

sK

F787

256

KF7

8752

6K

F787

445

KF7

8732

5K

F787

183

KF7

8736

9

Vans

omer

enia

:1sp

ecie

sw

orld

wid

eE

.Afr

ica

Cre

mat

ogas

ter

(M)

Faba

ceae

Ces

a:1

spec

ies

wor

ldw

ide

Som

alia

Unk

now

nU

nkno

wn

Cru

dari

a:3

spec

ies

wor

ldw

ide

Yes

Sout

hern

Afr

ica

Ano

plol

epis

(For

mic

inae

)Fa

bace

ae,Z

ygop

hylla

ceae

C.c

apen

sis

van

Son

AA

M-9

8-W

799

Gam

kaM

t.N

atur

eR

es.,

Sout

hA

fric

aY

es,D

Sout

hern

Afr

ican

Ano

plol

epis

(F)

Zyg

ophy

llale

sK

F787

273

KF7

8754

3K

F787

462

–K

F787

191

KF7

8738

8

C.l

erom

aW

alle

ngre

nA

H-9

5-Y

658

Lee

u-G

amka

,Sou

thA

fric

aS.

Afr

ican

,Zam

bezi

anA

nopl

olep

is(F

)Fa

bale

sK

F787

274

KF7

8754

4K

F787

463

KF7

8733

1K

F787

192

KF7

8738

9

C.w

ykeh

ami

Dic

kson

SQ-0

2-X

473

Witm

os,S

outh

Afr

ica

Sout

hern

Afr

ican

Ano

plol

epis

(F)

–K

F787

275

KF7

8754

5K

F787

464

–K

F787

193

KF7

8739

0

Chr

ysor

itis

:42

spec

ies

wor

ldw

ide

Yes

Sout

hern

Afr

ica

Cre

mat

ogas

ter,

Myr

mic

aria

Saun

ders

(M)

Faba

ceae

,A

mar

anth

acea

e,A

naca

rdia

ceae

,A

piac

eae,

Ast

erac

eae,

Bru

niac

eae,

Cra

ssul

acea

e,E

bena

ceae

,Eup

horb

iace

ae,M

yrsi

nace

ae,S

anta

lace

ae,Z

ygop

hylla

ceae

C.a

etho

nT

rim

enA

H-9

9-T

262

Gra

skop

,Sou

thA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)Sa

pind

ales

,Sa

xifr

agal

esK

F787

257

KF7

8752

7K

F787

446

––

–

C.a

ureu

sva

nSo

nA

H-9

5-Z

444

Gre

ylin

gsta

d,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Eri

cale

s,M

alpi

ghia

les

KF7

8725

8K

F787

528

KF7

8744

7–

––

C.b

rook

sibr

ooks

iR

iley

AH

-95-

Z41

5W

orce

ster

,Sou

thA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)Fa

bale

s,Sa

ntal

ales

,Z

ygop

hylla

les

KF7

8725

9K

F787

529

KF7

8744

8–

––

C.c

hrys

anta

sT

rim

enA

H-9

5-Z

431

Wal

lekr

all,

Sout

hA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)C

aryo

phyl

lale

sK

F787

260

KF7

8753

0K

F787

449

–K

F787

184

KF7

8737

0

© 2014 The Royal Entomological Society, Systematic Entomology, 40, 169–182

Phylogeny of the Aphnaeinae 171

Tab

le1.

Con

tinue

d

Gen

bank

acce

ssio

nnu

mbe

r

Taxo

nV

ouch

erL

ocal

ityA

phyt

o-ph

agya

Ran

geb

Ant

asso

ciat

ecH

ostp

lant

dC

OI

CA

DE

F1𝛼

GA

PD

HH

3W

g

C.c

hrys

aor

Tri

men

AH

-95-

Z45

4B

alfo

ur,M

P,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Ast

eral

es,F

abal

es,

Sapi

ndal

es,

Saxi

frag

ales

,Z

ygop

hylla

les

KF7

8726

1K

F787

531

KF7

8745

0K

F787

326

KF7

8718

5K

F787

371

C.d

icks

oniG

abri

elA

H-9

8-Y

698

Cap

eIn

fant

a,So

uth

Afr

ica

Yes

,ASo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)–

KF7

8726

2K

F787

532

KF7

8745

1–

KF7

8718

6K

F787

372

C.e

ndym

ion

Penn

ingt

onA

H-9

5-Y

759

Fran

schh

oek,

Sout

hA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)Sa

ntal

ales

KF7

8726

3K

F787

533

KF7

8745

2–

––

C.f

elth

amif

elth

ami

Tri

men

AH

-95-

Z46

0C

ape

Tow

n,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Zyg

ophy

llale

sK

F787

264

KF7

8753

4K

F787

453

––

–

C.l

yceg

enes

Tri

men

AH

-95-

Z92

1B

uum

baC

loud

-lan

ds,

Sout

hA

fric

a

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Ast

eral

es,E

rica

les,

Sapi

ndal

esK

F787

265

KF7

8753

5K

F787

454

––

–

C.l

yncu

rium

Tri

men

SJ-9

8-U

639

Xol

obe,

Sout

hA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)E

rica

les

KF7

8726

6K

F787

536

KF7

8745

5–

––

C.n

igri

cans

nigr

ican

sA

uriv

illiu

s

AH

-95-

Z44

2C

ape

Tow

n,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Ast

eral

es,

Sant

alal

es,

Zyg

ophy

llale

s

KF7

8726

7K

F787

537

KF7

8745

6–

––

C.o

reas

Tri

men

AH

-95-

Z91

1B

uum

baC

loud

-lan

ds,

Sout

hA

fric

a

Sout

hern

Afr

ican

Myr

mic

aria

(M)

Sant

alal

esK

F787

268

KF7

8753

8K

F787

457

KF7

8732

7K

F787

187

KF7

8737

3

C.p

yram

uspy

ram

usPe

nnin

gton

AH

-95-

Z95

1O

udts

hoor

n,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Ast

eral

es,

Sant

alal

esK

F787

269

KF7

8753

9K

F787

458

––

C.p

yroe

ispy

roei

sT

rim

enA

H-9

5-Y

627

Wor

cest

er,S

outh

Afr

ica

Sout

hern

Afr

ican

Cam

pono

us(F

),M

yrm

icar

ia(M

)Sa

ntal

ales

,Z

ygop

hylla

les

KF7

8727

0K

F787

540

KF7

8745

9K

F787

328

KF7

8718

8K

F787

374

C.t

hysb

eth

ysbe

Lin

naeu

sA

H-9

5-Z

447

Cap

eTo

wn,

Sout

hA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)A

ster

ales

,Fab

ales

,Sa

ntal

ales

,Z

ygop

hylla

les

KF7

8727

1K

F787

541

KF7

8746

0K

F787

329

KF7

8718

9K

F787

375

C.z

onar

ius

coet

zeri

Dic

kson

&W

ykeh

am

AH

-95-

Z42

3N

ieuw

oudt

ville

,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Ast

eral

esK

F787

272

KF7

8754

2K

F787

461

KF7

8733

0K

F787

190

–

Trim

enia

:5sp

ecie

sw

orld

wid

eY

esSo

uthe

rnA

fric

aA

nopl

olep

is(F

)A

phyt

opha

gous

T.ar

gyro

plag

aD

icks

onA

H-9

5-Y

631

Cal

vini

a,So

uth

Afr

ica

Yes

:A,B

Sout

hern

Afr

ican

Ano

plol

epis

(F)

–K

F787

294

KF7

8756

4K

F787

482

KF7

8734

4K

F787

212

KF7

8740

2

T.m

acm

aste

riD

icks

onN

P-99

-T47

4C

alitz

dorp

,Sou

thA

fric

aSo

uthe

rnA

fric

an–

–K

F787

295

KF7

8756

5K

F787

483

KF7

8734

5K

F787

213

KF7

8740

3

T.m

alag

rida

mar

yae

Dic

kson

&H

enni

ng

AH

-96-

Y73

3D

eHoo

pN

atur

eR

es.,

Sout

hA

fric

aY

es:A

,BSo

uthe

rnA

fric

anA

nopl

olep

is(F

)–

KF7

8724

2K

F787

512

KF7

8743

1K

F787

314

KF7

8717

0K

F787

404

T.m

.ced

rusm

onta

naD

icks

on&

Step

hen

AH

-98-

U48

7Sk

urw

eber

ge,W

C,

Sout

hA

fric

aY

es:A

,BSo

uthe

rnA

fric

an–

–K

F787

296

KF7

8756

6K

F787

484

eK

F787

214

KF7

8740

5

T.m

.paa

rlen

sis

Dic

kson

AH

-98-

U49

2Pa

arl,

Sout

hA

fric

aY

es:A

,BSo

uthe

rnA

fric

an–

–K

F787

297

KF7

8756

7K

F787

485

–K

F787

215

KF7

8740

6

T.w

ykeh

amiD

icks

onA

H-9

9-U

458

Kom

sber

g,So

uth

Afr

ica

Sout

hern

Afr

ican

––

KF7

8729

8K

F787

568

KF7

8748

6–

KF7

8721

6K

F787

407



Tab

le1.

Con

tinue

d

Gen

bank

acce

ssio

nnu

mbe

r

Taxo

nV

ouch

erL

ocal

ityA

phyt

o-ph

agya

Ran

geb

Ant

asso

ciat

ecH

ostp

lant

dC

OI

CA

DE

F1𝛼

GA

PD

HH

3W

g

Arg

yras

pode

s:1

spec

ies

wor

ldw

ide

Yes

,DS.

W.A

fric

aU

nkno

wn

Unk

now

n

Arg

yras

pode

sar

gyra

spis

Tri

men

AH

-95-

Z42

2W

alle

kraa

l,So

uth

Afr

ica

Yes

,DSo

uthe

rnA

fric

an–

–K

F787

243

KF7

8751

3K

F787

432

–K

F787

171

KF7

8736

3

Cig

arit

is:7

1sp

ecie

sw

orld

wid

eY

esA

fric

a,A

sia,

Mid

dle

Eas

tC

rem

atog

aste

r,P

heid

ole

(M)

Faba

ceae

,A

naca

rdia

ceae

,L

oran

thac

eae,

Ole

acea

e,Po

lygo

nace

ae,

Rub

iace

ae,

Ver

bena

ceae

,Z

ygop

hylla

ceae

C.c

rust

aria

Hol

land

RD

-98-

U04

5B

eni,

D.R

.Con

goC

ongo

lian

––

KF7

8728

6K

F787

556

KF7

8747

4K

F787

336

KF7

8720

4K

F787

376

C.e

lla

Hew

itson

AH

-04-

B54

6W

itsan

dN

atur

eR

es.,

Sout

hA

fric

aS.

Afr

ican

,Zam

bezi

an,

Som

alia

nC

rem

atog

aste

r(M

),P

heid

ole

(M)

Faba

les,

Sant

alal

esK

F787

250

KF7

8752

0K

F787

439

–K

F787

177

KF7

8737

7

C.e

parg

yros

Eve

rsm

ann

NK

-00-

P791

Bal

ta-K

ul,

Kaz

akhs

tan

Asi

aM

inor

wes

tto

Chi

na–

Faba

les

KF7

8723

6K

F787

506

KF7

8742

5–

KF7

8716

5K

F787

378

C.k

utu

Cor

bet

NP-

95-Y

326

Paha

ng,M

alay

sia

SEA

sia

––

KF7

8728

7K

F787

557

KF7

8747

5K

F787

337

KF7

8720

5K

F787

379

C.l

ohit

ase

nam

aFr

uhst

orfe

rM

WT

-93-

A02

2K

uala

Lum

pur,

Mal

aysi

aIn

dia

and

SEA

sia

–D

iosc

orea

les,

Faba

les,

Myr

tale

s,So

lana

les

KF7

8728

8K

F787

558

KF7

8747

6K

F787

338

KF7

8720

6K

F787

380

C.m

ozam

bica

Ber

tolin

iA

JG-0

7-N

619

Zam

bia

S.A

fric

an,Z

ambe

zian

,C

ongo

lian,

Som

alia

n–

Faba

les

KF7

8725

1K

F787

521

KF7

8744

0K

F787

320

KF7

8717

8K

F787

381

C.n

amaq

uaT

rim

enA

AM

-98-

U29

9G

arie

s,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Zyg

ophy

llale

sK

F787

289

KF7

8755

9K

F787

477

KF7

8733

9K

F787

207

KF7

8738

2C

.nat

alen

sis

Wes

twoo

dA

H-0

1-T

355

Hei

delb

urg,

Sout

hA

fric

aS.

Afr

ican

,Zam

bezi

anC

rem

atog

aste

r(M

)Fa

bale

s,G

entia

nale

s,L

amia

les,

Sant

alal

es

KF7

8729

0K

F787

560

KF7

8747

8K

F787

340

KF7

8720

8K

F787

383

C.p

hane

sT

rim

enA

H-0

4-B

550

Wits

and

Nat

ure

Res

.,So

uth

Afr

ica

S.A

fric

an,Z

ambe

zian

Cre

mat

ogas

ter

(M)

Faba

les,

Sant

alal

esK

F787

252

KF7

8752

2K

F787

441

KF7

8732

1K

F787

179

KF7

8738

4

C.s

yam

ate

rana

Fruh

stor

fer

MW

T-9

3-A

039

Kua

laL

umpu

r,M

alay

sia

SEA

sia

–Fa

bale

sK

F787

291

KF7

8756

1K

F787

479

KF7

8734

1K

F787

209

KF7

8738

5

C.t

akan

onis

Mat

sum

ura

JCC

-01-

P022

Seou

l,So

uth

Kor

eaY

es:A

,BJa

pan

Cre

mat

ogas

ter

(M)

–K

F787

292

KF7

8756

2K

F787

480

KF7

8734

2K

F787

210

KF7

8738

6

C.t

avet

ensi

sL

athy

DJM

-07-

A08

1N

gong

Hill

s,K

enya

Zam

bezi

an,C

ongo

lian,

Som

alia

nP

heid

ole

(M)

Faba

les

KF7

8729

3K

F787

563

KF7

8748

1K

F787

343

KF7

8721

1K

F787

387

Zer

itis

:6sp

ecie

sw

orld

wid

eA

fric

aU

nkno

wn

Unk

now

n

Z.s

orha

geni

iDew

itzA

JG-0

7-D

630

Kab

wel

uma

Falls

,Z

ambi

aZ

ambe

zian

––

KF7

8730

1K

F787

571

KF7

8748

9K

F787

348

KF7

8721

9K

F787

408

Axi

ocer

ses:

20sp

ecie

sw

orld

wid

eY

esA

fric

aC

rem

atog

aste

r,P

heid

ole

(M)

Faba

ceae

,Lor

anth

acea

e,O

leac

eae

A.a

man

gaW

estw

ood

AJG

-07-

N75

8Z

ambi

aS.

Afr

ican

,Zam

bezi

an,

Con

golia

n,So

mal

ian

Phe

idol

e(M

)Fa

bale

s,Sa

ntal

ales

KF7

8724

4K

F787

514

KF7

8743

3K

F787

315

KF7

8717

2–

A.b

amba

naG

rose

-Sm

ithR

D-9

8-U

072

Ben

i,D

.R.C

ongo

Zam

bezi

an–

Faba

les

KF7

8724

5K

F787

515

KF7

8743

4–

KF7

8717

3K

F787

364



Tab

le1.

Con

tinue

d

Gen

bank

acce

ssio

nnu

mbe

r

Taxo

nV

ouch

erL

ocal

ityA

phyt

o-ph

agya

Ran

geb

Ant

asso

ciat

ecH

ostp

lant

dC

OI

CA

DE

F1𝛼

GA

PD

HH

3W

g

A.c

oele

scen

sH

enni

ng&

Hen

ning

(?)

SC-9

9-T

171

Pang

ani,

Tanz

ania

S.A

fric

an,Z

ambe

zian

––

KF7

8724

6K

F787

516

KF7

8743

5K

F787

316

KF7

8717

4–

A.c

roes

usT

rim

enA

H-0

5-B

559

Mid

dlet

on,E

C,

Sout

hA

fric

aSo

uthe

rnA

fric

an–

–K

F787

247

KF7

8751

7K

F787

436

KF7

8731

7K

F787

175

–

A.h

arpa

xFa

bric

ius

TL

-96-

Y95

0C

ape

Coa

st,G

hana

Yes

:AZ

ambe

zian

,Con

golia

n,So

mal

ian,

Ara

bia

Cre

mat

ogas

ter

(M),

Phe

idol

e(M

)Fa

bale

sK

F787

248

KF7

8751

8K

F787

437

KF7

8731

8K

F787

176

KF7

8736

5

A.t

joan

eW

alle

ngre

nA

JG-0

7-N

847

Arc

turu

s,Z

imba

bwe

S.A

fric

an,Z

ambe

zian

–Fa

bale

sK

F787

249

KF7

8751

9K

F787

438

KF7

8731

9–

–

Alo

eide

s:57

spec

ies

wor

ldw

ide

Yes

S.E

.Afr

ica

Lep

isio

ta(F

),M

onom

oriu

mM

ayr,

Phe

idol

e(M

)

Faba

ceae

,Mal

vace

ae,T

hym

elae

acea

e,Z

ygop

hylla

ceae

A.a

rand

aW

alle

ngre

nA

P-98

-W75

7G

ouri

tsR

.Val

ley,

Sout

hA

fric

aS.

Afr

ican

,Zam

bezi

anP

heid

ole

(M)

Faba

les

KF7

8722

3K

F787

493

KF7

8741

2K

F787

302

KF7

8715

2K

F787

349

A.a

rida

Tite

&D

icks

onA

AM

-98-

Y99

6C

arol

usbe

rg,S

outh

Afr

ica

Sout

hern

Afr

ican

––

KF7

8722

4K

F787

494

KF7

8741

3–

KF7

8715

3K

F787

350

A.b

ampt

oniT

ite&

Dic

kson

AH

-06-

M54

2St

eink

opf,

Sout

hA

fric

aSo

uthe

rnA

fric

anL

epis

iota

(F)

Mal

vale

sK

F787

225

KF7

8749

5K

F787

414

–K

F787

154

–

A.b

arkl

yiT

rim

enA

H-0

6-M

543

Kam

iesk

roon

,So

uth

Afr

ica

Sout

hern

Afr

ican

––

KF7

8722

6K

F787

496

KF7

8741

5–

KF7

8715

5K

F787

351

A.c

arol

ynna

eD

icks

onA

H-0

0-T

122

Raw

sonv

ille,

Sout

hA

fric

aSo

uthe

rnA

fric

an–

Faba

les

KF7

8722

7K

F787

497

KF7

8741

6K

F787

303

KF7

8715

6–

A.l

utes

cens

Tite

&D

icks

onA

H-0

0-T

154

Wor

cest

er,S

outh

Afr

ica

Sout

hern

Afr

ican

–Fa

bale

sK

F787

228

KF7

8749

8K

F787

417

KF7

8730

4K

F787

157

KF7

8735

2

A.m

arga

reta

eT

ite&

Dic

kson

AH

-07-

T06

4L

ambe

rts

Bay

,So

uth

Afr

ica

Sout

hern

Afr

ican

––

KF7

8722

9K

F787

499

KF7

8741

8K

F787

305

KF7

8715

8K

F787

353

A.n

ollo

thiT

ite&

Dic

kson

AH

-06-

M55

6H

onde

klip

Bay

,So

uth

Afr

ica

Sout

hern

Afr

ican

Lep

isio

ta(F

)M

alva

les,

Zyg

ophy

llale

sK

F787

230

KF7

8750

0K

F787

419

KF7

8730

6K

F787

159

KF7

8735

4

A.p

alli

daT

ite&

Dic

kson

AH

-07-

C21

5B

eauf

ortW

est,

Sout

hA

fric

aY

es:C

Sout

hern

Afr

ican

Lep

isio

ta(F

)Fa

bale

sK

F787

231

KF7

8750

1K

F787

420

KF7

8730

7K

F787

160

–

A.p

enni

ngto

niT

ite&

Dic

kson

AH

-07-

P536

Dur

ban,

Sout

hA

fric

aSo

uthe

rnA

fric

an–

–K

F787

232

KF7

8750

2K

F787

421

eK

F787

161

KF7

8735

5

A.p

ieru

sC

ram

erA

H-9

5-Y

614

Red

hill,

Sout

hA

fric

aSo

uthe

rnA

fric

anL

epis

iota

(F)

Faba

les,

Mal

vale

s,Z

ygop

hylla

les

KF7

8723

3K

F787

503

KF7

8742

2–

KF7

8716

2K

F787

356

A.s

impl

exT

rim

enA

H-0

4-B

549

Wits

and

Nat

ure

Res

.,So

uth

Afr

ica

Sout

hern

Afr

ican

––

KF7

8723

4K

F787

504

KF7

8742

3K

F787

308

KF7

8716

3K

F787

357

A.t

hyra

Lin

naeu

s(?

)A

AM

-98-

U29

3L

ambe

rts

Bay

,So

uth

Afr

ica

Sout

hern

Afr

ican

Lep

isio

ta(F

)Fa

bale

s,M

alva

les

KF7

8723

5K

F787

505

KF7

8742

4–

KF7

8716

4K

F787

358



Tab

le1.

Con

tinue

d

Gen

bank

acce

ssio

nnu

mbe

r

Taxo

nV

ouch

erL

ocal

ityA

phyt

o-ph

agya

Ran

geb

Ant

asso

ciat

ecH

ostp

lant

dC

OI

CA

DE

F1𝛼

GA

PD

HH

3W

g

Eri

ksso

nia:

3sp

ecie

sw

orld

wid

eS.

Afr

ican

Lep

isio

ta(F

)Fa

bace

ae,T

hym

elae

acea

e

E.a

crae

ina

Tri

men

AJG

-07-

D58

3Z

ambi

aS.

Afr

ican

,Zam

bezi

anL

epis

iota

(F)

Mal

vale

sK

F787

276

KF7

8754

6K

F787

465

KF7

8733

2K

F787

194

KF7

8739

1E

.coo

kson

iDru

ceA

JG-0

7-D

644

Mun

dwiji

,Zam

bia

Zam

bezi

an–

–K

F787

277

KF7

8754

7K

F787

466

–K

F787

195

KF7

8739

2

Aph

naeu

s:22

spec

ies

wor

ldw

ide

Yes

Afr

ica

Cre

mat

ogas

ter

(M)

Faba

ceae

,Ana

card

iace

ae,C

onvu

lvul

acea

e,E

upho

rbia

ceae

,Lor

nath

acea

e,O

leac

eae,

Sapi

ndac

eae

A.e

riks

soni

mas

huna

eSt

empf

fer

AH

-99-

U51

7A

rctu

rus,

Zim

babw

eS.

Afr

ican

,Zam

bezi

anC

rem

atog

aste

r(M

)Fa

bale

s,So

lana

les

KF7

8723

7K

F787

507

KF7

8742

6K

F787

309

KF7

8716

6K

F787

359

A.fl

aves

cens

Stem

pffe

rA

JG-0

7-D

629

Mos

aH

ill,Z

ambi

aZ

ambe

zian

,Som

alia

n–

–K

F787

238

KF7

8750

8K

F787

427

KF7

8731

0K

F787

167

–

A.m

arsh

alli

Nea

veA

H-9

9-U

518

Arc

turu

s,Z

imba

bwe

Zam

bezi

an–

Faba

les

KF7

8723

9K

F787

509

KF7

8742

8K

F787

311

KF7

8716

8K

F787

360

A.o

rcas

Dru

ryR

D-9

8-U

020

Ben

i,D

.R.C

ongo

Zam

bezi

an,C

ongo

lian

–Fa

bale

s,M

alpi

ghia

les,

Sant

alal

esK

F787

240

KF7

8751

0K

F787

429

KF7

8731

2K

F787

169

KF7

8736

1

A.q

uest

iaux

iA

uriv

illiu

sA

JG-0

7-N

614

Zam

bia

Zam

bezi

an–

–K

F787

241

KF7

8751

1K

F787

430

KF7

8731

3–

KF7

8736

2

Tylo

paed

ia:1

spec

ies

wor

ldw

ide

Sout

hern

Afr

ica

Cre

mat

ogas

ter

(M)

Faba

ceae

,Ebe

nace

ae

T.sa

rdon

yxsa

rdon

yxT

rim

enN

P-99

-T46

3G

amka

Mou

ntai

n,So

uth

Afr

ica

Sout

hern

Afr

ican

–Fa

bale

sK

F787

299

KF7

8756

9K

F787

488

KF7

8734

7K

F787

217

–

T.s.

peri

ngue

yiD

icks

onA

H-9

7-Y

711

Citr

usda

l,So

uth

Afr

ica

Sout

hern

Afr

ican

Cre

mat

ogas

ter

(M)

Faba

les,

Ros

ales

KF7

8730

0K

F787

570

KF7

8748

7K

F787

346

KF7

8721

8–

Pha

sis:

4sp

ecie

sw

orld

wid

eSo

uthe

rnA

fric

aC

rem

atog

aste

r(M

)A

naca

rdia

ceae

,Mel

iant

hace

ae

P.cl

avum

Mur

ray

AH

-95-

Y64

3C

alvi

nia,

Sout

hA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)G

eran

iale

s,Sa

pind

ales

KF7

8728

1K

F787

551

KF7

8746

9–

KF7

8719

9K

F787

397

P.pr

ingl

eiD

icks

onA

AM

-98-

W26

2Sw

aarw

eerb

erg,

Sout

hA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)G

eran

iale

sK

F787

282

KF7

8755

2K

F787

470

KF7

8733

4K

F787

200

KF7

8739

8

P.th

ero

Lin

naeu

sA

AM

-98-

U29

5L

ambe

rts

Bay

,Sou

thA

fric

aSo

uthe

rnA

fric

anC

rem

atog

aste

r(M

)G

eran

iale

s,Sa

pind

ales

KF7

8728

3K

F787

553

KF7

8747

1K

F787

335

KF7

8720

1K

F787

399

Feni

seca

tarq

uini

usFa

bric

iusf

MW

T-9

6-Y

333

Pete

rsha

m,M

A,

USA

––

–K

F787

220

KF7

8749

0K

F787

409

–K

F787

149

KF7

8739

3

Orn

ipho

lido

tos

peuc

etia

penn

ingt

oniR

ileyf

SJ-9

7-Y

842

Ens

elen

iNat

ure

Res

.,So

uth

Afr

ica

––

–K

F787

221

KF7

8749

1K

F787

410

–K

F787

150

KF7

8739

6

Sim

iski

naph

eret

iaH

ewits

onf

NP-

95-Y

140

Peka

n,M

alay

sia

––

–K

F787

222

KF7

8749

2K

F787

411

–K

F787

151

–

aFo

rgen

era,

a‘y

es’i

ndic

ates

that

one

orm

ore

spec

ies

with

inth

ege

nus

exhi

bita

phyt

opha

gy.F

orm

ofap

hyto

phag

y,w

here

appl

icab

le,i

ssi

gnal

led

bya

lette

r.A

,tro

phal

laxi

s;B

,fee

dson

antb

rood

;C,f

eeds

onan

tegg

s;D

,pre

sum

edap

hyto

phag

ous,

butf

orm

isun

know

n.bFo

rge

nera

,the

loca

tions

liste

dar

eth

eco

untr

ies

orre

gion

sw

here

spec

ies

ofth

atge

nus

may

befo

und.

For

spec

ies,

the

rang

eis

desc

ribe

dby

the

biog

eogr

aphi

calr

egio

nsof

Lin

der

etal

.(20

12).

c Ada

shin

dica

tes

that

anta

ssoc

iatio

nis

unkn

own

for

this

spec

ies.

dFo

rge

nera

,ase

lect

ion

ofim

port

anth

ost-

plan

tfam

ilies

are

liste

d.Fo

rsp

ecie

s,al

lrec

orde

dho

st-p

lant

orde

rsar

elis

ted.

Ada

shin

dica

tes

that

plan

tass

ocia

tion

isun

know

nfo

rth

issp

ecie

s.e A

gene

frag

men

tles

sth

an20

0bp

inle

ngth

was

used

for

this

spec

imen

;bec

ause

Gen

bank

does

nota

ccep

tfra

gmen

tsof

this

size

,the

sequ

ence

isin

clud

edin

the

supp

ortin

gin

form

atio

n(T

able

S8).

f Out

grou

ps:t

hese

taxa

wer

eno

tinc

lude

din

the

ance

stra

lran

gere

cons

truc

tion,

and

thei

rch

arac

ter

stat

esw

ere

mar

ked

as‘u

nkno

wn’

inth

ean

cest

rals

tate

reco

nstr

uctio

ns.

Sour

ces:

Ack

ery

etal

.(1

995)

,C

hou

(199

4),

Cor

bet

&Pe

ndle

bury

(199

2),

D’A

brer

a(1

986)

,G

ardi

ner

&Te

rbla

nche

(201

0),

Hea

th(1

997)

,H

eath

etal

.(2

002,

2008

),K

roon

(199

9),

Lar

sen

(199

1),

Lew

is(1

973)

,Pi

erce

etal

.(2

002)

,Pri

ngle

etal

.(19

94),

Tuz

ov(1

997)

and

Will

iam

s(2

012)

.



Table 2. Summary of the major taxonomic rearrangements ofAphnaeinae.

Swinhoe (1911) Stempffer (1967) Heath (1997) Current genera

Aphnaeus Pseudaletis Pseudaletis PseudaletisLipaphnaeus Lipaphnaeus LipaphnaeusChloroselas Chloroselas ChloroselasDesmolycaena

Vansomerenia VansomereniaCrudaria Crudaria CrudariaChrysoritis Chrysoritis ChrysoritisPoecilmitisPhasis Phasis Phasis

Trimenia TrimeniaArgyraspodes ArgyraspodesTylopaedia Tylopaedia

Spindasis Jacksonia CesaCigaritis CigaritisSpindasis

ApharitisZeritis Zeritis ZeritisAxiocerses Axiocerses AxiocersesAloeides Aloeides AloeidesErikssonia Erikssonia ErikssoniaAphnaeus Aphaneus AphnaeusParaphnaeus

Horizontal rows demonstrate how sets of species changed generic nameover time. When Spindasis and Cigaritis were synonymized, Cigaritiswas found to be the older name, even though it is not one of Stempffer’sgenera.

observation). This enormous range in diet – among the mostdiverse for a group of its size within the Lepidoptera – makesAphnaeinae of particular interest in analysing factors contribut-ing to life-history evolution and the loss of herbivory. By deter-mining how such aphytophagous, parasitic species and generaare distributed throughout the phylogeny, we can gain informa-tion about the evolutionary dynamics of this parasitic behaviour:it allows us to ask does aphytophagy evolve frequently, but failto survive or diversify across evolutionary time; or do aphy-tophagous species evolve more rarely, but persist long enoughto speciate?

The systematics of this group has been revised several times(Table 2). Distant (1882–1886) proposed the name Aphnaria fora grouping of 24 diverse lycaenid genera based upon similaritiesin wing venation. This grouping remained unchanged untilSwinhoe (1911) erected the subfamily Aphnaeinae based onspecific morphological characters. Swinhoe made no mentionof Distant or the 24 genera placed in Aphnaria; instead heproposed only one genus with 27 species, Aphnaeus Hübner, inhis subfamily, having subsumed Cigaritis, Spindasis Wallengrenand Amblypodia Westwood under Aphnaeus.

Swinhoe’s one-genus subfamily remained unchanged untilStempffer (1967) proposed the adoption of 16 genera within it,namely: Aphnaeus, Paraphnaeus Thierry Mieg, Apharitis Riley,Spindasis, Lipaphnaeus Aurivillius, Chloroselas Butler, ZeritisBoisduval, Desmolycaena Trimen, Axiocerses Hübner, PhasisHübner, Aloeides Hübner, Poecilmitis Butler, Chrysoritis

Butler, Crudaria Wallengren, Erikssonia Trimen andPseudaletis Druce. Stempffer’s selection was based on adultmorphological characters, particularly the male genitalia.

Eliot (1973) relegated this same group of genera to tribal levelwithin Theclinae, separating Pseudaletis into one section withinthe tribe and the remainder into an Aphnaeus section. During thefollowing decade several new genera were added to the tribe.

Heath (1997) reviewed the Aphnaeini based on morphologi-cal characters, and abandoned Eliot’s two sections, although heconsidered Pseudaletis to be sister to the remaining genera. Hedid not provide a formal phylogenetic analysis as part of hisrevision, but arranged the taxa in his systematic classificationto reflect what he considered to be the most likely evolution-ary relationships among groups. He also synonymized severalof the existing genera as follows: Apharitis with Spindasis, andPoecilmitis (as well as two additional genera erected sinceEliot’s revision – Bowkeria Quickleburge and Oxychaeta Tite& Dickson) with Chrysoritis, and Argyrocupha Tite & Dick-son with Trimenia Tite & Dickson. He proposed several newcombinations by shifting all but one of the species withinDesmolycaena into the genus Chloroselas and creating a newgenus, Vansomerenia Heath, for the remaining species. In a laterwork, Heath et al. (2002) synonymized the genus Spindasis withCigaritis. The resulting genera were thus: Pseudaletis, Lipaph-naeus, Chloroselas, Vansomerenia, Jacksonia Heath, Crudaria,Chrysoritis, Trimenia, Argyraspodes Tite & Dickson, Cigaritis,Zeritis, Axiocerses, Aloeides, Erikssonia, Aphnaeus, Tylopae-dia Tite & Dickson and Phasis. Jacksonia was subsequentlyrenamed Cesa Seven after it was found to be a previouslyoccupied name.

With respect to higher level relationships, Eliot (1973) haddifficulty placing Aphnaeini within the Theclinae, originallyassigning it to a somewhat derived position near the Iolainiand Cheritrini, but he later placed it – along with Polyommatiniand Lycaenini – as sister to Theclinae. More specialized (andpresumably derived) characters that made initial diagnosis ofthis group difficult include the hind wing shape, which is oftentailed and lobed; the absence of a precostal vein on the hindwing; absence of tibial spurs; and an ungirdled pupa. Insteadof the male fore tarsus being segmented with double claws, itis fused into a single segment ending in a single claw (exceptfor Aphnaeus where it is blunted). Eliot also remarked uponthe extraordinary tufts of highly specialized scales found at thetip of the female abdomen (1973). These are found in certaingenera of the aphnaeines and several genera of the Hesperiidae,in which he presumed they must have evolved convergently.Some of these scales adhere to the freshly laid egg, appearing toserve as camouflage for the egg. Newly hatched larvae of somespecies have been observed to consume these scales, presumablygaining some benefit from doing so (Heath, 1997). The tribeAphnaeini was later placed in subfamily Lycaeninae by Scott(1985). We refer to the clade as a subfamily here in light ofa forthcoming phylogeny of the Lycaenoidea which recoversthe aphnaeines as a basally branching clade along with otherestablished subfamilies (N.E. Pierce et al., in preparation).

Apart from a phylogenetic study of the genus Chrysoritiswhich used several outgroups from other aphnaeine genera



(Rand et al., 2000), no work has been done to elucidate therelationships within Aphnaeinae using molecular characters.Using data from mitochondrial DNA, Rand and his colleaguesfound Aloeides+Argyraspodes to be weakly supported sistergenera, with Trimenia sister to that clade. Sister to Tri-menia+ (Aloeides+Argyraspodes) was the grouping Tylopae-dia+Phasis. The relationships among these, Chrysoritis andCrudaria were not determined.

Phylogenetic analyses are important prerequisites for under-standing the evolution of life-history variation, and the Aph-naeinae are of particular value in this regard because of theremarkable variability among species in both feeding strategyand ant association. The focus of this study is to use molecu-lar characters from both mitochondrial and nuclear markers toinfer relationships within and among the main lineages of Aph-naeinae, providing a rigorous framework for further explorationof life-history evolution. We also make the first contributiontoward such exploration by using ancestral state reconstructionsto infer biogeographical history and the evolution of host plantassociations in the group.

Materials and methods

Taxon sampling

Samples studied here included 79 ingroup taxa representing15 of the 17 genera recognized in Heath’s review of theaphnaeines (Heath, 1997; Heath et al., 2002). Unfortunately,no specimens of the rare monotypic genera Vansomerenia andCesa could be obtained. Three taxa from the closely relatedlycaenid subfamilies Poritiinae and Miletinae (sensu Eliot,1973) were used as outgroups. Specimens were collected freshinto 90–100% ethanol and stored at −20 or −80∘C prior to DNAextraction. Wings were kept separately in glassine envelopes,and vouchers of all samples are deposited in the DNA andTissues collection of the Museum of Comparative Zoology.

Molecular protocols

DNA was extracted from butterfly legs or thoracic tissue usinga Qiagen DNEasy Blood and Tissue Kit (Qiagen, Inc., Valencia,CA, U.S.A.). Fragments were amplified from a mitochondrialgene, cytochrome oxidase I (COI), and five nuclear gene regions,histone 3 (H3), elongation factor 1 alpha (EF1𝛼), wingless(wg), glyceraldehyde-3-phosphate dehydrogenase (GAPDH)and carbamoyl-phosphate synthetase 2, aspartate transcar-bamylase, and dihydroorotase (CAD). These mitochondrial andnuclear genes were selected due to their utility in reconstructingthe phylogeny of other insect groups of similar age and diver-sity (e.g. Wahlberg et al., 2005; Vila et al., 2011; Talavera et al.,2013).

Amplifications were done by standard polymerase chain reac-tions (PCR) mostly using published primers (Table S1). For eachsample, 25-μL reactions were subjected to a 3 min initial denat-uration at 94∘, then cycled through a 50 s denaturation at 94∘,

followed by a variable annealing phase (Table S2), and an 80 sextension phase at 72∘. The last cycle was followed by a finalextension phase of 5 min at 72∘.

PCR products were purified by adding 1.0 μL Antarctic Phos-phatase, 1.0 μL Antarctic Phosphatase buffer and 0.6 μL Exonu-clease I (New England Biolabs, Ipswitch, MA, U.S.A.) and incu-bating at 37∘C for 35 min followed by 20 min at 80∘C. Sampleswere then amplified for sequencing using BigDye chemistryv3.1 (Applied Biosystems, Foster City, CA, U.S.A.). Ampli-fied fragments were sequenced in both directions on an AppliedBiosystems 3130xl Genetic Analyzer using specified reactionconditions (Table S3).

Phylogenetic analyses

Sequences were edited in Sequencher 4.8 (Gene Codes, AnnArbor, MI, U.S.A.) and manually aligned. No indels were foundin any of the six markers.

Phylogenetic analyses were performed using Maximum Par-simony, Maximum Likelihood and Bayesian Inference-basedmethods. Bayesian inference was performed using MrBayesv3.2 (Ronquist & Huelsenbeck, 2003). Molecular data were par-titioned as in the maximum likelihood analysis (below). Foreach partition, a model was chosen using jModelTest (TableS4), using that program’s Akaike information criterion methodto select the highest-weighted model that is provided for inMrBayes (Posada, 2008). We used two runs, each with fourchains, and ran the analysis for 25 million generations, samplingevery 10 000 generations. Likelihoods were viewed using Tracerv1.5.0 (Rambaut & Drummond, 2007) and a burn-in set at 1 mil-lion generations before summarizing the sampled trees.

We also performed a parsimony analysis in TNT v1.1(Goloboff et al., 2008). The data included concatenatedsequences of all six markers. We used a New Technology Searchwith Sectorial Search and Tree Fusing, driving the search untila minimum length had been found 100 times. We then madea strict consensus tree of all the most parsimonious trees andcalculated nodal support with 1000 standard bootstraps.

Maximum likelihood analysis was performed using RAxMLv7.7.5 (Stamatakis, 2006; Stamatakis et al., 2008). The datawere partitioned by first, second or third codon position foreach of the six markers, for a total of 18 partitions. RAxMLdetermined the tree with the highest likelihood using a gen-eral time-reversible model with the Gamma model of rate het-erogeneity and an estimated proportion of invariant sites, andperformed a rapid bootstrap analysis with 100 bootstraps.

Ancestral range reconstruction

The chronopl function in the ape package in R (Paradiset al., 2004) was used to produce an ultrametric version of thesummary tree produced by MrBayes, setting lambda= 0, andremoving the three outgroup taxa before transforming the tree.Ancestral ranges were estimated on this tree using lagrangev20130526 (Ree & Smith, 2008). Each species was coded as



belonging to one or more of four of the biogeographical regionsof sub-Saharan Africa as defined in Linder et al. (2012; weprovide an illustration in Fig. 2), and/or as having a range outsideof sub-Saharan Africa (see Table 1). The Southern Africanregion was set as nonadjacent to the Somalian region and to thenon-African region; all other possible area combinations werepermitted.

Ancestral state reconstruction

We used BayesTraits v1.3 to estimate ancestral ant asso-ciates and host plants (Pagel et al., 2004). Species were codedas associating with one or more of six genera of ants, andwith one or more of 18 orders of host plants (Table S5).The reversible-jump MCMC algorithm of the BayesMultiStatesmodel was used along with an ultrametric phylogeny (trans-formed as for lagrange, but with the outgroup taxa included)and priors chosen to produce an acceptance rate around 30%.For the analysis of ant associates, the ratedev parameter wasset to 1.2 and a reversible-jump hyperprior set to exp 0 30. Forthe analysis of host plants, the same hyperprior was used witha ratedev parameter of 0.103. The analysis was run for 5.05million generations, sampling every hundred generations aftera 50 000 generation burn-in.

Results and discussion

Phylogenetic hypotheses of Aphnaeinae and Systematics

Monophyly of genera. With one exception, the generadescribed in Heath’s (1997) organization of the aphnaeineswere recovered as monophyletic with strong support in all threeanalyses, although only a single representative of the genusZeritis was included in the phylogeny (see Fig. 1; FiguresS1 and S2). The single exception was Aloeides. In his 1997review, Heath suggested that Aloeides and Erikssonia maybe congeneric, although he did not find sufficient reason tosynonymize them. Our analysis is similarly ambiguous as to therelationship between the two genera. Aloeides was recovered asmonophyletic, but the support for this arrangement was weak inthe maximum likelihood and Bayesian analyses. Based on ourdata, we cannot rule out the possibility that Erikssonia, althougha monophyletic group itself, is nested within Aloeides. Addi-tional markers and/or sample taxa will be needed to confirmthat the two groups are indeed reciprocally monophyletic.

Our analysis also provides support for several earlier syn-onymizations. When Riley (1925) erected the genus Apharitis,he noted that it differed from Spindasis only in its coloration;he acknowledged that structurally – in venation and other char-acters – the genera were similar. Stempffer described and illus-trated all the structural features in both genera and provided nodistinguishing characteristics (1967). Heath illustrated the malegenitalia of Cigaritis, Spindasis and Apharitis, showing all threeto be of the same type and arguing that their wing markingswere also of a common pattern. He synonymized Apharitis with

Spindasis on structural grounds, considering coloration to be aninsufficient basis for retaining Apharitis (1997). In a later publi-cation, Heath et al. (2002) synonymized Spindasis with Cigari-tis for similar reasons. Our analysis provides support for this, asCigaritis epargyros, a representative of the former genus Aphar-itis, is found nested within the rest of the genus Cigaritis. Heath(1997) also synonymized the genus Argyracupha with Trimenia.He found the male genitalia, the wing markings and other char-acteristics to be similar, differing only in the shape of the malehind wing. In this molecular phylogeny, T. malagrida Wallen-gren (formerly Argyrocupha) is closely related to other (original)Trimenia species.

We present here the Bayesian phylogenetic hypothesis, whichwe used for ancestral range and state reconstructions. Themaximum-likelihood and Bayesian trees recover the samewell-supported relationships among genera, although with a dif-ferent branching order at one weakly-supported node, and withminor differences in branch length and node support through-out the tree (Fig. 1; Figure S1). The parsimony tree (FigureS2) differs in several key respects from the likelihood-basedmethods, but many taxa show the same relationships under allthree methods. The most important difference among the treesis that parsimony recovers Aphnaeus as sister to the rest of thesubfamily, and Axiocerses+ Zeritis as sister to the remainder,although these clades are nested within one of two larger cladesin the Bayesian and likelihood analyses. These three genera haveamong the longest branch lengths of any genera in the Bayesianand likelihood analyses. The different arrangement of the parsi-mony tree may be influenced by long branch attraction, and sowe focus here on the Bayesian tree.

The six genera Pseudaletis, Chrysoritis, Crudaria, Cigaritis,Chloroselas and Lipaphnaeus form a single clade (markedClade J on Fig. 2). These data support Heath’s (1997) deci-sion to abandon Eliot’s (1973) placement of Pseudaletisin a section separated from the other genera. Pseudaletisis sister to the other five genera, and Chrysoritis sister tothe remaining four. The relationships among the remain-ing four genera are not well supported. The Bayesianphylogeny places Crudaria+Cigaritis sister to Chlorose-las+Lipaphnaeus, but the only well-supported relationship isLipaphnaeus+Chloroselas forming a single clade. This Lipa-phnaeus+Chloroselas clade is also supported by similaritiesof the genitalia, as is the grouping of those two genera withPseudaletis and Crudaria (Heath, 1997). Heath also groupedthese four genera with Chrysoritis, as is seen in our phyloge-nies. Although he did not include Cigaritis, he did include themonotypic genera Vansomerenia and Cesa in this grouping.

Clade J is sister to the remaining lineages, marked asClade B. Among these, Tylopaedia+Phasis is sister tothe rest of Clade B, but with poor support. Within thisclade, Aphnaeus is sister to Zeritis+Axiocerses, Trimenia+Argyraspodes+ (Aloeides+Erikssonia) form a single clade,and Trimenia is sister to Argyraspodes+Aloeides/Erikssonia.However, there is no strong support for the relationshipsbetween these groupings. This clade was not predicted bythe morphological analyses of either Eliot (1973) or Heath(1997), but was recovered by Rand et al. (2000), with low



Fig. 1. Phylogenetic hypothesis of Aphnaeinae: Bayesian Inference. ‘*’ marks nodes with 1.00 posterior probability in the Bayesian analysis and100% bootstrap support in the maximum likelihood analysis. Nodes with less support are labeled with Bayesian posterior probability (first number) andmaximum likelihood bootstrap support (second number). Nodes not recovered by the maximum likelihood analysis have a ‘–’ in place of the bootstrapsupport. Aphytophagous species are marked with an arrow.



Fig. 2. Life-history evolution of Aphnaeinae. The colour of the first block to the right of each butterfly genus indicates which ant genus had the highestprobability of association with the common ancestor of the given aphnaeine genus, as determined by the BayesTraits reconstruction. The colour of thesecond block indicates the order of host plant BayesTraits assigned as most likely to have associated with the common ancestor of the same aphnaeinegenus. The colour of each branch indicates the most plausible ancestral range along that branch. The inset map of Africa (after Linder et al., 2012) showsthe boundaries of each region. Note that aphnaeines are not found in several regions of Africa. Grey branch lines indicate that lagrange recoveredmultiple zones with a >0.2 probability that the range of the common ancestor contained at least that zone. For specific detail regarding each node, seeTables S5–S7.



support. Within the group, the clades of Tylopaedia+Phasisand Zeritis+Axiocerses are recovered with strong support.Genitalic similarities further support Zeritis+Axiocerses, andargue for the clustering of Trimenia and Argyraspodes (Heath,1997); the clade of Tylopaedia+Phasis was also recovered byRand et al. (2000).

Ancestral range reconstruction

lagrange provided evidence for a Southern African distri-bution of the ancestor of the taxon labelled Clade B in Fig. 2,with a 0.81 probability that the ancestral distribution includedSouthern Africa, and a 0.21 probability of ancestors distributedoutside Southern Africa (Fig. 2 and Table S5).

There was less evidence for the distribution of the ancestorof the Clade J. The probability that the ancestral distributionincluded at least Southern Africa was 0.64, and the probabilitythat it included at least the Congolian region was 0.55. Themost likely single possibility was a strictly Southern Africandistribution, with a probability of 0.21.

lagrange thus provides evidence for a probable SouthernAfrican origin of Aphnaeinae, with the possibility of a moreextensive ancestral distribution. Aphnaeine butterflies appear tohave radiated outside Africa only once, because the Orientalspecies of Cigaritis appear to form a single radiation. A fewother species have also independently colonized Arabia, includ-ing Chloroselas arabica Riley, C. esmerelda Butler (Larsen,1991), and Axiocerses harpax Fabricius. However, the South-ern African origin – as well as the ancestral ranges at internalnodes of the Aphnaeinae phylogeny – is not well supported. Oursampled taxa represent about a quarter of the overall aphnaeinediversity (79 of 278 species, 28%), and even less in some impor-tant genera, such as Cigaritis (12 of 71, 17%) or Pseudaletis (2 of24, 8%). Increased taxon sampling will be necessary to uncovermore about the history of the apparent single colonization ofAsia by Cigaritis and to clarify the within-Africa evolutionaryhistory of the rest of the subfamily. However, it is also possi-ble that Linder et al.’s African regions (2012) are simply toorestricted in area to produce a strong signal among relativelymobile insects, whose ancestral range history could conceivablyinclude many expansions, contractions, and reintroductions.

Ancestral state reconstruction

Ant assocations. Almost all aphnaeine butterflies associatewith ants: ant association has been recorded in 14 of the 17genera, and life histories are still unknown for the remainingthree, Zeritis and the monotypic genera Argyraspodes and Cesa.All aphnaeines with known life histories (approximately onethird of the total) interact in some way with ants, and theseassociations are obligate in 97% of cases (Pierce et al., 2002).

As to which ants have been the associates of Aphaenini,BayesTraits predicted that the aphnaeine common ancestorassociated with Crematogaster Lund (Formicidae: Myrmicinae)species (P= 0.70). The most likely scenario for ant association

is an association with Crematogaster in the common ancestor,with a few subsequent switches of ant associate: (i) to PheidoleWestwood (also a myrmicine) by the common ancestor of Zeri-tis+Axiocerses; and (ii) to Anoplolepis Santschi (Formicinae)by the common ancestor of the genus Crudaria. In addition, thecommon ancestor of Clade E likely associated with one of theformicine genera Lepisiota Santschi (P= 0.51) or Anoloplepis(P= 0.27). Either of these species may have been the associatesof the common ancestor of Clade D (P= 0.31 and 0.14, respec-tively), with a subsequent return to association with Cremato-gaster by the common ancestor of Clade H (P= 0.41); however,there is little support for any particular scenario in this part ofthe tree. The common ancestors of all other genera were mostlikely to have associated with Crematogaster (Table S6).

As with the ancestral range reconstructions, missinglife-history data prevent us from drawing strong conclu-sions regarding the overall frequency of ant switching and itseffect on the evolution of the group.

Plant associations. The evolution of host plant use is lessclear. Most species in Aphnaeinae feed and/or oviposit on hostplants in the order Fabales, and this trait was likely sharedby the aphnaeine common ancestor (P= 0.66). As shown inTable S7, the most likely host plant order was Fabales formost common ancestors at the genus-level and higher, with theexceptions of Phasis (Sapindales/Geraniales, P= 0.51/0.46),Phasis+ Tylopaedia (Sapindales/Geraniales, P= 0.24/0.20),Erikssonia (Malvales, P= 0.76), Psuedaletis (algae, P= 0.96),Chrysoritis (Zygophyllales, P= 0.53), and Lipaphnaeus (Eri-cales, P= 0.45). However, support for Fabales is weak inmany of the internal branches, possibly because BayesTraitsrequires internal nodes to have a single character state. Lowsupport values for any single character state could be a resultof ancestral aphnaeines feeding on host plants from multipleorders. This is a trait certainly observed in modern species, foraphnaeines consume a broad diversity of host plants outside ofFabales. Life-history records include host plants from at least 23families in 19 orders, 16 of which are represented in our ingrouptaxa (Table 1). Species in the larger genera, such as Aloeides,Chrysoritis and Cigaritis, all associate with several other plantorders, and several individual species have been recorded tofeed on four or more orders of host plants. This great diversityof host plants, as well as limited life-history information [hostplant is unknown for 20 (25%) of our ingroup taxa] and limitedtaxon sampling, makes interpreting the evolutionary history ofplant associations extremely difficult.

Aphytophagy

A strikingly high proportion of the genera of Aphnaeinaehave species recorded as parasites of ants, either eating ant eggsand brood in the nest, or inducing trophallaxis from the antworkers (‘cuckoo feeding’). This specialization on ants is incontrast to the closely related Miletinae. All of the some 120species of miletines are aphytophagous, but the majority feedon insects associated with ants, such as aphids and scale insects,



and a smaller proportion feed on the ants themselves (Pierceet al., 2002). Although aphytophagy is widespread across theLycaenidae and Riodinidae – almost certainly as a consequenceof larval associations with ants – parasitic and predaceous lifehistories appear to be more common in at least two of theearlier-branching lineages of the Lycaenidae, the Miletinaeand Aphnaeinae. Life histories of the third early branchinglineage, the Poritiinae, are still too sparse to assess. Despiteits representation amongst these older lineages, this strategydoes not appear to be particularly persistent over evolutionarytime, as suggested by the ‘tippy’ distribution of aphytophagousspecies across the phylogeny of Aphnaeinae (Fig. 2) as well asthroughout the Lycaenoidea as a whole (Pierce et al., 2002).

Conclusion

A phylogenetic hypothesis based on molecular characters sup-ports the classification of Heath (1997) based on morphologicalanalysis, including the monophyly of most of the aphnaeine gen-era. The phylogeny also underscores the great variety of lifehistories within the Aphnaeinae, with many species having com-plex larval ant associations and diet and host-plant preferences.Plant feeding appears to have been lost multiple times withinthe group, typically by individual species within otherwise phy-tophagous genera. As we discover more information about thelife histories of these butterflies, this phylogeny will provide theframework necessary to explore how ant association and loss ofphytophagy may have affected subsequent diversification in theAphnaeinae and the Lycaenidae as a whole.

Supporting Information

Additional Supporting Information may be found in the onlineversion of this article under the DOI reference:10.1111/syen.12098

Figure S1. Phylogenetic hypothesis of Aphnaeinae: maxi-mum likelihood. Nodes are labelled with bootstrap support.

Figure S2. Phylogenetic hypothesis of Aphnaeinae: maxi-mum parsimony. Nodes are labelled with bootstrap support,with unlabelled nodes indicating less than 50% bootstrapsupport.

Table S1. Genes and primers used in molecular analyses.

Table S2. Annealing conditions used in PCR.

Table S3. Thermocyling program for BigDye reactions.

Table S4. Substitution models used in Bayesian Inference.

Table S5. Ancestral range reconstruction information forinternal nodes (as marked in Fig. 2) above the genus level,and the first node of Cigaritis, which separates African fromnon-African species.

Table S6. Ant associate ancestral state reconstruction infor-mation for common ancestors of each genus and otherinterior nodes marked in Fig. 2.

Table S7. Host-plant ancestral state reconstruction informa-tion for common ancestors of each genus and other interiornodes marked in Fig. 2.

Acknowledgements

We thank A.J. Berry, J.C. Choe, S.C. Collins, C. Congdon, R.Ducarme, A.J. Gardiner, S. Joubert, N.P. Kandul, T.B. Larsen,D.J. Martins, A.A. Mignault, A. Plowes, S.P. Quek, M.W. Tan,S.E. Woodhall and the late I. Bampton for providing materialand field assistance. We are grateful to a number of Nature Con-servation bodies for permission to collect specimens used inthis study, particularly those of the Western Cape and North-ern Cape Provinces of South Africa. We took advantage of theOdyssey cluster for many of our computations, supported bythe FAS Science Division Research Computing Group at Har-vard University, and of the Willi Hennig Society for the use ofTNT. J.H.B. was supported by a Smith Family Graduate Fellow-ship from the Department of Organismic and Evolutionary Biol-ogy of Harvard University; Z.A.K. was supported by an NSFGraduate Research Fellowship. T.R.S. was supported by a Rad-cliffe Research Partnership Grant and a Research Experiencesfor Undergraduates Fellowship from the National Science Foun-dation; M.E. was supported by a postdoctoral fellowship fromthe Research Council of Norway. This research was supportedby grants from the Baker Foundation, the Putnam ExpeditionaryFund of the Museum of Comparative Zoology, the TempletonFoundation (FQEB), and NSF DEB-0447242 to N.E.P.

References

Ackery, P.R., Smith, C.R. & Vane-Wright, R.I. (1995) Carcasson’sAfrican Butterflies: An Annotated Catalogue of the Papilionoidea andHesperioidea of the Afrotropical Region. CSIRO, Canberra.

Bogdanowicz, S.M., Wallner, W.E., Bell, J., Odell, T.M. & Harrison,R.G. (1993) Asian Gypsy Moths (Lepidoptera: Lymantriidae) inNorth America: evidence from molecular data. Annals of the Ento-mological Society of America, 86, 710–715.

Brower, A.V.Z. & DeSalle, R. (1998) Patterns of mitochondrial versusnuclear DNA sequence divergence among nymphalid butterflies: theutility of wingless as a source of characters for phylogenetic inference.Insect Molecular Biology, 7, 73–82.

Cho, S., Mitchell, A., Regier, J.C., Mitter, C., Poole, R.W., Freidlander,T.P. & Zhao, S. (1995) A highly conserved nuclear gene for low-levelphylogenetics: elongation factor-1 alpha recovers morphology-basedtree for heliothine moths. Molecular Biology and Evolution, 12,650–656.

Chou, I. (ed) (1994) Monographia rhopalocerorum Sinensium, Vol. 2.Henan Science and Technology Publishing House, Zhengzhou.

Colgan, D.J., McLauchlan, A., Wilson, G.D.F. et al. (1998) Histone H3and U2 snRNA DNA sequences and arthropod molecular evolution.Australian Journal of Zoology, 46, 419–437.

Corbet, A.S. & Pendlebury, H.M. (1992) The Butterflies of the MalayPeninsula. United Selangor Press, Kuala Lumpur.

D’Abrera, B. (1986) Butterflies of the Oriental Region: Lycaenidae &Riodinidae. Hill House Publishers, Melbourne.

Distant, W.L. (1882-1886) Rhopalocera Malayana: A Description of theButterflies of the Malay Peninsula. West, Newman, London.



Eliot, J.N. (1973) The higher classification of the Lycaenidae (Lep-idoptera): a tentative arrangement. Bulletin of the British Museum(Natural History) Entomology, 28, 371–505.

Folmer, O., Black, M., Hoeh, W., Lutz, R. & Vrijenhoek, R. (1994) DNAprimers for amplification of mitochondrial cytochrome c oxidasesubunit I from diverse metazoan invertebrates. Molecular MarineBiology and Biotechnology, 3, 294–299.

Gardiner, A.J. & Terblanche, R.F. (2010) Taxonomy, biology, biogeog-raphy, evolution and conservation of the genus Erikssonia Trimen(Lepidoptera: Lycaenidae). African Entomology, 18, 171–191.

Goloboff, P.A., Farris, J.C. & Nixon, H.C. (2008) TNT, a free programfor phylogenetic analysis. Cladistics, 24, 774–786.

Heath, A. (1997) A review of African genera of the tribe Aphnaeini(Lepidoptera: Lycaenidae). Metamorphosis, 8(Suppl. 2), 1–60.

Heath, A., Newport, M.A. & Hancock, D. (2002) Butterflies of Zambia.Lepidopterists’ Society of Africa & A.B.R.I., Nairobi, Kenya.

Heath, A., McLeod, L., Kaliszewska, Z.A., Fisher, C.W.S. & Corn-wall, M. (2008) Field notes including a summary of trophic andant-associations for the butterfly genera Chrysoritis Butler, AloeidesButler and Thestor Hubner (Lepidoptera: Lycaenidae) from SouthAfrica. Metamorphosis, 19, 127–148.

Kroon, D.M. (1999) Lepidoptera of Southern Africa: Host-plants &Other Associations: A Catalogue. Lepidopterists’ Society of Africa,Jukskei Park, South Africa.

Larsen, T.B. (1991) The Butterflies of Kenya and their Natural History.Oxford University Press, New York, NY.

Lewis, H.L. (1973) Butterflies of the World. Follett Publishing Co.,Chicago, IL.

Linder, H.P., de Klerk, H.M., Born, J., Burgess, N.D., Fjeldsa, J. & Rah-bek, C. (2012) The partitioning of Africa: statistically defined biogeo-graphical regions in sub-Saharan Africa. Journal of Biogeography, 39,1189–1205.

Monteiro, A. & Pierce, N.E. (2001) Phylogeny of Bicyclus (Lepi-doptera: Nymphalidae) inferred from COI, COII and EF-1alpha genesequences. Molecular Phylogenetics and Evolution, 18, 264–281.

Pagel, M., Meade, A. & Barker, D. (2004) Bayesian estimation ofancestral character states on phylogenies. Systematic Biology, 53,673–684.

Paradis, E., Claude, J. & Strimmer, K. (2004) APE: analysis ofphylogenetics and evolution in R language. Bioinformatics, 20,289–290.

Pierce, N.E., Braby, M.F., Heath, A., Lohman, D.J., Mathew, J., Rand,D.B. & Travassos, M.A. (2002) The ecology and evolution ofant association in the Lycaenidae (Lepidoptera). Annual Review ofEntomology, 47, 733–771.

Posada, D. (2008) jModelTest: phylogenetic model averaging. Molecu-lar Biology and Evolution, 25, 1253–1256.

Pringle, E.L., Henning, G.A. & Ball, J.B. (1994) Pennington’s Butterfliesof Southern Africa, 2nd edn. Struik Winchester, Cape Town.

Rambaut, A. & Drummond, A.J. (2007) Tracer v1.5 [WWW document].URL http://beast.bio.ed.ac.uk/Tracer [accessed on 31 August 2011].

Rand, D.B., Heath, A., Suderman, T. & Pierce, N.E. (2000) Phy-logeny and life history evolution of the genus Chrysoritis within theAphnaeini (Lepidoptera: Lycaenidae), inferred from mitochondrialcytochrome oxidase I sequences. Molecular Phylogenetics and Evo-lution, 17, 85–96.

Ree, R.H. & Smith, S.A. (2008) Maximum likelihood inference of geo-graphic range evolution by dispersal, local extinction, and cladogen-esis. Systematic Biology, 57, 4–14.

Riley, N.D. (1925) The species usually referred to the genus Cigari-tis Boisd. [Lepidoptera: Lycaenidae]. Novitates Zoologicae, 32,70–95.

Ronquist, F. & Huelsenbeck, J.P. (2003) MrBayes 3: bayesian phyloge-netic inference under mixed models. Bioinformatics, 19, 1572–1574.

Scott, J.A. (1985) The phylogeny of butterflies (Papilionoidea andHesperioidea). Journal of Research on the Lepidoptera, 23,241–281.

Stamatakis, A. (2006) RAxML-VI-HPC: maximum likelihood-basedphylogenetic analyses with thousands of taxa and mixed models.Bioinformatics, 22, 2688–2690.

Stamatakis, A., Hoover, P. & Rougemont, J. (2008) A rapid bootstrapalgorithm for the RAxML web servers. Systematic Biology, 57,758–771.

Stempffer, H. (1967) The genera of the African Lycaenidae (Lepi-doptera: Rhopalocera). Bulletin of the British Museum (Natural His-tory) Entomology, (Suppl. 10), 1–322.

Swinhoe, C. (1911) Lepidoptera Indica, Vol. 9. Lowell Reeve & Co.,London.

Talavera, G., Lukhtanov, V.A., Pierce, N.E. & Vila, R. (2013) Estab-lishing criteria for higher-level classification using the systematics ofPolyommatus blue butterflies (Lepidoptera: Lycaenidae). Cladistics,29, 166–192.

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S.(2011) MEGA5: molecular evolutionary genetic analysis using max-imum likelihood, evolutionary distance, and maximum parsimonymethods. Molecular Biology and and Evolution, 28, 2731–2739.

Tuzov, V.K. (1997) Guide to the Butterflies of Russia and AdjacentTerritories: Lepidoptera, Rhopalocera. Pensoft, Sofia.

Vila, R., Bell, C.D., Macniven, R. et al. (2011) Phylogeny andpalaeoecology of Polyommatus blue butterflies show Beringia wasa climate-regulated gateway to the New World. Proceedings of theRoyal Society B, 278, 2737–2744.

Wahlberg, N., Braby, M.F., Brower, A.V.Z. et al. (2005) Synergisticeffects of combining morphological and molecular data in resolvingthe phylogeny of butterflies and skippers. Proceedings of the RoyalSociety B, 272, 1577–1586.

Williams, M.C. (2012) Afrotropical Butterflies and Skippers – A Digi-tal Encyclopaedia. [WWW document]. URL http://atbutterflies.com[accessed on 6 July 2012].

Accepted 1 July 2014First published online 20 October 2014


phylogeny of the aphnaeinae: myrmecophilous african butterflies … · 2015. 12. 1. · ah-95-z951...

Documents