Research Collection

Doctoral Thesis

Floral specialization in solitary bees A case study of the osmiine bees

Author(s): Praz, Christophe

Publication Date: 2008

Permanent Link: https://doi.org/10.3929/ethz-a-005697352

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library Diss. ETH r. 17800

loral speialiatio i solitary ees A ase stdy o the osmiie ees

Christophe Pra

DISS ETH Nr 178

Floral specialization in solitary bees A case study of the osmiine bees

A dissertation submitted to

ETH RICH

for the degree of

octor of Sciences

presented by

Christophe Praz

Dipl Biol niversity of Bern Born 2th August 1979 Citien of Nenda, alais

accepted on the recommendation of

Prof. r. Silvia orn, examiner Prof. r. Alex Widmer, co-examiner r. Andreas Müller, co-examiner

2008

able of contents

1. Summary p. 1

2. Rsum p.

. eneral introduction p. 7

. Specialized bees fail to develop on non-host pollen Do plants chemically protect their pollen p. 15

5. Host recognition in a pollen-specialist bee evidence for a genetic basis p.

6. hylogeny and biogeography of bees of the tribe smiini (Hymenoptera Megachilidae) p. 8

7. atterns of host-plant choice in bees of the genus Chelostoma the constraint hypothesis of host-range evolution in bees p. 7

8. eneral discussion p. 108

9. Cited references p. 11

10. Acknowledgment p. 12

11. Curriculum vitae p. 125

mma

1. Summary

eeseloolleaetaotheeotooathethese lat ots the easl st loes a ae ths essetal ollatosoalaeootootheloelatsasoltaee sees estt the olle oa to a e elate lat sees ololeteesheeasothesaeeealstsolleteesollet s osee aataeos e eeee o a lmte meoollesoesttheatosseletoololetoat to mata t ema lael lea he tatoal elaato o ololet ostlates that loal sealato ees s e teseomettohshothesssaseotheoseatothat seesh ee ommtes hae the hhest ooto o ololet ee sees t assmes that ololet sees ese om ollet aestos a hee that ees hae seale the eolto hs hothess lasemaleee a ees theeamato heamotheesetthessstoestatheoaeoltoo ololetaseleteooeesthesmatooeaette estaoloalhoeseeseeal

he ttoalaltoollemaeleloalsealato eeslesteshaeestatethelaaleomaesosolta ee sees he eae o eet olle ets a the eslts ae otatoa o ot eale a lea ealato o the ls etee ololet a olle alt o aess ths ametal aset o osmeeeseeseahsealetoaeet hosteeoeto eelo o ohost olle ets hathe es o oe sees ee taseeototheolleaetaososoaotheseesolle ets o steaeae a Ranunculus alaeae oe to e aeate o all ee sees teste eet those seale o these latstheolleoSinapisassaeaeaEchiumoaaeae ale to sot laal eelomet oe ee sees seale o Campanula Camalaeaeheseesltsstolsestthatolleo theseotaoomosossessesoteteoetesthathame estoheeoeollehemalomostosostlatetostol leetheollehostsetmoees

. u

n ese eeiens ee e ee secies Heriades truncorum cu ee n n nns en ies inicing e nuiin qui en es n enie ccun igec in is secies. is sein ises in quesins eging e neugic cess uneing s ecgniin n seecin in si ees. is e e gn n nns en i ece us e nee een in cnc i e en ei seciic s. is enes e inesigin ecuen quesin in s ecgniin in si ees ne ee eeience uing e e ie sges inuences cices in u ees iining e. ecn i eins unce ee ees H. truncorum i cec suie nn s en in e sence ei n ss e Asecee. ese secs ee inesige cecie in cice eeiens in cges cing e eeences us ee n s en i se iniius gn n ieen es nns en Echium n Campanula. ees ee e nes in cges ee s n nns es ee ie. A ees egess ie esice en cecin ei s ug e isie e es s n nns ns nec. en ee n e nn s en suce ees cese nesing ciiies. s cice in Heriades truncorum us ee s suising cnsee i s n inuence eeience n igec sees e se n neugic e n nuiin cnsins in is secies.

igec in si ees is eee eine ieen ecniss sigic cnsins ee en igesin n neugic iiins ee secgniin e ning. ssess e signiicnce ese secs in n euin cne e gen e siini s ecnsuce. e einsis eeen e in gene e siini ie ee inesige sing in cnesies eging e cssiicin is ee ie. e genus Chelostoma eege s e s iiie inege e siini. en nses secies Chelostoma eee is genus is in cse igecic secies ie cing e incesing eience igec n n ec is e nces se in ees. ue e ineence e

2 1

35 Chelostoma Chelostoma

3 . u

2

i nourrint ur r c du on t du nctr t ur iit incnt d nt ur ont d d oinitur nti. or di oitir ont utnt cii dn ur coi oru iitnt our ur rcot d on un u nr ou un u i d nt. ont c oioctiu. utr c dit octiu ont nrit t iitnt un u rnd rit d ur. i nt outi oci un ri nrit donc oin dndnt dun u rourc innnt cint rit ctur ui ont ntrn ou ui intinnnt ciition or rtnt rnt incori. rditionnnt ciition t intrrt co un rt d rourc ntr d c di n cotition donc co un rocu dcount d ort cotition ntr c. n t cot u ric n c di ritnt u ut roortion dc oioctiu. rnt tri tnt ducidr oriin d ciition or u in dun rou di oi t ini d iu corndr coi oru c i n nr.

uit nutriti du on t otntint un ctur iortnt our iur ciition or d i. ndnt u dtud ont nc ur ur nutriti du on our i oitir. our tntr ducidr c oint utr c doi ccun cii ur un dirnt t d ur ont t orc dor ur du on nont u dun c urnt dc ur roiion d on t d nctr rcot r un d utr c. ucun r n u dor ur on d coo trc t d rnoncu nuncuc u c d c rctint cii ur c du nt. n outr un c di n u dor ur on d outrd d c ricc t d irin orinc. rutt urnt u on d c utr rou d nt od un cooition ciiu ui rnd diici dition t on iition r i. r conunt ur nutriti ou

s cst c ct rbabt s c ra c s abs stars

ar ctr s scs css eriades truncorum a s r sr rts ts t c sr a scasat c ctt sc t as tr t ar a cst rstat r sts tcat a rcassac a att c s abs stars rrt s ars t ra sr t t rr s ats t aas t ctact ac r t sc c rt tr c rc acs rat s rrs stas t sr s rrcs ras s ats a st st sar s s s eriades truncorum t rctr c t trtt aat r rs ars absc r att abt s rrcs ras ats s sr t t t t t cars rs rc ca s s t cstrr rs s as s cas rsc atsts s Astraca t ts s caas Campanula s rs Echium ts s s crs cs s sr t t rct t sr r t ra absc r t abt s s t aba r t rs rctr sr t ccs s c ra eriades truncorum rrstt crtt tt trs csr b t ss c artssa rat s rrs stas sct a scasat aarat c ctt sc c t ar ss caacts rs artssa rcassac t tt ctr ar a at

a scasat s abs stars st c rbabt tt cass rr s aat trat a st t r tcat a c t A trr rtac rat cs ascts t t a s ss t t s rats tr s rca rs t t tabs a a a c rt rsr crtas ctrrss sr a casscat cs

. sum aeilles. e enre apparat comme tant le enre le plus ancien au sein des osmies. es analses palnoloiues pour espces de indiuent ue ce enre est constitu en rande maorit despces oliolectiues, ce ui renforce lhpothse ue la spcialisation est un caractre primitif che les aeilles. e plus, ltalissement des relations phlontiues entre ces espces permet de suire lolution de leurs choi florau. es choi apparaissent comme hautement consers les chanements dhtes sont rares et se passent dune faon trs prcise, ce ui indiue ue seul un choi limit de fleurs peut conenir au eiences phsioloiues et neuroloiues dune espce daeille. eu contraintes lies la diestion du pollen et la reconnaissance de lhte rissent proalement les choi florau des aeilles du enre .

a comprhension de ces mcanismes dpasse la comprhension de la spcialisation che les aeilles. es mmes contraintes ouernent les choi florau des aeilles dune manire nrale, aussi ien des espces oliolectiue ue pollectiues. Par consuent, le prsent traail offre de nouelles pistes pour mieu comprendre la lonue histoire olutie des fleurs et des aeilles.

3

3. General introduction

31

15000 2007

2007 2000 1996 2005 1996

2005 95 Campanula rapunculus 4 2006

7 . ntoction

ollen eieents o secies o solitay ees an shoe that the aoity o the eie the entie content o oe than loes to oision a single oo cell ths to oce one osing. he lage ollen eieents o ees as illstate y these to sties hae iotant ilications on the nestaning o eeloe elationshis.

hogh ollen collection ees act at the sae tie as ollinatos an heioes. s a conseence eeloe elationshis ae not eely talistic t ae ette escie as a alance tal eloitation estea . eeloe elationshis ange o a total eloitation o ees y loes e. g. in ecetie ochis hich attact ees t o not ea the to a colete eloitation o the loes y the ees in cases hee ees steal loal esoces ithot ollinating the lant e. g. in sall ee secies that aoi contact ith the stiga. otless the aoity o the inteactions lies soehee eteen these to etees. oee each atne is eecte to ty to ti the alance in its on ao as ollination an ollen collecting hae a ccial iact on the eoction o oth loes an ees.

olitay ees isit loes to collect ollen an necta theey acting as heioes an ollinatos. his ee secies oliis adna is oligolectic it stictly esticts ollen collection to loes o the ies gloss im lare. Picte y neas lle.

3

2006 Salvia Salvia 1989 10 1995 2000 2007 Salvia Lasioglossum convexiusculum Salvia 10 1997

32

1994 1925 1958 1958 1972 1980 1989 1996 2005 2008 1996 2006

9 3

Salvia, Salvia asioglossum convexiusculum Salvia 1989

10 . ntrouction

ligolect an pollect are foun in all iportant ee lineages as ell as in all ee counities inestigate ichener 2 olene estrich suggesting that oth strategies ust e successful in eolutionar ters. p to 6 of the ee species are oligolectic in soe ecosstes ichener ut the eolutionar forces shaping hostrange an ietreath in ees reain largel unclear. nerstaning hich floers are isite hich ees an h is of oious iportance for the coprehension of the eolution of eefloers relationships an hence of pollination in general.

.. R

n the asence of a phlogenetic fraeor earl stuies hae sought eolutionar eplanations for pollen specialiation in the liite eience proie faunistic an iogeographic consierations the proportion of oligolectic species is higher in ari ecosstes here the ee fauna is rich ut the floral resources are liite than in teperate or tropical ecosstes hich hae less ee species ut higher floer suppl an iersit ichener olene . his oseration has supporte the traitional ie that floral specialiation as rien interspecific copetition eteen ee species speciesrich ee counities oul ten to partition resources through specialiation to arious hosts Roertson 2 insle horp 6 ichener cislo an ane 6. his hpothesis relies on the unteste assuption that oligolectic ees hae specialie on their hosts or in other ors that the eolution goes fro pollect to oligolect olene .

oeer the initial stuies on ee phlogen reeale a surprising pattern contraicting this iportant assuption the priitie eers of the aorit of all ee failies appeare to e preoinantl copose of oligolectic species esterap ratochil estrich cislo an ane 6 hich as soon interprete as an inication that oligolect ight e the priitie trait in ees an not pollect. nee the first stu tracing floral specialiation in a group of ees ith a phlogenetic approach has reeale three transitions fro oligolect to pollect ut none fro pollect to oligolect ller 6. ince then

3

2006 2006 2007 2008 2008 2003 2006 1996 2006 2007 2008 2008 1975 1980 2006

1996 2006 1 2 3 4 5 1996

12 . Introduction

.4. OTIN O T PRSNT TSIS

The Osmiini present two features that mae them outstanding model organisms for the study of bee-flower relationships. irst, they constitute a species-rich tribe, with numerous species showing very diverse host specialization. Many species are floral specialists on various plant taxa (Westrich 18, Mller et al. 17). Other species are generalists, whereas many still show a distinct preference for a certain plant family (Mller et al. 17, Amiet et al. 2004). The high richness and variety in foraging behaviors offers many opportunities to investigate the evolution of floral associations in the Osmiini. Second, many osmiine bees nest under natural conditions in existing burrows in dead wood, in contrast to the maority of other solitary bees, which dig burrows in soil or construct surface nests on rocs. onseuently, several osmiine bee species accept artificial nesting sites, which facilitates encaged rearing and allows the manipulation of cell contents (e. g., egg transfer between pollen loads).

In the present thesis, the physiological mechanisms associated with host specificity are investigated in selected osmiine bee species. Oligolectic species were reared in cages to allow for observations and experiments on their oligolectic behavior. A phylogenetic approach is then used to place these physiological and behavioral findings into an evolutionary context. In addition to the reconstruction of the historical relationships between species, phylogenetic inference enables the historical reconstruction of selected traits, in this study floral specialization. ence the present thesis combines nowledge gained on the physiological basis underlying host selection in bees with a phylogenetic approach. In the past, such studies have proved extraordinary useful to unravel host specialization in herbivorous insects (anz et al. 2001).

In the first chapter, ioecticeesaitodeeoonnonostoen doantscemicayrotectteiroen a ey aspect of host-specialization in bees will be investigated, namely pollen nutritive value. Although strong evidence exists that pollen widely differs in its nutritive value to the most pronounced polylectic bee nown, the honeybee (reviewed by erbert 12, Roulston and ane 2000), the role of pollen chemical composition in

1 3

ost-recognition in a specialist bee species: evidence for a genetic basis

Phylogeny and biogeography of bees of the tribe Osmiini,

The evolution of host-plant relationships within bees of the genus helostoma helostoma 50 helostoma

14 ialist bs on nonost olln its

. ecle ees l t eel n nnst llen D lnts cecll tect te llen1

s ruir lar amounts of olln for tir on rroution il sral moroloial flor traits ar non to a ol to rott lants aainst ssi olln arstin b bs littl is non on o sltion to minimi olln loss ats on t mial omosition of olln n tis stu tra t laral lomnt of four solitar b sis a siali on a iffrnt olln sour n rar on nonost olln b transfrrin unat s of on sis onto t olln roisions of anotr sis Polln its of straa an anunulaa ro to b inauat for all b sis tst t tos siali on ts lants urtr olln of rassiaa an orainaa fail to suort laral lomnt in on b sis siali on amanulaa Our rsults stronl sust tat olln of ts four taonomi rous ossss rotti rortis tat amr istion an tus alln t nral i of olln as an astous rotin sour for flor isitors

OO

rat maorit of florin lants rl on insts or otr animals for ollination is intration as sa t olution of bot t aniosrms an tir ollinators sin t ris of t florin lants in t arl rtaous oltis t al mon insts bs ar t most imortant ollinators ain robabl oriinat in t rtaous anfort t al t sar a lon an intimat olutionar istor it t aniosrms or t rlationsis btn bs an flors ar not mrl mutualisti nou stram a or rin t al but ar bttr i as a balan mutual loitation stram

as on Pra llr an orn olo

4. Specialist bees on non-host pollen diets

ees are above all herbivores. They store pollen and nectar as the exclusive food source for their larvae. The quantity of pollen withdrawn from flowers for bee reproduction is huge. For example, as much as 95.5% of the pollen produced by flowers of amanla ranls was removed by bees, while only 3.7% contributed to pollination (Schlindwein et al. 2005). n another study, 85% of 41 bee species examined were found to require the whole pollen content of more than 30 flowers to rear a single larva, and some species even needed the pollen of more than 1000 flowers (ller et al. 2006).

ees not only collect large amounts of pollen, they also collect it very efficiently (Westrich 1989, ller 1996b), which frequently conflicts with the successful pollination of the flowers. n some proterandric flowers (i.e., those with stamens coming to maturity before the pistil) for example, female bees restrict their foraging to flowers in the male phase, thereby scarcely contributing to pollination (ller 1996a). Similarly, many bee species act as pollen thieves due to morphological incongruences between the flowers and the bees or inappropriate bee behavior (incley and Roulston 2006, and references therein). n addition, bees carefully groom their body after having visited a flower and transfer the pollen grains into specialied hair brushes, maing them generally inaccessible for pollination (Westeramp 1996). Consequently, specialied bee flowers must balance the need to attract bees for pollination, on the one hand, and to restrict pollen loss to bees, on the other. Plants are thus expected to evolve adaptations to minimie pollen loss by narrowing the spectrum of pollen feeding visitors. ndeed, several morphological flower traits can be viewed as adaptations preventing excessive pollen harvesting by bees: heteranthery, where showy anthers provide fodder pollen while inconspicuous anthers produce pollen for fertiliation (ogel 1993); concealment of the anthers in nototribic flowers (i.e., flowers in which the stamens and style are placed below the upper lip in order to come into contact with the dorsal surface of the foragers body; ller 1996a), in narrow floral tubes (e.g., oraginaceae; Thorp 1979, ller 1995), or in eel flowers (e.g., Fabaceae; Westeramp 1997b); concealment of the pollen in poricidal anthers (i.e., anthers releasing pollen through a distal opening; uchmann 1983, arder and arclay 1994); and progressive pollen release

16 peialist ees on nonhost pollen diets to ore pollinators to repeatedl isit the loers rar and eins hlindein et al

eletion ma as ell at on pollen toiit or on pollen ntritional alit to preent eessie pollen olletion Althoh the presene o seondar omponds in netar has reentl reeied onsiderale attention e Adler rin et al Adler and rin ohnson et al little is non o ho the hemial omposition o pollen inlenes pollen se ees or other insets t see etel and Win ernal and rrie ndiations eist that pollen is not an eas tose protein sore readil diestile or all loer isitors ein and Hada olston and ane oo et al n at srprisinl e inset taa rel on pollen as a sole protein sore renn et al rthermore i pollen ere a mere reard to loer isitors it shold ontain etra protein Hoeer pollen protein ontent has een ond to e assoiated ith the need or pollen te roth rather than to reard pollinators olston et al

an ee speies are pollen speialists olioleti and restrit their pollen orain to e related plant speies elonin to the same amil Westrih Wislo and ane ane and ipes n ontrast polleti ees hae a roader host rane that enompasses at least to plant amilies Hoeer man polleti ees still sho a restrited rane o pollen sores Westrih ller a ane and ipes pollen ere an easil diestile sore o protein the larae o oth olioleti and polleti ees shold e ale to deelop on nonhost pollen Onl a e oserations on the perormane o ee larae on non host pollen hae een reported o larae o the olioleti ee re normall on nonhost pollen oen a ee speies stritl speialied on Onaraeae deeloped on pollen o aaeae ohart and osse and the larae o the Asteraeae speialist deeloped on pollen o Hdrophllaeae and rassiaeae Williams h sparse eidene led to the tentatie sestion that loral speialiations in ees are not lined to the hemial omposition o the pollen Wislo and ane inle and olston Hoeer

4. pecialist bees on non-host pollen diets the findings that the larvae of Osmia liaria failed to develop on five different non-host diets (evin and Haydak 157) and that the larvae of eachilerotudata failed to grow on pollen of Asteraceae (Guirguis and Brindley 174) suggest that this assumption is possibly premature.

In the present study, the larvae of four oligolectic osmiine bee species (egachilidae Osmiini) were forced to feed on non-host pollen. Given that strict oligolectic bees may refuse to harvest pollen in the absence of their specific host plants (e.g., trickler 17, illiams 2003), we removed unhatched eggs from the brood cells and transferred them onto non-host pollen and nectar collected by one of the other species. From the observed patterns in larval survival, we infer possible protective properties of pollen and discuss their potential implications to our understanding of bee- flower relationships in general.

4.3. THO

4.3.1. Bee species and nest establishment

e selected five bee species (Table 1) belonging to the tribe Osmiini (Apoidea, egachilidae). They are widespread and common throughout urope, and their foraging behavior is well documented (estrich 1, and references therein). All species are strictly oligolectic, restricting pollen collection to a limited number of related plant species. Heriadestrucorum is an Asteraceae specialist preferring flowers of the Asteroideae. Chelostoma rauculi, Chelostoma florisome, and Holitis aduca, which are all oligolectic at the plant genus level, collect pollen exclusively on Camaula (Campanulaceae), auculus (Ranunculaceae), and chium (Boraginaceae), respectively. Osmia reicoris is a broad oligolege of the plant family Brassicaceae. ach bee species was allowed to build nests separately, in a cage made of gauze (10 x 70 x 120 cm), outdoors in urich, witzerland. The bees originated from several different localities in witzerland. e provided potted plants of suitable host species as pollen and nectar sources (Table 1) and hollow bamboo stalks as nesting sites. To avoid mixed provisions, we provided only one plant species in each cage at a given time.

1 . eialis ees n nns llen ies

B . e ie selee ee seies ei lal seialiain e lan seies se nes esalisen an e nns llen n i e laae ee e eel.

Bee seies eii s lan seies nes ns llen ese lan esalisen seaeae aanlaeae innaes annlaeae Bainaeae Bassiaeae

seaeae eeleie annlaeae Bainaeae Bassiaeae

seaeae innaes aanlaeae Bassiaeae

seaeae ane Bassiaeae aiis

.3.. anse ae laal eane n nns llen e ansee nae es e ells ne seies e llen an nea isins ane seies. e aelly e ea e i a in sala n llen a e a aen ne nes an eisly lae in an aiiial ell a eille lay l ae i aain i an Bs . llen aae e e as ee ai nainain. e esiae e llen eieens ea e ieenly sie ee seies y ain e y y asses e al eales aa lle e al. . n e ae ases ee e laa nse e enie isin ee enin is eelen aiinal llen as ie. e se nae es e ness esalise in e ae aes an aiinally ness llee in e il ness ee llee in ee ses ae s a leeens in esen ielan an ness an a sals a ieen laliies in nen ielan.

. pecialist ees on nonhost pollen diets

... onhost pollen tested

We compared the laral performance of . ror and . rapi on the same fie pollen tpes (ale ). arae of . orioe ere reared on three different nonhost pollen diets and larae of . aa on pollen of Bphha (Asteraceae). We otained too fe es of . reiori to assess its deelopment on different pollen diets and ths sed this species as a sorce of rassicaceae pollen onl. As controls e transferred es of each ee species onto its host pollen folloin exactl the same procedre as for e transfer onto nonhost pollen.

he pollen for or experiments oriinated from the ee nests in the aed caes. Additionall e sed pollen of a ari hi gare and Asteraceae from nests collected in the ild of . orioe .aa and . ror respectiel (see ethods E transfer). he se of pollen of Asteraceae from nests in the ild as restricted to control diets for . ror. Wheneer pollen from natral nests as sed the rood cell proisions ere analed microscopicall to confirm pollen prit. he pollen as stored at ntil se. We acnolede that or method confonds the impact of pollen and that of nectar on laral performance. Hoeer as the pollen proisions collected . ror . rapi . orioe and . reiori contain onl little amonts of nectar e postlate that the impact of nectar on laral performance as neliile. n contrast the pollen proisions of . aa contain considerale amonts of nectar. n this species e cannot exclde that nectar miht hae inflenced laral performance.

... aral deelopment

E hatchin and laral deelopment too place in the artificial cells in a climate chamer in the dar at and relatie hmidit. he es and larae ere checed eer to or three das. At each occasion e recorded hether the e had hatched hether the lara as alie and feedin and hether the lara had defecated or started to spin a cocoon. Es that did not hatch ere exclded from all analses. ead larae ere also exclded if the had ndotedl died from external factors sch as mites fnal roth in the pollen proision or mared chanes in the

4. pecialist bees on non-host pollen diets consistency of the pollen. As bee larvae go through several instars that are difficult to separate, we only distinguished four developmental stages (1)feeding, non-defecating; (2) feeding, defecating; (3) non-feeding, spinning cocoon; and (4) immobile, diapausing in completed cocoon. Cocoon spinning was discriminated from the production of silk strands to fix feces during the feeding phase. After the cocoon had been completed, cells were kept for 1 days in the climate chamber and then stored at 4C for overwintering. Although illiams (2003) recommended larval weight as the best surrogate measure of fitness to assess the performance of bee larvae on different pollen diets, we refrained from weighing the larvae as preliminary trials indicated that such handling induced higher mortality.

4.3.. Data analysis

For each individual, we determined the hatching date as the average of the two observation dates between which the egg hatched. e followed a similar protocol to assess the date of the onset of cocoon spinning and the date at which the larva entered diapause. was calculated as the number of days between hatching and onset of cocoon spinning, and as the number of days between hatching and diapause. For those individuals that died before the end of their larval development, pre-diapause life length was calculated as the number of days between hatching and death.

e used aplan-Meier statistics to compare survival of a bee species on host and non-host pollen following ee and ang (2003). e considered the parameter prediapause life length as censored data individuals that died before the end of their larval development were the exact observations for which the event (death) occurred, while those that reached the diapausing stage were the censored observations. The latter were thus withdrawn from survival calculations once the development was completed, which reflects the fact that diapausing larvae are much less exposed to mortality risk than feeding larvae. To test for differences between survival distributions, we used the log-rank test when comparing two groups and the -sample test implemented in (command ) when comparing more than two groups. For statistical analyses, we used

21 4. pecialist bees on non-host pollen diets the statistical package ( evelopment ore Team 2006) and P 11 (P 2005) for Macintosh O .

4.4. ET

In total, we transferred 405 bee eggs, of which 15 (77.7%) hatched. The survival of the individuals transferred onto their host pollen (controls) amounted to 80% for each of the four species (Table 2), indicating that egg transfer had little negative impact on larval development. The development time on host pollen significantly differed between the bee species (NO with uncan post hoc tests, F 29.821, df , 57, P 0.001; Fig. 1). It was longest for the steraceae specialist eiaes tcm, by far the smallest of the bee species tested, and shortest for estma aci and itis aca, the latter being the largest species investigated.

4.4.1. eiaes tcm

arval survival in the steraceae specialist . tcm significantly differed between pollen diets (aplan-Meier analysis, -sample test, 2 104, P 0.001; Table 2). No single larva reached the diapausing stage when reared on acs pollen (Fig. 2), and survival was significantly lower than on any other pollen diet (Table 2). The larvae fed acs pollen grew normally for an initial period of 4 days, then mostly turned green and eventually died after 57 days. In contrast, larval survival did not significantly differ between diets of host pollen and of those of the other non-host pollen species, namely of amaa, cim, and Siais (Table 2). On these non-host diets, the larvae were not visibly different from the controls in their size or general appearance. evelopment time was significantly longer on pollen of inapis than on the three other pollen types (NO with uncan post hoc tests, F 7.526, df , 7, P 0.001; Table ).

22 4

2

Heriades Asteraceae 21 19 905 4567 253 0001 trcorm Camala 30 23 767 3497 261 acls 19 0 00 684 078 chim 19 15 789 3550 164 iais 25 18 720 3861 431

Chelostoma 23 19 826 2026 149 0001 racli hthalmm 18 0 00 1242 180 acls 20 0 00 483 055 chim 14 0 00 864 071 iais 21 1 48 1381 124

Chelostoma 25 21 840 2522 153 0001 lorisome aacetm 18 0 00 942 074 Camala 25 2 80 1510 049 rassica 11 6 545 2847 208

Holitis 7 6 857 1875 210 0001 adca hthalmm 19 0 00 545 043

Holitisadca 005 Chelostomalorisome

1 Heriades trcorm Chelostoma racli Chelostoma lorisome Holitis adca 005

23 Specialist ees on nonhost pollen diets

A aral deelopment time numer of das from e hatchin to onset of cocoon spinnin for all ee species hen reared on host shon in oldface and nonhost pollen

eelopment time d Group heteroeneit ee species Pollen diet No ees ean S edian rane P Groups

Asteraceae a a a

ampaula

auculus a a

Echium

Notes or normall distriuted data the mean alue S is ien for data not normall distriuted the median and rane are ien Group heteroeneit as tested either an ANOA ith uncan post hoc tests or a rusalallis test ith pairise annhitne test and onferroni correction or each ee species, roups sharin the same letter did not differ sinificantl at P

G Cumulatie surial of larae of the Asteraceae specialist hen reared on host pollen control and on four nonhost pollen diets Crosses indicate indiiduals that completed deelopment and entered diapause censored data

4

442 Chelostoma rapunculi

Campanula C. rapunculi 2 H. truncorum k- 2 938 0001 2 Sinapis 25 14 3 6 10 uphthalmum Sinapis 30 3 uphthalmum anunculus chium 3

3 Campanula Chelostoma rapunculi

25 eiaiseessedies

arasriae seiais siiia diered eee e dies aaeier aasis sae es aeediesseraeaeresedi eesarasriaaraeedereaiedisi sa aded a rae e ad died s aer e iiiai deeairasearaeddeeese ad ae e deee ie e as siiia srer a eier s e r e rsaais es ae siiia diere r e deee ie ees rreedd Heer aerse reisiisaraesddedied i a ss aeari e erise irred d a s ia r ard disease i siar ees ired s ersa iai e erere sae a e is a arriae die r e sered raieiaariaasedaiei

e seiais aied dee e aeearaeaedadsaredeedadied iiaedaserereaiedeeaisae

issdearssa edesrerese aiersa eiae resre r ees e seraeae did a r e deee ad adeasaiadeaedier ad addii aied dee eadseriiairiediere aaiiessrdierediissiaiiasaradreesad ereeseiesdieredieiraiisesereir deee

4

2000 47955 2000 chium vulgare 2006 117344 244 2000 inapis anunculus 1985 2000 2006 Campanula C. rapunculi H. truncorum H.truncorum H. truncorum C. rapunculi H.adunca Campanula chium

H. truncorum anunculus 1985 C. rapunculi chium inapis uphthalmum asioglossum zephyrum 20 39 2002

2000

27 4

451

. truncorum . rapunculi Ranunculus Ranunculus Ranunculus 1941 2004 Ranunculus acris 1995 2004 12 Ranunculus Pulsatilla Anemone sylvestris Ranunculus 2004 Osmia cornuta Osmia bicornis Ranunculus 1989 1997 Ranunculus .florisomne Ranunculus

28 4. ecast ees hst e ets

the e rtae s the a ate cue the et erstr et a. a sustace cat the e ras a at taa. he eree t hch the et s este ees ares a seces s a e a reereces there. s a e u that the et the e as cete este arae . ctrast as 3 sueste that the et e eatheae steraceae r checas th t cu terere th the utret assat rcess reer ts est arae cut. ur stu the aure t ee e a ht sar reate t the hh auts et tca steraceae a rasscaceae as 3. he e ras the aauaceae the secc hst ats cta tte et . er uushe ata. he t s rece that the ature the et ctrutes t the aheseess the e resut the rat e cus r a etter aherece t the the atrs reee ac a esse the uct the ate cus reset the et s t u uerst. s these ates are resse r the secc e r the are assue t uct ether as a checa cue t attract atrs s r as a eterret aast at e eeers ete a 3 s a erstr . he act that these ates ccur see hh ccetrats the e se eeate ers usess ht suest a eese echas aast ecesse e cect ees rather tha a aatat t attract the. ar the ets ccurr ether the et r the e a ht ctrute t t the rtect the e ras aast u actera r raat tae a ses 4 ut as t e eese aast e eeers. ee the arae th a ate a rae hue he rase steraceae e suest that the e ets ha accuuate ther . Further stues are eee heer t car t hat etet the aure these t seces steraceae e as ue t these ets.

. Specialist bees on non-host pollen diets

Similar pollen defense mechanisms may also underlie the rapid mortality of the larvae of rapunuli fed pollen of ium ular. This pollen is particularly rich in protein (Somerville and Nicol 200), but contains toxic pyrrolizidine alkaloids in very high concentrations (Boppré et al. 200). The fact that H trunorum could successfully develop on these provisions again demonstrates the varying physiological abilities of different bee species to use non-host pollen.

..2. ack of essential nutrients

Essential nutrients might be lacking or underrepresented in some pollen types, thereby limiting full development of the bee larvae. ack of essential sterols in the pollen was postulated to account for the observation that honey bees did not forage on rbutus uno (Ericaceae) for pollen, although they collected nectar from it (Rasmont et al. 200). Similarly, pure pollen diets of araxaum oiinal (Asteraceae) proved to be inadequate for both honey bee adults and larvae (reviewed in Roulston and Cane 2000). The low nutritional value of araxaum pollen is probably due to deficiencies in several essential amino acids, as pure araxaum diets that were experimentally supplemented with the essential amino acid arginine proved to be appropriate for larval development of the honey bee (Herbert et al. 170).

To determine whether similar deficiencies in the content of essential amino acids occur in the pollen of other species of the Asteraceae, we analyzed the large data set of Wille et al. (1), which provides the concentrations of nine essential amino acids in the pollen of plant taxa belonging to different plant families. Of the 1 Asteraceae taxa included, nine were severely deficient in arginine, including all members of the Asteroideae. Among the 20 taxa having the lowest arginine content, 1 were Asteraceae. Similarly, the content of phenylalanine in the pollen of Asteraceae was in the lower range of the observed values. Thus, deficiencies in the content of essential nutrients, possibly of arginine, may substantially contribute to the failure of the larvae of rapunuli, lorisomn, and H auna to develop on Asteraceae pollen. ndeed, the long survival of these larvae in combination with their very slow growth

0 4

453

1994 1994 1978 1994 1997 Taraacum 1985 Taraacum C. rapunculi C.florisomne H. adunca

454

1970 1985 2006 2005 1957 1974 153 Helianthus 86 1980 60 Colletes 1330

31 4

1980 1989 1996

455

1989 1996 2006 2008 1996 Chelostoma 7 Ceramius caucasicus 2006 Lasioglossum 2003 2006 60 2006

32 ecialist ees on nonhost ollen iets constraint reenting ees from ecoming ollectic Hence e urge careful reconsieration of the assumtion that ees are not secialie ue to the chemical comosition of ollen cislo an Cane Mincle an oulston s alrea suggeste oson an eng oth the nutritional alue of the ollen or the ailit to metaolie toic ollen chemicals ma unerlie floral associations in ees

CC

he general ie of ollen as an eastouse rotein source that is reail igestile for all floer isitors shoul e consiere ith caution he results of our rearing eeriments suggest that ollen might e rotecte chemicall seconar comouns eg the lac of essential nutrients eg steraceae anor structurall ollen alls resistant to igestion e ostulate that the enormous ollen reuirements of ees ma hae selecte for such rotectie roerties in ollen hese conclusions oen a ne fiel of research in the stu of insectfloer relationshis n future more attention shoul e ai to oth the chemical comosition of the ollen an the hsiological caailities of the ollinators to igest an utilie the ollen

Host reoitio i

5. Host recognition in a pollen-specialist bee: evidence for a genetic basis2

BC

o iestiate the effet of laral polle diet o floral hoie i a speialied bee speies e ompared the floral preferees of idiiduals of (eahilidae) reared o host polle ith those of idiiduals reared o to differet tpes of ohost polle emales ere alloed to est i aes here both host ad ohost floers ere aailable ll females reardless of laral diet restrited polle olletio to their host althouh the isited the floers of both host ad ohost plats for etar he offered ol the ohost polle soure females eased esti atiities ales reared o ohost polle elusiel restrited their patrolli flihts to floers of their ormal host his stud proides the first empirial iestiatio of the impriti theor i olioleti bees ad uambiuousl suests that host reoitio has a eeti basis i e disuss the impliatio of this fidi for the uderstadi of beefloer relatioships

OCO

he maorit of herbiore isets feed o a restrited rae of plats (Beras ad Chapma ) ad thus must hae the abilit to idetif their hosts at some stae of their life most ases host speifiit is mediated b plat hemials hih ma either attrat isets stimulate feedi or eliit oipositio (Beras ad Chapma ) arl obseratios hae poited out that isets ofte sho a preferee for those plats o hih the deeloped (alsh Hopis ) his has led to a importat uestio i the field of platiset iteratios amel to hat etet the resposes of adult isets are eetiall fied or are iflueed b leari or oditioi eeral hpotheses hae bee formulated to eplai ho these preferees a be idued irst the Hopis host seletio priiple (ethier Barro ) assumes the persistee of eural haes amel sesitiit to a hemial

Based o Pra C ller ad or submitted to

5. Host recognition in rias rncorm compound, from the larval stadium to the adult stage. This mechanism remains controversial (reviewed by Barron 2001) and it has been demonstrated in very few cases only (Rietdorf and Steidle 2002, Villagra et al. 2007) including a hymenopteran species (Gandolfi et al. 2003). Second, the neo-Hopkins principle (Jaenicke 1983) and the chemical legacy hypothesis (Corbet 1985) suggest that the conditioning occurs during the early lifetime of the imago. This can be at or shortly after emergence, provided the pupa is in close contact with the host plant or remains of it. There is substantially more evidence for early adult conditioning, either for host-plant preferences (e. g., Solarz and Newman 2001, Olsson et al. 2006) or for other responses to volatiles (Jaisson 1980, Cortesero et al. 1995, Rse et al. 1997, Bjorksten and Hoffmann 1998, Breed et al. 1998).

Bees rely entirely on plant products for their reproduction, and are thus herbivores. Many bee species are floral specialists and restrict their pollen foraging to a few related host plants. These species, the so-called oligolectic bees, show narrow relationships with their host, which is not only the lone food source for their offspring, but also often acts as rendez- vous place for mating. The males of many oligolectic bees patrol the flowers of their females pollen hosts to locate conspecific females (Westrich 1989). Host recognition is therefore a key process in oligolectic bees, yet it has not been fully elucidated. In the light of earlier studies that demonstrated the importance of pollen odors in flower selection by the pollinating insects (e. g., Plateau 1897, von risch 1919), Linsley (1958) suggested that host recognition by oligolectic bees could rely on pollen odors rather than, or in combination to, floral color or morphology. Indeed, this was confirmed later for several oligolectic bee species (Dobson 1987, Dobson and Bergstrm 2000, Dtterl et al. 2005, Dtterl and Schffler 2007). The prediction of Linsley relied on the idea that the only prior experience which the newly emerging solitary oligolege has had with its appropriate pollen source is the pollen and nectar which were stored by its parent (Linsley 1958). This has led the same author to the formulation of the imprinting theory in solitary bees, suggesting that bees may develop sensitivity to specific pollen volatiles either during the larval period or as emerging adults in the brood cell (Linsley 1961). Since then, the imprinting theory is repeatedly mentioned in bee studies (e. g. Stephen et al. 1969,

35 5 Heriades truncorum

1997 2006 1969 1979 1978 1972 1989 1996 1969 1994 2002

2000 smia bicornis ( smia rufa) Brassica napus Brassica Megachile rotundata Medicago sativa Daucus 1996 1979 2003 2003 4

Heriades truncorum 1758 1 1981 1989 1989 H. truncorum Echium Campanula

36 ost rconition in

ampanulaca an rassicaca haptr o tst th imprintin thory compar th loral prrncs o aults o that lop as lara ithr on polln an nctar o thir normal host th straca or on polln an nctar o to nonhost plants an rspctily orc th lara to lop on ths nonhost polln proisions y transrrin unhatch s onto polln an nctar proisions that ha n collct y to irnt oliolctic spcis thus proi th irst stuy on imprintin in oliolctic s an iscuss th implications o th rsults or our unrstanin o th olution o lor rlationships

h spcis is oliolctic on straca his mal collcts polln on stroia Pictur y nras llr

Hosreoiioi

H

aradeeopeoosadoospoedies

eidiidaso iesiaediisaperare oseaeeredro earaesedieeperiesdesriedi aperaed eso ere raserred i a i spaaooepoeadearproisiosoeed aer i is srioioei o oraiaeae ad epeeieriesiearesspoeo apaaeaeoeproisioso ereoeed ro rap ess a oe oai i orer ierad ere ase oospaaaiaeoseo oriiaed ro a eaed reari ere as e o poe sore e es o ere oaied o ro a eaed reari i ad oseraeaeasospasadrorapessoeedadiere oaiies i ierad ara deeope oo pae i ariiiaes predried a os oaed i paraiorio ad os i a iaeaeriedaraadreaieidiearae ereepieariiiaesroo eireiredeeope i adeereeiapasiaraeieseariiiaeseresoredia od roo a or oerieri e oed odoors io a sa iseaeadeoaesieeseoreesaro e eperies i e eered ads ere ared idiidaoeoraieaepaisiaodeosisiosi diere oors oeper e a o ar e ees e ere ioiied i e od roo a ad e iediae raserred ioaareodooroseraio aeadeoae oreperies

e ariiiaes oaii e diapasiarae eresored ia od roo ad ere e eered e ads ere areda ad raserredodoorsioeoseraioaesiiess aoeda eidiidaaredeeesoeoraieaepaisia odeosisiosidiereoors

5. Host recognition in eiaestom

As control in the experimental phase with Camala (see below), we used females of tom which had been transferred as eggs onto provisions of Asteraceae pollen and reared following the same procedure as for those grown on non-host pollen. As control for the experiments with him, we used freshly hatched females and males from trap nests collected in nature. Individuals grown on each of the different pollen types were kept in a separate polyester box. Thus, larvae grown on non-host pollen were never in direct contact with pollen of the Asteraceae.

Seven viable females and four males of tom emerged from a total of 15 pupae grown on pollen of him lae, eight females and seven males from 23 pupae reared on pollen of Camala otiolia, and five females and three males from 11 pupae grown on pollen of Asteraceae (see chapter 4). In addition, eight females obtained from trap nests were used as controls for the experiments.

5.3.2. Choice experiments

The marked bees were allowed to fly, mate and build nests outdoors in the observation cages. Males and females were introduced together into the cage to ensure mating. The cages were provided with potted plants as pollen and nectar sources, hollow bamboo stalks as nesting sites and coniferous resin (isspec. and ieaspec.) in Petri dishes as material for nest construction. We used two different cages, one for the experiments with the him-grown bees (hereafter the him cage), and one cage with the individuals reared on Camala (hereafter the Camala cage).

In a first phase (choice phase), we offered both host (Asteraceae) and non-host plants (either him or Camala) in equal abundances. In the him cage, we provided him lae as non-host plant and hthalmm saliiolim and aaetm lae as host plants in a random distribution. In the Camala cage, eleim atmale and aaetmlae were offered as host and Camalaoteshlaiaa a Cotiolia as non-host plants. Larvae of tom were found to be equally able to develop on both Camala species (Chapter 4). In contrast to the him cage, we presented the plants here in patches: two

39 5 Heriades truncorum

H. autumnale T. vulgare C. portenschlagiana C. rotundifolia Campanula 46

533

H. truncorum 1997 23 Echium 100 Campanula 20

H. truncorum

40 Host recognition in

ollen composition o rood cell provisions

e microscopically analysed the pollen provision o each rood cell ilt in the oservation cages y the emales o o prevent the larvae rom consming the pollen provisions dring the ongoing eperiments e repeatedly removed the amoo stals containing completed rood cells rom the cage his as done in the late evening hen the emales ere already sleeping ithin their nests ter cooling the amoo stals don to C e split them longitdinally ith a nie and removed the pollen provisions rom the completed cells e then carelly closed the nest ith adhesive tape placed the ee ac into the nest and the nest ac into the cage his procedre did not aect the nesting emales as cold e dged y the act that the emales normally resmed their nesting activities the olloing morning he entire content o each rood cell as emedded into aisers glycerine gelatine erc armstadt ermany on si dierent slides e thoroghly mied the pollen ith a needle eore covering it ith a covering slip to achieve a random distrition o the pollen grains o determine the composition o the pollen provisions e randomly selected ive spots on each slide one in each adrant and one in the centre and conted all grains at a magniication o on average grains per spot e calclated the ratio o host pollen to nonhost pollen in each rood cell y averaging the nmer o pollen grains over all conts the provisions ere composed o pollen o oth host and nonhost plants the percentages o the nmer o pollen grains ere corrected y their volme sing the volme vales given in ller et al

cage

a ales hree ot o or males o reared on the nonhost pollen o patrolled loers in the cage to locate emales ach o these three males strictly restricted its patrolling lights to the speciesspeciic host plants the steraceae o single approach to the loers o as oserved oth host and nonhost loers ere reglarly visited or

5. ost recognition in Heriades trncorm nectar, though. The fourth male did not perform patrolling flights but rested on the cage wall during the whole observation period and was thus excluded from the experiments.

b. Females hoice phase nly two out of the seven chim-grown females provisioned brood cells. The first female completed one cell and was subsequently lost. The provision of this cell consisted of .3 % Asteraceae and 0.7 % chim pollen (Table 1). The second female completed eight cells, all consisting of more than % pollen of Asteraceae with only trace amounts of chim pollen. ut of the eight control females reared on Asteraceae pollen, only two nested. The first female built two nests with four and six cells, respectively (Table I). All four cells of the first nest contained almost pure host-pollen provisions (all over %). In the second nest, two cells contained considerable amounts of chim pollen (% and 11%, respectively). This was most likely due to pollen contaminations during very frequent nectar visits to chim triggered by an accidental shortage of flowering Asteraeae in the cage. This female was never observed to actively collect pollen on the flowers of chim. nce new Asteraceae were introduced into the cage, the female again exclusively harvested host pollen: the provisions of subsequent cells were composed of more than % Asteraceae pollen. The second control female constructed four cells, each containing over % Asteraceae pollen (Table 1).

onchoice phase Immediately after all host plants were removed from the cage and replaced by additional flowering chim plants, all three females still nesting at this time (one chim-grown and two controls) discontinued provisioning their brood cells. ithin three days, two females (one chim- grown and one control) sealed the entrance of the nest with a plug of resin, although the bamboo stalk they nested in still had enough space for additional brood cells. uring this non-choice phase no attempt to collect pollen on chim was observed though all three females regularly visited chim for nectar uptake.

42 5. ost recognition in Heriades truncorum

TABLE 1. The pollen composition of brood cell provisions in the choice experiment, where both host plants (Asteraceae) and non-host plants (either Echium or amanua were available. The average proportion of Asteraceae pollen (in volume) per nest is given. The bees were grown as larvae either on host pollen or on non-host pollen.

Non-host plant Number of Asteraceae in emale Larval diet Nest tested cells in nest pollen provision

Echium 1 Asteraceae 1 4 99.5 2 6 96.9 2 Asteraceae 1 4 99.3 2 1 99.5 3 Echium 1 1 99.3 4 Echium 1 8 99.8

amanua 5 Asteraceae 1 7 100.0 6 amanua 1 4 100.0 7 amanua 1 1 100.0

5.4.2. amanua cage

a. Males our out of the seven males grown on the non-host pollen of amanua, exhibited patrolling flights. These four males exclusively patrolled Asteraceae during 100 of the observation time. No single approach to the flowers of amanua was registered. Nectar uptake predominantly took place on the Asteraceae, nectar visits to amanua flowers were very rarely recorded.

b. emales hoice hase Two out of the eight females grown on amanua started nest construction. They built four cells and one cell, respectively. The provisions all consisted of pure Asteraceae pollen (Table 1). Only one out of the five control females built a nest. This was composed of seven cells, each containing pure pollen provisions of Asteraceae (Table 1). We did not detect

43 . Host recognition in Heriades truncorum any traces of Campanula pollen in the cells of these three females hich is in line ith the oservation that the females made very fe nectar visits to the floers of Campanula.

Non-choice phase After the host plants ere removed from the cage all three females to Campanulagron and one control ceased nesting activity and ithin three days sealed their nests ith a plug of resin. Again none of their amoo stals as completely filled ith cells at this time. All three females ere regularly oserved to gather nectar on Campanula floers ut no single attempt to collect pollen as oserved.

.. SSSON

Our study provides the first empirical investigation of the imprinting theory in an oligolectic ee species. oth females and males of Heriades truncorum recognised their host even hen they had never een in contact ith its pollen in the natal cell contrasting the prediction of insley . hus neither preimaginal learning nor early adult conditioning is liely to e the ase of host recognition in this species.

t may appear surprising that so fe females started to construct nests in our eperiments. his lo nesting rate as oserved oth in the individuals gron on nonhost pollen four out of females uilt a nest and in the controls three out of . n nature solitary ees sho the tendency to nest close to their old nests and often prefer to renest in the type of material from hich they emerged Stephen et al. . Similarly megachilid ees are attracted to the odors of old nests PittsSinger and females of Megachile rodundata prefer to nest in already occupied nesting sites rather than in ne ones airey and ieverse . n our eperiments, the lo rate of nest estalishment y H. truncorum may e due to the asence of occupied nests in the cage as ell as to the fact that the tested ees had een rought to the cage as adults and not in the preimaginal phase. n fact Maciel e Almeida orreia reported oservations that individuals of this species did nest in a cage only hen they had hatched and developed in the same cage. Hoever such an

Host reoitio i Heriades truncorum approah as ot ompatile ith the prpose o the rret stdy that reired e traser ad idiidal mari additio ooos o H.truncorum are ery loose ad a ot e traserred to sitale esti sites ithot damai the ees

hoh the mer o rood ells proisioed y the Echium ad ampanularo emales as admittedly small e eertheless thi that or reslts allo reetio o the impriti hypothesis i H.truncorum or three reasos irst eah ell proisioed y a ee represets a lare mer o loral isits: emales o H. truncorum reire orai trips to proisio oe ell ad eah trip osists o loral isits o aerae aiel e Almeida Correia eod the emales ro o ohost polle osistetly eased to proisio their ells he oered oly this polle althoh they had eperieed it i the rood ells impriti had ileed the ees loral hoies these emales old hae ee epeted to ollet polle at least partially rom the e hosts otrast to illiams ho osered the olioleti ee speies Osmia californica to ollet polle o ohost loers as lo as the speiesspeii host plats ere preset e did ot osere a sile attempt to harest ohost polle either dri the hoie phase or dri the ohoie phase hird oth the Echium ad ampanula ro males o H. truncorum elsiely searhed or emales at the iloresees o the Asteraeae terestily all idiidals o H.truncorum isited the ohost loers or etar idiati that they are ale to pereie these loers olsio the preseted reslts stroly idiate that the preeree or Asteraeae i H. truncorum is iate ad thereore mst hae a eeti odatio

Host reoitio i solitary ees ees host reoitio proaly relies o olatory or isal es or o a omiatio o oth isley islo ad Cae he role o polle or loral odors i host reoitio has ee dometed or seeral olioleti ee speies oso oso ad erstrm tterl et al tterl ad hler t old ot e deiitely sho or other speies oso ad erstrm otrast little is o o the importae o isal es i host reoitio y solitary ees isal

HosreoiioiHeriades truncorum

esih ideooraersoreeioororashaedeed oors ere sho o iee hos hoie i o oei seies o Megachile iheer oe ad rad ad a iae reeree or eo oers as od i hree oioei ee seies osoadersr

H. truncorum hos reoiio is ie o re o isa es aoe his seies oraes o a arie o diere seies o he seraeae oe hoss i era roe ide ao a ohers Buphthalmum salicifolium Tanacetum vulgare ad Achillea millefolium sai seroideae Centaurea sp. ad Cirsium sp. ardoideae as easCichorium sp. ihorioideaeesrihhesehossdisi dier i shae ad oor hs e osae ha hos reoiio i H.truncorum s re redoia o oaor es so ar o ear reerees or oe or oer odors od e sho i ae H.truncorumosoadersr

iaiosoreeoerreaioshis

hoh oe reerees are eera hih osered ihi seiaiseeadesosereaedseiesereodoeoioeio reaed ora hoss i soe ieaes iso ad ae hese seeiaseiaissahaedesededroaoioei aesor haireaiedhearrodiereadhsihedhosiead osohisheoeohasiiiaeeiedaseideeha ierseii oeiio dries seiaiaio hroh resore ariioi e his hohesis reais esed ad oroersia eras ad iso ie ad oso hor oraeda hohesisoseiaioediaedhrohhossihesi ros o ose reaed ee seies o he ses Diandrena his odea oioei ee seies aide a e hos iois oe diedeoashoraeoisorahoshisasiiisideedosered i soe oioei ees ie ad oso ad reerees hereiseeodiioiiheaaeasoioraiai hishosshiiseedieeraiosadhsoriesoheirs ses oardrerodieisoaioas he seiioe hossoea as redeosaesoraes ad eaes irii does ori

5. Host recognition in Heriades truncorum solitary bees, this scenario could happen in a particularly short time and, in theory, even in sympatry (Bush 1). However, the results of the present study do not support imprinting in bees. Therefore, the above scenario of a sympatric speciation mediated through host shift appears to be unlikely. Other isolating mechanisms must be involved, such as geographic isolation and adaptations to the new hosts under disruptive selection, for eample foraging efficiency, digestion efficiency or phenological synchronisation (Thorp 16, Wcislo and Cane 16). The scenario would thus approach the general mosaic theory proposed by Thompson (1).

An additional important aspect that may influence diet breadth and host shift is the sensory capabilities of the adult bee, which were suggested to influence host breadth in phytophagous insects (reviewed in Bernays 1). In our eperiments, H. truncorum refused to collect non-host pollen in spite of its suitability to support larval development. This suggests that neural constraints are more important than nutritional constraints in shaping the host range of this species. This neural fiation may lead to fitness advantages that compensate for the dependency on a limited range of host plants (Bernays 1).

5.5.3. Future research

Though the pollen specialization in H. truncorum has most probably a genetic basis, imprinting can not be ecluded to occur in other bee species. As suggested by Dobson and Peng (1), a hypothetical conditioning to pollen volatiles during early life stages may strongly vary among different bee-flower pairs. Many lipid-soluble volatiles, which might contribute to imprinting, are included in the pollenkitt (Dobson 1, 1). The amount of pollenkitt strongly varies depending on the plant taon, as does the degree to which solitary bees digest it (Dobson and Peng 1, Dobson and Bergstrm ). Thus, the presence or absence of pollen volatiles in the natal cell may strongly influence the possibility of conditioning to occur at the time of adult emergence. Therefore, the imprinting theory has to be tested in other bee species before it can be generally rejected.

smiini plogen

6. Phylogeny and biogeography of bees of the tribe Osmiini (Hymenoptera: Megachilidae)3

1 AA

e smiini egaciliae constitte a taonomicall an iologicall ierse trie of ees o resole teir generic an sprageneric relationsips e inferre a plogen ase on tree nclear genes Elongation factor-1 -roopsin an A appling ot parsimon an aesian metos r plogen ic incles osmiine species representing 1 of te 1 crrentl recognie genera is ell resole it ig spport for most asal noes e core osmiine genera ere fon to form a ell-spporte monopletic grop t for small genera oteriaes Afroeriaes seoeriaes an possil creriaes formerl incle in te smiini o not appear to elong itin tis trie r plogen reslts in te folloing taonomic canges tenosmia an oplosmia are rece to sgeneric ran in oplitis an smia respectiel icreriaes is recognie as a sgens in oplitis an te sgens astosmia is transferre from oplitis to smia e inferre a iogeograpic scenario for te smiini appling maimm lielioo inference an moels of caracter eoltion e proie eience tat te smiini originate in te alearctic an tat etensie ecanges occrre eteen te alearctic an te earctic e latter fining ma relate to te fact tat man osmiine species nest in oo or in stems facilitating ispersal oerseas transport of te nests

e smiini constitte a er ierse trie itin te megacili ees it oer 1 species crrentl recognie icener ngrict et al in press e occr on all continents ecept Astralia an ot America eing especiall ierse an nmeros in regions it eiterranean an eric climates of te alearctic sotern Erope nortern Africa ile East an central Asia ort America

ase on ra ller A anfort risol imer A an orn smitte to oleclar logenetics an Eoltion

6. Osmiini phylogeny

(Southestern Deserts, California) and southern Africa (Cape Province, amibia).

The biology of the Osmiini is astonishingly diverse and has fascinated entomologists for ell over a century (Raumur 1748, Fabre 1886, Ferton 1923). In particular, the nesting biology of the Osmiini is highly varied and encompasses much of the diversity observed in other bees (Malyshev 1937, Westrich 1989, Mller et al. 1997, Cane et al. 2007). Depending on the species, osmiine bees build their nests in holes in the ground, belo stones, on rock surfaces, in pithy stems, galls or in beetle borings in dead ood. Many species are knon to nest exclusively in abandoned snail shells (Mller 1994, Bellmann 1997, Haeseler 1997). Diversity in nesting site is matched by materials used in nest construction, such as sand or mud, masticated plant material, bright-colored petals or resin. The Osmiini are thus model organisms to trace the evolution of nesting behavior in bees in general (Bosch et al. 2001). The Osmiini are also famous for their specific relationships ith floers. Several osmiine genera and subgenera are predominantly composed of pollen specialists (e. g., elsma Westrich 1989, Michener 2007 see also chapter 7 Apsmia and pliis subgenus reriaes: Michener 2007, and references therein), hile other taxa are mainly composed of generalists (e. g., rsmia and smia subgenus rsmia: A. Mller, unpublished). The Osmiini are of particular interest for studies of the evolution of floral relationships because some species can easily be reared in trap nests or under caged conditions. This feature has enabled studies on the acceptance of novel pollen hosts by adult bees (Strickler 1979, Williams 2003 chapter 5) and on pollen digestion by larvae (Levin and Haydak 1957, Suare-Cervera et al. 1992, Dobson and Peng 1997, Williams 2003 chapter 4). Studies combining experiments on the physiological capacities of bees ith a phylogenetic approach are most promising to elucidate underlying mechanisms of bee-floer relationships in general (Williams 2003).

Unfortunately, the phylogeny of the Osmiini remains controversial. First, the monophyly of this tribe is contentious. Its traditional recognition has relied on symplesiomorphies relative to the other tribes of the Megachilidae (Michener 1941, 2007), hich has led several authors to

49 6. Osmiini phylogeny suspect that the Osmiini are paraphyletic (Engel 1999, 2001, Ascher et al. 2001, Michener 2007). Second, the suprageneric subdivision of the Osmiini is much debated. The Osmiini have long been divided into two groups, the Heriades group and the Osmia group. owever, species with combinations of characters from both groups exist, leading to the view that these two groups may merge (Michener 2007). For instance, the genus helostoma was long included in the Heriades group, but is currently placed in the Osmia group due to its similarity to some members of the genus Hoplitis (Michener 2007). Lastly, important controversies exist on the generic classification of the Osmia group. Griswold and Michener (1997) recognied eight genera within the Osmia group. Michener (2007) retained this classification, but acknowledged that the Osmia group is one of the most problematic bee taxa due to the fluent boundaries between the genera. In contrast, most European authors recognise only one large genus Osmia sensu lato for all members of the Osmia group (Westrich 1989, Schwar et al. 1996, Amiet et al. 2004) or even for all osmiine bees (Warncke 1991, Westrich and Dathe 1997, Westrich 2006). This important difference in the treatment of the Osmiini by American and European authors may relate to the high diversity of osmiine bees in the Old World, rendering the clear segregation of taxonomic groups in the Palearctic more difficult than in orth America. Indeed, most American members of Osmia and Hoplitis can unambiguously be assigned to one of the two genera by a typical combination of morphological characters (Michener 2007). American Osmia species are rather robust (chalicodomiform), possess punctiform parapsidal lines and are mostly metallic blue or green in coloration, whereas the species of Hoplitis are more slender (hoplitiform), have linear parapsidal lines and are non-metallic. In the Palearctic, however, many species exist, which present intermediate combinations of these traits, such as Osmia cephalotes, O. rufohirta or O. andrenoides, which have linear parapsidal lines. Moreover, many Palearctic species of Hoplitis have a chalicodomiform rather than a hoplitiform body.

In the present study, we provide the first molecular phylogeny of the Osmiini. It is based on three nuclear genes and includes 95 osmiine species. Our aim is to assess the monophyly of the osmiine bees and to clarify their suprageneric and generic classification. We further use our

50 6. Osmiini phylogeny phylogenetic framework to develop a hypothesis of the biogeographic history of the Osmiini.

6.3. METODS

6.3.1. Taxon sampling

The 95 osmiine species included in our phylogeny (Table 1) represent 18 of the 19 genera worldwide and 62 of the 82 subgenera currently recognied (Michener 2007). The only genus not included is eroheriades, a monotypic genus that has been assigned to the eriades group (Griswold 1986a). For widespread subgenera (those present in more than one of the three main biogeographic ones Palearctic, earctic and Afrotropic), we included species from each continent whenever possible. The nomenclature and classification of the tribe Osmiini follows Michener (2007) and ngricht et al. (in press). As outgroups, we selected 12 species representing all other megachilid tribes, namely the Fideliini and Pararhophitini (subfamily Fideliinae), and the Lithurgini, Anthidiini, Dioxyini and Megachilini (subfamily Megachilinae). According to phylogenies based on morphology (Roig-Alsina and Michener 1993) and DA sequences (Danforth et al. 2006b), the Fideliinae and the Megachilinae are sister groups, and the Lithurgini are sister to the other tribes of the Megachilinae, including the Osmiini.

6.3.2. DA sequencing

DA was extracted from bees preserved in 100 ethanol and from a few pinned specimens up to five years old. Whenever possible, we preferred males because they are haploid, which simplifies sequencing of nuclear markers. We extracted DA from the head and conserved the rest of the specimen as voucher as osmiine males can usually be identified by the examination of their metasoma. For females, we extracted DA from the thorax, keeping the other parts as a voucher. For minute species, we used the entire body for DA extraction selecting another specimen of the same species as voucher. ouchers are deposited in the Entomological Collection of the ET urich. Our DA extraction protocol follows Danforth (1999), except that we did not use RAse.

51 6. Osmiini phylogeny

TABE 1. Species and locality information included in this study. (Collectors: BD, B. Danforth CP, C. Pra AM, A. Mller TG, T. Griswold).

Taxon ocality Collector

Outgroup ieliopsis maor South Africa, Clanwilliam BD Pararopies aras Tunisia, Nefta CP irs rysrs taly, Abruen, Massa AM Alaoapis rienaa Switerland, eneggen CP Aniim pnam Switerland, Weiach AM elis pnlaissima Switerland, ohtenn AM rasa yssina Switerland, Splgen AM eaile alisea taly, Massa Maritima AM eaile piliens Switerland, Weiach AM eaile parieina Switerland, ohtenn AM oelioys afra Switerland, Weiach AM

ngroup Afroeriaes prims South Africa, North Cape TG Afroeriaes sp n South Africa, Nieuwoudtville TG Asmeaiella Arooila imerlaei USA, CA, Mariposa Co. . kerd Asmeaiella Asmeaiella arila USA, UT, Garfield Co. TG Asmeaiella ilosima roonaa Mexico, Sonora R. Minckley Asmeaiella ionaa enomasa USA, UT, Cane Co. . untinger Asmeaiella sosmia riana Mexico, Sonora R. Minckley Aoposmia Aoposmia elonaa USA, CA, Mariposa Co. E. Stephens Aoposmia Aoposmia eiis USA, CA, Tuolumne Co. TG Aoposmia remosmia mirifia USA, N, Clark Co. S. igbee Aoposmia remosmia sp n aff aleae USA, UT, ane Co. . kerd Aoposmia remosmia imerlaei USA, N, Clark Co. S. igbee Aoposmia eosmia opelania USA, CA, Mariposa Co. TG elosoma eraeriaes lamellm China, unan Province C. Sedivy elosoma elosoma florisomne Switerland, Chur E. Steinmann elosoma oelosoma areoinm Thailand, Chiang Mai C. Sedivy elosoma oeosmia alifornim USA, CA, Mariposa Co. TG elosoma oeosmia ampanlarm Switerland, Winterthur AM elosoma yroromella rapnli Switerland, Fully CP elosoma Proelosoma pilaelpi USA, MD, Pr. Georges Co. S. Droege aeosmia irmena UEA, Sharah Desert Park T. van arten eriaes eriaes rnorm Switerland, Winterthur AM eriaes ienerella pnlifera Greece, Rhodos, Stegna AM eriaes eorypees rifer USA, A, Chiricahua Mts TG offeria smieenei Greece, Chimara CP opliis Aliamea leomelana Switerland, ohtenn AM opliis Aliamea miis Switerland, eneggen CP opliis Aliamea pilosifrons USA, N, Thompkins Co. . Ascher opliis Aliamea rienaa taly, Aosta, StPierre CP opliis Annosmia annlaa reia Greece, Rhodos, Afandou AM opliis Anoopa isla Greece, Rhodos, Stegna AM opliis Anoopa emispaeria ordan, Wadi Muib CP, AM, C. Sedivy opliis Anoopa sp South Africa, Nieuwoudtville . Timmermann opliis Anoopa illosa Switerland, Sntis, Flalp D. Dietiker opliis liopliis illsris Turkey, Ankara E. Scheuchl opliis liopliis sp n aff onyopora ordan, Wadi el asa CP, AM, C. Sedivy opliis yrosmia yporia USA, CA, Tuolumne Co. TG opliis asyosmia isellae USA, CA, Riverside Co. . Ascher opliis ormiapis rosa Switerland, isperterminen CP

52 siini yleny

B nined

Hoplitis (Hoplitina) moavensis la iee Hoplitis (Hoplitis) adunca aly sa Hoplitis (Megahoplitis) tigrina ey naa el Hoplitis (Micreriades) antalyae eee ds and Hoplitis (Micreriades) lebanotica dan adi i ediy Hoplitis (Microhoplitis) paralias nisia ea Hoplitis (Monumetha) albifrons na sa se Hoplitis (Monumetha) tuberculata ieland asseaen Hoplitis (Nasutosmia) nasuta ane eilles le Hoplitis (entadentosmia) moricei ie Hoplitis (entadentosmia) villiersi nisia ea Hoplitis (enteriades) incanescens ny ene Hoplitis (latosmia) platalea ie Hoplitis (rionohoplitis) brachypogon aly sa iee Hoplitis (roteriades) zuni aield nine Hoplosmia (Hoplosmia) spinulosa ieland eisasen Hoplosmia (Odontanthocopa) scutellaris eee ds alnas Hoplosmia (aranthocopa) pinguis nisia asa Noteriades sp. ailand ian ai ediy Ochreriades fasciatus dan adi ay ediy Osmia (Acanthosmioides) integra a ane nine Osmia (Allosmia) rufohirta aly sa iee Osmia (Allosmia) sybarita eee alia Osmia (Cephalosmia) montana anan ilsn Osmia (Erythrosmia) andrenoides ieland enn Osmia (Euthosmia) glauca alinia adea ed Osmia (Helicosmia) aurulenta ieland asseaen Osmia (Helicosmia) coloradensis anan ilsn Osmia (Helicosmia) niveata ieland enn Osmia (Hemiosmia) anceps nisia ea Osmia (Hemiosmia) difficilis ey naa el Osmia (Melanosmia) uxta anan ilsn Osmia (Melanosmia) xanthomelana ieland enn Osmia (Metallinella) brevicornis eany nsan eann Osmia (Monosmia) apicata eee laania andss Osmia (Mystacosmia) nemoris na sa anie Osmia (Neosmia) bicolor ieland dlinen Osmia (Neosmia) tingitana nisia asa Osmia (Osmia) cornuta ieland i Osmia (Osmia) lignaria ane Osmia (yrosmia) ferruginea nisia Bi el aey Osmia (yrosmia) gallarum ieland enn Osmia (Tergosmia) rhodoensis ey neaa e el Osmia (Tergosmia) tergestensis ieland enn Osmia (Trichinosmia) latisulcata a ane essine Othinosmia (Megaloheriades) globicola ia iedille ae in Othinosmia (Othinosmia) sp. aff. securicornis ia ieseld rotosmia (Chelostomopsis) longiceps eee ds and rotosmia (Chelostomopsis) rubifloris lne rotosmia (Nanosmia) minutula ieland d rotosmia (rotosmia) humeralis dan adi ay ediy seudoheriades moricei lan an aen Stenoheriades asiaticus eee al Stenosmia aravensis enia alandyan Be Stenosmia minima nisia ea Wainia (Caposmia) eremoplana dan adi al asa ediy

6. Osmiini phylogeny

PCR amplifications were performed in 5 l reactions with 2.5 mM

MgCl2, 2M dNTP, . M of each primer and Ta polymerase (PromegaTM, GoTa) using a thermocycler (GeneAmp, Applied Biosystems). We used nested PCR with specific primers for some pinned specimens with highly degraded DNA. We always added a blank sample to each PCR and used this blank sample in the second, nested PCR. In those rare cases where PCR products were detected in these controls, we discarded all samples. PCR conditions are given in Table 2. PCR products were purified using DNA-purification kits (GFTM) or MultiscreenTM 6- wells plates when processing more than 1 samples at the same time. In the latter case, PCR products were bound with HCL-Guanidine buffer (M Guanidine-HCL; 2 mM MES Buffer, pH 5.6), then washed twice with ethanol and eluted with TE buffer (1 mM Tris-HCL, pH ; .1 mM Na2EDTA). Prior to seuencing, the PCR products were filtered through SephadexTM columns. Automated seuencing of the PCR products was performed on an ABI Prism 1xl seuencer using BigDye technology. We used internal primers to seuence Elongation factor-1 and the PCR primers for the other genes.

6... Genes analysed

longation factor-1 Elongation factor-1 (hereafter EF) has been widely used to infer bee phylogeny (e. g., Danforth et al. 2, 26b, Larkin et al. 26, Schwarz et al. 26, Cameron et al. 2, Patiny et al. 2). We amplified two overlapping fragments of the F2 copy of this gene (Danforth and Ji 1). For the large fragment (approximately 12 bp), we used the primer pair HaF2For1 and F2Rev1 (Danforth et al. 1a). Based on initial seuences, we designed a modified reverse primer specific to the Megachilidae, F2Rev1- Meg (Table 2). For the short fragment (approximately 6 bp), we replaced For (Danforth et al. 1a) by the F2-specific forward primers For and Fora (Table 2). This F2-specific primer site was chosen in a region showing a highly differing amino acid composition between the F1 and F2 copies in Apis mellifera (GenBank accession numbers: F1, 52; F2, AF1526). For and Fora yield together with Cho1 (Danforth et al. 1a) one single bright band corresponding to the F2 paralog. We designed F2 - specific

5 smiiniyoeny

ae Primers sedor te tree enes onation actor arodosin andonseredAPase Domain eir osition on ised seences or Apis mellifera is ien enan accession nmersAAandDorosinandADresectiey Primer eence PositiononApis

onationactor

aor AAAAAA ee AAAAA or AAAA ora AAAA o AAA ntrone AAAAAAAA ntronor AAAAA one AAAAA onor AAAAAAAAA one AAAAAAAAA ore AAAAA onor AAAAA e AAAAAAA

Pconditions aoreme oro

odosin

sorsm AAAAAA sine AAAAA sine AAAA sinor AAA sinora AAA sine AAAA sinea AAA

Pconditions sinorosmsine sinorsine

AD

ADor AAAAAA ADee AAAA

Pconditions ADorADeosm

6. Osmiini phylogeny seuencing primers (F2-Intron-Rev and F2-Intron-For, Table 2) in a highly conserved region found in both introns 2 and 3, thus appropriate for the large and the small fragment. Additionally, we used the primers Exon2Rev, Exon2For, Exon3Rev, For3-Meg, Exon4For and EF-Rev (Table 2) as seuencing primers or primers for nested PCR. The assembled fragment has a length of approximately 1600 bp.

rhodopsin The phylogenetic utility of -rhodopsin (hereafter opsin) to infer bee phylogeny has been widely discussed (Mardulyn and Cameron 1999, Ascher et al. 2001, Cameron and Mardulyn 2003, Danforth et al. 2004). This gene shows a comparatively high rate of non-synonymous substitutions (Danforth et al. 2004). Ascher et al. (2001) suggested opsin to be suitable for resolving low-level taxonomic relationships, namely within tribes and genera. Based on the published opsin seuence of Osmia icornis (O rufa GenBank accession number A572828 Spaethe and Briscoe 2004), we modified the widely applied primer OpsinFor ( -RhF Mardulyn and Cameron 1999) by the more specific OpsinFor-Osm (Table 2), which we used in combination with specific reverse primers designed for the Osmiini, OpsinRev3 and OpsinRev3b (Table 2). As the primers OpsinRev4 and OpsinRev4a (Danforth et al. 2004) failed to amplify the 3 end of opsin in the Megachilidae, we designed two new reverse primers OpsinRev5 and OpsinRev5a by aligning published seuences of short-tongued bee species (Danforth et al. 2004). These primers were used with specific forward primers designed for the Megachilidae, OpsinFor5 and OpsinFor5a (Table 2). The two fragments overlap and yield together a fragment of approximately 1200 bp. e performed BAST searches in Genbank to verify that all seuences included in our data set represent the first of two opsin genes described in bees by Spaethe and Briscoe (2004).

Consered APase omain CA CAD has recently been used in analyses of family-level relationships in bees (Danforth et al. 2006a) and yielded promising results. e seuenced a 450-bp fragment of the approximately 1200-bp fragment used by Danforth et al. (2006a). e applied the primers CADFor4 ( ApCADFor4 Danforth et al. 2006a) with the modified reverse primer CADRev1-Meg

56 6. Osmiini phylogeny

able which we designed by aligning published megachilid seuences anforth et al. 006a. his primer pair yielded one bright band in all megachilid species investigated corresponding to the eon 6 of A anforth et al. 006a.

6.3.. euence editing

he seuences were trimmed and assembled using the software euencher . for acintosh Gene odes orp. and aligned applying lustal 1.3 hompson et al. 1. he alignments were corrected visually in aclade .0 for acintosh O addison and addison 005. Reading frame and introneon boundaries were determined by comparison with published seuences for pis mellifera GenBan accession numbers AF0156 AU606 and 061 for EF opsin and A respectively. All introns were removed prior to analysis in aclade. he coding seuences of the three genes were converted into amino acid seuences to ensure that the correct reading frame had been found. o stop codons were detected in any of the three genes. Alignments of the translated proteins further confirmed that our data sets consist of orthologous seuences for all three genes.

6.3.5. Phylogenetic analyses

arsimony analyses e first performed parsimony analyses of the coding seuences of each gene separately using Paup .0b for acintosh wofford 00 with the following parameter settings unweighted analysis heuristic search 100 random seuence additions with four trees held at each step maimum of 500 trees retained and BR branch swapping. e performed 100 bootstrap replicates with 10 random seuence additions to assess the robustness of the clades. e then combined the three gene seuences into a single matri and analysed the combined dataset performing 1000 bootstrap replicates with 10 random seuence additions settings as above.

5 6. Osmiini phylogeny

Bayesian analyses For the Bayesian analyses, the three genes were analyed collectiely under two different partitioning regimes. First, we partitioned the data set by gene, resulting in three partitions. econd, we partitioned the data set by gene and codon position, resulting in a total of nine partitions. e ran MrModeltest 2.2 Nylander 2004 to determine the best model of sequence eolution for each partition Table . Two approaches are implemented in MrModeltest, the hierarchical lielihood ratio test RT and the aie information criterion I. s the RT approach may prefer different substitution models according to the hierarchy used, we choose in these cases the most comple model, i.e., the one with the most parameters. e ran two analyses for each partitioning regime, one with the models suggested by the RT approach and one with those suggested by the I approach. dditionally, we performed a fifth analysis applying a separate

GTR model to each codon position GTR + R. e calculated Ic and BI scores as described in McGuire et al. 200 to select for the best of these fie analyses.

TB . lternatie partitioning regimes and N substitution models for the Bayesian analyses. ubstitution models were selected according to the two criteria implemented in MrModeltest the hierarchical lielihood ratio test RT and the aie information criterion I. Three partitions riterion for model selection RT I F GTR + I + G GTR + I + G Opsin GTR + I + G + I + G GTR + I + G 0 + I + G

Nine partitions riterion for model selection RT I F nt1 + I + G + I + G F nt2 + I GTR + I F nt + I + G + I + G Opsin nt1 GTR + I + G + I + G Opsin nt2 GTR + I + G GTR + I + G Opsin nt GTR + I + G + I + G nt1 GTR + I GTR + I + G nt2 + I + G M + I nt 0 + G GTR + I + G

smiini pyloeny

aro ain one arlo analyses ere neraen sin rayes elsene an onis e ran a leas for million eneraions samplin rees an parameers eery eneraions o eermine e rnin reion e ee for saionariy of all moel parameers sin e sofare raer ama an rmmon Preliminary analyses iniae a e efal emperare reime for eain resle in infreen sappin eeen e ol an e ree eae ains ene e folloe ire e al an fie e emperare parameer o i resle in appropriae ales of sae sap freeny in irally ienial rees n all analyses e alloe e raes o ary eeen pariions i e prse raeprariale omman in rayes e isare e rees sae rin e rnin perio omine e remainin rees of o rns an ompe maoriyrle onsenss rees sin Pap

econstrctionoancestralgeographicrange e inferre anesral eorapi ranes sin maimm lielioo inferene an moels of araer eolion e se of araer moels o infer anesral eorapial ranes sffers from some limiaions espeially ease sae ransiions ofen ae immeiae effes on speiaion if saes are eorapi ranes onis ee e al ee an mi oeer e se of araer moels is iely applie for ioeorapy reonsrion eporoeff e al ey e al ire e al Pereira e al Pereira an aer as i presens many aanaes oer parsimonyase meos araer moels ae pyloenei nerainy an ran lens ino aon an allo isin raes eeen ifferen saes eir rren se in ioeorapy inferene learly represens a ransiional sep nil more omple moels e ee an mi eome roinely aailale

e sore ioeorapy as a reesae aeorial araer for Paleari for eari an for Afroropi ine ery fe osmiine speies are presen in ropial Asia e i no apply a for sae orresponin o e rienal one e isriion aa ere ran from iener n ases in i erminal aa ere presen in more an one eorapi one e alloe polymorpisms e also ran e

6. Osmiini phylogeny same analyses without polymorphisms by attributing to such taxa the geographic range where the highest species diversity is found. As such a scoring had no substantial influence on the results, we allowed polymorphisms in all analyses. The subgenera Pentadentosmia and Annosmia were scored as alearctic although few species of these taxa enter the Afrotropic zone in Sudan. In addition, we assigned state 1 to one of the two included alearctic species of the subgenus Pyrosmia as the morphologically highly similar subgenus iceratosmia, not available for our study, occurs in North America.

We inferred ancestral states with the software BayesTraits (agel et al. 2004, agel and Meade 2006) at nine highly supported nodes (all with posterior probabilities of 100 in the Bayesian analysis) corresponding to important lineages within the Osmiini. We used a sample of 1000 trees saved in the favored Bayesian analysis. This analysis was allowed to run for 1 million generations after convergence. We sampled trees every 2000 generations and combined the trees from both runs. We explored two models of character evolution applying the maximum likelihood approach implemented in BayesTraits. First, rates were all restricted to be equal, corresponding to the simple model Mk1 (Lewis 2001) usually used in biogeography reconstruction (Dubey et al. 2007, McGuire et al. 2007). Second, we allowed three free rates of biogeographic exchange: Afrotropic versus alearctic, alearctic versus Nearctic, Afrotropic versus Nearctic. The rates between two geographic regions were constrained to be equal in both directions using the command restrict. reliminary analyses with six free rates failed to correctly estimate the rates and parameters, suggesting that our data did not contain enough information to estimate as many rates.

To assess the robustness of the ancestral reconstructions, we successively constrained ancestral states at each node to each of the three states using the fossil command in BayesTraits. We calculated the differences in lnlikelihood for each tree and averaged them over all trees. A difference of two logunits is conventionally taken as evidence for a significant difference (agel 1).

60 6. Osmiini phylogeny

6.4. ESUTS

6.4.1. DNA seuences

Our dataset is complete except for the opsin seuences of Osmia montana (entirely missing) and of rotosmia ruifloris (400 bp fragment only, seuenced with OpsinFor5Opsinev5). After removal of the introns, our dataset comprised 1111 bp for EF, 682 bp for opsin and 448 bp for CAD, of which 333 (30.0%), 275 (40.3%) and 163 (36.4%) were parsimony informative, respectively. Only EF showed a biased base composition across taxa due to the third nucleotide position.

6.4.2. arsimony analyses

arsimony analyses performed for each of the three genes separately (data not shown) indicated that opsin was the most useful gene, followed by EF and CAD, as udged by the number of nodes with a bootstrap support of over 50%. e noticed very little incongruence among the genes (Table4). First, EF indicated that the eriades group was derived from within the Osmia group (59% bootstrap support), whereas the other genes suggested a monophyletic Osmia group. Second, CAD was the only gene supporting the monophyly of Othinosmia (63% bootstrap support). Third, aetosmia was sister group of a clade comprising ainia, toposmia, shmeadiella, oplosmia and Osmia in the analysis with opsin alone (67% bootstrap support), whereas the two other genes placed it as sister to the whole Osmia group (CAD 50% bootstrap support EF 55%). These conflicting topologies are minor and no incongruence was supported by bootstrap values above 70%. e therefore combined the three genes into a single matrix for all subseuent analyses.

arsimony analysis of the combined dataset yielded one island with 134 most parsimonious trees. The 50%-bootstrap consensus tree is highly resolved with most basal nodes supported by bootstrap values 70% (Fig.1). The Osmiini appeared monophyletic with a bootstrap support of 100% (Table 4) with the exclusion of the four small genera Ochreriades, froheriades, seudoheriades and Noteriades. The eriades group and the Osmia group were each recovered as monophyletic with bootstrap

61 Osmiiihloe suortsofadresectielaleiththeecetioofthe four eera metioed aoe ad the eus helostoma his eus turedouttoethesistertoallotherOsmiiiithaootstrasuortof

aleummarofsuortmeasuresforthemailieaesadeeraoftheOsmiii

diidualees Comiedaalses a Osi CA Parsimo aesiaaalses Parsimo Parsimo Parsimo Osmiii oteriadeseachilii eriadesrou Osmia rou arahletic rotosmia eriades Othinosmia arahletic arahletic arahletic arahletic arahletic Atoposmia arahletic arahletic Ashmeadiella Osmia arahletic arahletic arahletic arahletic arahletic arahletic Osmia oplosmia arahletic arahletic oplitis

eeraAfroheriades, oteriades, Ochreriades adseudoheriadesecluded eeraAfroheriades, oteriades ad seudoheriades ecluded eerahelostomaadOcheriades ecluded cludieusStenosmia

aesiaaalses

ACc ad C scores oth faored the reimes ith ie artitios ecettheaalsishichshoedtheloestllielihoodofall aalsesaleHoeereamiatioofthearameterestimatiosfor thesetofaoredreimesiththesoftareraceridicatedthattheto arametersaddidotcoereforseeralartitioshichsuests oerarametriatioNladeretalcotrasteerarameteri theaalsiscoeredaddifferedfromtheriorssuesti

o oerarametriatio Comii these facts ith the ACc ad C criterio e selected the aalsis ith a differet model aliedtoeachofthethreeeeartitiosasthefaoredaalsisi adfortherecostructioofioeorahichistor

Osmiini phyloeny

Pararhophites quadratus (Outgroup) Fideliopsis major (Outgroup) Lithurgus chrysurus (Outgroup) Aglaoapis tridentata (Outgroup) Ochreriades fasciatus 100 Noteriades sp. 88 Coelioxys afra (Outgroup) 77 Megachile albisecta (Outgroup) Megachile parietina (Outgroup) Megachile pilidens (Outgroup) 100 Trachusa byssina (Outgroup) 87 Stelis punctulatissima (Outgroup) 69 Anthidium punctatum (Outgroup) 100 Afroheriades primus Afroheriades sp. n. Pseudoheriades moricei Chelostoma (Eochelostoma) aureocinctum 100 Chelostoma (Foveosmia) californicum Chelostoma (Foveosmia) campanularum 61 Chelostoma (Gyrodromella) rapunculi Chelostoma (Chelostoma) florisomne 81 Chelostoma (Prochelostoma) philadelphi

Chelostoma (Ceraheriades) lamellum Chelostoma 100 Protosmia (Nanosmia) minutula 99 Protosmia (Protosmia) humeralis 57 Protosmia (Chelostomopsis) longiceps 100 100 Protosmia (Chelostomopsis) rubifloris 91 Othinosmia (Megaloheriades) globicola 100 Hofferia schmiedeknechti 99 Stenoheriades asiaticus 81 Othinosmia (Othinosmia) sp. aff. securicornis 100 Heriades (Heriades) truncorum 67

Heriades (Neotrypetes) crucifer Heriades group Heriades (Michenerella) punctulifera Haetosmia circumventa Hoplitis (Megahoplitis) tigrina 100 99 Hoplitis (Annosmia) annulata Hoplitis (Hoplitis) adunca 100 Hoplitis (Chlidoplitis) illustris Hoplitis (Chlidoplitis) sp. n. aff. onychophora 97 Hoplitis (Anthocopa) bisulca 99 Hoplitis (Anthocopa) sp. 95 Hoplitis (Anthocopa) hemisphaerica Hoplitis (Anthocopa) villosa 100 Stenosmia aravensis 53 Stenosmia minima 74 100 100 Hoplitis (Pentadentosmia) moricei Hoplitis (Pentadentosmia) villiersi 86 Hoplitis (Penteriades) incanescens 51 Hoplitis (Proteriades) zuni 100 Hoplitis (Hoplitina) mojavensis 100 Hoplitis (Micreriades) antalyae 60 Hoplitis (Micreriades) lebanotica 67 Hoplitis (Microhoplitis) paralias 81 Hoplitis (Platosmia) platalea Hoplitis (Formicapis) robusta 75 Hoplitis (Prionohoplitis) brachypogon Hoplitis (Alcidamea) tridentata 100 73 Hoplitis (Dasyosmia) biscutellae Hoplitis (Cyrtosmia) hypochrita 67 Hoplitis (Alcidamea) mitis Hoplitis (Alcidamea) leucomelana 72 Hoplitis (Alcidamea) pilosifrons 99 Hoplitis (Monumetha) tuberculata Hoplitis (Monumetha) albifrons Wainia (Caposmia) eremoplana Atoposmia (Hexosmia) copelandica Atoposmia (Eremosmia) mirifica 70 Atoposmia (Eremosmia) sp. n. aff. daleae 74 69 Atoposmia (Eremosmia) timberlakei 100 Atoposmia (Atoposmia) elongata 88 Atoposmia (Atoposmia) hebitis 66 Ashmeadiella (Isosmia) hurdiana 100

Ashmeadiella (Ashmeadiella) aridula Osmia group 93 Ashmeadiella (Cubitognatha) xenomastax 62 Ashmeadiella (Chilosima) rhodognatha Ashmeadiella (Arogochila) timberlakei 96 Osmia (Erythrosmia) andrenoides 99 Osmia (Tergosmia) rhodoensis Osmia (Tergosmia) tergestensis 97 Osmia (Hemiosmia) difficilis Osmia (Hemiosmia) anceps Hoplitis (Nasutosmia) nasuta 100 Hoplosmia (Paranthocopa) pinguis 68 56 Hoplosmia (Hoplosmia) spinulosa 95 Hoplosmia (Odontanthocopa) scutellaris 55 Osmia (Neosmia) bicolor 100 Osmia (Neosmia) tingitana 64 Osmia (Allosmia) rufohirta Osmia (Allosmia) sybarita Osmia (Metallinella) brevicornis 99 Osmia (Pyrosmia) ferruginea Osmia (Pyrosmia) gallarum 56 Osmia (Euthosmia) glauca 100 100 Osmia (Monosmia) apicata 56 Osmia (Osmia) lignaria Osmia (Osmia) cornuta 99 Osmia (Trichinosmia) latisulcata Osmia (Cephalosmia) montana 63 66 Osmia (Helicosmia) aurulenta 100 Osmia (Helicosmia) coloradensis 99 Osmia (Helicosmia) niveata 100 Osmia (Melanosmia) juxta 56 Osmia (Mystacosmia) nemoris 62 Osmia (Melanosmia) xanthomelana Osmia (Acanthosmioides) integra

FE Parsimony ootstrap consensus tree o the comined dataset All nodes ith less than ootstrap support ere collapsed ootstrap alues are indicated aoe the nodes

. Osmiini phloen

Pararhophites quadratus Aglaoapis tridentata Fideliopsis major Lithurgus chrysurus Ochreriades fasciatus 100 Trachusa byssina 100 Stelis punctulatissima Anthidium punctatum 89 100 Afroheriades primus Afroheriades sp. n. 96 98 Pseudoheriades moricei Noteriades sp. 100 100 Coelioxys afra 100 Megachile albisecta 85 Megachile parietina Megachile pilidens Chelostoma (Eochelostoma) aureocinctum 77 100 Chelostoma (Chelostoma) florisomne 84 51 Chelostoma (Foveosmia) campanularum Chelostoma (Gyrodromella) rapunculi 57 Chelostoma (Foveosmia) californicum 100 Chelostoma (Prochelostoma) philadelphi

Chelostoma (Ceraheriades) lamellum Chelostoma 100 Protosmia (Nanosmia) minutula 100 Protosmia (Protosmia) humeralis 98 Protosmia (Chelostomopsis) longiceps 100 100 Protosmia (Chelostomopsis) rubifloris 100 Othinosmia (Megaloheriades) globicola 100 Hofferia schmiedeknechti 100 Stenoheriades asiaticus 96 Othinosmia (Othinosmia) sp. 100 Heriades (Michenerella) punctulifera

50 Heriades (Heriades) truncorum Heriades group Heriades (Neotrypetes) crucifer Haetosmia circumventa 100 100 Hoplitis (Anthocopa) bisulca 100 Hoplitis (Anthocopa) sp. 72 Hoplitis (Anthocopa) hemisphaerica Hoplitis (Anthocopa) villosa 100 Stenosmia aravensis 100 74 Stenosmia minima 100 Hoplitis (Pentadentosmia) moricei Hoplitis (Pentadentosmia) villiersi 75 Hoplitis (Megahoplitis) tigrina 58 100 100 Hoplitis (Chlidoplitis) illustris Hoplitis (Chlidoplitis) sp. n. aff. onychophora 100 Hoplitis (Hoplitis) adunca Hoplitis (Annosmia) annulata 89 100 Hoplitis (Proteriades) zuni 82 Hoplitis (Penteriades) incanescens 100 Hoplitis (Hoplitina) mojavensis 100 Hoplitis (Micreriades) antalyae 56 100 Hoplitis (Micreriades) lebanotica 98 Hoplitis (Microhoplitis) paralias 100 Hoplitis (Platosmia) platalea 67 Hoplitis (Formicapis) robusta 71 Hoplitis (Prionohoplitis) brachypogon Hoplitis (Alcidamea) tridentata 100 Hoplitis (Dasyosmia) biscutellae 52 Hoplitis (Cyrtosmia) hypochrita 100 Hoplitis (Alcidamea) mitis 100 Hoplitis (Monumetha) tuberculata 100 Hoplitis (Monumetha) albifrons 78 Hoplitis (Alcidamea) leucomelana Hoplitis (Alcidamea) pilosifrons Wainia (Caposmia) eremoplana Atoposmia (Hexosmia) copelandica 89 Atoposmia (Eremosmia) mirifica 100 Atoposmia (Eremosmia) sp. n. aff. daleae 100 99 Atoposmia (Eremosmia) timberlakei 100 Atoposmia (Atoposmia) elongata 100 Atoposmia (Atoposmia) hebitis 97 Ashmeadiella (Isosmia) hurdiana

100 Ashmeadiella (Ashmeadiella) aridula Osmia group 100 Ashmeadiella (Cubitognatha) xenomastax 91 Ashmeadiella (Chilosima) rhodognatha Ashmeadiella (Arogochila) timberlakei 100 Osmia (Erythrosmia) andrenoides 87 100 Hoplosmia (Paranthocopa) pinguis 84 Hoplosmia (Hoplosmia) spinulosa 100 Hoplosmia (Odontanthocopa) scutellaris Hoplitis (Nasutosmia) nasuta 65 100 Osmia (Allosmia) sybarita Osmia (Neosmia) bicolor 58 Osmia (Neosmia) tingitana 100 Osmia (Allosmia) rufohirta 100 Osmia (Tergosmia) rhodoensis 57 Osmia (Tergosmia) tergestensis 100 Osmia (Hemiosmia) difficilis Osmia (Hemiosmia) anceps 100 Osmia (Pyrosmia) ferruginea 99 97 Osmia (Pyrosmia) gallarum 99 Osmia (Metallinella) brevicornis 100 Osmia (Euthosmia) glauca 100 Osmia (Osmia) lignaria 100 60 Osmia (Monosmia) apicata Osmia (Osmia) cornuta 100 Osmia (Trichinosmia) latisulcata Osmia (Cephalosmia) montana 100 97 Osmia (Helicosmia) aurulenta 100 Osmia (Helicosmia) coloradensis 100 Osmia (Helicosmia) niveata 100 Osmia (Melanosmia) juxta 99 Osmia (Mystacosmia) nemoris 97 Osmia (Melanosmia) xanthomelana Osmia (Acanthosmioides) integra i. . Maorit rule consensus o trees ,, in the aesian analsis applin a separate model to each o the three enes. odes inconruent beteen this and an o the our other aesian analses are indicated ith solid suares.

6. Osmiini phylogeny

Table 5. n-likelihood values for the five different Bayesian analyses of the combined dataset. Model selection criteria (AICc and BIC values) were calculated following McGuire et al. (2007). Model likelihood Number of

Model (harmonic mean) parameters AICc BIC 3 partitions, RT -2932.72 1 5899.01 6079.11 3 partitions, AIC -29669.39 27 59393.6 61159. 9 partitions, RT -28539.31 82 5728.93 59323.59 9 partitions, AIC -2858.93 93 57112.00 5927.69 9 partitions, GTR SSR -29973.20 100 60155.8 62330.2

In spite of these differences in likelihood, trees yielded by the five Bayesian analyses (Table 5) were almost identical topologically. In Fig. 2, all conflicting topologies in the 50-majority rule consensus trees are indicated by black squares. The following relationships varied among the five different Bayesian analyses. Ochreriades was sister group of all other Osmiini in the three analyses under nine partitions (98 posterior probability in the GTRSSR model), but not related to the Osmiini in the analyses with three partitions. Haetosmia was sister to the Osmia group under both three-partition regimes, but sister group of the clade ainia, Atoposmia, Ashmeadiella, Osmia and Hoplosmia in the three analyses under nine partitions. astly, two subgenera of Osmia, Hemiosmia and erosmia, were monophyletic in all analyses under three partitions and the GTRSSR model, but formed a grade in the two other nine-partition analyses.

6...Biogeography We inferred the ancestral geographic range for nine selected nodes in the Bayesian tree (A-I in Fig. 3) The analysis allowing for three free rates of biogeographic exchange had a higher average ln-likelihood (-6.) than the analysis under the Mk1 model with only one rate (-69.9), this difference being significant (ln-likelihood ratio 10.9, d.f. 2, p 0.00 Pagel 1999). The estimations of the rates in the three-rate model suggest higher exchanges between the Palearctic and the Nearctic than between the Palearctic and the Afrotropic zones. Ancestral state reconstruction differed little between both models (Fig. 3). Both analyses indicated a Palearctic origin for nodes A, B, C, F, G and I, and an African origin for node . Pairwise comparisons with each node successively constrained to each of the three

65 smiiihloge

*

P Chelostoma (Eochelostoma) *0.97/0.98* P Chelostoma (Foveosmia) californicum N Chelostoma (Foveosmia) campanularum P Chelostoma (Chelostoma) P Chelostoma (Gyrodromella) N Chelostoma (Prochelostoma) P Chelostoma (Ceraheriades) ) , P Protosmia (Nanosmia) P Protosmia (Protosmia) *0.61/0.43 N Protosmia (Chelostomopsis) rubifloris 0.86/0.90* P Protosmia (Chelostomopsis) longiceps A Othinosmia (Megaloheriades) P Hofferia A, P Stenoheriades A Othinosmia (Othinosmia) - A, P Heriades (Heriades) N Heriades (Neotrypetes) 0.85/0.87 A, P Heriades (Michenerella) P Haetosmia circumventa P Hoplitis (Anthocopa) bisulca A Hoplitis (Anthocopa) spec. P Hoplitis (Anthocopa) villosa P Hoplitis (Anthocopa) hemisphaerica P Hoplitis (Pentadentosmia) P Stenosmia P Hoplitis (Hoplitis) + P Hoplitis (Annosmia) P Hoplitis (Chlidoplitis) P Hoplitis (Megahoplitis) *0.90/0.82* N Hoplitis (Proteriades, Penteriades and Hoplitina) / P Hoplitis (Micreriades) P Hoplitis (Microhoplitis) P Hoplitis (Platosmia) *0.99/1.00* P, N Hoplitis (Formicapis) P Hoplitis (Prionohoplitis) P Hoplitis (Alcidamea) tridentata . P Hoplitis (Alcidamea) mitis N Hoplitis (Dasyosmia) biscutellae N Hoplitis (Alcidamea) hypochrita *0.85/0.88* P, N Hoplitis (Monumetha) P Hoplitis (Alcidamea) leucomelana N Hoplitis (Alcidamea) pilosifrons A, P Wainia N Atoposmia N Ashmeadiella P Osmia (Erythrosmia) P Hoplosmia P Hoplitis (Nasutosmia) P Osmia (Allosmia) 0 P Osmia (Neosmia) P Osmia (Tergosmia) P Osmia (Hemiosmia) 0.60/0.58* P, N Osmia (Pyrosmia and Diceratosmia) P Osmia (Metallinella) N Osmia (Euthosmia) 1 P Osmia (Monosmia) N Osmia (Osmia) lignaria *1.00/1.00* P Osmia (Osmia) cornuta N Osmia (Trichinosmia) Palearctic N Osmia (Cephalosmia) P Osmia (Helicosmia) aurulenta N Osmia (Helicosmia) coloradensis 11.8 2.3 P Osmia (Helicosmia) niveata N Osmia (Melanosmia) juxta 0.0 N Osmia (Mystacosmia) P Osmia (Melanosmia) xanthomelana Nearctic Afrotropic N Osmia (Acanthosmioides)

igMaimmlielihoodecostctiooacestalgeogahicageoieselectedcladeso osmiie ees hee geogahic oes ee ecogied aleactic hite eactic ge ad otoiclachealesattheieselectedodesgietheoailitohaigthemostliel stateatthatodedeto models o chaacteeoltio theeideedetatesletale oall ates ealight aleie diagams eesettheoailit o eacho the thee states de the theeatemodelheasteissidicatethataalsescostaiigthemostlielstatehadsigiicatl highellielihoodalesthaaalsesithothalteatiestatescostaiedoisaliatiothe tee as da i MacClade ith asimo ecostctio o acestal ages he tee toolog coesods to the maoit le cosess tee i the aoed aesia aalsis eea ad sgeeaetielestictedtooegeogahicoeeesmmaiedtooetemialtao

. siini phylogeny states significantly supporte one state over the others asteriss in ig. for all noes ecept either uner both oels noes an or only uner one oel an .

..

n cobination the three selecte arers provie a strong phylogenetic signal for the taonoic level consiere in our stuy. psin as the ost useful gene an our results confir the preiction of scher et al. that this gene is particularly suitable to resolve recent ivergences in bees naely ithin tribes an genera. as surprisingly inforative in spite of the sall sie of the fragent inclue. t shos coparable properties as opsin such as unbiase base coposition an substantial nt an nt variation anforth et al. a an thus appears highly proising for the reconstruction of bee phylogeny. perfore coparatively poorly even for the basal noes consiering its high nuber of inforative sites. his arer is highly conserve at the aino aci level an uch of the silent nucleotie variation shos relatively high levels of hoplasy. evertheless these three genes copleente each other to yiel a ell-supporte phylogeny. ur results resolve any issues on the systeatics an taonoy of the siini. hey constitute an iportant first step toars the evelopent of a stable classification of this tribe an toars the evolutionary reconstruction of the iverse biology observe aong these bees.

... ysteatics

oohlasaeelassatoothesm ase on our stuy the onophyly of the siini can be sounly assesse for the first tie. our sall an rare genera o not appear to be relate to the siini. he genus oteaes as foun to be the sister group of the egachilini in all analyses. his placeent as reveale by each of the three genes able 4 an highly supporte in both parsiony bootstrap support an ayesian analyses posterior probability of the cobine ataset. ase on the orphology risol alreay suspecte a close relationship beteen oteaes an the egachilini. ost egachilini are characterie by the absence of arolia

7 6. Osmiini phylogeny between the claws, although two exceptions exist (Peters 1970, Baker and ngel 2006). oteiaes is thus a further example of a Megachilini with an arolium. Its basal position within the Megachilini makes it a suitable outgroup for phylogenetic studies of this tribe. In contrast, the placement of the two small genera oheiaes and seuoheiaes remains elusive. These two genera formed a well-supported monophyletic group in all analyses. However, they are not allied to the eiaes group as previously assumed (Michener 2007). Though they never appeared to be related to the Osmiini, we refrain from excluding them definitively from this tribe because their phylogenetic position varied in our analyses (Parsimony analysis sister to the Anthidiini Bayesian analysis sister to the Megachilini). Similarly, the position of the genus heiaes remains unsolved, being alternatively sister to the Osmiini (nine partitions regime) or sister to the clade Megachilini, Anthidiini and Osmiini (three partitions regime). These three genera do not seem to be closely related to any other tribe of the Megachilidae. To elucidate their position, analyses with other markers, additional megachilid species and a different outgroup are currently underway (B. Danforth, T. riswold and C. Praz, unpublished data).

Apart from these four small taxa, the remaining genera of the Osmiini, comprising approximately 99 of all known osmiine species, formed a well-supported monophyletic group in all analyses. Our data enable a clearer suprageneric subdivision of the Osmiini than that suggested by the current classification (Michener 2007). The genus Chelostoma, whose placement switched from the eiaes group to the smia group in the past, did not appear to be related to any of these two groups but clearly emerged as the sister group of all other Osmiini. Hence, it would deserve the same rank as the eiaes and the smia group, often recognized as subtribes Heriadina and Osmiina, respectively (ngel 2005, Ungricht et al. in press). The monophyly of the genus Chelostoma is also supported by several uniue morphological traits such as the orientation of the third and fourth labial palpi, the nonfringed labrum of females as well as the presence of a comb-like structure on sternum 5 of males (Michener 2007 chapter 7). The exclusion of Chelostoma and heiaes from the smia group, and of oheiaes, oteiaes and seuoheiaes from the eiaes group, sharpens the distinction between these two

6

Heriades ProtosmiaOthinosmia, Stenoheriades HeriadesHeriades Osmia Chelostoma Osmia Heriades Chelostoma Osmia Hoplitis Atoposmia AshmeadiellaHaetosmiaOsmia

Generic classification StenosmiaHoplitis 4 Hoplitis Hoplosmia Osmia Hoplosmia Osmia Osmia Micreriades Hoplitis Alcidamea asutosmiaHoplitis Osmiaasutosmia Hoplitis Hoplitis asutosmia Osmia Hoplosmia asutosmia

smiii hloge hitheto thoght deed thee ae seveal othe secies ithi the ges Osmia hich have liea aasidal lies the sgeea llosmia ad rythrosmia i Osmia Pyrosmia cephalotes as ell as i seveal descied secies o dieet sgeea o Osmia Mlle lished data hese secies eseciall i the emale se caot easil e distigished om secies o Hoplitis Michee sggested emoval o the sgeea llosmia ad rythrosmia om the ges Osmia to shae its deiitio hloge does ot sot sch eclsio ad the oits ot the vage mohological odaies etee the geea o the Osmiago e acoledge that megig all cetl ecogied geea o the Osmia go ito oe sigle ges Osmia as is doe ma oea athos estich cha et al miet et al is ot icoget ith o hloge oeve e sggest the etetio o the geeic classiicatio o Michee ith the thee ecessa alteatios detailed aove is classiicatio hich is ased o the detailed std o isold ad Michee has alead gaied ide accetace ad cetail ill emai the stadad eeece i the te heeoe a sstatial chages old theat omeclatal stailit deed the idetiicatio es i Michee allo the amigos assigmet o ove o all osmiie secies to the geeic ad sgeeic level

Future research he olloig hlogeetic elatioshis emai solved ad call o the aalses i he ges Othinosmia aeaed aahletic i all aalses t moe secies shold e iclded eoe a classiicato chages ae made ii hloge ailed to ecove the moohl o the ges toposmia aticla the ositio o the sges Hexosmia emais clea iii cotast to asimo that laced the ges Haetosmia at the ase o the Osmia go some aesia aalses laced it as siste to the clade Wainia-toposmiashmeadiellaOsmia he latte lacemet is soted the stcte o the male stea hich ae simila i Haetosmia Wainia toposmia ad Osmia iv l oe sges o Wainia as iclded i o std s the othe sgeea sstatiall die om it the moohl o this ges shold e coimed icldig additioal secies v imilal coimatio o the moohl o

6. Osmiini phylogeny

tenoheriades is needed. Our study includes a representative of only one of the four species groups within tenoheriades These groups span morphologies that otherwise are used to differentiate genera (Griswold 1985). vi. epresentatives of the rare central Asian taxa umoia and aartinula presently placed as subgenera of olitis (Michener 2007), could not be included in our study. Some species of these taxa show an odd combination of morphological characters rendering them intermediate between smia and olitis.

6.5.2. Biogeography

Our data indicate a Palearctic origin for the Osmiini (ig. 3), a result largely supported by the high genus and species diversity observed in the Old World osmiine bees. Only the eriades group seems to be eually diverse in sub-Saharan Africa and the Palearctic (Griswold, 1985). Our data indeed suggest an African origin for one clade of this group. urther, we found significantly more exchanges between the Palearctic and the Nearctic than between the Palearctic and the Afrotropic zones (ig. 3), a result that has important implications for the understanding of the biogeography of the Osmiini and of bees in general. The comparatively low dispersal between the Palearctic and sub-Saharan Africa might reflect barriers formed by both deserts and tropical ecosystems. Indeed, the Osmiini are rare in tropical regions (Michener 2007). The extended tropical areas of both Central America and the Indo-Malaysian region might have prevented the Osmiini from colonizing South America and Australia, respectively, where this group of bees is completely lacking. On both continents, regions with Mediterranean climates exist that would provide suitable habitats for the osmiine bees.

Our phylogeny reveals at least 15 exchanges between the Old and the New World, and this value will certainly increase once detailed phylogenies of holarctic subgenera are available (e. g., onumetha, smia s. str, elanosmia, Alidamea). or the maority of these exchanges, our data imply colonization of North America from the Palearctic, but the reverse pattern of colonization is possible in some holarctic groups eually diverse in North America and the Palearctic (e. g., elanosmia). The high number of exchanges between the Palearctic and North America calls for an

71 smhloe elaato as ths sesalate s lealhhe tha that osee othe os oees es et al aoth et alat et al t see Cameo et al es ome olaate holatosmeleaeshhhaeaoealsttoothasa aoth meaeMelanosmiaMonumethaFormicapismahae ee ale to oss the etat elateleet tmes mle eesmostsesaleetseteealeataeathaeoe laesoolaateseesmost lelthelast emllos easesmlalseealmammalaleaesoalosse a la e etee lasa a ea mllos eas ao t aeaeetalotasttotheoealosmeleaes maothmeaosmetaaaeesttetoohaetheeteo stto a eos o the sotheste eseall the aea eo hee e Chelostoma (Foveosmia), eroheriades Hoplitis (Proteriades), Atoposmia a Ashmeadiella hs sttoattemhtateolesesaleetsomtheeaste hemsheeetheleoeeasoseetolaeso lessolaatemleeesaeseteles a shes eet aloee ao the loeeoeeoat

teestlmaothmeaosme taa lestsea ooostemsheeasthealeatelatesshoahheest o est ehaos l est the o hs s est eemle the es Hoplitis hee the seose oest seea o the l ol ae aset om oth mea e Anthocopa, PentadentosmiaAnnosmiasteotheestto the aleathe maot o the mea sees o Hoplitis est ooostemsheehesameattesoseeProtosmia aOsmiaHelicosmia)heeseesothesetaaesetthee ol est oo as a as o sol Cae et al heeheeas the lol elatesmost lel the o om hh the hae eole sho a mh moe aale est ehaoheelleaCalshesseste heeooesteesaemoeleltoossate aestha oesteesas oeseas aae oests oo

smiii phloe ma ehace ispesal o, the hih pecetae o ooesti taa amo the osmiie ees miht eplai h the me o echaes etee the aleactic a oth meica is hihe tha that osee i othe ops o ees, hich pimail est i the o aoth et al 24, ati et al 2 imilal, isct istitios osee i othe aoeo esti ee taa, e , i the Chelostomoides op o seea o Megachile ichee o i lloapii cha et al 2, miht also e patl eplaie oeseas ispesal o ests

3 7. ost-plant choice in eosto

7. Patterns of host-plant choice in bees of the genus Chelostoma: the constraint hypothesis of host-range evolution in bees

7.1. ASA

o trace the evolution of host-plant choice in bees of the genus eosto (Megachilidae), we assessed the host plants of 3 Palearctic, North American and ndomalayan species by microscopically analyzing the pollen loads of 634 females and reconstructed their phylogenetic history based on four genes and a morphological dataset, applying both parsimony and ayesian methods. All species except two were found to be strict pollen specialists at the level of plant family or genus. hese oligolectic species together exploit the flowers of eight different plant orders that are distributed among all major angiosperm lineages. ased on ancestral state reconstruction, we found that oligolecty is the ancestral state in eosto and that the two pollen generalists evolved from oligolectic ancestors. he distinct pattern of host broadening in these two polylectic species, the highly conserved floral specializations within the different clades, the exploitation of unrelated hosts with a striing floral similarity as well as a recent report on larval performance on non-host pollen in two eosto species clearly suggest that floral host choice is physiologically or neurologically constrained in bees of the genus eosto. ased on this finding, we propose a new hypothesis on the evolution of host range in bees.

7.2. NODON

ees are the major pollinators of angiosperms in most ecosystems (Michener 2007). hey provision their brood cells with large amounts of pollen and nectar, which maes the bees indispensable mutualists of flowering plants on the one hand and very effective herbivores on the other (esteramp 1996, Mller et al. 2006). n their natural habitats, bees are often confronted with a dazzling array of different flowers from which

ased on Sedivy, ., Praz, ., Mller, A., idmer, A. and Dorn, S., submitted to outon. he contribution of . Praz consisted of the supervision of the phylogenetic reconstruction and molecular wor carried out by the diploma student . Sedivy, of the inference of ancestral state, and of part of the redaction.

74 7. Host-plant choice in Chelostoma they have to mae the most rewarding choice. In fact, while some bee species exploit a wide range of different flowers, others restrict their flower visits to closely related plant taxa. obertson (1925) was the first to recognize that this floral specificity is limited to the collection of pollen but not to the uptae of nectar. He introduced the terms oligolectic for pollen specialists and polylectic for pollen generalists. ligolectic bees are characterized by consistently collecting pollen from flowers of a single genus, subfamily or family (insley and MacSwain 1958, Westrich 1989, Cane and Sipes 2006, Müller and Kuhlmann in press). In contrast, polylectic bees exploit flowers of more than one plant family.

Polylectic and oligolectic species coexist in all investigated bee faunas. Therefore, both polylecty and oligolecty obviously represent successful evolutionary strategies. Whereas polylecty is considered advantageous in reducing dependence upon a limited number of pollen hosts (Moldene 1975, icwort and insberg 1980), the ecological and evolutionary factors that select for oligolecty or act to maintain it remain subect of several, mostly untested hypotheses. ne traditional assumption is that oligolecty has evolved to reduce interspecific competition for pollen (obertson 1899, 1925, insley 1958, Michener 1954, 1979, Thorp 1969). This hypothesis is based on the observation that pollen specialists are especially abundant in species-rich bee faunas, with up to 60% oligoleges in Californian deserts (Michener 1979, Mincley and oulston 2006). If competition is the most important factor, closely related bee species are expected to harvest pollen on different plant taxa. However, several studies that combined bee phylogenies with pollen preferences suggest that close relatives are generally specialized on the same pollen hosts (Müller 1996b, Wcislo and Cane 1996, Michez et al. 2004, 2008, Sipes and Tepedino 2005, Mincley and oulston 2006, Patiny et al. 2007, arin et al. 2008). Another assumption addresses the possibly higher foraging efficiency of specialist bees compared to generalists that selects for host specificity (ovell 1913, 1914). Indeed, some studies demonstrated that specialist bees are actually more efficient in pollen harvesting than generalists (Stricler 1979, Cane and Payne 1988, averty and Plowright 1988). However, a comparison of foraging rates of oligolectic and polylectic bees on eao sata (Fabaceae) indicates that specialists are not always faster than generalists

75 7. Host-plant choice in Chelostoma at utilizing shared hosts (Pesenko and adchenko 1993, Minckley and oulston 2006). Similarly, no differences were observed in the flower handling techniques of oligolectic versus polylectic anthidiine bees (Müller 1996b).

Traditionally, it has been a widely accepted assumption that oligolectic bees have evolved from polylectic ancestors (Michener 1954, Linsley 1958, MacSwain et al. 1973, Iwata 1976, Moldenke 1979, Hurd et al. 1980). Indeed, there do exist some clear examples of transitions from polylecty to oligolecty, e.g., in the genus Lasioglossum where oligolectic species have evolved twice within clades of polyleges (Danforth et al. 2003). However, growing evidence suggests that many generalist bee species have evolved from oligolectic ancestors. The basal clades of most bee families include a high proportion of pollen specialists (Westrich 1989, Wcislo and Cane 1996, Patiny et al. 2007). The Dasypodaidae and Melittidae, which are probably the most basal bee families, are predominantly composed of oligoleges, suggesting that oligolecty might be the ancestral state in bees (Danforth et al. 2006b, Michez et al. 2008). Oligolecty is also assumed to be the plesiomorphic condition in the genus Andrena, with polylecty having independently evolved several times (Larkin et al. 2008). Furthermore, polylecty appears to be a derived trait in several anthidiine bees as well as in a pollen-collecting masarine wasp (Müller 1996b, Mauss et al. 2006). iven the huge pollen quantities needed to rear a single bee larva (Schlindwein et al. 2005, Müller et al. 2006), strong selection should act on oligolectic bees to reduce their heavy dependence upon a limited number of host plants. However, pollen specialists are widespread and outnumber the generalists in numerous bee clades as well as in some habitats (Westrich 1989, Minckley and oulston 2006, Michener 2007). Therefore, oligolecty in bees is possibly best considered as an evolutionary constraint that has been repeatedly overcome in many polylectic bee lineages (Müller 1996b, Larkin et al. 2008).

ecently, two possible constraints have been identified that might prevent oligolectic bee species from becoming polylectic or from switching hosts, i.e., constraints linked to pollen digestion and neurological (including cognitive) constraints. First, the failure of several specialized bee species to

76 7. Host-plant choice in Chelostoma develop on non-host pollen clearly indicates that the pollen of some plant taxa possesses unfavorable or protective properties that render its digestion difficult (Chapter 4). Similarly, the pattern of use of Asteroideae pollen by bees of the genus Colletes suggests that this pollen has chemical properties which interfere with its digestion by generalists (Müller and Kuhlmann in press). Therefore, physiological adaptations might be needed to overcome the protective properties of some pollen types. This in turn may constrain the bees capability to use other pollen types similar to herbivorous insects, where adaptations to the secondary chemistry of their hosts may result in a lower capability to exploit alternative hosts (Strauss and angerl 2002, Singer 2008). Second, constraints in recognizing or handling non-host flowers are likely to prevent the pollen-specialist bee Heriades truncorum from becoming polylectic (Chapter 5). This species, which exlusively collects pollen on Asteraceae in nature, was found to be able to develop on several types of non-host pollen. However, the females refused to collect non-host pollen despite its suitability for larval development even in the absence of the normal host. This finding suggests that neurological limitations are more important than nutritional constraints in shaping the host range of this species.

Phylogenetic inference is a powerful method to uncover patterns of host-plant choice and to test hypotheses on the evolution of host-plant associations (Harvey 1996). So far, only few studies applied phylogenetic inference to analyse bee-flower relationships at species level (Müller 1996b, Michez et al. 2004, Sipes and Tepedino 2005, Larkin et al. 2008). Most of these studies, however, were restricted to one biogeographical region and did not include the whole diversity of the bee taxon under investigation. In the present study, we used phylogenetic inference to analyse patterns of host-plant choice in bees of the genus Chelostoma (Megachilidae, Osmiini). Though this genus is assumed to consist of mainly oligoleges (Michener 2007), solely the pollen preferences of the few central European and orth American species are known so far (Hurd and Michener 1955, Moldenke and eff 1974, Krombein et al. 1979, Parker 1988, Westrich 1989, 1993, Gogala 1999, Amiet et al. 2004, Michener 2007). y including a substantial proportion of species from all three biogeographical regions where the genus is known to occur, we provide

77 . ostlant hoie in Chelostoma the irst stu o eeloer relationshis on a orlie sale. eiiall e aresse the olloing uestions hih are the loer reerenes o the ierent Chelostoma seies s oligolet the anestral state in the genus Chelostoma an hae olleti seies eole rom oligoleti anestors or ie ersa s hostlant hoie a onsere trait ith memers o the same lae haing the same host reerenes re the osere atterns o hostlant hoie in line ith the hothesis that the oligoleti hait is onstraine hsiologial or neurologial limitations ase on our inings e outline a ne hothesis on the eolution o host range in ees.

..

... ee seies

he genus Chelostoma hih is iie into si sugenera ihener is reresente esrie seies in the alearti region nine seies in orth meria an a single seies in sutroial an troial southeast sia ihener ngriht et al. sumitte. he entre o iersit is situate in the eastern eiterranean area o uroe an estern sia. here is strong eiene that the genus Chelostoma is monohleti an sister to all other seies o the trie smiini Chater.

or the resent stu e selete a total o seies an suseies taa rom the alearti eight rom the earti an one rom nomalaa ale or hih enough ollen samles ere aailale or assessing loral host range. hese seies reresent all si sugenera urrentl reognie an enomass most o the morhologial ariailit ithin Chelostoma. our seies are ne to siene the are reerre to as species 2, 3, 23 an 24 resetiel. heir esrition is in rearation .ller unulishe. ouher seimens o all ee taa selete or the resent stu are eosite in the ntomologial Colletion at the urih. he nomenlature ollos romein et al. an ngriht et al. sumitte or the genus Chelostoma ngriht et al. sumitte or the smiini an har et al. or the other megahili seies.

ost-plant choic in Chelostoma

al 1 h nin otgrop spcis and th Chelostoma spcis incldd in this std or spcis lacing localit data onl orphological data r aailal Collctors A A llr D Danforth C Cra C Cladio di

pcis Localit Collctor Outgroup Lithurgus chrysurus Ital Arn assa A Anthidium punctatum itrland Wiach A Megachile pilidens itrland Wiach A Ochreriades fasciatus ordan Wadi ha C C A Hofferia schmiedenechti rc Chiara C C Hoplitis adunca Ital Aosta A Osmia cornuta itrland rich A Protosmia minutula itrland Ed C Heriades truncorum itrland Wintrthr A Subgenus Ceaheaes Chelostoma lamellum China nan roinc C Subgenus Chelostoma Chelostoma carinulum - - Chelostoma diodon rc Lsos A rac Chelostoma edentulum orocco oss A Chelostoma emarginatum rc latania tandfss Chelostoma florisomne itrland Chr E tinann Chelostoma grande itrland Erschatt C Chelostoma mocsaryi rc latania C C Chelostoma nasutum rc Andhritsna C C Chelostoma species 2 rc Cprs C chid-Eggr Chelostoma species 3 ordan Dad a C C A Chelostoma transversum rc achloro C C Subgenus ohelostoma Chelostoma aureocinctum hailand Chiang ai C Subgenus oeosma Chelostoma bytinsii ordan Wadi i C C A Chelostoma californicum A CA ariposa Co risold Chelostoma campanularum itrland Wintrthr A Chelostoma cocerelli A CA olon Co is Chelostoma distinctum itrland Edt A Chelostoma foveolatum Ital oscana A Chelostoma garrulum - - Chelostoma hellenicum rc agtos ts C C Chelostoma incisulum A CA olon Co L rst Chelostoma isabellinum - - Chelostoma laticaudum rc Andhritsna C C Chelostoma m. marginatum - - Chelostoma m. incisuloides A CA olon Co W Ird Chelostoma minutum A CA olon Co risold Chelostoma phaceliae A CA olon Co L rst Chelostoma species 23 - - Chelostoma species 24 - - Chelostoma styriacum rc ichas-Laoata C C Chelostoma tetramerum - - Chelostoma ventrale r Anara E chchl Subgenus oomella Chelostoma rapunculi itrland ll C Subgenus ohelostoma Chelostoma philadelphi A D r orgs co Drog

. ostplant choice in helostoma

... ost plants

o assess the pollen hosts o the helostoma taxa selected or this study e analyed the scopal contents o emales rom museum or priate collections y light microscopy using the method outlined y estrich and chmidt . or each species e sampled specimens rom as many localities as possile to account or potential dierences in pollenhost use o dierent populations. eore remoing pollen rom the adominal scopae e estimated the degree to hich they ere illed. he amount o pollen as assigned to ie classes ranging rom ull load to illed only to oneith. he pollen grains ere stripped o the scopae ith a ine needle onto a slide and emedded in glycerine gelatine. e estimated the percentages o dierent pollen types y counting the grains along our lines chosen randomly across the coer slip at a magniication o x. ollen types represented y less than o the counted grains ere excluded to preent potential ias caused y contamination. n loads consisting o to or more dierent pollen types e corrected the percentages o the numer o pollen grains y their olume. ter assigning dierent eights to scopae according to ho illed they ere ull loads ere ie times more strongly eighted than scopae illed to only one ith e summed up the estimated percentages oer all pollen samples o each species. he pollen grains ere identiied at a magniication o x or x ith the aid o the literature cited in estrich and chmidt and an extensie reerence collection. dentiication o the pollen samples rom the orth merican species as acilitated y Constance and Chuang ho gie a surey on the pollen morphology o the ydrophyllaceae. n general e identiied the pollen grains don to amily or i possile to genus leel those o the steraceae don to the suamilies steroideae and Cichorioideae respectiely.

o characterie dierent degrees o hostplant association among the helostoma species inestigated e used the to categories oligolecty and polylecty sensu Mller and uhlmann in press. e did not dierentiate eteen the sucategories o oligolecty and polylecty as deined y Cane and ipes and Mller and uhlmann in press respectiely. o classiy a helostoma species as oligolectic e applied to

7. Host-plant choice in Chelostoma different approaches introduced by Mller (16) and Sipes and Tepedino (2005), respectively. A species was designated as oligolectic if i) 5 or more of the pollen grain volume belonged to the same plant family or genus, or ii) if 0 or more of the females collected pure loads of one plant family or genus. Both approaches yielded exactly the same categorizations for all Chelostoma species analysed. To infer the host range of those species for which only a small number of pollen loads was available, both the literature and unpublished field data were also considered (Appendix 1).

7.3.3. Molecular phylogeny

seuences reshly collected material allowing for DNA extraction was available for 28 of the 35 Chelostoma taxa included in the present study (Table 1). As outgroup species, we selected five representatives of the tribe Osmiini (eiaes tuncoum offeia schmieenechti oplitis aunca smia conuta and otosmia minutula), one species each of the tribes Lithurgini (ithuus chsuus), Anthidiini (nthiium punctatum) and Megachilini (eachile piliens), and cheiaes fasciatus, a megachilid bee originally assigned to the Osmiini (but see chapter 6). We generated a DNA matrix composed of 3018 aligned sequences from four genes: the three nuclear genes Elongation factor-1 (2-copy hereafter E), Long-wave rhodopsin (opsin), Conserved ATPase domain (CAD) and the mitochondrial gene Cytochrome oxidase subunit 1 (COI). Preliminary phylogenetic analyses indicated that the coding sequence of E was too conserved for the phylogenetic level considered here: there was almost no variation in codon positions 1 and 2 in the ingroup and only little, mostly silent, variation in position 3, which was AT-biased and had a biased base composition across species. We therefore sequenced only the two introns (approximately 200 and 240 bp, respectively) included in the 1600 bp fragment often used to infer bee phylogeny (Danforth et al. 2004). or opsin, we included both the coding sequence (600 bp) and three introns (approximately 80-100 bp each). The fragment used for CAD (448 bp) had no introns and corresponds to exon 6 in the fragment described by Danforth et al. (2006a). or these three nuclear markers, we used primers designed for bees in general (Danforth et al. 2004, 2006a), for the Osmiini (chapter 6) as well as new

81 7. Host-plant choice in Chelostoma primers specific to Chelostoma (Table 2). The fragment of CI consisting of 1219 bp was amplified with universal primers for insects (hang and Hewitt 1996 Table 2). For five Chelostoma species, chromatograms of CI seuences revealed several double peaks indicating the presence of pseudogenes (ensasson et al., 2001). As these double peaks concerned less than 1 of the complete seuence for two species and less than for two others, and as no indels were found, we did not use cloning techniues to separate the different copies but coded all base pairs with double peaks as N. The seuence of C. emarginatum contained too many double peaks and was therefore excluded from the analyses.

NA was extracted from bees preserved in 100 ethanol and from a few pinned specimens up to five years old following the extraction protocol of anforth (1999). CR-amplification products were purified using GF NA purification kit. Automated seuencing of the CR products was performed on an AI rism 3130xl seuencer using igye technology. The locality data for the specimens used for NA extraction are listed in Table 1.

Alignments for all genes were performed manually using acClade version .08 for S (addison and addison 200). Some regions in the ingroup as well as all outgroup intron seuences could not be aligned unambiguously and were excluded. e initially coded five indels that could be unambiguously aligned as additional characters, but as they did not influence the phylogenetic results, we excluded them from all subseuent analyses. To ensure that the correct reading frame of each gene was found, the coding seuences were converted into amino acid seuences. No stop codons were found in any of the exons of the genes. The complete alignment will be deposited in TreeAS (www.treebase.orgtreebase index.html).

82 Hostplant choice in Chelostoma

ale Primers used for the four genes longation factor, Lrhodopsin, Consered APase Domain CAD and Cytochrome Oidase CO

Primer equence eference

longation factor Haor AA A C CAA A C Danforth et al eMeg AA CA CA CAC CC ora AC CA AA C C Cho AC C AC H CC A C Danforth et al ntrone AAA AA CC CC AA AC onor CC AC AA CC ACA AC AAA C C one A CAA A A CAA CCA AC C PC conditions HaoreMeg C, C, C oraCho C, C, C onoreMeg C, C, C

hodopsin OpsorOsm AA A A ACA Opsine CC AA A CAC C CA C Opsinorc C C ACC A A C Opsine CC A A AA C PC conditions OpsinorOsmOpsine C, C, C OpsinorcOpsine C, C, C

C CADor AA A A AC A C Danforth et al a CADeMeg CC AC AC C CC A CC C A PC conditions CADorCADeMeg C, C, C

CO A A AC A CCC AC AA AAA AA Lunt et al A A A CC N CAN CA A A A Lunt et al A CAA C AC A C CC AA Lunt et al COChele ACN CA A A A COChelor A A A A A A PC conditions AA C, C, C AA C, C, C ACOChele C, C, C COChelor A C, C, C

hloeeticaalses e performed parsimony analyses of each gene separately in Paup ersion for Macintosh offord, ith the folloing parameter settings uneighted analysis, heuristic search, ranch sapping, ootstrap replicates, random sequence additions, four trees held at each step, maimum of trees retained As ery little incongruence as osered eteen the four genes and as no conflicting topology as supported y ootstrap alues aoe , e comined the four genes into a single matri and analyed the comined dataset ith

7. Host-plant choice in Chelostoma the same parameter as above performing 1000 bootstrap replicates. e performed analyses with and without the third nucleotide position of COI, which was strongly AT-biased and hence prone to high level of homoplasy (Danforth et al. 2003).

For the Bayesian analyses, the four genes were analyed collectively under four different partitioning regimes. First, we partitioned the dataset by gene, yielding four partitions. Second, we partitioned opsin into two partitions, the coding seuence and the introns, resulting in a total of five partitions. Third, we partitioned COI into three partitions for the first, second and third nucleotide position, which yielded seven partitions. Lastly, we performed a fully partitioned analysis with eleven partitions, with a separate TR model applied to each gene and codon position (TR SSR). Analyses by MrModeltest (ylander 2004) identified the following best models of seuence evolution for each partition: EF intron, HY opsin, TR I opsin coding seuence, HY I opsin intron, TR I CAD, 80 COI, TR I COI nt1, TR I COI nt2, TR I COI nt3, TR . A posteriori examination of parameter plots with Tracer version 1.4 (Rambaut and Drummond, 2003) indicated that the proportion of invariant sites (I) and the shape () parameters could not be properly estimated for the three site-specific partitions of COI, and hence we applied the TR model to these three partitions. To select between these partitioning regimes, we calculated Bayes factors (ylander et al. 2004) and Akaike Information Criteria (AIC) using the formula of Mcuire et al. (2007).

Markov Chain Monte Carlo analyses were performed using MrBayes 2.1.1 (Huelsenbeck and Ronuist 2001). e performed two simultaneous runs with one cold and three heated chains each (temperature parameter fixed to 0.05) for two million generations, sampling trees and parameters every 100 generations. The onset of stationarity was determined by an examination of plots for log-likelihood values and for all parameters using Tracer. All trees sampled before stationarity (usually 10%) were discarded and the remaining trees from both runs were combined into a single maority rule consensus tree in Paup.

84 ostlatchocees

g gen n sue nss or see o the es seces clded ths stdy oly morholocal data ere aalaleo clde these see taa to the hyloeyecomedthemoleclaradthemorholocaldatasetto a totaledece sermatr ero ad Gatesy o collect morholocal charactersoth males ad emales o all es seces ere eamed eterally s a dssect mcroscoe addtoedsmemered the adomeo the males toet arorate esotheothersehddesteraadtheetalaademeddedthe scoalhars o the emales lycere elate or mcroscocal stdy hesearchormorholocalcharactersasacltatedythelcatos ocheer rd ad cheerarce adcharadGseleterhemorholocalaalyssyelded characters see ed ad e dd ot code morholocal characters or the otro seces as homoloy roed mossle to esre e selected uenu as a otro or aaly the morholocaldataset aloeas ths secesamoslyaeared as sster to all other es seces the moleclar hyloeetc aalyses e rst erormed arsmoy aalyss o the morholocal dataset aloe Pa th the ollo arameter settsall characters ehted ad treated as ordered herstc search rach sa ootstra relcates radom seece addtosor trees held at each ste mamm o trees retaed hee comed the morholocal ad the moleclar data sets ad erormedarsmoyadayesaaalysestholythosesecesor hch oth datasets ere aalaleor the morholocal artto e aledasmlecharactermodelthalltrastoratesealadaed roorto ocharacterstates amma shae dstrtoo rates as otttedtothemorholocalarttoasrelmaryaalysesaledto correctlyestmatethsarameterastlyeraaayesaaalyssthall secescldedcodoth thelacmoleclardataor thesee addtoalsecesadthelacmorholocaldataortheotroas mssdatahssermatrasaalyedthrayesder the aoredarttoremeormoleclardataadaaddtoalartto comosed o the morholocal data set aly the same model as

7. Host-plant choice in Chelostoma above. We ran 5 millions generations and constrained the ingroup (Chelostoma) to be monophyletic.

Evolution of host-plant choice To reconstruct the evolution of host-plant choice within the genus Chelostoma, we first applied parsimony mapping in MacClade, using the topology of the majority rule consensus of trees saved in the Bayesian analysis of the supermatrix. As parsimony reconstruction of ancestral state does not take branch length into account, we used maximum likelihood inference of ancestral character states as implemented in BayesTraits (Pagel et al. 2004, Pagel and Meade 200). Transition rates between all states (i.e., pollen hosts) were assumed to be equal using the restrictall command in BayesTraits. We used two samples of trees saved during Bayesian analyses of the supermatrix: first, the analysis including only those 28 species for which both molecular and morphological data were available, and second, the analysis with all 35 species. In the second supermatrix analysis, the length of branches leading to the seven species without molecular data could not satisfactorily be estimated due to the missing DNA data. However, as these species were all closely related to other species and well nested within the clades for which ancestral state reconstruction was performed, we postulate that the biased branch lengths did not substantially affect the results. The Bayesian analyses were allowed to run for 2 millions generations after convergence, saving trees every 4000 generations in both runs resulting in a total of 1002 trees. We excluded the outgroup taxa in Mesquite for OS (Maddison and Maddison 2007).

7.4. RESLTS

7.4.1. Host plants

We classified 2 out of the 35 Chelostoma taxa selected for this study as oligolectic based on the microscopical analysis of pollen loads (Table 3) as well as on the literature and unpublished field data (Appendix 1). These specialized species restrict pollen harvesting to Campanulaceae (10 species), Hydrophyllaceae (), Ranunculaceae (3), Asteraceae (2), Dipsacaceae (2), Brassicaceae (1), Ornithogalum (Hyacinthaceae) (1) and

8 7. Host-plant choice in Chelostoma

hiladelphus (Hydrangeaceae) (1). In six additional species, the pollen loads exclusively consisted of pollen of ampanulaceae (3 species), Ranunculaceae (1), Allium (Alliaceae) (1) and chima (Theaceae) (1). The small number of pollen loads available for study and the lack of literature or field data did not allow to unambiguously classify these six species as pollen specialists (Table 3). However, because the closest relatives of the putative ampanulaceae specialists C. isabellinum, C. garrulum and C.species are strictly oligolectic on ampanulaceae, and as the presumed Ranunculaceae specialist C. species is a member of a clade which contains several Ranunculaceae oligoleges, these four species are most probably pollen specialists as well. Similarly, C. tetramerum and C.aureocinctum are treated as oligolectic on Allium and chima, respectively. Four out of the five pollen loads of C. tetramerum originated from different localities and the nine females of C. aureocinctum available for study were collected at seven different localities in India, epal, Thailand and hina, which clearly points to a pollen specialization. Two Chelostoma species turned out to be polylectic harvesting pollen on at least five (C. species 3) and three plant families (C. minutum), respectively (Table 3). The host range of C. lamellum, which is known only from Sichuan, unnan and Gansu province in hina, is still unknown. The only two pollen loads available for study contained pollen grains that could not be differentiated microscopically from pollen samples of hiladelphus (Hydrangeaceae). Small pollen packages removed from the labrum and clypeus of two females, which have been collected at two different localities in unnan province in 1992 and 1993, respectively, were also composed of hiladelphus grains. In addition, the only two individuals of C.lamellum we collected in hina in uly 2006 were observed flying near flowering hiladelphus shrubs. Therefore, we hypothesize that hiladelphus is an important or even the exclusive pollen host of C. lamellum. This assumption is supported by the finding that one pollen load of the closely related hinese species C. sublamellum, which could not be included in our phylogeny due to the lack of specimens for study, entirely consisted of hiladelphus pollen.

87 . ostplant choice in Chelostoma

able 3 ostplant preerences and inerred categor o host range in 3 Chelostoma species and subspecies. n total number o pollen loads number o pollen loads rom dierent localities. Abbreiations o plant taa AAlliaceae APApiaceae ASAsteraceae Arassicaceae CACampanulaceae CSCistaceae Pipsacaceae Aacinthaceae drophllaceae Adrangeaceae Aanunculaceae Sesedaceae heaceae gophllaceae. he nomenclature o Chelostoma ollos ngricht et al. submitted or the Palearctic and ndomalaan species and rombein et al. or the eartic species. Subgeneric classiication o Chelostoma according to ichener 2. Species or hich literature and unpublished ield data ere used to iner host range in addition to the results o the microscopical pollen analsis are mared ith an asteris see Appendi or a compilation o published and unpublished loer records.

Subgenus and species n esults o microscopical analsis ost range o pollen loads pollen grain olume pure loads o preerred host ubgenus Ceaheaes Chelostoma lamellum 2 2 A c. hiladelphus unnon possibl oligolectic on hiladelphus A see tet ubgenus Chelostoma Chelostoma carinulum A oligolectic on A Chelostoma diodon 3 AS Asteroideae .4 .2 oligolectic on AS AS Cichorioideae .2 A .4 Chelostoma edentulum 2 2 A .3 S 3.3 A .3 2. oligolectic on A AS Asteroideae .2 Chelostoma emarginatum 32 2 A oligolectic on A Chelostoma florisomne 4 42 A oligolectic on A Chelostoma grande 2 P oligolectic on P Chelostoma mocsaryi 32 22 A c. rnithogalum 3. oligolectic on rnithogalum . A . AP . Chelostoma nasutum 24 CA oligolectic on CA Chelostoma species 2 4 4 A probabl oligolectic on A Chelostoma species 3 3 CS 4. A 3.2 A .4 pollectic 24. S 3.4 . Chelostoma transversum 2 P oligolectic on P

ubgenus ohelostoma Chelostoma aureocinctum Schima probabl oligolectic on Schima

ubgenus oeosma Chelostoma bytinskii 4 3 CA oligolectic on CA Chelostoma californicum 2 22 hacelia oligolectic on hacelia Chelostoma campanularum 34 24 CA oligolectic on CA Chelostoma cockerelli 4 oligolectic on Chelostoma distinctum CA oligolectic on CA Chelostoma foveolatum CA oligolectic on CA Chelostoma garrulum CA oligolectic on CA Chelostoma hellenicum CA oligolectic on CA

7. Host-plant choice in Chelostoma

TABLE 3. (continued)

Chelostoma inisulum 29 19 HD (Phaelia) 95.2, ALL 96.6 oligolectic on Phaelia (cf. Allium) 4.8 Chelostoma isaellinum 2 2 CA 100 100 probably oligolectic on CA Chelostoma latiaudum 15 11 CA 100 100 oligolectic on CA Chelostoma m. marinatum 5 2 HD (Phaelia) 100 100 oligolectic on HD Chelostoma m. inisuloides - - - oligolectic on HD Chelostoma minutum 18 12 HD 69.9, ALL (cf. llium) 70.6 polylectic 23.6, unknown 6.5 Chelostoma phaeliae 46 17 HD (Phaelia) 100 100 oligolectic on Phaelia Chelostoma speies 1 1 CA 100 100 oligolectic on CA Chelostoma speies 2 1 CA 100 100 probably oligolectic on CA Chelostoma styriaum 11 5 CA 100 100 oligolectic on CA Chelostoma tetramerum 5 4 ALL (cf. llium) 100 100 probably oligolectic on llium Chelostoma ventrale 34 25 AST (Asteroideae) 100 100 oligolectic on AST Subgenus Gyrodromella Chelostoma rapunuli 36 31 CA 100 100 oligolectic on CA Subgenus Prochelostoma Chelostoma philadelphi 21 11 HDRA (cf. Philadelphus) 100 oligolectic on Philadelphus 100

7.4.2. Phylogeny

aximum parsimony bootstrap analysis of the combined molecular data set with the exclusion of the third codon position of C yielded an almost completely resolved tree (26 of the 27 nodes with bootstrap support above 50%, see values in Figure 1). The inclusion of the strongly AT-biased third codon position of C had only minor influences on the tree topology, however the tree was slightly less well supported (22 of the 27 nodes with bootstrap support above50%). Parsimony analysis of the morphological data set alone yielded 385 most parsimonious trees. The topology of the strict consensus tree (Appendix 4) was largely congruent with the molecular trees based on parsimony. Parsimony analysis of the combined molecular and morphological dataset produced a well resolved consensus tree (values in Figure 1) highly similar to that of the molecular data alone but with slightly better support.

n the Bayesian analyses, log-likelihood values and AC scores favored the partitioning regime with five partitions (harmonic means of likelihood values: -23072, -23033, -24364, -24256 and -23574, for the analyses under

89 . ostplant choice in Chelostoma

our ive si seven partitions and the analysis respectively. All ive analyses yielded consensus trees that ere almost ully resolved and virtually identical in their topology. he maorityrule consensus tree in the avored analysis had only to polytomies and 2 o the 2 other nodes had posterior probabilities o above values in igure . Adding the morphological dataset as a supplementary partition to this analysis resulted in similar or slightly higher supports or all nodes apart rom one ith substantially loer support igure . verall only ive nodes ere not recovered by all our analyses parsimony and ayesian analyses each ith and ithout morphology igure C. marginatum incisuloides as sister to C. cockerelli and C. minutum in one parsimony analysis C. bytinskii and C. laticaudum ere sister group o C. foveolatum in both parsimony analyses but not in the ayesian analyses there as a polytomy ith C.ventrale yrodromella and Chelostoma s. str. in one ayesian analysis and C. ventale as sister to yrodromella in one parsimony analysis lastly C. diodon and C. mocsaryi had a dierent position in one parsimony analysis.

he totalevidence ayesian analysis o the supermatri resulted in a tree igure 2 ith no conlict in topology ith the combined molecular phylogeny igure . All seven species or hich only morphological data ere available clustered ith the same sister species in both the total evidence phylogeny and the strict consensus tree based on the morphological data alone Appendi 4.

.4.3. volution o hostplant choice

ased on the totalevidence phylogeny and the assumption o parsimony oligolecty is the ancestral state in the genus Chelostoma. Parsimony mapping o pollen hosts igure 2 reveals a derived position o the to polylectic species C. minutum and C. species 3 hich provides solid evidence that polylecty independently arose tice ithin Chelostoma. he to polylectic species broadened their diet under maintenance o the ancestral host ithin the clade rom hich they are derived able 3 C.minutum added pollen o Allium to the ydrophyllaceae diet and C.species3 broadened the anunculaceae diet ith pollen o mainly Cistaceae and rassicaceae. Classiying the si species as polylectic or

7. Host-plant choice in Chelostoma which only few pollen samples were available (C. aureocinctum, C.tetramerum, C. isabellinum, C. garrulum, C. species 24, C. species 2) would still result in two independent transitions from oligolecty to polylecty. However, due to the basal position of C. aureocinctum, the ancestral state of host range (oligolecty vs. polylecty) in the genus Chelostoma would be euivocal.

Lithurgus chrysurus Anthidium punctatum Megachile pilidens Ochreriades fasciatus Hoplitis adunca Outgroup Osmia cornuta Protosmia minutula Hofferia schmiedeknechti Heriades truncorum Chelostoma aureocinctum Eochelostoma

100/100 Chelostoma phaceliae 96/96 Chelostoma californicum 100/100 100/100 100/100 99/100 Chelostoma incisulum New World 100/100 -/63 Foveosmia 100/100 Chelostoma m. incisuloides 98/99 66/68 Chelostoma cockerelli 63/64 Chelostoma minutum 99/100 86/82 100/100 Chelostoma philadelphi Prochelostoma 100/99 Chelostoma lamellum and Ceraheriades Chelostoma foveolatum

100/100 100/100 Chelostoma bytinski 53/51 93/95 59/59 100/100 Chelostoma laticaudum 83/75 Old World -/- Chelostoma styriacum 100/100 Foveosmia 92/97 Chelostoma hellenicum 99/100 100/100 57/74 69/75 82/76 Chelostoma distinctum 52/<50 Chelostoma campanularum Chelostoma ventrale

100/100 Chelostoma nasutum 100/100 Gyrodromella 92/95 96/99 Chelostoma rapunculi -/88 Chelostoma grande -/<50 100/100 100/100 Chelostoma transversum Bayesian 100/100 posterior 100/100 Chelostoma mocsaryi probability -/60 DNA / DNA + Morph. 56/<50 Chelostoma florisomne DNA / DNA + Morph. 100/100 Chelostoma 86/55 Chelostoma diodon 99/56 s. str. Parsimony 65/- Chelostoma edentulum 100/100 bootstrap Chelostoma species 2 92/95 100/100 99/99 95/92 Chelostoma species 3 76/77 Chelostoma emarginatum

FIRE 1. hylogenetic relationships within the genus Chelostoma. The tree shown is the 50- majority rule tree of trees 5000-50000 in the favored Bayesian analysis of the combined dataset (four genes divided into five partitions plus morphology; CI nt included). alues above branches give the posterior probabilities for the Bayesian analyses (left without morphology; right with morphology). alues below branches give the parsimony bootstrap values without CI nt (left without morphology; right with morphology). Missing values (-) indicate clades not recovered in the analysis.

91 7. Host-plant choice in Chelostoma

Chelostoma aureocinctum Chelostoma californicum 99 57 Chelostoma phaceliae

100 Chelostoma tetramerum * 100 70 Chelostoma minutum

100 Chelostoma cockerelli Chelostoma incisulum 100 97 Chelostoma m. incisuloides Node A Chelostoma m. marginatum * 100 100/99 99 Chelostoma philadelphi Chelostoma lamellum Chelostoma isabellinum * 85 99 Chelostoma foveolatum Chelostoma garrulum * 52 97 100 Chelostoma laticaudum Chelostoma bytinskii 71 Chelostoma styriacum

97 Chelostoma species 24 *

100 Chelostoma distinctum 56 Chelostoma campanularum

64 Chelostoma hellenicum Chelostoma species 23 * Chelostoma ventrale

Node B 98 100 Chelostoma nasutum 97/98 Chelostoma rapunculi 89 100 Chelostoma grande

100 Chelostoma transversum Chelostoma mocsaryi Node C 60 48/49 Chelostoma florisomne 100 Chelostoma diodon 51 Chelostoma edentulum 95 Chelostoma carinulum * Node D 64 49/52 53 Chelostoma species 2 61 Chelostoma emarginatum Node E Chelostoma species 3 83/88

Schima Allium Campanulaceae Dipsacaceae Ranunculaceae Hydrophyllaceae Philadelphus Asteraceae Ornithogalum Brassicaceae Figure . Majority-rule consensus tree of trees 5000-50000 in the Bayesian analysis of the “supermatri including those seven Chelostoma species (indicated by asteriss) for which only the morphological dataset was available. Outgroup species are omitted from the figure. The floral hosts of the 33 oligolectic species are mapped onto the tree using the criterion of maimum parsimony. Both polylectic species (grey branches and underlined) as well as Chelostoma lamellum, whose pollen preferences are not definitely nown, were coded as “missing data. The values at the five selected nodes -E give the average probabilities of having the most-liely state at his node in maimum lielihood ancestral state reconstruction (left value: reconstruction with those species for which both molecular and morphological data were available; right value: reconstruction with all 35 species included). The pie diagrams represent the ancestral state reconstructions for all 10 pollen hosts for each of the five nodes.

9 . ostplant choice in Chelostoma

ncestral reconstruction o hostplant choice at the ie selected nodes iure conirmed the results ased on parsimon. ielihood alues o inerred hosts did not sustantiall dier eteen analses ith and ithout the seen taa lacin molecular data alues in iure . The ancestor o the merican Foveosmia species node as most liel a specialist o drophllaceae and the common ancestor o the Palearctic Foveosmia species yrodromella and Chelostoma s. str. node a specialist o Campanulaceae. The ancestral hosts or nodes C and D ere less clear iure . The proale ancestral host at node as anunculaceae ith lielihood alues o . and . or the analses ith and species respectiel iure . These relatiel lo alues are liel due to the unstale position o the steraceae specialist C. diodon ithin this clade in o the trees sister to all memers o clade in sister to all memers o clade ecept C. florisomne. To circument this prolem e applied the most common recent ancestor approach. The common ancestor o the our anunculaceae specialists ith or ithout C. diodon dependin on the tree sampled as most liel a anunculaceae specialist lielihood alues . and . in analses ith and species respectiel. Similarl the most common recent ancestor approach reealed that the specialiations to each o the other pollen hosts had occurred onl once ecept or the steraceae drophllaceae . and . in analses ith and species respectiel Campanulaceae . and . Dipsacaceae . and .. The most common ancestor o C. ventrale and C. diodon as unliel to e an steraceae specialist . and . ut rather a Campanulaceae specialist . and . indicatin to independent sitches aa rom the Campanulaceae. These to independent sitches are urther conirmed lielihood comparisons o analses ith the ancestor o node C successiel constrained to e specialied on either Campanulaceae steraceae or anunculaceae aerae lielihood alues oer trees . . and . respectiel. s a dierence o to lounits is conentionall taen as stron eidence Pael there is sustantial support that the ancestor at node C as a Campanulaceae specialist. e did not iner the ancestral host o the ancestor o C. philadelphi and C.lamellum as the hostplant spectrum o the latter species is not deinitel non. oeer hiladelphus the eclusie pollen host o C.

7. Host-plant choice in Chelostoma

hlaelh (Figure 2, Table 3), is also a pollen host of C lamellum, which strongly suggests that a specialization to hlaelhus had occurred only once.

7.5. DSCSS

7.5.1. Phylogeny

The present study provides a well supported phylogenetic hypothesis for 35 Chelostoma species enabling the evolutionary reconstruction of host-plant associations within this bee genus on a worldwide scale. ur phylogeny differs from the current subgeneric classification of the genus Chelostoma (ichener 2007; see Table 3) in three respects. ost interestingly, these three divergences are strongly corroborated by floral host choice. First, the subgenus oeosma is not monophyletic as previously assumed but was found to consist of three distinct clades: a orth American clade closely associated with flowers of the family Hydrophyllaceae, a Palearctic clade comprising all oeosma species specialized on Campanulaceae, and C etale, an oligolege of the Asteraceae. Second, the orth American C hlaelh and the eastern Palearctic C lamellum are closely related and visit the flowers of hlaelhus to collect pollen. Thus, their inclusion into the two different subgenera ohelostoma and Ceaheaes is no longer ustified. Third, Casutum, classified as a member of the subgenus Chelostoma s. str., is a member of the subgenus oomella. t turned out to be a specialist of Campanulaceae as is its close relative C auul.

7.5.2. Evolution of host-plant choice

The genus Chelostoma mainly consists of oligolectic species. nly two out of the 35 taxa investigated were found to be pollen generalists. These two polylectic species evolved from oligolectic ancestors. n both species, the evolution of polylecty followed a distinct pattern. First, both species maintained the exclusive pollen host of their closest relatives in their polylectic diet (Hydrophyllaceae in C mutum, Ranunculaceae in Csees). The fact that three of four transitions from oligolecty to polylecty in the western Palearctic anthidiine bees and seven of eight cases

94 Hostplant choice in Chelostoma o host roadening in orth merican iadasia species occurred under maintenance o the original pollen hosts ller ipes and epedino indicates that this pattern o diet roadening might e idespread in ees econd some o the additional pollen hosts incorporated into the diets o the to pollectic Chelostoma species are alread utilied closel related species n C. minutum the additional host llium is the pollen source o C. tetramerum hereas loers o the rassicaceae one o the ne hosts o C. species 3 are the eclusie pollen source o C.edentulum imilarl alternatie host use in ees o the genus iadasia Emphorini is strongl iased toards host amilies that are alread eploited other iadasia or Emphorini species as primar hosts ipes and epedino Host sitches constrained to plants that are used related species ere also ound in phtophagous insects eg in the eetle genus phraella utuma et al and in the utterl trie mphalini an et al

t has een repeatedl shon that loral specialiations in ees are highl consered ith sister species generall eploiting the same host ller cislo and Cane iche et al ipes and epedino incle and oulston atin et al arin et al iche et al hlogeneticall consered host associations ere also ound in the genus Chelostoma Ecept or to independent specialiations to the steraceae sitches to all other host plant taa happened onl once ost remarale in this respect is the utiliation o the same host ie Philadelphus oth the orth merican C. philadelphi and the Chinese C. lamellum indicating a loral host choice that might hae een consered or seeral million ears ater a dispersal eent proal rom the eastern to the estern hemisphere had occurred Hines chapter hus loral host choice in the genus Chelostoma does not appear to e a laile trait easil shaped selectie orces as or eample loer suppl or interspeciic competition n act in southeastern Europe up to ie dierent Chelostoma species can e osered to simultaneousl eploit the same Campanula patch together ith seeral Campanula oligoleges o the genera ndrena and oplitis C edi Cra ller unpulished data his supports the ie championed incle and oulston that rather than restricting oraging to

. Host-plant choice in plants not exploited by other specialist bees, oligoleges are often specialized on widely used host plants where competition for pollen appears to be especially severe (Hurd and insley 1, Hurd et al. 1, Sipes and olf 1, Mincley et al. ). Hence, oligolecty in bees seems to be maintained or selected for by specific plant traits rather than by the avoidance of interspecific competition alone.

The oligolectic species exploit the flowers of ten different plant families, which belong to eight orders. These eight orders are distributed among all maor angiosperm lineages from the more basal ones to the most derived ones (Soltis et al. ), i.e., the monocots (sparagales) and the eudicots, the latter including the anunculales, the rosids (Brassicales), the ornales, the Ericales as well as both the euasteridsI (Boraginales) and the euasterids II (ipsacales, sterales). Similarly, the host-plant taxa newly added by the two polylectic species are not related to their ancestral hosts (sparagales in addition to Boraginales in Malvales and Brassicales in addition to anunculales in ). Host switches to distantly related hosts were also observed in bees of the genus (Sipes and Tepedino ). These findings show that host shifts in bees do not necessarily involve switches to closely related plants and indicate that other factors than host-plant phylogeny might underlie floral host specialization.

Indeed, visual appearance is striingly similar across flowers of several plant taxa exploited by bees of the genus . The flowers of many species among those host-plant taxa that have been newly incorporated by into its polylectic diet (istaceae, Brassicaceae) are as brightly yellow as the flowers of , its presumed ancestral host. The multistaminate androecium of the flowers of both the anunculaceae and the istaceae additionally contributes to their highly similar visual appearance. urthermore, the flowers of both (Theaceae) and (Hydrangeaceae) are of similar size, have a conspicuous white corolla and possess many yellow stamens. In addition, both taxa are shrubs or trees. Several genera among the Hydrophyllaceae (e.g., , , ) contain species characterized by distinctly bell-shaped and often blue- or purple-colored flowers that are surprisingly similar to those

7. Host-plant choice in helostoma of the ampanulaceae, though the mechanism of pollen presentation is completely different in these two plant families. ur phylogeny does not reveal direct switches between Schima and hiladelphus and between Hydrophyllaceae and ampanulaceae. However, the support for the phylogenetic position of . philadelphi and . lamellum is weak (igure 1, 2). Based on our morphological data alone (ppendix ), these two species are more basal forming the sister group of all other helostoma species except . aureocinctum, which is also supported by the plesiomorphic morphology of the labial palpus they have in common with . aureocinctum (Michener 2007). more basal position of . philadelphi and . lamellum, would result in direct switches between Schima and hiladelphus and between Hydrophyllaceae and ampanulaceae. ision is a key sensory modality in host-plant recognition in hymenopteran species including bees (ischer et al. 2001, iurfa and ehrer 2001). Thus, the presented cases of a striking floral similarity in otherwise non-related hosts might point to an important role of floral shape, morphology or color in directing the selection of new hosts in bees in general. This hypothesis is supported by the visually very similar but unrelated hosts of two closely related sister species of the anthidiine bees (Müller 1996b). Trachusa dumerlei is a strict specialist of knapweeds and thistles (steraceae), whereas Trachusa interrupta exclusively collects pollen on ipsacaceae. Both plant taxa have mostly red- to blue-colored flowers that are concentrated into compact inflorescences resulting in a very similar visual appearance. The use of visually surprisingly similar but unrelated floral hosts is also found in pollen specialist bees of the genera Macrotera and iadasia (anforth 1996, Sipes and Tepedino 2005). In both genera, several species exploit the similarly shaped and colored flowers of some actaceae (e.g., puntia) or Malvaceae (e.g., Sphaeralcea), respectively.

7.5.. Evolutionary constraints

Several findings detailed above indicate that evolutionary constraints have strongly influenced host-plant choice in bees of the genus helostoma. irst, host broadening in the two polylectic helostoma species appears to have been far from an accidental process. Its distinct pattern suggests that the newly added hosts might necessitate similar physiological or neurological (including cognitive) capabilities to cope with

97 ostlathoeChelostoma as the aestal host o that these aaltes ee hete om a ommoaestosaesltolalmteset oloesmalllthe eeseemetshhmaelahthetoolletChelostoma sees le hosts to the ets aleatle loselelate sees eo the hhl osee loal sealatos o the es Chelostoma as ell as ma othe ee leaes ate ltes esa om the ololet haththe seleto o elatehosts th a stloalsmlatsest that ees o the es Chelostoma mht e eoloall lmte to elot o etet loesotooeshaesmoholoesoolos

heostatsatohostaeeesotheesChelostoma maelassetototeshsoloalostatselatetoolle esto a eoloal ostats elate to the eoto o hal o loesee o oth tes o ostats omes om ea eemets ote th C. rapunculi a C. florisomne hate C. rapunculi hh elsel ollets olle o Camalaeae ale to eelo o o eet ets o ohost olle amel uphthalmum steaeae Ranunculus alaeae Sinapis assaeae a Echium oaaeae sest a sto lmtatohostaeassoatetholleestohesttheeo these o olle hosts ae elote memes o the ses Chelostoma s st hh ates that these sees hae eole hsoloal aatatos tosesslltle them ata sest that ths hsoloal ostat has ee oeome to tmes eeetloeC. ventraleaoetheaestooChelostoma sstoeethelatteasetheathasoetoseale o seealothe hostssh as assaeae saaeae rnithogalum a alsotastC. florisomnehhs sttlsealeo alaeaeasotoe aletoeelo otoohost olle etsamelCampanula a rassica hate stshos that oe o the memes o the ses Chelostoma s st l C.florisomne elot Camalaeae o olle althoh the eole omaestosthateeololetothslatamloesaC e hs lealots toostatsthat aeot elateto tto t athe to host eoto o loe hal h

7. Host-plant choice in Chelostoma information-processing constraints are actually assumed to be the reason why the solitary bee eaes tom refused to harvest pollen on Camala and hm in the absence of its specific host, the Asteraceae, though both types of non-host pollen support larval development (chapter5). Neurological constraints might explain why related species in the genus Chelostoma as well as in other groups of bees tend to specialie on flowers that are similar in shape, morphology or color. In fact, there is evidence that adult bees have limited memory and learning capacities for shapes and colors (Bernays and cislo 14, hittka et al. 2001, iurfa and Lehrer 2001).

7.5.4. The constraint hypothesis

The results of our study strongly support the hypothesis that oligolecty in the genus Chelostoma is evolutionary constrained. Based on this finding, we propose a new hypothesis on the evolution of host range in bees (Box 1). This constraint hypothesis is related to the oscillation hypothesis of host-plant range recently postulated for herbivorous insects (Jan and Nylin 200). Indeed, patterns of host use in Chelostoma as well as in other bee lineages display striking similarities to those of phytophagous insects (Sipes and Tepedino 2005, Mller and uhlmann in press).

ur constraint hypothesis distinguishes five consecutive stages of host-range evolution (Box 1). lolet hase: Numerous oligolectic bee species appear to be highly adapted to their specific hosts and, as shown above, are probably constrained by physiological or neurological limitations (including vision or possibly also olfaction as cognitive sensory modalities) rendering switches to or incorporations of other hosts difficult. In fact, many host-specific herbivorous insects were found to be physiologically adapted to the secondary chemistry of their host plants, but less adapted to utiliing other hosts (Slansky 13, Strauss and angerl, 2002, ornell and Hawkins 2003, Singer 200). In other phytophagous insects, limited information-processing abilities, i.e. neurological or cognitive constraints, are assumed to underlie host-plant specialiation (Bernays 1, 2001).

7 Chelostoma

t

s t o

l a h e

i t t

a t l

i

s

i a n t a i i o t i i h n a

t

i h i s i

h l i

s

l i a c e e o s t

u o n c e

i o

e e e

e e n e i

e l h e m n l c i t n

e t i l

h

n

a t

e c t h e o o l

c s a n t l i i e t s

s t h l

e t e h t

o s t o a t

a e

l n

t o o o t i

l n

h o i n i h s a s

t h

i o

t i i t e s t t a

t e s s t s i s e s e s s

o o e h

h a

n o h e h t h h e

e

o h a o c

c n a h n t

o t t n o e e h c

a l l o t

e o l e

l

l n

a t o l i n o

m o h i l t a o s

e o t c n i l

i a

s t t l

o t i

c s e i n o i s m i t l

e e u t i n c t i c

c s o l e i

c

e a o i n n c l n e n c e

i t

a

e s o

e s h s a s l t s

a l n o n

n n o i i o e i o o h a o t n

t o h o i

c

c

c e t a

n m s u l i e l e n a n h e i i s a i l

c t o

m

l a c a

s n o i o n i e l e a e i l a c l o i

o e i i

l o c

a e t t c c l

a i e a i e

s

i o t c n l o m

n i i e i

o i

e a c

t s h n u e h

i c

e

i o e e o h l

t i

o e t m i n t

o n t t t s e

i s

i o i

o e e l t a e i c o o o n u o

l

e o e

n t u e c l

e a e i a a n u

n h

i

c o e m e t i e n o t l o c i

o t

i o n t s e e

o e l o c t o l

t o t a

l

s a o c e s o o e s h

l i

o o n l l n l i c l

i

a

e a m o h a s

o i

a a c t

c c

s t h i c o c i e i n i

i n c h i a

i t

o n e o n e i o s c i o a

i h

o l l l

a

s e l s a l o u a o e i t n a h i i o i t

s l t o c o s a i c s c t s i i n e i s u o e i e i l e

l m o e h e h h e i i h s n s o i

n e e

o s i n a t a o h t i u

l t c c

i o i a o c t t i h c c i t s l e

e e c

l l a n e e e e e i e o l o s e s s s c l a a a a l l e i i n l l h h h h o o

a s

i

s t e s h o t h o

o

h

t

s e t

t i o e s

n i n i a

t e a t t a a n e i h t i h n a t t

s

c a s

c

a l i n o t e

a h t o

n n c c

i m c o a c i

e

i s t a o e l i e t e

e c

n c h t a l i e s e h i e s e n o a t

t l e n

n e t i c o e o a

o e o h

a e

e l t c t

i

e o l e h e o s n s t

i o i

100 Chelostoma

ollen shortage phase 200 200 olylectic phase rtica 2001 rtica 2001 200 ligolectic phase P

1 1 1 200

101 7. Host-plant choice in Chelostoma on the direction of selection. ost important the phylogenetic traces of a host-range epansion can e lost with time i.e. a period of epanded host range followed y specialiation on one of the new hosts will in retrospect appear as a clean host switch once all traces of the additional hosts are lost an and Nylin 008. his proaly also applies to the ten seemingly clean host switches in the genus Chelostoma which are strongly epected to hae proceeded oer a shorter or longer transitional phase of epanded host range.

In conclusion host-plant choice in ees appears to e a dynamic process enaling transitions oth from a narrower to a roader diet and ice ersa aser et al. 1 Sipes and epedino 005. Howeer floral host choice does not appear to e a highly fleile trait that can e easily changed y selectie forces. Instead it appears to e eolutionary much more constrained than hitherto thought.

10 . oslan coice in Chelostoma

oslan reerences o ees o e enus Chelostoma ased on e lieraure and on unulised ield daa

uenus and secies ieraure and unulised ield daa uenus Chelostoma Chelostoma emarginatum oliolecic on Ranunculus ie e al. Chelostoma florisomne oliolecic on Ranunculus esric 8 ie e al. Chelostoma grande oliolecic on isacaceae esric ie e al. Chelostoma mocsaryi oliolecic on Ornithogalum Goala . edi .. ra and . ller unulised ield daa ro al rance and Greece Chelostoma transversum oliolecic on isacaceae . edi .. ra and . ller unulised ield daa ro Greece uenus Foveosmia Chelostoma bytinskii oliolecic on Campanula . edi .. ra and . ller unulised ield daa ro ordan Chelostoma californicum oliolecic on acelia oldene and e roein e al. .Grisold in li. Chelostoma campanularum oliolecic on Campanula esric 8 ie e al. Chelostoma cockerelli oliolecic on riodictyon oldene and e roein e al. .Grisold in li. Chelostoma distinctum oliolecic on Campanula esric 8 ie e al. Chelostoma foveolatum oliolecic on Campanula ie e al. Chelostoma hellenicum oliolecic on Campanula . anduss in li. . edi .. ra and .ller unulised ield daa ro Greece Chelostoma incisulum oliolecic on Phacelia oldene and e roein e al. .Grisold in li. Chelostoma laticaudum oliolecic on Campanula . anduss in li. . edi .. ra and .ller unulised ield daa ro Greece Chelostoma m. marginatum oliolecic on drollaceae roein e al. oldene and e Chelostoma m. incisuloides oliolecic on drollaceae roein e al. oldene and e Chelostoma minutum oliolecic on Phacelia oldene and e ollecic on Phacelia llium and Sedum arer 88 Chelostoma nasutum oliolecic on Campanula . anduss in li. . edi .. ra and ller unulised ield daa ro Greece Chelostoma phaceliae oliolecic on Phacelia oldene and e roein e al. .Grisold in li. Chelostoma species 2 oliolecic on Campanula . ller unulised ield daa ro odos Chelostoma styriacum oliolecic on Campanula anduss in li. . edi and .. ra unulised ield daa ro Greece uenus yrodromella Chelostoma rapunculi oliolecic on Campanula esric 8 ie e al. uenus Prochelostoma Chelostoma philadelphi oliolecic on Philadelphus icener

7. Host-plant choice in Chelostoma

APPENDI 2: Morphological character and character states used in the cladistic analysis of the genus Chelostoma. If not otherwise stated, the characters refer to both sexes. The terminology of bee morphology follows Michener (2007).

1) ertex with a distinct preoccipital ridge developed at least medially (1) rounded, without a distinct preoccipital ridge (0). 2) Antenna of male reddish-colored (1) uniformly dar-colored (0). ) Antennal segment 3 of male with long and erect bristle-lie hairs (1) only microscopically haired (0). 4) Antennal segment 3 of male 1.5x to 2x as long as pedicel (1) shorter to slightly longer than pedicel (0). 5) Base of labrum of female strongly bent (1) more or less straight (0). 6) Apex of labrum of female tripartite, with a distinct proection on both sides of the median proection (1) rounded, truncated or emarginate (0). 7) Hypostomal carina of male forming a distinct tooth (1) straight to evenly rounded (0). 8) Lower side of mandible of male with a distinct pilosity, which is as long and nearly as dense as that on the adacent genal area (1) with a less distinct pilosity composed of rather short and scattered hairs (0). 9) Inner margin of mandible of female more or less straight without a prominent tooth behind the apical two teeth (1) with a prominent, triangular tooth (0). 10) Segment 3 of labial palpus flattened, its axis a continuation of that of segment 2 (1) not flattened, its axis directed laterally (0). 11) Number of segments of maxillary palpi: (1) 3 (0). 12) Parapsidal line short, less than 0 of length of tegula (1) long, more than 50 of length of tegula (0). 1) Pronotal lobe of female rounded all around (1) slightly eeled or bulging at its base (0). 14) Tegula of female punctured only apically and along inner margin (1) densely punctured across its whole surface (0). 15) ugal lobe connate to the hindwing for less than two-thirds of its length (1) connate to the hindwing for two-thirds of its length or more (0). 16) Hind coxa with a carina along inner ventral margin (1) not carinate (0). 17) Posterior margin of basal area of propodeum eeled or thicened along its whole width (1) rounded or only laterally eeled (0). 18) Propodeum densely chagreened, dull (1) polished (0). 19) Apical hair fringes on terga of female strongly developed (2) wealy developed (1) absent (0). 20) Metasomal tergum 7 of male with dorsal pit (1) without dorsal pit (0). 21) Metasomal tergum 7 of male evenly rounded (0) of different shape (1). 22) Metasomal tergum 7 of male a single rounded proection (1) of different shape (0). 2) Metasomal tergum 7 of male with two broad and truncated teeth (1) of different shape (0). 24) Metasomal tergum 7 of male with two pointed teeth, incision about as broad as one tooth (1) of different shape or incision much broader or much narrower than one tooth (0). 25) Metasomal tergum 7 of male three-toothed (1) of different shape (0). 26) Metasomal tergum 7 of male four-toothed (1) of different shape (0). 27) Median teeth of the four-toothed metasomal tergum 7 of male fused (1) not fused or absent (0). 28) Lateral tooth of metasomal tergum 7 of male distinctly curved inwards (1) not curved inwards or absent (0). 29) Scopal hairs of female distinctly tapered towards the apex (1) apically blunt (0). 0) Scopal hairs of female with spiral swellings clearly visible at a magnification of 00x (1) with smooth surface (0). 1) Scopal hairs of female with short side-branches (1) unbranched (0). 2) Metasomal sternum 2 of male with a distinct median elevation or a hump (1) without proection or hump (0). ) Median elevation on metasomal sternum 2 of male half-elliptically shaped (1) of another shape (0). 4) Median elevation on metasomal sternum 2 of male distinctly concave (1) not distinctly concave (0). 5) Base of metasomal sternum 3 of male densely covered with plumose hair (1) without a dense cover of plumose hair (0). 6) Metasomal sternum 3 of male with two spots of blac bristles developed near the centre of the sternum (3) between the centre and the apical margin (2) at the apical margin (1) blac bristles lacing (0).

10 7. Host-plant choice in Chelostoa

Membraneous flaps at apical margin of metasomal sternum of male nearly as long as disc of sternum or longer (1) half as long as disc of sternum or less (0). Plumose hair on metasomal sternum of male densely covering the whole sternal surface (3) loosely covering the sternal surface (2) developed only lateroapically (1) absent (0). Apical margin of metasomal sternum of male with a dense and uninterrupted fringe of hairs, which are bent at right angles to the sternal surface (2) with a dense but medially interrupted fringe of such hairs (1) without a fringe of such hairs (0). Apical margin of metasomal sternum 5 of male apically fringed (1) not fringed (0). ateral margin of metasomal sternum 5 of male lifted and distinctly keeled, at least in the apical half (1) flat and normally rounded (0). Hair comb at the apical margin of metasomal sternum 5 of male dense, short and developed along the whole sternal width (slightly interrupted medially in Chelostoahelleic) (1) of other shape or absent (0). Apical margin of metasomal sternum 5 of male with a bowl-shaped comb of hairs (1) different (0). Comb hairs at apex of metasomal sternum 5 of male shaped like a pearl necklet (3) wavy (2) zigzagged (1) of other shape or absent (0). Apex of metasomal sternum 6 of male carinate laterally, resulting in a triangular to rounded projection (1) of different shape (0). Apical margin of sternum 8 of male truncated to slightly emarginate, with a tuft of hairs medially (1) of other shape and without a median tuft of hairs (0). Gonostylus apically clubbed and beset with long hairs on inner and outer side (2) apically clubbed and hairless or only microscopically haired (1) apically slender (0). nner margins of penis valves distinctly divergent (1) more or less parallel, lying close together (0).

105 7. Host-plant choice in Celostoma

APPENDI orphological data matrix used in the cladistic analysis of the genus Celostoma. Unknown states are coded as ? and polymorphic states with P.

Celostoma iladeli 000001000000010010001000000000010000000000000000 Celostoma aceliae 0000000001P0010100211000100100010000000100000000 Celostoma caliornicum 000000000100010100211000100100010000000100000000 Celostoma cocerelli 00000000010001010001100001010001000000010000000 Celostoma minutum 000000000100010100011000010100010000000100000100 Celostoma m marinatum 000000001100010100011000010100010000000100020000 Celostoma m incisuloides 00000000110001010001100001110001000000010002000 Celostoma tetramerum 000000001110010100111000100100010000000100000000 Celostoma incisulum 01000000110001010001100001110001000000010000000 Celostoma aureocinctum 000000000010000100200000000000000000000100030000 Celostoma entrale 100000000100010110211001000011010000000100000001 Celostoma diodon 100010011101011100211000000011010002020110000120 Celostoma edentulum 011100011101011100211010000011011113031110101120 Celostoma emarinatum 011100011101011100211010000011011113031110101120 Celostoma secies 011100010101011100211010000011011113031110101120 Celostoma lorisomne 011100011101011111211010000010111103032110131120 Celostoma carinulum 011100011101011101211010000011011113031110101120 Celostoma secies 011100011101011110211010000011011110031110101110 Celostoma transersum 010110111111011100211010000011010011110110130120 Celostoma rande 010110111111011100211010000011010011110110131120 Celostoma mocsari 111100010101011110211000000011010112031110101120 Celostoma raunculi 100000000100010100211000000011010001000110110011 Celostoma nasutum 100000000100011100211010000011010001000110110111 Celostoma oeolatum 000000001100010100011100000011010000000000000000 Celostoma laticaudum 000000000100110100111100000011010000000100000001 Celostoma arrulum 000000001100010100011100000011010000000000000000 Celostoma tinsii 000000000100110100101100000011010000000000000001 Celostoma isaellinum 000000001100010100011000000011010000000100000000 Celostoma secies 000000000100010110111001000011010000000100000000 Celostoma camanularum 000000000100010110011001000011110000000101000000 Celostoma distinctum 000000000100010110011001000011010000000100000000 Celostoma ellenicum 000000000100010110011001000011010000000101000000 Celostoma striacum 000000000100010110011001000011010000000100000000 Celostoma secies 000000000100010110011001000011010000000101000000 Celostoma lamellum 000001001010010100211000000001010000000000000000

10 tantiinCest

titnnu tuaainiutinuaitainanai tiaatat

Chelostoma aureocinctum Chelostoma philadelphi Chelostoma lamellum Chelostoma phaceliae Chelostoma californicum Chelostoma tetramerum Chelostoma cockerelli Chelostoma minutum Chelostoma m. marginatum Chelostoma m. incisuloides Chelostoma incisulum Chelostoma isabellinum Chelostoma foveolatum Chelostoma garrulum Chelostoma species 24 Chelostoma distinctum Chelostoma styriacum Chelostoma campanularum Chelostoma hellenicum Chelostoma species 23 Chelostoma laticaudum Chelostoma bytinskii Chelostoma ventrale Chelostoma rapunculi Chelostoma nasutum Chelostoma diodon Chelostoma grande Chelostoma transversum Chelostoma edentulum Chelostoma emarginatum Chelostoma species 3 Chelostoma florisomne Chelostoma carinulum Chelostoma species 2 Chelostoma mocsaryi

. Discussion

8. General discussion

The present thesis is a case-study of the relationships between bees and flowers, exemplified in a selected group of bees, the Osmiini. It largely focuses on pollen specialization in solitary bees oligolecty and brings a new hypothesis on the factors underlying floral choices in bees in general. Floral choices appear as a highly conserved trait that is little influenced by ecological factors such as interspecific competition or pollen supply. Rather, floral choices are strongly constrained by the physiological capabilities of the bees.

In the first chapter, the view of pollen as an easy-to-use protein source for bees is challenged. Several oligolectic bee species failed to develop on non-host pollen, thus indicating that pollen strongly varies in its nutritional value for bees. This surprising result suggests that plants protect their pollen through secondary compounds, exactly in the same way as they defend other tissues against herbivores. This finding has two important implications for the interpretation of bee-flower relationships. First, pollen cannot be generally considered as a reward offered to the bees. Pollen is an essential plant tissue carrying the reproductive gametes and conseuently, it is expected to be protected by plants. Pollen may operate as a reward to pollinators in few cases only, e. g., in nectarless plant species Jrgens and Dtterl or in plants that rely on pollen-foraging bees for pollination e. g. , Medicago, Dorn and eber . Second, pollen chemical composition is postulated to strongly influence the pollen spectrum of bees, contrasting previous assumptions cislo and Cane , Minckely and Roulston .

In chapter two, a genetic basis of host-recognition was demonstrated for the oligolectic bee Heriades truncorum. Individuals grown as larvae on non-host pollen restricted their foraging to their usual host as adults. Hence host choice appeared as a surprisingly conserved trait that was not influenced by experience, as sometimes suggested the imprinting theory. Moreover, the bees gave up nesting in the absence of their normal pollen host and refused to collect pollen from non-host plants, in spite of the suitability of this pollen for the larvae. This observation points to

iscussion neurological limitations preventing this ee species to escape from oligolecty Hence oligolecty in Heriades truncorum appears to e ased on ehavioral or neurological rather than nutritional constraints

In chapters three and four the phylogenetic relationships ithin the Osmiini ere investigated The genus Chelostoma emerged as the most primitive lineage of the Osmiini Pollen analyses for species of Chelostoma revealed that this genus is mainly composed of oligolectic species orldide matching the increasing evidence that oligolecty is the ancestral state in ees anforth et al The inference of the phylogenetic relationships eteen these species alloed for the evolutionary tracing of floral choices The pattern oserved led to the formulation of a ne hypothesis on the evolution of hostrange in ees the constraint hypothesis This hypothesis represents the core of the present thesis It is ased on the oservation that floral choices are highly conserved over time in Chelostoma and ees in general Mller Sipes and Tepedino Changes in the pollen diet or host sitches are rare and far from eing a random process strongly suggesting that only a limited array of floers matches the physiological reuirements of a given ee species To types of constraints could e distinguished difficulties in pollen digestion as demonstrated in chapter and neurological limitation as documented in chapter In conclusion the present thesis presents the first hypothesis on the evolution of host range in ees that integrates oth empirical and phylogenetic evidence The significance of the constraint hypothesis goes eyond the evolutionary eplanation of oligolecty This ne hypothesis provides ne ideas to unravel the long coevolutionary history eteen ees and floers

THE IMPLICATIONS O THE CONSTAINT HPOTHESIS

The appearance of the ees

Bees originated at least millions years ago Poinar and anforth proaly after the appearance of the angiosperms see discussions in Engel and Michener hich originated at least millions years ago Soltis et al The scenario of the transition eteen prey hunting sphecid asps and the first ee the protoee remains

. Discussion conectural Radcheno and Peseno Michener . The growing evidence that oligolecty and not polylecty is the primitive state in bees Danforth et al. references in chapter may indicate that the initial relationship between the proto-bee and flowers was specialied. The constraint hypothesis largely supports this view the transition from a carnivorous diet to a phytophagous of plant origin diet must have been a comple step. Sphecid wasps from which bees stem are essentially carnivorous. As such they assimilate many essential nutrients directly from their animal diets for eample essential amino acids or certain sterols. Such compounds may lac in plant tissues which have a higher CN ratio and a different sterol composition than animal tissues. The first pollen host of the proto-bee must have had by chance a distinct chemical composition maing it a suitable diet for a carnivorous wasp. Such nutritional suitability is certainly not universal in pollen see discussion in chapter which supports the view of a specialied initial scenario of the appearance of bees. Possibly the latest sphecid ancestor of the bees was a specialied prey-hunter see discussion in Müller b. The great maority of the sphecid wasps restrict their foraging to definite arthropod preys e. g. certain taonomic groups or development stages Bohart and Mene wata . A tempting hypothesis is that the ancestor of the bees was specifically hunting for preys on one type of flowers rendering the accidental but persistent transport of pollen more liely. Had by chance the pollen of this flower constituted a suitable diet not much time would pass until the prey could be omitted and the diet could merely be composed of pollen.

... Bee radiation

A simple view of early bee evolution may postulate a rapid diversification and speciation radiation soon after the dawn of the bees as a free ecological niche pollen collecting was opened. This view is repeatedly uoted in studies that intuitively match the early radiation of the bees with the radiation of the angiosperms e. g. Engel during the second half of the Cretaceous - millions years ago. t is certainly an oversimplification of the true pattern of diversification as pointed out by Danforth et al. b. First the constraint hypothesis demonstrates the difficulties associated with host switches in bees and hence

. Discussion angiosperm diversification is unlikely to have immediate conseuences on bee diversification. Coevolution beteen bees and floers has been at most diffuse sensu Herrera , contrasting a pervasive idea in pollination biology that pollination systems are specialized see revie in Waser et al. . Second, the same transition from preyhunters to pollen collectors also occurred in the pollen asps Masaridae. This group did not undergo intense radiation the orldide number of species is approimately , although this group is presumably as old as bees ess and ess . Conseuently, although angiosperm radiation has obviously been forerunner to the diversification of bees, intrinsic factors relating to bee biology may have played a more important role than angiosperm diversity in the early radiation of bees. Such factors may include nesting behavior, the raise of sociality, or biogeography.

... olylecty

Many polylectic bee species, possibly the great majority of them, still sho a restricted pollen spectrum Mller b, Cane and Sipes , Mller and Kuhlmann . This pattern suggests that similar mechanisms as those described in oligolectic species may shape their floral choices. To polylectic species of Megachile ere found to collect pollen on a variety of hosts shoing similar color Michener , olpen and Brandt, . Similarly, it is ell possible that the chemical composition of the pollen also influences host choice in polylectic bees, although no study has ever analysed pollen types found in the diet of a polylectic bee. Mller and Kuhlmann convincingly shoed that some pollen types e. g., the Asteroidea are generally avoided by polylectic bees of the genus Colletes, in spite of their preponderance in most ecosystems. Therefore polylectic bees do not collect pollen from all floers available. Many polylectic bees may either have gradually integrated suitable pollen hosts into their diets, or may be constrained on some hosts due to visual or olfactive cues. An eception is found in social bees, especially the highly social honeybee. This species has epanded polylecty so much as to virtually collect pollen from any floer, even from gymnosperms or indpollinated angiosperms. loral choices in social species are more comple than in solitary bees. oraging is strongly affected by the colonys needs, by individual learning and even sometimes

. Discussion by recruitment, or training of new workers by eperienced workers (the bee dance, see for eample Tautz ).

.. OUTLOOK AND FUTURE RESEARCH

... Pollen chemical composition

This study has documented the failure of several bee species to develop on non-host pollen, pointing to an important nutritional component shaping bee-flower relationships. The abundant literature on honeybee nutrition allows for the formulation of several hypotheses eplaining the observed pattern, e. g., pollen toicity, lack of essential nutrients or difficulties in the etraction of nutrients from the pollen grain. In evolutionary terms, toic pollen would have a different significance from pollen lacking essential nutrients. The former points to an active defense mechanism in flowers, whereas the latter could merely represent a passive process related to the physiological abilities of the bees. The unambiguous demonstration of pollen toicity would have intriguing consequences for the field of pollination. Only one study could so far demonstrate a relationship between pollen toicity and the mode of pollination (Jürgens and Dtterl ).

... Host recognition

A better understanding of bee-flower relationships goes together with a comprehension of the cues underlying host selection in bees. In chapter two, evidence was presented that odor recognition has a genetic base in Heriades truncorum. To what etent visual cues such as shape, color or U-reflecting patterns influence host-selection in bees, remains an important question (Dobson, , ). Lastly, how do polylectic bees choose their host Do specific chemical compounds act as attractant, other as repellent for generalist bees How can be eplained that the maority of the known polylectic bees of the genus Colletes avoid pollen collection from the Asteroideae Have these bees learnt, in the course of evolution, to avoid this low-nutritive pollen Or rather, were the Asteroideae never definitely included into their pollen diets This thesis has largely focused on oligolectic bees, as oligolecty has been the state requiring attention for

iscussion almost a century owever polylecty is actually the state that necessitates most physiological adaptations and a better understanding of bee-flower relationships will be gained in future from studies of polylectic species

Mellitophily or the pollination by bees

he dilemma faced by flowers to attract bees for their pollination and to protect their pollen from ecessive collection opens many uestions on bee pollination in general he uantitative fate of pollen in insect- pollinated flowers remains poorly nown A comparison of the pollination efficiency the percentage of pollen grains contributing to pollination of female bees male bees and other insects would be essential t is documented that nectar foraging bees are more efficient pollinators than pollen foraging female bees in some flowers e g in the apple tree Westeramp Larsson compared pollination efficiency of specialist bees generalist bees and other insects in Knautia flowers Pollen collecting female bees accounted for of all visits to Knautia flowers but only for of overall pollen transfer due to their preference to pollen- presenting flowers rather than to flowers in the female phase Similar studies are much needed in other study systems e g in those flowers particularly adapted to bees Lamiaceae Campanulaceae Such studies in a wide range of flowers would further sharpen our nowledge on the particular status of bees - concurrently pollinators and herbivores

9. References

. Cited references

Adler, L. S. (2000). The ecological significance of toxic nectar. Oios 91: 409-420. Adler, L. S., and Irwin, R. E. (2005). Ecological costs and benefits of defenses in nectar. Ecology 86: 2968-2978. Adler, L. S., Wink, M., Distl, M., and Lentz, A. J. (2006). Leaf herbivory and nutrients increase nectar alkaloids. Ecology Letters 9: 960-967. Amiet, F., Herrmann, M., Müller, A., and Neumeyer, R. (2004). Apidae 4: Anthidium Chelostoma Coelioxys Dioxys eriades Lithurgus Megachile Osmia Stelis. CSCF, Neuchtel. Ascher, J. S., Danforth, B. N., and Ji, S. . (2001). Phylogenetic utility of the major opsin in bees (Hymenoptera: Apoidea): a reassessment. Molecular Phylogenetics and Evolution 19: 76-93. Baker, D. B., and Engel, M. S. (2006). A new subgenus of Megachile from Borneo with arolia (Hymenoptera : Megachilidae). American Museum Novitates 3505: 1-12. Barron, A. B. (2001). The life and death of Hopkins host-selection principle. Journal of nsect Behavior 14: 725-737. Bellmann, H. (1997). Bienen in Schneckenhäusern. Biologie in unserer Zeit 27: 106-113. Bensasson, D., Zhang, D. X., Hartl, D. L., and Hewitt, G. M. (2001). Mitochondrial pseudogenes: evolutions misplaced witnesses. Trends in Ecology Evolution 16: 314-321. Bernays, E. A. (1998). The value of being a resource specialist: behavioral support for a neural hypothesis. The American Naturalist 151: 451-464. Bernays, E. A. (2001). Neural limitations in phytophagous insects: Implications for diet breadth and evolution of host affiliation. Annual Review of Entomology 46: 703-727. Bernays, E. A., and Chapman, R. F. (1994). Host-plant selection by phytophagous insects. Chapman Hall, London. Bernays, E. A., and Wcislo, W. T. (1994). Sensory capabilities, information processing and resource specialization. The uarterly Review of Biology 69: 187-204. Bjorksten, T. A., and Hoffmann, A. A. (1998). Plant cues influence searching behaviour and parasitism in the egg parasitoid Trichogramma nr brassicae. Ecological Entomology 23: 355-362. Bohart, R. M., and Menke, A. S. (1976). Sphecid Wasps of the World - A generic revision. University of California Press, Berkeley. Boppr, M., Colegate, S. M., and Edgar, J. A. (2005). Pyrrolizidine alkaloids of Echium vulgare honey found in pure pollen. Journal of Agricultural and ood Chemistry 53: 594-600. Bosch, J., Maeta, Y., and Rust, R. (2001). A phylogenetic analysis of nesting behavior in the genus Osmia (Hymenoptera : Megachilidae). Annals of the Entomological Society of America 94: 617-627. Breed, M. D., Leger, E. A., Pearce, A. N., and Wang, Y. J. (1998). Comb wax effects on the ontogeny of honey bee nestmate recognition. Animal Behaviour 55: 13-20. Bush, G. L. (1994). Sympatric speciation in animals - new wine in old bottles. Trends in Ecology Evolution 9: 285-288. Cameron, S. A., Hines, H. M., and Williams, P. H. (2007). A comprehensive phylogeny of the bumble bees (Bombus). Biological Journal of the Linnean Society 91: 161-188. Cameron, S. A., and Mardulyn, P. (2003). The major opsin gene is useful for inferring higher level phylogenetic relationships of the corbiculate bees. Molecular Phylogenetics and Evolution 28: 610-613. Cane, J. H., Griswold, T., and Parker, F. D. (2007). Substrates and materials used for nesting by North American Osmia bees (Hymenoptera : Apiformes : Megachilidae). Annals of the Entomological Society of America 100: 350-358. Cane, J. H., and Payne, J. A. (1988). Foraging ecology of the bee abropoda laboriosa (Hymenoptera: Anthophoridae), an oligolege of blueberries (Ericaceae: Vaccinium) in the southeastern United States. Annals of the Entomological Society of America 81: 419-427. Cane, J. H., and Sipes, S. (2006). Characterizing floral specialization by bees: analytical methods and a revised lexicon for oligolecty. Pages 99-122 in N. M. Waser and Ollerton, J., editors. Plant-pollinator interactions: from specialiation to generaliation. The University of Chicago Press, Chicago and London. Chittka, L., Spaethe, J., Schmidt, A., and Hickelsberger, A. (2001). Adaptation, constraint, and chance in the evolution of flower color and pollinator color vision. Pages 106-126 in L. Chittka and Thomson, J. D., editors. Cognitive ecology of pollination. Cambridge University Press, Cambridge.

114 9. References

Constance, L., and Chuang, T. I. (1982). SEM survey of pollen morphology and classification in Hydrophyllaceae (Waterleaf family). American Journal of Botany 69: 40-53. Cook, S. M., Murray, D. A., and Williams, I. H. (2004). Do pollen beetles need pollen The effect of pollen on oviposition, survival, and development of a flower-feeding herbivore. Ecological Entomology 29: 164-173. Corbet, S. A. (1985). Insect chemosensory responses - a chemical legacy hypothesis. Ecological Entomology 10: 143-153. Cornell, H. V., and Hawkins, B. A. (2003). Herbivore responses to plant secondary compounds: A test of phytochemical coevolution theory. American aturalist 161: 507-522. Cortesero, A. M., Monge, J. P., and Huignard, J. (1995). Influence of two successive learning processes on the response of Eupelmus vuilleti Crw (Hymenoptera, Eupelmidae) to volatile stimuli from hosts and host plants. Journal of Insect Behavior 8: 751-762. Cruden, R. W. (1972). Pollination biology of emophila menziesii (Hydrophyllaceae) with comments on the evolution of oligolectic bees. Evolution 26: 373-389. Danforth, B. N. (1996). Phylogenetic analysis and taxonomic revision of the Perdita subgenera Macrotera, Macroteropsis, Macroterella and Cocerellula (Hymenoptera: Andrenidae). The niversity of ansas Science Bulletin 55: 635-692. Danforth, B. N. (1999). Phylogeny of the bee genus Lasioglossum (Hymenoptera : Halictidae) based on mitochondrial COI seuence data. Systematic Entomology 24: 377-393. Danforth, B. N., Ascher, J., Anderson, R. S., and Farrell, B. D. (1999b). Flowers and Insect Evolution. Science 283: 143 Danforth, B. N., Brady, S. G., Sipes, S. D., and Pearson, A. (2004). Single-copy nuclear genes recover Cretaceous-age divergences in bees. Systematic Biology 53: 309-326. Danforth, B. N., Conway, L., and Ji, S. . (2003). Phylogeny of eusocial Lasioglossum reveals multiple losses of eusociality within a primitively eusocial clade of bees (Hymenoptera : Halictidae). Systematic Biology 52: 23-36. Danforth, B. N., Fang, J., and Sipes, S. (2006a). Analysis of family-level relationships in bees (Hymenoptera : Apiformes) using 28S and two previously unexplored nuclear genes: CAD and RNA polymerase II. Molecular Phylogenetics and Evolution 39: 358-372. Danforth, B. N., and Ji, S. . (1998). Elongation factor-1 alpha occurs as two copies in bees: Implications for phylogenetic analysis of EF-1 alpha seuences in insects. Molecular Biology and Evolution 15: 225-235. Danforth, B. N., Sauuet, H., and Packer, L. (1999a). Phylogeny of the bee genus Halictus (Hymenoptera : Halictidae) based on parsimony and likelihood analyses of nuclear EF-1 alpha seuence data. Molecular Phylogenetics and Evolution 13: 605-618. Danforth, B. N., Sipes, S., Fang, J., and Brady, S. G. (2006b). The history of early bee diversification based on five genes plus morphology. Proceedings of the ational Academy of Sciences of the nited States of America 103: 15118-15123. Dethier, V. G. (1954). Evolution of feeding preferences in phytophagous Insects. Evolution 8: 33-54. Dobson, H. E. M. (1987). Role of flower and pollen aromas in host-plant recognition by solitary bees. Oecologia 72: 618-623. Dobson, H. E. M. (1988). Survey of pollen and pollenkitt lipids - chemical cues to flower visitors. American Journal of Botany 75: 170-182. Dobson, H. E. M., and Ayasse, M. (2000). Does larval pollen diet influence host-flower selection in solitary bees American Journal of Botany 87: 42. Dobson, H. E. M., and Bergström, G. (2000). The ecology and evolution of pollen odors. Plant Systematics and Evolution 222: 63-87. Dobson, H. E. M., and Peng, . S. (1997). Digestion of pollen components by larvae of the flower- specialist bee Chelostoma florisomne (Hymenoptera: Megachilidae). Journal of Insect Physiology 43: 89-100. Dorn, M., and Weber, D. (1988). Die Luzerne-Blattschneiderbiene und ihre Verwandten in Mitteleuropa. Die Neue Brehm-Bücherei, A. Ziemsen Verlag, Wittenberg Lutherstadt. Dötterl, S., Füssel, U., Jürgens, A., and Aas, G. (2005). 1,4-Dimethoxybenzene, a floral scent compound in willows that attracts an oligolectic bee. Journal of Chemical Ecology 31: 2993-2998. Dötterl, S., and Schäffler, I. (2007). Flower scent of floral oil-producing Lysimachia punctata as attractant for the oil-bee Macropis fulvipes. Journal of Chemical Ecology 33: 441-445.

115 .eeence

e . ndllet . 00.ot ce n lnteedngnect ndte motnce n mtc ecton. Philosophical Transactions of the Royal Society of ondon Series BBiological Sciences 1. e.lmn.dc..ee.ndogel.00.olecllogenetc o ew mml ocdeeeltmngo tncontnentlcolonton.Molecular Phylogenetics and Evolution11. ckwot..ndneg..10.ogng ndmtngeonAode. Annual Review of Entomology1. ngel . . 1. Megachile glaesaria te t megcld ee ol om me menote egclde.American Museum ovitates11. ngel..001.Amonogoteltcmeeendeoltono teAode menote. Bulletin of the American Museum of atural History1. ngel..00.mlgonmeoeemenoteAode.American Museum ovitates 1. .nden.1.otonedolleneleendtendomeoecondolleneentton nteCampanulalesAsteralescomle.Flora10. e..1.oenentomologe.omee.elge. e..ndeee.A..1.ellodctontelllecttngeeMegachile rotundata .nnewndedwoodndoltenenetngmtel.ournal of Applied Entomology10 11. eton.1.ede elle et de geoee coegoe et nnote tenne detnocd.on. ce . met . cke . . nd on . 001. ntecton o tonl nd l ce n todotlocton.ournal of Comparative Physiology aeuroethology Sensory eural and Behavioral Physiology11. c..11.edenecnn deene ndeneltenologceedetng. Zoologische ahrbücher. Abteilung für allgemeine Zoologie und Physiologie der Tiere1. tm..eee..ndnk..1.enetcontntoncoeoltonteoltono otAltonnteeeetleenphraella.Evolution0. tm..eee..ndcee..1.enetccontntndtelogenonectlnt octoneoneophraellacommunaoleoteomeldetootlntot congene.Evolution0. tm..l..oton.nk..ndeee..1.enetctonnlogenetc conteteoneotwoecledleeetleoleoteomeldetootlnto tecongene.ournal of Evolutionary Biology11. ndol . ttcc . nd on . 00. emgnl lenng detemne dlt eone to cemcltmlntcw.Proceedings of the Royal Society of ondon Series BBiological Sciences0. e..1.eollenw.eecologndntltootene.dnet emdgeA. . nd ee . 001. oneee on nd lol dl om detecton to cloe ecognton.ge1in.ttkndomon..edto.Cognitive ecology of pollination. mdgenetemdge. oglA. 1. tlgenec eon o te ee te mnmenote egclde. Acta Entomologica Slovenica1. oglA.1.eenoloenceckltoecemenoteAode.Scopolia1. olen..ndndt.A.1.Alllowecolooctedwtdeentleedetlectte ee.Agronomy ournal00. wold. . 1. A genec nd genec eon o te Heriades gengo menote egclde.octole.tttenetognt. wold. . ndcene .. 1.e clcton o te mno te eten emee menoteegclde.ournal of the Kansas Entomological Society00. eele.1.smia rutilawldeeoccngntecotleoteotwetmedtenen weetnowndngeo etncton.in.cowod...ndd .. edto.The ecology and conservation of European dunes.cenc neto ell ell. de . . nd cl . . . 1. e nctonl gncnce o ocdl nte nd ollntoncontolledollenemolomDodecatheon.Functional Ecology01.

11 . References

Harvey, P. 1. Phylogenies for ecologists. The ournal of Animal Ecology : 2-24. Herbert, E. W., Bickley, W. E., and Shimanuk.H. 1. The brood-rearing capability of caged honey bees fed dandelion and mixed pollen diets. Journal of Economic Entomology : 21-21. Herrera, C. M. 1. Determinants of plant-animal coevolution: the case of mutualistic dispersal of seeds by vertebrates. Oikos 44: 12-141. Hines, H. M. 2. Historical Biogeography, Divergence Times, and Diversification Patterns of Bumble Bees Hymenoptera: Apidae: Bomus. Systematic Biology : - . Hopkins, A. D. 11. A discussion of H. G. Hewitts paper on Insect Behaviour. Journal of Economic Entomology 1: 2-. Huelsenbeck, J. P., and Ronquist, F. 21. MRBAES: Bayesian inference of phylogenetic trees. Bioinformatics 1: 4-. Hunt, R. M. 24. Global climate and the evolution of large mammalian carnivores during the later Cenozoic in North America. Bulletin of the American Museum of Natural History 2: 1-1. Hurd, P. D., aBerge, W. E., and insley, E. G. 1. Principal sunflower bees of North America with emphasis on the southwestern nited States Hymenoptera: Apoidea. Smithsonian Contriutions to oology 1: 1-1. Hurd, P. D., and insley, E. G. 1. The principal Larrea bees of the southwestern nited States Hymenoptera: Apoidea. Smithsonian Contriutions to oology 1: 1-4. Hurd, P. D., and Michener, C. D. 1. The megachiline bees of California Hymenoptera: Megachilidae. Bulletin of the California Insect Survey : 1-14. Inouye, D. W. 1. The terminology of floral larceny. Ecology 1: 121-12. Irwin, R. E., Adler, . S., and Brody, A. K. 24. The dual role of floral traits: Pollinator attraction and plant defense. Ecology : 1-111. Irwin, R. E., Brody, A. K., and Waser, N. M. 21. The impact of floral larceny on individuals, populations, and communities. Oecologia 12: 11-1. Iwata, K. 1. Evolution of instinct. Comparative ethology of Hymenoptera. Amerind, New Delhi. Jaenike, J. 1. Induction of host preference in rosophila melanogaster. Oecologia : 2-2. Jaisson, P. 1. Environmental preference induced experimentally in ants Hymenoptera, Formicidae. Nature 2: -. Janz, N., Nyblom, K., and Nylin, S. 21. Evolutionary dynamics of host-plant specialization: A case study of the tribe Nymphalini. Evolution : -. Janz, N., and Nylin, S. 2. The oscillation hypothesis of host-plant range and speciation. Pages 2-21 in K. J. Tilmon, editor. Specialization speciation and radiation The evolutionary iology of herivorous insects. niversity of California Press, Berkeley. Johnson, S. D., Hargreaves, A. ., and Brown, M. 2. Dark, bitter-tasting nectar functions as a filter of flower visitors in a bird-pollinated plant. Ecology : 2-21. Kratochwil, A. 11. Blüten-Blütenbesucher-Konnexe: Aspekte der Co-Evolution, der Co-Phnologie und der Biogeographie aus dem Blickwinkel unterschiedlicher Komplexittsstufen. Annali di Botanica 4: 4-1. Krenn, H. W., Plant, J. D., and Szucsich, N. . 2. Mouthparts of flower-visiting insects. Arthropod Structure evelopment 4: 1-4. Krombein, K. V., Hurd, P. D., Smith, D. R., and Burkers, B. D. 1. Catalog of Hymenoptera in America north of Mexico. Smithsonian Press, Washington. arkin, . ., Neff, J. ., and Simpson, B. B. 2. The evolution of a pollen diet: host choice and diet breadth of Andrena bees Hymenoptera: Andrenidae. Apidologie . arsson, M. 2. Higher pollinator effectiveness by specialist than generalist flower-visitors of unspecialized Knautia arvensis Dipsacaceae. Oecologia 14 : 4-4. averty, T. M., and Plowright, R. C. 1. Flower handling by bumblebees: a comparison of specialists and generalists. Animal Behaviour : -4. ee, E. T., and Wang, J. W. 2. Statistical methods for survival data analysis. John Wiley Sons Inc., New ork. evin, M. D., and Haydak, M. H. 1. Comparative value of different pollens in the nutrition of Osmia lignaria. Bee orld : 221-22. ewis, P. O. 21. A likelihood approach to estimating phylogeny from discrete morphological character data. Systematic Biology : 1-2. eys, R., Cooper, S. J. B., and Schwarz, M. P. 22. Molecular phylogeny and historical biogeography of the large carpenter bees, genus ylocopa Hymenoptera : Apidae. Biological Journal of the Linnean Society : 24-2.

11 . eerece

ile . . 1. e ecl litar ee. Hilgardia . ile . . 11. e rle ler eciicit i te eluti litar ee. ae in . nternationaler ongress für Entomologie ie. ile . . 1. emral atter ler iitati litar ee it articular reerece t te uteter ite tate. ournal of the ansas Entomological Society 1. ile . . a acai . . 1. e iiicace lral ctac am ee te eu Diadasia metera triae. Evolution 1 1. ell . . 11. e rii te litric ait am ee metera. Entomological news 111. ell . . 11. e rii litrim. Entomological news 11. ut . . . . a . . mura a . . eitt. 1. e iect ctcrme iae ee elutiar atter a cere rimer r leetic tuie. nsect Mol. Biol. 11. aciel e lmeia Crreia . . . 11. Ctriuti ltue e la ilie Heriades truncorum . metera iea eaciliae. . ect tliue. Apidologie 1 1. acai . . ae . . a r . . 1. Cmaratie eair ee a araceae. V. Claria ee te eter ite tate. University of California Publications in Entomology 1. ai . . a ai . . . acclae . r . iauer ciate c. uerla aacuett. ai . . a ai . . . euite .. mular tem r elutiar aali. ale . . 1. eeecicte er mie smia atr.. Zoologische ahrbücher, Abteilung für Systemati, Geographie und Biologie der Tiere 11. arul . a Camer . . 1. e mar i i ee ecta metera rmii uclear ee r ier leel leetic. Molecular Phylogenetics and Evolution 1 11. au V. uller . a ri . . ati eti a ler aciati te eat eiterraea lle a Ceramius bureschi i reece metera Veiae aariae. Entomologia Generalis 1. au V. ller . a ilirim . . irt ctriuti t te imic te lle a Ceramius caucasicus r 1 metera Veiae aariae i ure. ournal of the ansas Entomological Society 1. cuire . . itt C. C. ltuler . . a eme . V. . leetic tematic a iera ummiir aeia a maimum lieli aale artitie ata a electi a arriate artitii trate. Systematic Biology . iceer C. . 1. merica ee te eu Chelostoma. The Pan-Pacific Entomologist 1 . iceer C. . 11. eeric reii te merica miiae it ecritie te l rl eera metera eaciliae. American Midland Naturalist 11. iceer C. . 1. amic te te eera Chelostoma a Ashmeadiella metera eaciliae. Entomological news 1. iceer C. . 1. e il a leacutter ee Megachile brevis a it aciate. The University of ansas Science Bulletin 11. iceer C. . 1. ecr a ecriti rt merica meacili ee. ournal of the ansas Entomological Society . iceer C. . 1. iera te ee. Annals of the Missouri Botanical Garden . iceer C. . . e ee te rl ec eiti. e i ierit re altimre. iceer C. . a ettemeer C. . 1. e etl Andrena erythronii it cmaratie ata ter ecie metera reiae. The University of ansas Science Bulletin . ice . ati . amt . immerma . a Vereece . . . le a tlat eluti i elittiae .l. metera iea. Apidologie . ice . er . a amt . . lie iera et ci lrau e aeille lilectiue u ere Dasypoda atreille 1 metera iea elittiae. Annales de la Société Entomologiue de rance n. s. 1. icle . . Cae . . a eri . . rii a eclical ceuece lle ecialiati am eert ee. Proceedings of the oyal Society of London Series B-Biological Sciences 1.

11

26 6 in lantpollinator interactions from specialization to generalization 26 6 in lantpollinator interactions from specialization to generalization 1 Oecologia 21 2122 1 hytologia 1 1 11 Scheizerische Bienenzeitung 6 1 Osmia spinulosa 12 erffentlichungen fr Naturschutz und Landschaftspflege BadenWrttemberg 66 21 1 Entomologia eneralis 2 16 Biological Journal of the Linnean Society 222 16 Ecological Monographs 66 22 26 Biological Conservation 1 661 1 Colletes Biological Journal of the Linnean Society 2 sychotria Systematic Biology 2 2 2 2 in 2 Systematic Biology 6 26 Ephestia cautella lodia interpunctella Animal Behaviour 2 1 2 lora 2 1 1 Systematic Biology 612622 26 American Naturalist 16 2 2 Systematic Biology 66 1 Chelostoma anacific Entomologist 6 1 2 Cladistics 2 11 1 hysiological Entomology 1 2 2 Molecular hylogenetics and Evolution 6

11 9. References

Pereira, S. L., Johnson, K. P., Clayton, D. H., and Baker, A. J. (2007). Mitochondrial and nuclear DNA sequences support a cretaceous origin of columbiformes and a dispersal-driven radiation in the paleogene. Systematic Biology 56: 656-672. Pernal, S. F., and Currie, R. W. (2001). The influence of pollen quality on foraging behavior in honeybees (Apis mellifera L.). Behavioral Ecology and Sociobiology 51: 53-68. Peters, D. S. (1970). Die Stellung von Heriadopsis Cockerell 1931 (Insecta: Hymenoptera: Apoidea: Megachilidae). Senckenbergiana biologica 51: 199-203. Pitts-Singer, T. L. (2007). Olfactory response of megachilid bees, Osmia lignaria, Megachile rotundata and M. pugnata, to individual cues from old nest cavities. Environmental Entomology 36: 402-408. Plateau, F. (1897). Comment les fleurs attirent les insectes. Recherches expérimentales 5. Bulletins de lAcadémie Royale des Sciences, des Lettres et des Beau-Arts de Belgique 34: 847-880. Poinar, G. O., and Danforth, B. N. (2006). A fossil bee from Early Cretaceous Burmese amber. Science 314: 614. ueiroz, A. d., and Gatesy, J. (2006). The supermatrix approach to systematics. Trends in Ecology and Evolution 22: 34-41. R Development Core Team. (2006). R: A language and environment for statistical computing. in. R Foundation for Statistical Computing, Vienna, Austria. Radchenko, V. G., and Pesenko, Y. A. (1994). Protobee and its nests: A new hypothesis concerning the early evolution of Apoidea (in Russian; English translation in Entomological Review 75: 140-162, 1996). Entomologischeskoe Obozrenie 73: 913-933. Rambaut, A., and Drummond, A. J. (2003). Tracer, MCMC Trace Analysis Tool Version v1.4. (http: tree.bio.ed.ac.uk softwaretracer). Rayner, C. J., and Langridge, D. F. (1985). Amino acids in bee-collected pollens from Australian indigenous and exotic plants. Australian Journal of Eperimental Agriculture 25: 722-726. Réaumur, R. A. F. (1742). Mémoires pour servir à l'histoire des insectes. Tome sixime. Imprimerie Royale, Paris. Ree, R. H., Moore, B. R., Webb, C. O., and Donoghue, M. J. (2005). A likelihood framework for inferring the evolution of geographic range on phylogenetic trees. Evolution 59: 2299-2311. Ree, R. H., and Smith, S. A. (2008). Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Systematic Biology 57: 4-14. Rietdorf, K., and Steidle, J. L. M. (2002). Was Hopkins right Influence of larval and early adult experience on the olfactory response in the granary weevil Sitophilus granarius (Coleoptera, Curculionidae). Physiological Entomology 27: 223-227. Robertson, C. (1899). Flowers and insects. Botanical Gazette 28: 27-45. Robertson, C. (1925). Heterotropic bees. Ecology 6: 412-436. Roig-Alsina, A., and Michener, C. D. (1993). Studies of the phylogeny and classification of long-tongued bees (Hymenoptera: Apoidea). University of Kansas Science Bulletin 55: 123-173. Ronquist, F. (1997). Dispersal-vicariance analysis: A new approach to the quantification of historical biogeography. Systematic Biology 46: 195-203. Rse, U. S. R., Alborn, H. T., Makranczy, G., Lewis, W. J., and Tumlinson, J. H. (1997). Host recognition by the specialist endoparasitoid Microplitis croceipes (Hymenoptera: Braconidae): Role of host- and plant-related volatiles. Journal of Insect Behavior 10: 313-330. Roulston, T. H., Cane, J. H., and Buchmann, S. L. (2000). What governs protein content of pollen: Pollinator preferences, pollen-pistil interactions, or phylogeny Ecological Monographs 70: 617-643. Schlindwein, C., Wittmann, D., Martins, C. F., Hamm, A., Siqueira, J. A., Schiffler, D., and Machado, I. C. (2005). Pollination of Campanula rapunculus L. (Campanulaceae): How much pollen flows into pollination and into reproduction of oligolectic pollinators Plant Systematics and Evolution 250: 147-156. Schwarz, M., and Gusenleitner, F. (2000). Weitere Angaben zur Bienenfauna sterreichs sowie Beschreibung einer neuen Chelostoma-Art aus der Westpalarktis. Vorstudie zu einer Gesamtbearbeitung der Bienen sterreichs IV (Hymenoptera, Apidae). Entomofauna 21: 133-164. Schwarz, M., Gusenleitner, F., Westrich, P., and Dather, H. H. (1996). Katalog der Bienen sterreichs, Deutschlands und der Schweiz (Hymenoptera, Apidae). Entomofauna 8: 1-398. Schwarz, M. P., Fuller, S., Tierney, S. M., and Cooper, S. J. B. (2006). Molecular phylogenetics of the exoneurine allodapine bees reveal an ancient and puzzling dispersal from Africa to Australia. Systematic Biology 55: 31-45. Schwarz, M. P., Gusenleitner, F., Westrich, P., and Dathe, H. H. (1996). Katalog der Bienen sterreichs, Deutchlands und der Schweiz (Hymenoptera, Apidae). Entomofauna Supplement, 8: 1-398.

120 9. References

Singer, M. S. (2008). Evolutionary ecology of polyphagy. Pages 29-42 in K. J. Tilmon, editor. Specialization speciation and radiation he evolutionary biology of herbivorous insects. University of California Press, Berkeley. Sipes, S. D., and Tepedino, V. J. (2005). Pollen-host specificity and evolutionary patterns of host switching in a clade of specialist bees (Apoidea : Diadasia). Biological Journal of the Linnean Society 86: 487-505. Sipes, S. D., and Wolf, P. G. (2001). Phylogenetic relationships within iadasia, a group of specialist bees. Molecular Phylogenetics and Evolution 19: 144-156. Slansky, F. (1993). Nutritional ecology: the fundamental quest for nutrients. Pages 29-91 in N. E. Stamp and Casey, T. M., editors. Caterpillars: ecological and evolutionary constraints on foraging. Chapman and Hall, New York. Solarz, S. L., and Newman, R. M. (2001). Variation in hostplant preference and performance by the milfoil weevil, Euhrychiopsis lecontei Dietz, exposed to native and exotic watermilfoils. Oecologia 126: 66-75. Solits, D. E., Solits, P. S., Endress, P. K., and Chase, M. W. (2005). Phylogeny and evolution of angiosperms. Sinauer, Sunderlans. Somerville, D. C., and Nicol, H. I. (2006). Crude protein and amino acid composition of honey bee-collected pollen pellets from south-east Australia and a note on laboratory disparity. Australian Journal of Eperimental Agriculture 46: 141-149. Spaethe, J., and Briscoe, A. D. (2004). Early duplication and functional diversification of the opsin gene family in insects. Molecular Biology and Evolution 21: 1583-1594. Stanley, R. G., and Linskens, H. F. (1974). Pollen - Biology, Biochemistry, Management. Springer-Verlag, Berlin, Heidelberg, New York. Stephen, W. P., Bohart, G. E., and Torchio, P. F. (1969). The biology and external morphology of bees. Agricultural Experiment Station, Oregon State University, Corvallis, USA. Strauss, S. Y., Siemens, D. H., Decher, M. B., and Mitchell-Olds, T. (1999). Ecological costs of plant resistance to herbivores in the currency of pollination. Evolution 53: 1105-1113. Strauss, S. Y., and Zangerl, A. R. (2002). Plant-insect interactions in terrestrial ecosystems. Pages 77-106 in C. Herrera and Pellmyr, O., editors. Plant-animal interactions: an evolutionary approach. Blackwell, Oxford, UK. Strickler, K. (1979). Specialization and foraging efficiency of solitary bees. Ecology 60: 998-1009. Suarez-Cervera, M., Marquez, J., Bosch, J., and Seoane-Camba, J. (1994). An ultrastructural study of pollen grains consumed by larvae of Osmia bees (Hymenoptera: Megachilidae). Grana 33: 191-204. Swofford, D. L. (2002). Paup.4.0b10 for Macintosh. Sinauer Associates Inc., Sunderland. Tautz, J. (2007). Phänomen Honigbiene. Elsevier, München. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F., and Higgins, D. G. (1997). The Clustal windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids esearch 25: 4876-4882. Thompson, J. N. (1994). The coevolutionary process. The University of Chicago Press, Chicago and London. Thorp, R. W. (1969). Systematics and ecology of bees of the subgenus iandrena (Hymenoptera: Andrenidae). niversity of California Publications in Entomology 52: 1-146. Thorp, R. W. (1979). Structural, behavioral, and physiological adaptations of bees (Apoidea) for collecting pollen. Annals of the Missouri Botanical Garden 66: 788-812. Thorp, R. W. (2000). The collection of pollen by bees. Plant Systematics and Evolution 222: 211-223. Toepfer, S., Gu, H., and Dorn, S. (1999). Spring colonisation of orchards by Anthonomus pomorum from adjacent forest borders. Entomologia Eperimentalis et Applicata 93: 131-139. Torchio, P. F., and Bosch, J. (1992). Biology of ricrania stansburyi, a Meloid beetle cleptoparasite of the Bee Osmia lignaria propinua (Hymenoptera, Megachilidae). Annals of the Entomological Society of America 85: 713-721. Ungricht, S., Müller, A., and Dorn, S. (Submitted). A taxonomic catalogue of the palaearctic bees of the tribe Osmiini (Hymenoptera: Apoidea: Megachilidae). ootaa. Van der Made, J., Morales, J., Sen, S., and Aslan, F. (2002). The first camel from the Upper Miocene of Turkey and the dispersal of the camels into the Old World. Comptes endus Palevol 1: 117-122. Villagra, C. A., Pennacchio, F., and Niemeyer, H. M. (2007). The effect of larval and early adult experience on behavioural plasticity of the aphid parasitoid Aphidius ervi (Hymenoptera, Braconidae, Aphidiinae). Naturwissenschaften 94: 903-910. Vogel, S. (1993). Betrug bei Pflanzen: Die Täuschblumen.

121 9. References von Aufess, A. (1960). Geruchliche Nahorientierung der Biene bei entomophilen und ornithophilen Blten. eitschrift fr vergleichende Physiologie 43: 469-498. Walsh, B. D. (1864). On phytophagic varieties and phytophagic species. Proceedings of the Entomological Society of Philadelphia 3: 403-430. Warncke, K. (1991c). Die Bienengattung Osmia Panzer 1806, ihre Systematik in der Westpaläarktis und ihre Verbreitung in der Turkei. 7. Die Untergattung oveosmia nov. Linzer Biologische Beitrge 23: 267-281. Waser, N. M., Chittka, L., Price, M. V., Williams, N. M., and Ollerton, J. (1996). Generalization in pollination systems, and why it matters. Ecology 77: 1043-1060. Wcislo, W. T., and Cane, J. H. (1996). Floral resource utilization by solitary bees (Hymenoptera: Apoidea) and exploitation of their stored foods by natural enemies. Annual Review of Entomology 41: 257-286. Weingartner, E., Wahlberg, N., and Nylin, S. (2006). Dynamics of host plant use and species diversity in Polygonia butterflies (Nymphalidae). Journal of Evolutionary Biology 19: 483-491. Westerkamp, C. (1996). Pollen in bee-flower relations - Some considerations on melittophily. Botanica Acta 109: 325-332. Westerkamp, C. (1997a). Flowers and bees are competitors - Not partners. Towards a new understanding of complexity in specialised bee flowers. Pages 71-74 in K. W. Richards, editor. Seventh International Symposium on Pollination - Pollination from Theory to Practise. Westerkamp, C. (1997b). Keel blossoms: Bee flowers with adaptations against bees. lora 192: 125-132. Westrich, P. (1989). Die Wildbienen Baden-Wrttembergs. Eugen Ulmer GmbH Co, Stuttgart. Westrich, P. (1993). ber die Verbreitung und Bionomie der Scherenbiene Chelostoma grande (Nylander) (Hymenoptera, Apoidea, Megachilidae). Linzer Biologische Beitrge 25: 97-111. Westrich, P. (2006). ur Nomenklatur von Chelostoma mocsaryi Schletterer, 1889, Osmia mocsaryi Friese, 1895 und Heriades gasperinii Schmiedeknecht, 1890 (Hymenoptera, Apidae). Entomofauna 27: 297-300. Westrich, P. (2008). Flexibles Pollensammelverhalten der ansonsten streng oligolektischen Seidenbiene Colletes hederae Schmidt Westrich (Hymenoptera: Apidae). Eucera 1: 17-29. Westrich, P., and Dathe, H. H. (1997). Die Bienenarten Deutschlands (Hymenoptera, Apidae): Ein aktualisiertes Verzeichnis mit kritischen Anmerkungen. Mitteilungen - Entomologischer erein Stuttgart 32: 1-34. Westrich, P., and Schmidt, K. (1986). Methoden und Anwendungsgebiete der Pollenanalyse bei Wildbienen (Hymenoptera, Apoidea). Linzer Biologische Beitrge 18: 341-360. Wille, H., Wille, M., Kilchenmann, V., Imdorf, A., and Buhlmann, G. (1985). Pollenernte und Massenwechsel von drei Apis mellifera-Völkern auf demselben Bienenstand in zwei aufeinanderfolgenden Jahren. Revue Suisse de oologie 92: 897-914. Williams, N. M. (2003). Use of novel pollen species by specialist and generalist solitary bees (Hymenoptera : Megachilidae). Oecologia 134: 228-237. ang, D. ., and Hewitt, G. M. (1996). Assessment of the universality and utility of a set of conserved mitochondrial COI primers in insects. Insect Molecular Biology 6: 143-150.

122

10. Acknowledgment

irst would lie to than Andreas Müller for the design and initiation of this fascinating proect and for the excellent supervision of my thesis Andreas has introduced me into the intriguing world of bees which has rapidly become a new passion for me e has motivated me when needed it and has contained my ardor when it happened to be excessive ith Andreas have learned a lot not only about bees am very much looing forward to the continuation of our friendly collaboration

am most grateful to Silvia Dorn for the supervision of this proect for giving me the opportunity to wor in her group for her trust throughout this thesis and for the pertinent and quic revisions of my drafts also want to than her for creating a very professional woring atmosphere in the Applied Entomology group have benefited from the best research conditions and full support during my PhD at ET

am also indebted to Alex idmer for giving me the occasion to wor in his lab for the supervision of the molecular wor and above all for introducing me into his very friendly research group also benefited from the excellent advises from Claudia Michel and Catharine Aquino ournier

During this thesis have been very lucy to be teamed with Claudio Sedivy e proved to be not only an outstanding diploma student but also a very friendly colleague always enthusiastic to discuss and to plan new trips to collect bees

Claude ornalla has not counted his time to help me solving any technical problem or installing obsolete phylogeny pacages on my computer Urs Guyer and Christian Bni generously helped me with plant breeding also want to than all actual and former members of the Applied Entomology group for their friendliness especially Sile ein Dominique Mai arsten Mody Ana Rott Marion Schmid Antonia Zurbuchen for the numerous advises and discussions

The phylogenetic part of this thesis was only possible thanks to the very positive collaboration with Bryan Danforth and Terry Griswold. The Walter Hochstrasser Foundation generously founded two trips to collect bees. Many people provided bee specimens for the phylogeny. I want to thank in particular Werner Arens, John Ascher, Sara Bangerter, Oistein Berg, Sam Droege, Andrew Grace, Antonius van Harten, Michael Kuhlmann, Charles Michener, Robert Minckley, Sbastien Patiny, Severin Roffler, Erwin Scheuchl, Christian Schmidt-Egger, Maximilian Schwarz and Kim Timmermann.

The rearings of the bees were only possible because the following people provided bee nests from their private garden: Felix Amiet, Mike Herrmann, Karl Hirt, Albert Krebs, Sabine Oertli, R. Prosi, E. Steinmann, Hansueli Tinner, the botanical Garden of Zürich and the Neolithic village of Gletterens. Many thanks to all for welcoming me very warmly at each visit.

Enfin, aimerais remercier chaleureusement mes amis et ma famille, ui mont accompagn tout au long de ces uatre annes: ma deuxime famille, Vrne et Mathias Hirt, pour la conception des blocs de cramiue Jake Alexander, pour la correction des textes anglais Jrme Dayer, ui ma accompagn au milieu du Sahara pour chercher une minuscule abeille Gilles Carron, pour son intrt constant pour ce proet Denis Michez et Nicolas Vereecken pour les discussions passionnes Maria Domenica Moccia, Ccile Thonar et Herv Vanderschuren, pour avoir rendu la vie Zurich plus latine mes parents et ma famille, pour leur soutien sans relche enfin, Krisztina, pour son amour et sa comprhension.

124

11. clmte

Christophe Praz Born 2 August 1 Place of Birth Sion, alais

22 PhD Student at ETH Zurich Applied Entomology Group Supervision Prof. Silvia Dorn and Dr. Andreas Müller

23 Master in biology at the University of Berne

2223 Master thesis in vegetation ecology Supervision Prof. David ewbery

123 Student at the University of Berne

1 Maturit fdrale, Baccalaurat B

131 High school at the yce Collge des Creusets, Sion

12