Life-history strategies of a facultative cooperative breeder in a fragmented Afrotropical cloud forest – Life-history strategie¨envan een facultatief co¨operatieve broedvogel in een gefragmenteerd Afrotropisch nevelwoud

Dries Van de Loock

2019

Dissertation submitted in fulfillment of the requirements for the degree of Doctor (Ph. D.) in Science: Biology

This dissertation is part of a joint-PhD degree between Ghent University and University of Antwerp

Ghent University, Faculty of Sciences, Department of Biology, Terrestrial Ecology Unit University of Antwerp, Faculty of Sciences, Department of Biology, Evolutionary Ecology Group

Supervisors: Prof. Dr. Luc Lens Prof. Dr. Erik Matthysen Members of the examination committee: Prof. Dr. Dries Bonte (chairman) Dr. Liesbeth De Neve (secretary) Prof. Dr. Wendt Muller Dr. Beata Apfelbeck Dr. Job Aben Dr. Matthieu Paquet

The author and supervisors give the authorisation to consult and copy parts of this work for personal use only. Every other use is subjected to copyright laws. Permission to reproduce any material contained in this work should be obtained from the author or supervisors.

Please refer to this work as: Van de Loock, D. (2019). Life-history strategies of a facultative cooperative breeder in a fragmented Afrotropical cloud forest. Ph.D. thesis, Ghent University & University of Antwerp, Belgium. Acknowledgements

Het dankwoord. Terwijl dit doctoraat enkel mijn naam draagt, heb ik zelden het gevoel gehad er alleen voor te staan. En dat is aan volgende mensen te danken. Luc, bedankt om me dit avontuur toe te vertrouwen en om een goede promotor te zijn. Van Kenya tot in Belgi¨e,je stond altijd paraat om met raad en advies bij te staan, zaken te reviewen en knopen door te hakken wanneer nodig. Het immer aanstekelijk enthousiasme en de standvastige consequentie waarmee mails beantwoord worden doen maar weinig mensen je na. Erik, bedankt om Luc te vertrouwen over zijn vertrouwen in mij. Bedankt voor alle reflecties en feedback tijdens momenten van Antwerpse bezinning. Ik vond het bijzonder fijn samenwerken met jullie. Lies, jij was er reeds vanaf het begin bij, toen nog voor mijn masterthesis. In juli 2011, toen we samen om 22u op de luchthaven van Nairobi aankwamen, de reisgids opensloegen op zoek naar een hotel en beseften dat we dit v´e´elvroeger hadden moet doen. We besloten toen met onze ’Rough Guide’ in de hand, opengeslagen op de pagina met een hotel dat ons wel fijn leek, een telefoonkaart-verkoper aan te spreken en hem te vragen even te bellen. Het duurde niet lang voor hij de situatie doorhad : ”dat hotel bestaat niet meer, ik kan je een veel beter voorstellen”. Even later spendeerden we onze eerste nacht in downtown Nairobi - ergens in een obscuur hotelletje in de charmante wijk rond het busstation. Het begin van een heel fijn avontuur, waarbij ik toen nog niet kon inbeelden dat het zich zo zou ontplooien. Bedankt om samen de TaitaBase te overmeesteren, de resultaten te verteren, en over het leven te reflecteren. Diederik, jij kwam maar tijdens een later bedrijf op scene, maar je had het script gelezen. Heel erg bedankt voor je uitgebreid bijstand met deskundige raad en daad. Ze hebben significant bijgedragen tot wat het eindresultaat nu is. De collega’s van Terec, Limno en Eon aan de UGent. Bedankt voor al het vossen en voor al die prima donna’s (met kappertjes zonder mais). Ik was trouwens altijd in hogere sferen wanneer ik tussen jullie kon vertoeven op het elfde. In’t bijzonder : Hans, bedankt voor al de zwans (en alle andere exquisie-teit). Ik hoop dat de grijze haren de zwans-o-meter nog wat sparen. Viki, bedankt om die vaste waarde te zijn. Flashback maart

Page iii 2009, toen het ’negende’ nog het ’negende’ was. Ik zat naast jou, lichtjes indommelend bij de video’s die ik voor mijn bachelorproef aan het doorscrollen was. Jij diende je gsm op te laden en plugde de stekker doodleuk in de verdeeldoos van je computer, onbewust van het feit dat je door deze handeling die ook meteen had uitgezet. De verbazing op je gezicht was groot toen bleek dat je scherm wel heel erg zwart was geworden en de muis bitter weinig respons gaf. Toen hoorde ik, voor zover ik mij kan herinneren, mijn eerste M`aAleeeee. Angelica, bedankt voor het minutieus en onvoorwaardelijk verwerken van al mijn receipts. Tenminste, totdat bleek dat er een gapend gat in de begroting was. Toen werd het helaas wat minder onvoorwaardelijk. Bedankt ook om even penningmeester van de TEREC-motorclub te zijn. Ik wist dat het een kwestie van volharden was alvorens ik een perfecte match zou vinden. Die bleek uiteindelijk op 2-takt te draaien en 175cc te bezitten. Lionel, merci pour ˆetretoujours l`aquand j’avais une crise de stats, merci pour ˆetremon ami franais, et merci pour ˆetreun bon gagneur/perdant dans le ping-pong (`a toi pour barrer ce qui ne pas correct). Laurence, thanks for fighting for the same causes, and sharing the same burdens. Bram, the Muffinman, en Dean the Banean, ik vond het een waar genoegen 5 jaar lang liefde en leed te delen in de witruimte die we onze bureau noemden. Was er een wordcloud van onze gesprekken dan zou die ongetwijfeld ’pintje’ - bij uitbreiding ’spinnenkopje’ - en ’koffietje’ doen uitlichten, al dan niet met vraagteken of uitroepteken. Bedankt voor de uitstapjes en de occasionele spontane twitch of verlengd weekendje - een doctoraat schrijf je toch niet van ’nine’ tot ’five’ en leent zich uitstekend tot telewerken. Jorunn, bedankt om een toevluchtsoord aan te bieden voor schrijfelijk en statisch verderf. Ik vond het er geweldig ! Er was ooit een wederupstanding of the infamous grouse die je net miste. Bij deze sluit ik een wederupstanding of the infamous DJ Virage niet uit. De collega’s van EVECO aan de UAntwerpen. Bedankt om me welkom te doen voelen in jullie grote stad, ik vond het er telkens heel fijn vertoeven. Stijn, bedankt om van de vele praktische afhandelingen een droom te maken. Frank, bedankt voor de toegangscodes tot de EVECO-Mainframe. Die Antwerpse Goliath kon de onzekere Bayesian wel meester. Bedankt Joris om me iedere keer weer met ´e´enof andere (vogel) anekdote te doen inzien dat ik meer moet buiten komen. Ook bedankt om de Taita Heuvels (met volledige telemetrie outfit) te trotseren en daarbij alle veldobservaties niet neer te schrijven, maar zwoel in een dictafoon in te fluisteren. ”Pink metal aan de rechterpoot, en volgens mij heeft em ook nen zender aan...... zender niet 100% zeker”. Parental advisory, explicit content. De master- en bachelorstudenten, bedankt voor jullie inzet en toewijding. Lissa, ik vond het heel fijn een birdseye view te krijgen van de understory. Jos, de Carcassonne- koning in wording! Bedankt voor het fantastische gezelschap, het lianen-geslinger en om me te laten inzien dat het niet per se vanzelfsprekend is om met een brommer rond te rijden in de hills, of met een antenne en GPS rond te lopen in het bos op zoek naar een

Page iv zender. Alex, thanks for your telemetry-dedication in the field and your pasta- dedication in the kitchen. Ruben, bedankt voor je toewijding bij het feeding experiment! Alexander, bedankt voor het verslinden van de video’s! Team Fantastic - that is Peter, Adam, Lawrence and Oliver - thanks for being absolutely fantastic in the field!!! I’ve built this thesis on your foundations. Credit where credit is due, and that’s without a doubt you. Thanks to Henderson, Cosmas and Josai too, you were great additions to the team. Mwangi, thanks for saving us that first Nairobi morning in July, and the many times thereafter. Being it with permits, fieldwork related issues, motorbike issues, financial issues, relationship issues, philosophical issues, wildlife issues, you were always there and in full support. Asanti sana for the Tuskers (malt) baridi, both at and in the pool. Pete and Gorm, thanks for all the petrinizing discussions, the sundowners at Rocks and for being delightful housemates. The Chovu family, Mama Grace, Janice& Martha, thanks for making it feel like returning home after a day in the field. And special thanks to Max and Liesbeth for actually building that home. Thanks to the staff of the Helsinki research centre for their hospitality and for making an authentic Scandinavian experience possible. Lawrence, Vincent and Sylvester, thanks for helping me out in the field and sorting out the occasional paperwork, thanks for the great company and the good laughs. Wendy, thanks for showing me the leafy side of town, for the occasional glass of red wine and for sharing a passion for . Colin, thanks for creating a small paradise on earth and for sharing your expertise and knowledge on ringing. Thanks to the jury members, for dusting of the thesis, eliminating unnecessary side tracks and keeping my conclusions clear. Thanks to the Kenya Forest Service for permitting this research in the Taita Hills, and for Jonam in particular for his cooperation in the field. I also want to acknowledge the National Commision for Science, Technology and Innovation for granting my research license. Finally, my greatest appreciation goes to the ornithology section of the National Museums of Kenya for having me as an affiliated researcher, and to Peter Njoroge, for helping me sort all this out. De vrienden, vele van hen ook mede-doctoraatsstrijders in binnen en buitenland, bedankt voor de momenten van ontspanning wanneer het werk zich te veel opdrong. Op caf, in het bos, langs de schelde, aan de kust, in de Viroin, in de Gaume en overal daartussenin. Al fietsend, al wandelend, al lopend, al dansend, al zittend, al liggend en alles daartussenin. (sorry als het weer maar even over het doctoraat ging) In’t bijzonder wil ik Laura bedanken om me meermaals te redden van alweer een avond fastfood, en me te behoeden voor een hapjes-loze receptie. Ottolenghi in’t kwadraat. Stefan, bedankt om mijn horizon te verbreden op het moment dat die zich wel heel beangstigend aan het vernauwen was. Moeke, bedankt voor de moederlijke steun en toeverlaat. En vake, voor het aanleveren van jouw muzikale passie en technische genen. Lode, bedankt voor de Zwitserse retraite. Bedankt Klaas, om die fantastische broer te zijn, en Karen om het beste in hem

Page v boven te halen. Marc, om zonder veel woorden te weten wanneer het nodig is alles te laten vallen en te hulp te schieten. Hilde, om de geografisch dichtste moeder te zijn, en de Maesekes, omdat er wel eens wat mag gedronken worden. Syby, je verdient een reuze bloemenkrans! Zo van die lokaal gedroogde, die eeuwig meegaan. Je hebt een monumentale rechtstreekse en onrechtstreekse bijdrage geleverd aan dit werk. Van de ontspanningsdansjes in huis tot de fietstochten ver daarbuiten. Van momenten van totale overgave aan zwakzinnigheid tot momenten van diepe reflectie. Bedankt om me te lokaliseren, toen ik mijn co¨ordinaten was verloren. Bedankt om aan de rem te trekken toen de carrousel bleef draaien. Bedankt om voor mij uit te zoomen toen die functie haperde. Bedankt om dit avontuur samen aan te gaan, te ondergaan, te doorstaan, en af te ronden. Voor ons volgend avontuur houd ik me alvast vast aan de mantra : ”the sky is the limit, the horizon is our destination”. Of, in de woorden van een ander groot filosoof : ”To infinity and beyond !”.

Dries

Gent, 31/01/2019

Page vi Table of Contents

Acknowledgements iii

Summary xi

Samenvatting xv

1 General introduction1 1.1 Anthropogenic Habitat Change...... 3 1.2 Cooperative Breeding...... 4 1.2.1 Costs and benefits of cooperation...... 5 1.2.2 Ecological drivers of cooperation...... 7 1.3 Cooperative breeding under anthropogenic habitat change...... 8 1.4 The Placid Greenbul in the Taita Hills...... 9 1.4.1 The Taita Hills...... 9 1.4.2 The Placid Greenbul...... 10 1.4.3 Effect of habitat change on the Taita Hills’ avian community, includ- ing the Placid Greenbul...... 13 1.4.4 General methodology...... 14 1.5 Aims and Outline...... 15

2 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore 23 2.1 Abstract...... 24 2.2 Introduction...... 24 2.3 Methods...... 26 2.3.1 Study species and area...... 26 2.3.2 Survey and sampling design...... 27

Page vii 2.3.3 Nest-site Selection...... 29 2.3.4 Reproductive correlates of nest-site selection...... 29 2.4 Results...... 32 2.5 Discussion...... 34 2.6 Acknowledgements...... 37

3 Cooperative breeding of an Afrotropical bird in a degraded and frag- mented landscape 39 3.1 Abstract...... 40 3.2 Introduction...... 40 3.3 Methods...... 43 3.3.1 Study area and species...... 43 3.3.2 Data collection & Genotyping...... 44 3.3.3 Statistical Analysis...... 45 3.4 Results...... 46 3.4.1 Cooperative breeding and group composition...... 46 3.4.2 Degree of philopatry...... 48 3.5 Discussion...... 48 3.6 Acknowledgements...... 52

4 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape 55 4.1 Abstract...... 56 4.2 Introduction...... 56 4.3 Methods...... 59 4.3.1 Study area & species...... 59 4.3.2 Data collection & handling...... 60 4.3.3 Statistical analysis...... 61 4.4 Results...... 64 4.4.1 Maternal Investment...... 64 4.4.2 Reproductive success & nestling condition...... 65 4.5 Discussion...... 67 4.6 Acknowledgements...... 72

5 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird 75 5.1 Abstract...... 76 5.2 Introduction...... 76 5.3 Material and Methods...... 77 5.3.1 Study system...... 77

Page viii 5.3.2 Resighting data...... 78 5.3.3 Statistical modeling...... 78 5.4 Results...... 80 5.5 Discussion...... 81 5.6 Acknowledgements...... 83

6 General discussion 87 6.1 Fitness payoffs and ecological drivers of cooperative breeding...... 88 6.1.1 Subordinates benefit from cooperating...... 88 6.1.2 Breeding cooperatively benefits breeders...... 91 6.1.3 Does habitat change promote cooperative breeding?...... 95 6.2 Effect of anthropogenic habitat change...... 96 6.2.1 Increasing spatial replication...... 97 6.2.2 The importance of a reliable metric of habitat quality...... 97 6.2.3 Cooperative breeding may moderate effects of anthropogenic habitat change...... 99 6.3 General conclusions...... 102

7 Bibliography 105

A Appendix to Chapter 2 131 A.1 Vegetation Composition...... 131 A.2 Daily Survival Rates...... 132

B Appendix to Chapter 4 141

C Appendix to Chapter 5 147 C.1 Literature Review...... 147

Page ix

Summary

Anthropogenic, or human-induced, habitat change involves loss, fragmentation and degra- dation of pristine habitat and has a profound impact on bird species from every corner of the world. Though birds have been shown to vary in their sensitivity to these changes, it is not always clear how they will respond. For instance, cooperatively breeding species, in which several non-breeding individuals help a breeding pair during a reproductive attempt, are characterized by sex-specific, short-distance and delayed dispersal. While these traits may render them more sensitive to changes in the landscape, the large number of non-breeding individuals may, in contrast, also better buffer cooperatively breeding populations. At present, it remains unclear if, and to what extent, anthropogenic habitat change influences cooperatively breeding species, and it is hence difficult to judge their sensitivity, or robustness against such changes. The aim of this thesis is to advance our understanding of the influence of anthropogenic habitat change on cooperative breeders. To do so, I studied a facultative cooperatively breeding population of the Placid Greenbul ( placidus) from a severely degraded and fragmented landscape, the Taita Hills in SE Kenya. This medium- sized forages and breeds as pairs or small groups in cloud-forest, of which about a dozen remnant fragments are scattered across a landscape that is otherwise dominated by small-scaled subsistence agriculture. These forest fragments vary in size, degree of isolation and level of degradation, and together offer an interesting opportunity to evaluate how habitat change may influence the species life-history traits and cooperative breeding. I here present data from almost ten breeding seasons (since 2007), of which four where specifically studied for this thesis (from 2012 until 2016). In brief, during the whole of each breeding season, a team of local field assistants systematically searched all fragments for greenbul nests. These were further monitored to assess the outcome and record reproductive parameters. At the same time, a combination of field observations, targeted mist-net trapping, video-recording and radio-telemetry were used to infer cooperative breeding behaviour and other aspects related to breeding such as nest-site choice, maternal investment and post-fledging survival.

Page xi Across all forest fragments, a majority of pairs (64%) bred in groups together with non-breeding individuals (subordinates). Subordinates are represented by offspring from previous breeding attempts (in 73% of all groups), by unrelated birds of unknown origin (22%), or a combination of both (5%). Helping behaviour (i.e., providing food to the nestlings) was documented for roughly 50% of related subordinates, but never observed for unrelated immigrants. In chapter3 , I evaluated variation in delayed dispersal, group composition and cooperative breeding between the forest fragments. Although I expected to detect differences in delayed dispersal and cooperative breeding due to variation in habitat quality, constraints on independent breeding and/or disruption of dispersal, I did not find such differences. Delayed dispersal, cooperative breeding and most group characteristics remained constant across all fragments. However, I found a significant difference in subordinate sex-ratio across fragments which may be due to sex-specific mortality and timing of dispersal. Together, these results suggest that cooperative breeding is generally maintained in this fragmented Placid Greenbul population. Nest predation is the main cause of nest failure in open-cup nesting birds like the Placid Greenbul. Predation risk may be mitigated by tactically selecting safe sites for nesting, although this strategy may be compromised when habitat degradation alters the habitat makeup. In chapter2 , I compared the environment at sites used for nesting with randomly chosen sites without evidence of nesting and did this for two contrastingly degraded, but similarly sized, forest fragments. I inferred what the influence is of habitat degradation on nest-site selection and subsequently evaluated the reproductive consequences thereof. I found that Placid Greenbul nest sites vary between fragments, but did not detect a relationship between reproductive success and nest-site selection. Such variation in nest-site selection may be due to a plastic response of the species to fragment-specific conditions, or may be driven by fragment-specific variation in habitat composition. Given the high predation rates, better insight on the relationship between nest-site selection and reproductive success will likely come from better knowledge on predator communities. In a next chapter (4), I compared pre- and post-hatching maternal investment (i.e., variation in egg size and nestling provisioning, resp.) and reproductive success (i.e., fledging success and nestling body condition) across fragments and relate this to variation in group size (breeding pair and all subordinates) and number of helpers (subordinates that provision food to nestlings). Egg investment varied between fragments, with females either laying larger eggs in the presence of subordinate, or smaller ones. Egg investment was irrespective of the number and/or presence of helpers. In contrast, investment in provisioning decreased with number of helpers, irrespective of habitat degradation. This suggests that maternal investment may be modulated by the degree of degradation and/or other (fragment-specific) environmental conditions, but depends on the type of investment and the size and composition of the group. While nestling condition was not affected by any of the tested variables, fledging success increased with group size. Taken together,

Page xii the findings of this chapter illustrate that females may benefit from cooperative breeding through both flexible maternal investment as well as augmented reproductive success. Yet, it is now necessary to explore the long-term fitness consequences of such plastic maternal strategies in both degraded and pristine landscapes to judge whether this flexibility may actually mitigate the negative consequences of habitat degradation. Positive (or negative) fitness consequences of cooperation may also be detectable after fledging. While this is particularly relevant for cooperative breeders where offspring typically depend longer on their parents, these affects have rarely been assessed due to the difficulty of following mobile fledglings. In chapter5 , I used a novel, indirect, radio- telemetry approach to overcome this challenge and found that the survival probability is positively related to group size. This illustrates the importance of studying the early dependency period just after fledging to evaluate the effects of cooperative breeding on reproductive success. In conclusion, there is currently no evidence that anthropogenic habitat change interferes with cooperative breeding behaviour in this species, as few investigated rela- tionships varied across a fragmented landscape composed of differently degraded forest fragments. In fact, I here showed that cooperative breeding benefits breeders as breeding in larger groups is associated with higher fledging success, and increased survival of fledglings. Yet, it is now necessary to obtain information from populations in other, more pristine forest conditions, to infer whether the findings in the Taita Hills are representative for this, and perhaps other cooperatively breeding species. Then, further (experimental) investi- gation is required to confirm if, and to what extent, this breeding strategy may actually buffer this population against consequences of habitat change, as some observations suggest this may indeed be the case (e.g. flexible maternal investment and the large number of non-breeding subordinates).

Page xiii

Samenvatting

Overal ter wereld worden vogels geconfronteerd met de menselijke invloed op hun natu- urlijke biotoop. Terwijl de reactie van sommige vogelsoorten op menselijke vernietiging, versnippering en verstoring van hun leefgebied goed gekend is, is dit voor veel andere soorten, zoals coperatief broedende vogelsoorten, nog onduidelijk. Bij vogelsoorten met co¨operatief broedgedrag, waarbij een broedpaar hulp krijgt van niet broedende individuen tijdens een broedpoging, blijven mannetjes langer bij de ouders om vervolgens een territo- rium in de buurt te bemachtigen. Vrouwtjes daarentegen verlaten sneller hun ouderlijk territorium en gaan veel verder op zoek naar een partner. Door deze strategie worden co¨operatief broedende vogels geacht gevoeliger te zijn aan het verstoren en versnipperen van hun leefgebied. Door de grote groep niet-broedende individuen die populatieschommelingen kunnen opvangen, worden co¨operatieve vogels echter verondersteld juist beter bestand te zijn tegen biotoopsveranderingen. Hoe, en in welke mate, antropogene habitat verandering coperatief broedende vogelsoorten precies be¨ınvloed is dus nog niet goed gekend. Om meer inzicht te krijgen in het effect van antropogene verstoring op dergelijke vogels werd in deze thesis een facultatief co¨operatief broedende vogel, de Placid Greenbul (Phyllastrephus placidus), bestudeerd in een sterk verstoord en versnipperd gebied, de Taita Hills in Kenya. Als een facultatieve co¨operatief broedvogel zijn zowel paartjes als groepen te vinden in het nog resterende en sterk versnipperde nevelwoud dat verspreid ligt in een landschap dat voornamelijk bestaat uit kleinschalige landbouw. De resterende bosfragmenten verschillen in grootte, mate van isolatie en van verstoring, waardoor de Taita Hills een interessant gebied is om de invloed van antropogene verstoring op de soort en zijn broedgedrag te onderzoeken. Deze thesis maakt gebruik van data van ongeveer tien jaar (verzameld sinds 2007), waarvan data van vier jaar die specifiek voor deze thesis verzameld werd (2012 tot 2016). Doorheen deze periode werd er gedurende elk broedseizoen intensief en systematisch naar nesten gezocht, die daarna verder werden opgevolgd om te bepalen of het nest succesvol was of niet. Van elk nest werd er getracht zo veel mogelijk informatie te verzamelen door middel van observaties, vangsten, video opnames en radio-telemetrie. Zo konden we co¨operatief broedgedrag en andere facetten gelinkt aan reproductie in kaart

Page xv brengen. Het merendeel van de broedparen (64%) broedde in de Taita Hills in groep, samen met andere niet-broedende individuen of ’subordinates’. Subordinates bestonden in 73% van alle groepen exclusief uit jongen uit een vorig broedseizoen en zijn dus verwant aan het broedpaar, in 22% van de groepen exclusief uit niet-verwante individuen van onbekende origine, en in 5% van de groepen uit een mix van beide. De helft van alle verwante individuen hielp het broedpaar bij het voeden van de jongen in het nest, en worden daarom helpers genoemd. Doordat de bossen verschillen in grootte en kwaliteit verwachtte ik verschillen te vinden in co¨operatief broedgedrag. Echter, in tegenstelling tot mijn verwachting was de samenstelling van deze groepen, en het voorkomen van co¨operatief broedgedrag tussen de verschillende bosfragmenten onderling gelijk (Hoofdstuk3 ). Enkel de verhouding mannelijk/vrouwelijk subordinates in de groep was verschillend, wat zowel aan een verschil in sterfte tussen mannetjes en vrouwtjes, als aan een verschil in verspreidingsstrategie¨en kan liggen. Deze resultaten wijzen erop dat het co¨operatief broedgedrag van de soort niet is verstoord. Nest predatie, waarbij de eitjes of jongen uit het nest worden geroofd, is bij vogels zoals de Placid Greenbul de voornaamste reden waarom een broedpoging faalt. Alhoewel vogels dit risico kunnen verkleinen door hun nest op een veilige plaats te bouwen, kan dit in een sterk verstoord gebied niet altijd meer mogelijk zijn. In (Hoofdstuk2 ) vergeleek ik nestplaatskeuze tussen twee fragmenten die sterk verschillen in mate van verstoring. Vogels kozen in beide fragmenten verschillende plaatsen voor hun nest, alhoewel deze keuze niet bepalend was voor een succesvolle broedpoging. Het verschil in nestplaatskeuze tussen bosfragmenten is op twee manieren te verklaren. Enerzijds kunnen vogels actief kiezen om hun nest op een andere plaats te bouwen, en zich zo beter aan te passen aan de plaatselijke omstandigheden in het bos. Anderzijds kan dit een gevolg zijn van het verschil in de structuur en samenstelling van het bos, waardoor sommige plaatsen gewoonweg niet beschikbaar zijn. Om dit vraagstuk op te lossen, en om te achterhalen waarom er geen verband is tussen nestplaats keuze en broedsucces, is er meer kennis nodig over de roofdieren in de bossen. In een volgend hoofdstuk (Hoofdstuk4 ) bestudeerde ik de investering van het broedend vrouwtje in haar eitjes en jongen en bekeek in welke mate deze afhankelijk is van de aanwezigheid van subordinates en helpers, en tevens of deze verschilt tussen bosfragmenten. In aanwezigheid van subordinates legden vrouwtjes grotere eitjes in een groot en sterk verstoorde bosfragment, maar kleinere eitjes in alle andere bosfragmenten (n groot en minder verstoord en verschillende sterk verstoorde kleine fragmenten). Dit was onafhankelijk van het aantal of de aanwezigheid van helpers. Vrouwtjes voedden de jongen dan weer altijd minder bij een groter aantal helpers in alle bosfragmenten. Vrouwtjes leken dus hun investering te vari¨erenafhankelijk van de graad van verstoring en/of de verschillende lokale condities in de bosfragmenten en afhankelijk van de samenstelling van

Page xvi de groep waarin ze broedden. Zowel meer als minder investeren kan zinvol zijn aangezien het eerste de overlevingskans van de jongen kan verhogen, terwijl het tweede juist gepaard gaat met een betere levensverwachting voor het vrouwtje. In dit hoofdstuk toonde ik ook aan dat grotere groepen een beter broedsucces hebben in alle bosfragmenten, onafhankelijk van aantal helpers. Dit, samen met de flexibele strategie die vrouwtjes kunnen aannemen, wijst erop dat co¨operatief broeden voordelig is voor vrouwtjes. Het is echter nog niet duidelijk wat de lange-termijn voordelen zijn van zon flexibele strategie, en of die dan ook de negatieve gevolgen van habitatverstoring kunnen compenseren. In een laatste hoofdstuk bestudeerde ik de periode nadat de jongen het nest verlaten hebben (Hoofdstuk5 ). Jongen zijn tijdens deze periode nog afhankelijk van hun ouders voor voedsel en zeer vatbaar voor roofdieren. Alhoewel deze periode bij co¨operatief broedende vogels vaak lang duurt, en zeer belangrijk is voor de ontwikkeling van de jongen, is er nog weinig gekend omwille van de moeilijkheid om jongen te lokaliseren. Hier maakte ik gebruik van kleine zendertjes om jongen terug te vinden en vond dat de overlevingskans van jongen hoger is in groep. Dit illustreert dat er ook na het uitvliegen nog voordelen zijn van co¨operatief broeden, en benadrukt het belang om deze periode niet te overzien wanneer de kosten en baten voor deze vogels worden bepaald. Op basis van de bevindingen in deze thesis lijkt het co¨operatief broedgedrag van de soort stand te houden in dit sterk verstoorde gebied. Daarenboven lijkt co¨operatief broeden voordelig te zijn voor het broedpaar omdat grotere groepen zowel het broedsucces, als de overlevingskans van jongen na het uitvliegen verhogen. Alhoewel we dit pas kunnen uitsluiten wanneer we ook naar een populatie uit een intact gebied kijken, wijzen verschillende bevindingen, zoals de flexibele strategie van de vrouwtjes en de grote groep niet-broedende subordinates erop dat de Placid Greenbul beter bestand kan zijn tegen antropogene habitat verstoring dan niet-co¨operatief broedende soorten.

Page xvii

1

CHAPTER 1

General introduction

To meet the ever-increasing need of the Earth’s human population, humans gradually transformed natural ecosystems to such extent that their signature can be traced back in almost every natural phenomenon or process on the planet (Waters et al., 2016; Boivin et al., 2016). The currently arising Anthropocene - an epoch caused and influenced by humans - represents the universal impact from direct (e.g., through landscape modifications) or indirect (e.g., through climate change) human activities on our natural world (Venter et al., 2016; Sanderson et al., 2002; Parmesan & Yohe, 2002; Thomas et al., 2004). The Anthropocene is characterized by the dramatic and human-caused decline of hundreds of species, some of which are now at the brink of extinction (e.g., Spoon-billed Sandpiper Calidris pygmaea), while others are already gone for good (e.g., Great Auk Pinguinus impennis; Pimm & Raven, 2000). Though many different processes can cause the decline of these species, the exact causes are not always well understood, may interact and are often species-specific (Maxwell et al., 2016). Some species are primarily threatened by the introduction of exotic species, as is often the case for species endemic to islands (e.g. the now extinct Stephens Island Wren Traversia lyalli and Dodo Raphus cucullatus). Other species mainly suffer from hunting and poaching (e.g. the Black Rhino Diceros bicornis). A global threat that all species face is landscape modification, which results from destruction, degradation and disturbance of native, pristine habitat and is collectively referred to as anthropogenic habitat change (Pimm & Raven, 2000; Barlow et al., 2016; Laurance et al., 2009; Maxwell et al., 2016). How species respond to, and cope with, these changes depend on the characteristics of the modification process and species-specific life-history traits (Keinath et al., 2017; Bregman et al., 2014; Sekerciolu, 2011). Birds have been shown to vary in their sensitivity to anthropogenic habitat change in relation to their body size, life span, and other species-specific traits, which are

Page 1 General introduction

collectively referred to as life-history traits (Sodhi et al., 2004; Bregman et al., 2016; Keinath et al., 2017; Newbold et al., 2013; Khimoun et al., 2016). The specific combination of these 1 traits characterizes the species’ life-history strategy and is known to vary considerably across regions. Birds from temperate latitudes lay large clutches and fledge many offspring each nesting attempt, whereas birds from tropical regions lay small clutches that suffer great predation risks. Tropical species, however, may compensate for such low reproductive output by a prolonged breeding season that allows multiple nesting attempts, greater post-fledging care that leads to higher first year survival and greater longevity that allows several reproductive opportunities (Roper, 2005; Wikelski et al., 2003; Johnston et al., 1997; Stutchbury & Morton, 2001). Because many tropical species evolved these strategies in combination with a strong dependency on a particular diet or habitat-niche (ecological specialization, Futuyma & Moreno, 1988), they are believed to be more sensitive to anthropogenic habitat change (Bregman et al., 2014; Keinath et al., 2017). Understory insectivores, for instance, are a group of species which forage on a very limited suite of invertebrate resources and depend on a narrow range of environmental conditions found only in the understory of dense, dark, wet, tropical rainforest (Laurance et al., 2012). They may show low tolerance to the warmer, brighter and less humid conditions that often characterize edges and smaller, disturbed forest fragments (Stratford & Robinson, 2005; Bregman et al., 2014). Additionally, they may be reluctant to cross hostile gaps of non-forested habitat (Sodhi et al., 2004; Lees & Peres, 2009), or may even be physically incapable of flying over distances larger than a few hundred meters (Moore et al., 2008). Species may also vary in their sensitivity to anthropogenic habitat change due to their mating and breeding system. Social species with cooperative breeding strategies, for instance, differ from pair-breeding species in the number of individuals among which parental duties are shared. In the majority of these species, offspring from previous broods delay dispersal and remain on the natal territory, which they may take over, or use as a safe harbor from which to explore adjacent territories (Koenig et al., 1992; Koenig & Dickinson, 2016; Walters et al., 2004). Such delayed and short-distance dispersal results in clusters of related individuals, which causes fine-scale genetic structuring of cooperative breeding populations (Leedale et al., 2018). To avoid or reduce inbreeding from mating with a relative, sex-biased dispersal evolved in many cooperative breeding species and female offspring typically disperse sooner and farther (Koenig & Haydock, 2004; Rollins et al., 2012). However, severe habitat fragmentation may hinder dispersal to other fragments, whereas habitat degradation may alter the benefits of delayed dispersal for offspring. Both features (e.g. short-distance and sex-biased dispersal) influence population dynamics in avian populations and characterize the sensitivity of cooperative breeders to anthropogenic habitat change (Walters et al., 2004). In the following sections, I will briefly introduce the processes associated with anthropogenic habitat change and highlight the mechanisms that may affect individuals

Page 2 General introduction and populations faced by habitat change. Next, I will discuss the ecology, evolution and behavior of cooperative breeding and end with elaborating on the influence of anthropogenic habitat change on cooperative breeders, the study of which constitutes the main aim of 1 this thesis.

1.1 Anthropogenic Habitat Change

Habitat loss dramatically reduces the availability to species which depend on that specific habitat. This is one of the principle components of anthropogenic habitat change as this directly limits the number of individuals that can survive and reproduce, causing the population to decline (Gaston et al., 2003; Brooks et al., 2002). Alongside the process of habitat loss, remaining habitat may become fragmented and confined to smaller fragments isolated from each other by altered habitat or matrix - a process referred to as habitat fragmentation (Saunders et al., 1991; Lindenmayer & Fischer, 2007). Although populations of plants and may persist in these smaller, remnant fragments, they become more vulnerable to environmental and demographic stochastic processes, and suffer loss of genetic variability (Lande, 1988, 1993; Spielman et al., 2004). Due to their small size, there is a higher probability that any perturbation disrupts the demographic structure of a population, such as the availability of unrelated mates (Lande, 1993; Banks et al., 2007; Stacey, 1992). Incestuous mating may further aggravate the loss of genetic variability, potentially resulting in reduced reproduction and survival (inbreeding depression; Spielman et al., 2004; Frankham, 2005). The impact of these processes is inversely related to population size and may aggravate over time (Ewers & Didham, 2006; Haddad et al., 2015), but can be dampened by immigrants : unrelated individuals dispersing between fragments (Brown & Kodric-Brown, 1977). Yet, the likelihood that an individual disperses between fragments depends on the connectivity between fragments, which in turn depends on the level of isolation and hostility of the matrix (Fischer & Lindenmayer, 2007; Prugh et al., 2008; Watling et al., 2011; Baguette et al., 2013). Intrinsically associated with the process of habitat fragmentation is a reduction in core-to-edge ratio of remnant fragments and the creation of more edge habitat. Edge effects are caused by changes in biotic and abiotic conditions at the intersection of two habitat types, and reported as one of the most important patterns structuring ecological communities (Pfeifer et al., 2017; Woodroffe & Ginsberg, 1998; Murcia, 1995). A recent meta-analysis demonstrated that more than 80% of the studied species reacted either positively or negatively to edge habitat, whereas only a minority showed no response (Pfeifer et al., 2017). Although edge effects depend upon the structure of the edge habitat and decline with distance to the edge (Didham & Lawton, 1999; Watson et al., 2004), this meta-analysis highlights the universal impact of edges on species. Additionally, as smaller, remnant forests often become more accessible and more prone to human disturbance, for

Page 3 General introduction

instance through selective logging (Potapov et al., 2017), this further induces changes in forest structure and species composition (Burivalova et al., 2014). By now it has become 1 clear that population persistence in a fragmented landscape may primarily be influenced by edge effects per se rather than by the effects of small population size (Banks-Leite et al., 2010; Ewers & Didham, 2006; Pfeifer et al., 2017; Barlow et al., 2016). A major process affected by anthropogenic habitat change is the rate of nest predation, an important driver of reproductive success in birds (Martin & Roper, 1988). This is particularly affecting open-cup nesting , where up to 80% of nests may fail due to predation (Ricklefs, 1969; Newmark & Stanley, 2011). How habitat change may influence nest predation rates depends on the spatial scale addressed. At a landscape scale, increased fragmentation and disturbance is often associated with higher nest predation rates (Robinson et al., 1995; Tewksbury et al., 2006), while distance to the habitat edge may, depending on the context and species, lead to an increase or decrease in nest predation rates within the remnant habitat fragment (Lahti, 2001; Vetter et al., 2013). Ultimately, changes in the environmental characteristics of a nesting site, such as level of nest concealment, are known to influence nest predation rates at an even smaller scale (Borgmann & Conway, 2015; Lamanna et al., 2015). Such changes in nest predation rates are caused by a number of interactively acting mechanisms. For instance, predator communities may change in a fragmented landscape, as novel predator species can now venture deeper into remnant fragments (mesopredator release; Crooks & Soul´e, 1999). Likewise, sites traditionally used for nesting may become unavailable in degraded habitat, or be less protective against these new predators (ecological trap; Hale & Swearer, 2016). Hence, anthropogenic habitat change may affect population dynamics through changes in nest predation rates (Chalfoun et al., 2002).

1.2 Cooperative Breeding

Cooperative breeding, where more than two individuals jointly rear a brood and thus likely care for offspring that are not their own, poses an evolutionary paradox. Why would individuals forego the opportunity to breed themselves and perform the altruistic task of caring for the offspring of others instead (Darwin, 1859)? This is, however, not a rare phenomenon among birds, since up to 13% of the species are reported to do so - a percentage which may only increase when future research fills the gaps in our natural history knowledge (Griesser et al., 2017). The majority of cooperative species, like the Carrion Crow ( corone; Baglione et al., 2003) and the Florida Scrub- ( coerulescens; Woolfenden & Fitzpatrick, 1990), form groups when the offspring from one brood delay their dispersal, remain on the natal territory and help raise their younger relatives during subsequent breeding attempts. As such groups consist of family members, helpers gain indirect fitness benefits when caring for their non-descendent kin. Hence,

Page 4 General introduction kin selection provides a powerful explanation for cooperative breeding among related individuals (Hatchwell, 2009; Clutton-Brock, 2002; Baglione et al., 2003), which is further supported by the fact that care is only directed towards kin in many species (Griffin & 1 West, 2003; Green et al., 2016). However, cooperative breeding among unrelated individuals is more common than previously recognized and may occur in a variety of occasions (Riehl, 2013). For instance, solitary, unrelated individuals may join a family group, which is the case in the Pied Kingfisher (Ceryle rudis; Reyer, 1984), Iberian ( cyanus; Valencia et al., 2003) and Rifleman (Acanthisitta chloris; Preston et al., 2013). However, cooperation may also occur within complex alliances formed among relatives and non-relatives, as is the case in the Greater Ani (Crotophaga major; Riehl, 2011), the Taiwan Yuhinas (Yuhina brunneiceps; Yuan et al., 2004) and the Dunnock (Prunella modularis; Davies, 1985). In these species, non-breeding helpers are unable to gain indirect fitness benefits from joining a group, and kin selection alone falls short in providing a conclusive explanation for the evolution of cooperative breeding. However, individuals in social groups can gain several direct fitness benefits from cooperation, which provides the dominant explanation for this behavior among unrelated individuals (Shen et al., 2017; Kingma et al., 2014).

1.2.1 Costs and benefits of cooperation

Non-breeding individuals may help the breeding pair with a variety of tasks including nest building, nest defense, incubation and nestling provisioning. As all these tasks involve costly behavior such as the increased risk of mortality during nest defense, or increased energy expenditure from foraging and provisioning nestlings, helping is expected to accrue sufficient benefits that outweigh these costs (Heinsohn & Legge, 1999). Helpers in family groups may obtain such benefits indirectly by accumulating inclusive fitness, that is, from indirectly increasing the number of copies of their genes in the population (Hamilton, 1963; Oli, 2003; McGraw & Caswell, 1996). This can occur through an increased production and survival probability of the offspring, increased survival probability of the breeding pair, or a combination of both (Hatchwell et al., 2014). Using long-term data to quantify the life-time fitness gains of helpers in the Long-tailed Tit (Aegithalos caudatus), confirmed the idea that helping increases inclusive fitness through increased survival and recruitment of offspring (Hatchwell et al., 2004; MacColl & Hatchwell, 2004). Alternatively, when relatedness among group members is low, or even absent, helpers may gain fitness directly through increased reproductive success from access to mates or breeding position, acquisition of skills and better survival probability in groups (Cockburn, 1998; Kokko et al., 2001; Kingma et al., 2014; Duval, 2013; Langen, 1996). The relative importance of both indirect (e.g., higher inclusive fitness) and direct (e.g., access to mates, skill acquisition and better survival prospects) benefits for the

Page 5 General introduction

evolution of helping behavior among both related and unrelated group members, is still debated (Browning et al., 2012; Wright et al., 2010; Clutton-Brock, 2009; Kingma et al., 1 2011). Studies however increasingly highlight the importance of direct fitness benefits obtained from living in groups, even among species where groups are formed by delayed dispersal and consist of family members (Shen et al., 2017; Kingma et al., 2014). The Seychelles warbler (Acrocephalus sechellensis), for instance, is a species where females delay dispersal and help the breeding pair during subsequent breeding seasons, which suggests that groups consist of family members. Yet, in this species, the direct benefits of helping outweigh the indirect benefits for two reasons : (i) female helpers frequently gain parentage in the nests they help and (ii) high levels of extra-group paternity (i.e., offspring sired by individuals from outside the apparent family group) result in low levels of relatedness among helpers and non-descendent kin (Richardson et al., 2002). Helpers are also expected to have a positive effect on the breeders’ fitness (Rodrigues & Kokko, 2016; Kokko & Ekman, 2002). A straightforward effect would be the increase in reproductive success of a breeding attempt, for instance by producing more and/or heavier fledglings (Brouwer et al., 2012; Hatchwell, 1999; Emlen & Wrege, 1991). Yet, empirical studies often fail to detect such relationship, which may raise the question if, and through what mechanisms, helping behavior might benefit breeders (Magrath & Yezerinac, 1997; Legge, 2000b; Eguchi et al., 2002)? One likely possibility stems from the prolonged period of post-fledging care and parental dependency frequently observed in cooperative breeders (Langen, 2000). Because helpers may further contribute to offspring care after fledging, they may relieve breeders from post-fledging parental duties, as well as influence juvenile post-fledging survival (Rowley & Russell, 1990). As a consequence, breeders with helpers may be able to rear more clutches within a season and produce more offspring than breeders without help, and their offspring may have higher probabilities of surviving and reproducing in the future (Hatchwell et al., 2004). Another possibility through which helpers may affect breeders, is by allowing breeders to vary their brood investment. The effect of helpers on reproductive success may as such be masked by these different investment strategies. As birds are faced by a trade-off between investing in reproduction or survival, they are predicted to adjust their reproductive investment according to the current, and the expected future environmental and social conditions (Cunningham & Russell, 2000; Horvathova et al., 2012; Lima, 2009; Fontaine & Martin, 2006). For instance, in tropical and south-temperate regions where breeding seasons are extended and birds typically have long-life spans, individuals may have many future breeding opportunities. In such conditions, breeders may favor survival over current reproduction by reducing their reproductive investment (Hatchwell, 1999). While only breeding females may do so through laying smaller clutches or less voluminous eggs, both parents may lower provisioning effort of nestlings. As helpers may compensate for both reduction in egg and nestling investment, such parental investment strategy can

Page 6 General introduction be expected to be similar to the reproductive success of pair-breeders without helpers (Hatchwell, 1999). 1

1.2.2 Ecological drivers of cooperation

The majority of cooperatively breeding species do not require helpers to rear offspring, and are still able to breed successfully as a pair. Such facultative cooperation, as opposed to obligate cooperation, suggests that cooperative breeding is shaped by a number of ecological factors, such as the environmental and social conditions experienced throughout the year (Shen et al., 2017; Griesser et al., 2017). One process that nicely demonstrates the interaction among these factors is delayed dispersal, which is considered a pre-requisite for cooperative breeding (Griesser et al., 2017). Consider a recent fledgling faced with the choice of staying on the natal territory, or dispersing away from it. The bird may be required to stay when it fails to find a mate or a territory, for instance when the habitat is saturated and no empty territories are available (Emlen, 1982). Yet, it may also deliberately chose to stay if it can readily access resources within the safety of the natal territory (Stacey & Ligon, 1987, 1991; Baglione et al., 2005; Komdeur, 1992). However, that same bird may choose to leave when mates and empty territories are readily available, when the quality of its natal territory is low and/or holds few parentally defended resources, or when it needs to compete for those resources with several of its siblings. These two alternative routes demonstrate that the decision of an individual to delay dispersal, and the occurrence of cooperative breeding, will be determined by both social, as well as environmental conditions (Hatchwell & Komdeur, 2000; MacColl & Hatchwell, 2002; Kingma et al., 2016b; Baglione et al., 2002; Nelson-Flower et al., 2018). The idea that ecological factors drive delayed dispersal and cooperative breeding behavior is supported by several experimental, as well as a number of comparative, studies (e.g. Komdeur, 1992; Baglione et al., 2002; Rubenstein & Lovette, 2007. For instance, Carrion Crow nestlings from a pair-breeding colony in northern Italy which were translocated to a cooperative breeding colony in northern Spain delayed dispersal just as their foster colony did (Baglione et al., 2002). In contrast to the Italian colony, the ecological conditions in Spain (e.g., predictability and defensibility of trophic resources) facilitated year-round territoriality (Baglione et al., 2005). Hence, birds from groups in Spain could benefit from group-defended resources, which was not possible in the Italian colony where territoriality was more transient and not year-round (Baglione et al., 2005). Likewise, in a newly founded population of Seychelles warblers, the frequency of delayed dispersal and helping behaviour was positively related to population size, as gradual population growth resulted in saturation of high-quality territories and birds obtained higher fitness by staying and helping than by dispersing and breeding on inferior territories (Komdeur, 1992; Komdeur et al., 1995).

Page 7 General introduction

1.3 Cooperative breeding under anthropogenic habi- 1 tat change Anthropogenic habitat change can alter the ecological factors that drive cooperation. Habitat loss, fragmentation and degradation can affect the availability, accessibility and quality of breeding habitat, which are in turn expected to strongly influence the costs, benefits and constraints of independent breeding and delayed dispersal. For instance, the cost to cross a hostile matrix and disperse between isolated habitat fragments may impose a constraint on dispersal and independent breeding, promoting delayed dispersal and helping behavior. This idea was supported by a comparative study of two Long-tailed Tit populations from a continuous and a fragmented habitat, whereby individuals from the latter population were more likely to delay dispersal and help, resulting in a higher incidence of cooperatively breeding pairs (Russell, 2001). On the other hand, the benefits of delayed dispersal may diminish when the quality of the natal territory is lower. Western Bluebird (Sialia mexicana) offspring were more likely to disperse when food resources where experimentally lowered (i.e., creating lower quality territories; Dickinson & McGowan, 2005), whereas the opposite pattern was observed in Carrion Crows when food resources where experimentally increased (i.e., creating higher quality territories; Baglione et al., 2006). In this regard, an important question can be raised as to whether and how cooperative breeders might be more sensitive, or alternatively, more buffered to the effects of anthropogenic habitat change? From an individual perspective, cooperative breeding likely offers the opportunity to gain fitness from a suite of alternative possibilities such as delaying and helping kin, joining an unrelated group, or breeding independently. Although helping may be viewed as a best-of-a-bad job when compared to independent breeding, it may be a more valuable alternative over floating in search for a mate (Ekman, 2006; Cockburn, 1998; Marzluff & Balda, 1990). For instance, compared to group-living helpers, floaters in a population of cooperatively breeding Pied Babblers (Turdoides bicolor) lost weight relative to the number of days floating due to increased vigilance behavior and less foraging time (Ridley et al., 2008). From a population perspective, the excess of individuals in cooperatively breeding species may serve as a pool of replacement breeders and may dampen fluctuations in the size of the breeding population caused by demographic and environmental stochasticity (Walters et al., 2004). However, the predominantly short- distance, sex-biased dispersal strategy often observed in cooperative breeders (’Stay and Foray’ sensu Brown, 1987; Zack, 1990) may aggravate the disruption of demographic processes caused by habitat fragmentation. The impact of anthropogenic habitat change on cooperatively breeding populations will largely depend on the spatial configuration of the fragments and the suitability of the matrix, and is expected to be more severe in small and isolated populations. Immigrants may be unable to reach such isolated fragments,

Page 8 General introduction which may restrain breeding vacancies to be filled in. In contrast, the impact of habitat change might be less severe in larger remnant fragments where it is more likely that breeding vacancies can be filled by dispersing individuals. Cooperative breeding species 1 can hence be predicted to have lower risk of extinction than pair-breeding species in large, isolated habitats, but higher risks in small, isolated fragments (Walters et al., 2004).

1.4 The Placid Greenbul in the Taita Hills

The research for this thesis was carried out on a cooperatively breeding Placid Greenbul (Phyllastrephus placidus) population in the remnant cloud-forests of the Taita Hills in south-east Kenya (30°25’S, 38°20’E).

1.4.1 The Taita Hills

The Taita Hills represent a group of verdant, isolated mountain blocks (max altitude 2200 m.a.s.l) which rise abruptly from the semi-deciduous, lowland Tsavo plains (app. 700 m.a.s.l). Three mountain massifs are distinguishable : Dabida, the main and largest massif comprising a number of distinctive ridges and valleys with several remnant cloud-forest fragments; Mbololo, adjacent to Dabida and characterized by one large forested ridge; and Sagala, a satellite massif situated at least 10km southeast of Dabida and with hardly any remnant forest left. Due to their topography, they are isolated from similar montane habitat in every direction for at least 50km (Mount Kasigau being the closest). South-east trade winds bring in cloud precipitation from the Indian Ocean, which accumulate over the rugged hilltops. This creates almost permanent humid conditions which support cloud-forest formation (Thijs, 2015; Pellikka et al., 2009). Annual rainfall averages 1200 mm and is mostly concentrated in two distinct seasons: the long rains, from March to June; and the short rains, from October to December (Pellikka et al., 2009). These hills denote the northernmost part of the Eastern Arc Mountains. This chain of isolated mountains stretches from south-east Kenya through south-central Tanzania and features a common geological history (Fig. 1.1; Newmark, 2002). They are identified as a Biodiversity Hotspot due to their high degree of endemism and both rapid and severe loss of forest cover. Currently, it is estimated that only 30% of the originally forested area remains (Myers et al., 2000; Newmark, 2002). The Taita Hill forests in particular rank among the most severely impacted and threatened forests within the Eastern Arc mountain chain. It is estimated that over 90% of the original indigenous forest cover is currently lost to agriculture, while selective logging for firewood and charcoal production have additionally impacted the interior of the remaining fragments (Pellikka et al., 2009; Lovett & Wasser, 1993; Newmark, 2002). Due to the favorable climatic and edaphic conditions for agriculture, the Taita Hills have a long history of human settlement. Based on pottery

Page 9 General introduction

and other archeological remains, humans have resided in the area for at least 2000 years - perhaps already clearing forested land at that time. Forest loss and fragmentation was 1 first noted by early explorers during pre-colonial times (Pellikka et al., 2009), whereas a recent, detailed comparison between aerial photographs from 1955 and 2004 indicated that forest loss was also severe and rapid during the last decades. During this timeframe, over 50% of the indigenous forest cover was lost to agriculture and exotic plantations (Pellikka et al., 2009). Currently, most of the remaining forest fragments are protected although they continue to be under treat in this densely populated landscape (>500 people/km, Pellikka et al., 2009). The remnant cloud-forest fragments are separated by a mosaic of small-scale subsistence agricultural fields, isolated and exotic plantations (Fig 1.2). Cloud-forest cover totals 500ha and is confined to three large (>80 ha; Mbololo, Ngangao and Chawia), and a dozen small, isolated, remnant fragments (<15 ha; e.g. Fururu, Yale, Ndiwenyi, Macha, Mwachora, Msidunyi, Susu, Vuria), which all vary in level of degradation (see Table 1.1 for details). All fragments are isolated from each other by gaps spanning at least 500m of non-indigenous vegetation in all directions. The indigenous forest cover of all fragments, except Susu and Yale, is continuous and uninterrupted with sharply defined boundaries. Susu and Yale forests resemble a patchwork of several closely separate (<250m gap), indigenous forest patches. Forest degradation is characterized by shifts in species communities and changes in vegetation structure (Thijs, 2015). Degraded forests feature a more open canopy and a denser understory. They are more dominated by trees typically associated with forest gaps and margins, and suffer more from frequent cattle grazing and fire wood collection (Aerts et al., 2011; Thijs, 2015). In contrast, less degraded forests harbor more late-successional species and feature a more closed canopy, but a more open understory.

1.4.2 The Placid Greenbul

The Placid Greenbul (formerly considered a subspecies of the Cabanis’s Greenbul P. cabanisi) is a medium-sized understory insectivore that inhabits East Africa’s moist forests (Bennun et al., 1996; Fry & Keith, 2000). Small, noisy flocks are a common sight in the forest understory, where individuals generally forage by gleaning invertebrates from the bark and leaves of shrubs and climbers, and occasionally sieve through leaf litter or near Dorylus ant swarms on the forest floor (Fry & Keith, 2000). The species is sexually monomorphic, though males tend to be larger. Individuals are long-lived and highly sedentary, occupying the same area for multiple consecutive breeding seasons. Based on an extensive capture-mark recapture dataset (ongoing since 1996), the oldest individual recorded in the Taita Hills to date is at least 19 years old and adults are usually retrapped in the same neighborhood (average distance = 152m; range = 0 - 2760m, n = 85). The

Page 10 General introduction

1

Figure 1.1: The Taita Hills in South-east Kenya are the northern most part of the Eastern Arc Mountain Chain (dark grey, indigenous cloud-forest in dark green). The global range of the Placid Greenbul (Phyllastrephus placidus) and the previously considered conspecific sister species Cabanis’s Greenbul (P. cabanisi) are mapped by red and grey overlay, respectively. Lower right inset portrays the location of the three mountain massifs comprising the Taita Hills. Black line marks the area above 1200m a.s.l. Range data reproduced from the Xeno-Canto foundation - www.xeno-canto.org.

Page 11 General introduction

1

Figure 1.2: Location of the forest fragments on the main massif (Dabida) of the Taita Hills. Grey lines indicate the 15m interval height contours, black line marks the area above 1200m a.s.l. Both insets display detailed overview of the cluster fragments.

Page 12 General introduction birds are socially monogamous and mated pairs usually remain together during several seasons (longest pair-bond observed in the Taita Hills since 2007 was at least six breeding seasons). In the Taita Hills, the extended breeding season of the Placid Greenbul coincides 1 with the onset of the short rainy season in November and may last up to March. Nests are cup-shaped, built out of leaf-litter, dried grass and moss and positioned in shrubs, climbers and small trees at an average height of 1.3 m (Fry & Keith, 2000). Usually two eggs (mean ± SE : 1.94 ± 0.015 eggs, range 1 - 3, n = 677 clutches) are laid and incubated by the breeding female for about 15-17 days, but nestling provisioning and care are shared between the breeding pair and other individuals for 11-13 days until fledging. The majority of initiated nests (69%) fail due to predation (Spanhove et al., 2014), but pairs or groups generally re-nest after breeding failure, and occasionally after successful breeding. Although a conspicuously social species, cooperative breeding was first observed in the Taita Hills in 2007 and later published in 2012 (Callens, 2012). Up to now, this remains the only locality where the species is known to display this breeding strategy, even though this, and other Phyllastrephus species, were anticipated as cooperative breeders based on their social behavior as early as the seventies (Grimes, 1976).

1.4.3 Effect of habitat change on the Taita Hills’ avian commu- nity, including the Placid Greenbul

Former studies already addressed the impact of forest fragmentation and degradation on the viability and persistence of a number of bird species in the Taita Hills’ forests (e.g. Lens et al., 2002). The Placid Greenbul, among several other forest dependent species, experiences severe levels of environmental stress from habitat degradation. This stressor is noticeable in the degree of body asymmetry of individuals (Fluctuating Asymmetry,Parsons, 1992; Leung et al., 2000), which has been shown to increase with degree of habitat degradation (Lens et al., 1999). At a landscape scale, the population is genetically differentiated into three clusters that more or less correspond with the three largest fragments (e.g. Mbololo, Ngangao and Chawia; Callens et al., 2011). Also, mark-recapture data (Lens et al., 2002) and homing experiments (Aben et al., 2012, 2014) indicated low to moderate rates of mobility and among-fragment dispersal. Whereas this indicates loss in connectivity and reduced between-fragment dispersal over time, levels of genetic admixture have been shown to increase again over the last decade (Husemann et al., 2015). Finally, a comparative study along a range of increasingly fragmented forests within the Eastern Arc Mountains indicated that forest fragmentation increases Placid Greenbul nest predation rates (Newmark & Stanley, 2011). A more detailed nest predation analysis within the Taita Hills revealed that nest predation rates varied between fragments, and depended on the location within the fragments and timing of nest initiation. In general, predation risk is higher in Ngangao than in Chawia, but increases in both fragments with distance to the

Page 13 General introduction

fragment edge and throughout the breeding season (Spanhove et al., 2014). Taken together, several lines of evidence illustrate that the Placid Greenbul 1 population in the Taita Hills is affected by the severe historic habitat fragmentation and degradation. In spite of this, the Taita Hills population appeared to have increased over the past two decades, though the situation varied between individual fragments at a more local scale. While the (sub)population of some fragments has declined, the (sub)population at other fragments has increased, which suggests that the Taita Hills sustains a meta-population network (Husemann et al., 2015). Besides the fact that this population is sustained in the Taitas, this typical and common montane species can be expected to occur at far more intact and pristine sites within its global range (e.g. East Usambara Mountain range Newmark, 2002). We may therefore judge the Taita Hills as being representative of typical, albeit severely degraded habitat for this species.

1.4.4 General methodology

The fieldwork for this thesis was mainly conducted in ten forest fragments of the Dabida section of the Taita Hills (see Fig. 1.2, Table 1.1). These include all fragments in Dabida with known Placid Greenbul breeding activity, as well as those where current or historic mist-net traps or observations of the species indicated the potential to breed. On Dabida, ca. 75% of the species’ population is individually color-banded as a result of ongoing mist-netting since 1996 (Bytebier, 2001). Moreover, the geographical isolation from other mountain blocks and the species’ dependency on relatively accessible forest fragments increases the detection probability and accuracy of survival estimates. Finally, the Taita Hills area offers an interesting opportunity to evaluate variation in life-history traits and cooperative breeding across an anthropogenic degraded landscape due to the suite of fragments that differ in size, degree of isolation and level of degradation. In general, the data for this thesis was collected by monitoring Placid Greenbul breeding activity during four breeding seasons from 2012 until 2016. During the whole of each breeding season, a team of local field assistants systematically searched all ten fragments for greenbul nests. Nest searching began at the start of the breeding season in November, and usually lasted until mid-March, when no new nests were discovered any more. Upon detection, nests were further monitored to record both reproductive and social parameters. To do this, a combination of field procedures such as observation, mist-net trapping, video-recording and radio-telemetry, as well as genetic techniques such as parentage analysis were used. The data collected for this thesis supplements the species’ breeding ecology database that has been continuously collected since 2006. Table 1.1 shows details on monitoring intensity and fragment coverage since 2006. Not all data was collected during all ten years of fieldwork, mainly due to practical constraints and variation in employed field techniques. In addition, not all reproductive and social parameters

Page 14 General introduction could be extracted from each nesting attempt due to difference in timing of detection, or predation. For instance, there is no information on egg volume or clutch size for nests that were detected at a later (nestling) stage, whereas there is no information on cooperative 1 behavior for nests that were predation prior to video recording, with which cooperative behavior was assessed. More elaborate information on methodology, data collection and final sample sizes can be found in the individual chapters.

1.5 Aims and Outline

This thesis aims to advance our understanding of the effect of anthropogenic habitat change on the life-history strategies of a facultative cooperatively breeding bird species. Studies on the effect of anthropogenic habitat change on life-history strategies of cooperative breeders are scarce. Yet, without a good understanding on the impact of habitat change on these birds, it is impossible to predict whether cooperative breeding may buffer populations against these changes, or whether this strategy would make them more sensitive instead. At present, such ecological insight is particularly lacking from the tropical regions. As tropical habitat change occurs at severe and unprecedented rates, it is in these parts of the world that the effect of habitat change on long-term persistence of populations is particularly unclear. Unfortunately we cannot predict the response of tropical species on habitat change based on the knowledge gained from the better studied northern regions because of the contrasting life-history strategies between tropical and temperate species. Moreover, tropical and southern latitudes are disproportionally rich in cooperatively breeding species. Hence, we urgently need to increase our understand of the effects of tropical habitat change on cooperatively breeding species. This thesis deals with various processes and different life-history traits associated with reproduction in a chronological order. In each chapter I evaluate the effect of habitat change, the effect of cooperation and/or the interaction between the two, on these processes and traits. Nest predation is the main cause of nest failure in open-cup nesting birds like the Placid Greenbul. Predation risk may be mitigated by tactically selecting safe sites for nesting, although this strategy may be compromised when habitat degradation alters the habitat makeup. In chapter2 , I compare the environment at sites used for nesting with randomly chosen sites without evidence of nesting between two contrastingly degraded forest fragments. I infer what the influence is of habitat degradation on nest-site selection and subsequently evaluate the reproductive consequences thereof. Cooperative breeding is influenced by variation in habitat quality and/or constraints on independent breeding, because individuals vary their dispersal and reproductive strategy according to these conditions. Anthropogenic habitat change may affect both habitat quality (e.g. due to habitat degradation), as well as increase constraints on independent breeding (e.g.

Page 15 General introduction

resulting from habitat loss or fragmentation) and hence influence individual strategies and cooperative breeding behavior. In chapter3 , I use a comprehensive dataset of individual 1 resighting histories combined with information on individual reproductive strategies to evaluate variation in delayed dispersal, group composition and cooperative breeding across a distinct set of degraded forest fragments to infer whether, and to what extent, cooperative breeding is influenced by anthropogenic habitat change. It is generally acknowledged that breeding pairs attain benefits by breeding cooperatively, such as allowing breeding females to optimize their reproductive investment strategy. Cooperatively breeding females may, for instance, invest more under benign conditions but less when conditions deteriorate, whereas pair-breeding females do not have such flexibility. In turn, this may influence reproductive success of the females. In a next chapter (4), I evaluate maternal investment both before (i.e. egg volume) and after (i.e. provisioning rate) hatching to address whether both pair and cooperatively breeding females vary their investment strategies in response to habitat degradation. Positive (or negative) fitness consequences of cooperation may also be detectable after fledging. While this is particularly relevant for cooperative breeders where offspring typically depend longer on their parents, these affects have rarely been assessed due to the difficulty of following mobile fledglings. In chapter5 , I use a novel, indirect, radio-telemetry approach to overcome this challenge. Based on a detailed capture-mark resighting dataset I then infer early post-fledging survival probabilities and relate these to group size and helping behavior. In a final chapter (6), I integrate and elaborate further on the results of all previous chapters to discuss two broad topics. First, I explore the fitness payoffs of cooperative breeding for both subordinates and breeders, and discuss whether anthropogenic habitat change may act as an ecological driver for cooperative breeding. In a second part, I first discuss the impact of anthropogenic habitat change on this species and how this may affect, or interact with, cooperative breeding. I end this part on a more speculative discussion regarding the capacity of cooperative breeding to buffer this population against anthropogenic habitat change and formulate three general conclusions that can be drawn from this thesis.

Page 16 General introduction

1 Table 1.1: Overview and details of all remnant forest fragments in the Taita Hills known, or expected, to support breeding Placid . Susu and Yale represent a cluster of different patches (see text for details). First nests were detected in 2006 -2007, but systematic searches started in 2007 - 2008. From 2007 - 2008 onwards, nest searching and monitoring activity varied between years, due to logistical constraints or unknown breeding activity at that time. Fragments were only covered moderately during 2010 - 2011 and 2011 - 2012. The species has been observed in Wundanyi, Ronge and Sagala forest fragments, but no systematic or opportunistic nest search activity is ever undertaken. Nest searching in Mbololo is logistically highly challenging and only conducted opportunistically in 2007-2008 and 2015-2016. Placid Greenbul breeding units is the highest number of unique breeding females identified in any given breeding season. In small fragments (e.g. Susu, Ndiwenyi), this likely reflects the actual number of breeding groups or pairs, and hence number of territories held; while in larger fragments this is most likely a conservative minimum estimate. In spite of various multi-disciplinary research programs carried out in the Taita Hills over the past decades, further exploration of the area revealed a new, previously unknown indigenous fragment in 2012 (Msidunyi) and several small indigenous forest pockets adjacent to previously known fragments (Susu, Yale).

Indigenous Level of Nest searching and monitoring c Placid Fragment Abbr. forest degrada- This thesis Greenbul cover tion b breeding (ha) a units d 2013-2014 2006-2007 2014-2015 2015-2016 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2012-2013 Ngangao NG 120.1 Moderate ••••••••• 36 Chawia CH 86.3 Severe ••••••••• 21 Fururu FU 8.1 Severe ••••••••• 3 Ndiwenyi ND <1 Severe • .... •••• 2 Susu SU 9.2 NA/Severe North •• 2 East •••• 2 West ••• 4 Dabida Yale YA 15.7 Severe Main ••• 1 Private ••• 1 Msidunyi MS 9.9 NA/Severe •••• 4 Vuria VU <1 Severe •••• ? Macha MA 2.5 Severe ••• .. •••• ? Mwachora MW 2.3 Severe ••• .. •••• ? Wundanyi WU NA/Severe Number of Placid Greenbul units in Dabida 86

Mbololo MB 220 Least degraded •• Mbololo Ronge RO <1 Severe

Sagala Sagala SA Severe a Fragment sizes extracted from Pellikka et al., 2009; Aerts et al., 2011 and based on GIS calculations using Google Earth satellite imagery(Susu and Msidunyi). b Level of degradation obtained from Aerts et al., 2011; Thijs, 2015; Chege & Bytebier, 2005. NA/Severe = fragments (or sub fragments) that are not assessed with respect to forest structure but assumed to be severely degraded as well. c • = Nests were searched systematically, or . = sporadically d ? = never active nests found, although individuals have been observed in the fragment and old nests may have been found.

Page 17 The Placid Greenbul is a medium-sized, brownish, but nonetheless exciting, bird. It lives in small groups (upper left photo shows mother and 4 months old son), and builds open-cup nests in dense, humid, forest understory. Clutches are only incubated by the dominant breeding female, but care of nestlings shared among others. Upper right photo by Sybryn Maes.

CHAPTER 2 2

Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

Dries Van de Loock1,2,3 Diederik Strubbe1 Koen W. Thijs4 Thomas Van de Peer4 Liesbeth De Neve1 Mwangi Githiru3,5 Erik Matthysen2 Luc Lens1

Modified from: Van de Loock, D., Strubbe, D., Thijs, K.W., Van de Peer, T., De Neve, L., Githiru, M., Matthysen, E. and Lens, L. 2018. Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore. Ibis In Press.

1 Terrestrial Ecology Unit, Ghent University, K. L. Ledeganckstraat 35, 9000 Ghent, Belgium 2 Evolutionary Ecology Group, University of Antwerp, Campus Drie Eiken, Universiteitsplein 1, 2610 Wilrijk, Belgium 3 Department of Zoology, National Museums of Kenya, P.O. Box 40658-00100, Nairobi, Kenya 4 Division of Forest, Nature and Landscape, University of Leuven, Celestijnenlaan 200 E, Box 2411, 3001 Leuven, Belgium 5 Wildlife Works, P.O. Box 310-80300, Voi, Kenya

Page 23 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

2.1 Abstract

Human activities impact upon natural habitats used by birds for breeding and foraging, and lead to changes in the composition and spatial distribution of predator communities, mainly through loss, fragmentation and disturbance of formerly pristine habitat. Yet 2 possible fitness consequences of such changes through impacts on -site selection remain poorly known. Here we study nest-site selection and reproductive success of Placid Greenbuls Phyllastrephus placidus in the Taita Hills, SE Kenya. We show that habitat features associated with nest-site selection by this insectivorous, open-cup nesting bird species vary among forest fragments that are exposed to different levels of habitat disturbance. Such differences in sites selected for breeding result from a plastic response to fragment-specific conditions, or may be driven by fragment-specific variation in the distribution and availability of certain habitat features. Given the overall high nest predation rates in our study area, we expected variation in nest-site selection to correlate with reproductive success and nestling condition, but detected no such relationships. Because predator density and nest predation rates may vary strongly in space and time, a better understanding of spatio-temporal variation in predator communities is needed to assess the possible adaptive value of nest-site selection strategies for reducing the high predation rates that are typical for this, and many other open-cup nesting tropical passerines.

2.2 Introduction

The fitness of an individual strongly depends on its reproductive strategy (Morris, 1989; DeCesare et al., 2014; Losier et al., 2015), and factors affecting nest-site selection are therefore likely to be under selective pressure (Martin, 1998; Chalfoun & Schmidt, 2012). Predation on eggs and nestlings is a major cause of reproductive failure in birds (Wilcove, 1985), particularly in open-cup nesting passerines, where up to 80% of nest-losses can be attributed to predation (Ricklefs, 1969; Martin, 1993; Ib´a˜nez-Alamo´ et al., 2015). Open-cup breeders are hence expected to select safe nest sites that are less likely to be detected by - or accessible to - predators (Martin & Roper, 1988; Eggers et al., 2006; Chalfoun & Martin, 2009; Latif et al., 2012). For example, dense tree foliage may hinder the transfer of auditory, olfactory and visual cues, and hence reduce the likelihood of nest detection (Martin & Roper, 1988; Liebezeit & George, 2002; Borgmann & Conway, 2015). Because birds are known to trade off current reproductive success against their potential for future reproduction (Stearns, 1992; Roff, 2002), factors other than nest predation risk may drive adaptive nest site selection. For instance, incubating birds may increase their survival by selecting nest sites that allow early visual detection of approaching predators (Miller et al., 2007; Maga˜na et al., 2010), or they may require

Page 24 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore a nesting site allowing them to fulfil their energetic needs under harsh climatological conditions (Conway & Martin, 2000; Hart et al., 2016). Furthermore, chick development may be improved by nesting in resource-rich areas that may not necessarily coincide with areas of high nest concealment (Chalfoun & Martin, 2007; Crampton & Sedinger, 2011). Ultimately, avian nest-site selection reflects the interaction between several factors that optimize the trade-off between costs (e.g. nest predation, interspecific competition) and 2 benefits (e.g. access to resources, female survival) of a specific site (Chalfoun & Schmidt, 2012). Natural fluctuations of these factors in space and time (e.g. spatial variation in predation communities or inter-annual variation in food resources) may result in adaptive plasticity in nest-site selection, whereby breeders select different nest sites under different circumstances (Schaefer, 1976; Forstmeier & Weiss, 2004; Janiga & Viˇsovsk´a, 2004). For instance, under experimentally raised predation pressure, Siberian Jays infaustus chose nest sites with greater protective cover (Eggers et al., 2006). However, due to anthropogenic activities such as logging and agriculture, natural landscapes may change dramatically through large-scale loss, fragmentation and disturbance of formerly continuous, pristine habitat (Fischer & Lindenmayer, 2007; Lindenmayer & Fischer, 2007). Such large-scale habitat modification may affect the composition or spatial distribution of predator communities, for example due to changes in core-edge ratios of remnant habitat patches (Saunders et al., 1991; Chalfoun et al., 2002), while selective tree logging, cattle grazing or the invasion of exotic species may cause further small-scale changes in vegetation structure and micro-climate (Lahti, 2001; Chalfoun et al., 2002; Robinson & Sherry, 2012; Vetter et al., 2013). Because such anthropogenic changes in land-use and habitat structure occur much more rapidly than under natural regimes, they may result in non-adaptive nest site selection (Robertson & Hutto, 2006) as breeders may experience unfamiliar habitat (e.g. Bowman & Woolfenden, 2002) or be faced with novel predators (e.g. Crooks & Soul´e, 1999). Due to their high level of habitat specificity, low mobility and confinement to the forest interior, tropical forest understory insectivores are particularly sensitive to these habitat changes (Sekerciolu et al., 2002). Yet studies on nest site selection have been biased to the temperate region (Chalfoun & Schmidt, 2012), with a marked lack of studies from the Afrotropical region (Deikumah et al., 2014; Di Marco et al., 2017). To bridge this knowledge gap, we studied nest-site selection and reproductive success in two populations of a typical understory insectivore, the Placid Greenbul Phyllas- trephus placidus, that are exposed to different levels of past and current habitat disturbance. Both populations suffer high nest predation rates (up to 70% of nests fail due to predation, Spanhove et al., 2014) and nest-site selection may constitute an important mechanism through which individuals reduce predation probability. One population inhabits a moder- ately disturbed forest fragment of ca. 120 ha characterized by late-succession tree species and limited contemporary human disturbance (forest fragment Ngangao, NG). The second

Page 25 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

population is in a heavily disturbed forest fragment of ca. 80 ha that is exposed to ongoing firewood collection and cattle grazing, and characterized by few late-successional trees, but many pioneer species (forest fragment Chawia, CH, Fig. 2.1; Pellikka et al., 2009; Aerts et al., 2011). On a larger, landscape scale, this species prefers indigenous forest over exotic plantations, agricultural land and other landscape types (Aben et al., 2012), and can hence 2 be classified as a forest specialist which depends on the forest interior for breeding and survival (sensu Bennun et al., 1996). However, within the sharply defined boundaries of remnant cloud-forest fragments, little is known about how micro-habitat characteristics influence greenbul breeding success. As a forest specialist sensitive to habitat fragmentation and degradation (Bregman et al., 2014), such information is especially relevant for guiding conservation actions, particularly where improving habitat quality of extant, remnant patches is more realistic than the establishment of new, sufficiently large and natural forests (Aben et al., 2016). Therefore, we first compare nest-site selection within and between populations by comparing environmental variables between nesting sites and randomly chosen sites without evidence of nesting. We focus on habitat variables considered relevant for predator avoidance or resource availability. Second, we use a multi-model inference strategy to assign a probability score to each measured site, reflecting the likelihood that a certain site will be used as a nesting site based on its environmental characteristics (Nest Site Probability Score: NSPS). We then (1) assess whether reproductive success (i.e. hatching and fledging success) is positively related to NSPS to test our expectation that predation avoidance drives nest-site selection, and (2) assess whether nestling condition is positively related to NSPS to test whether resource access, rather than predation pressure, serves as a prime driver of greenbul nest-site selection.

2.3 Methods

2.3.1 Study species and area

The indigenous cloud-forests of the Taita Hills of SE Kenya (30°25’S, 38°20’E) are part of the Eastern Arc biodiversity hotspot and characterized by exceptionally high levels of both endemism and anthropogenic habitat disturbance (Lovett & Wasser, 1993; Myers et al., 2000; Burgess et al., 2007a). These forests have been under pressure since pre-colonial times (Lovett & Wasser, 1993), which has resulted in ca. 95% indigenous forest loss across the region (Pellikka et al., 2009). At present, 13 indigenous forest patches remain, interspersed by small-scale agro-forestry and exotic plantations (Pellikka et al., 2009). These patches all differ in size, level of disturbance and isolation, and nine of them host differently-sized populations of the Placid Greenbul (formerly considered a subspecies of the Cabaniss Greenbul Phyllastrephus cabanisi). In the Taita Hills, this medium-sized insectivore either breeds in pairs or in small cooperative groups consisting of the breeding

Page 26 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore pair and one to five subordinates, of which up to three subordinates may also feed the nestlings (Van de Loock, unpubl. data). As forest specialists (sensu Bennun et al., 1996), they depend on the forest interior where they nest and forage in the understorey on invertebrate prey (Fry & Keith, 2000). The onset of short rains in November dictates the start of the breeding season, which extends until March. Usually two eggs (range:1 - 3) are laid in an open, cup-shaped nest, built in the branch fork of a seedling or climber and 2 resembling trapped leaf debris. Eggs are incubated by the female for about 15 days and nestlings fledge 11 (10-13) days later. Natural predation pressure is high, with up to 70% nest failure (Spanhove et al., 2014), and fewer than 50% of the fledged young surviving until nutritional independence at about 55 days post-fledging (Van de Loock et al., 2017). While the species is currently not considered at risk in its global distribution range (Least Concern, IUCN 2014), structural body size measurements of the remaining Taita Hills populations suggest the species is stress-sensitive to forest disturbance (Lens et al., 1999). In addition, mark-recapture analysis, as well as homing experiments, have revealed low to moderate rates of mobility and among-fragment dispersal (Lens et al., 2002; Aben et al., 2012, 2014).

2.3.2 Survey and sampling design

We mapped Placid Greenbul nests in two populations (forest fragments CH and NG) from 2006 and monitored these intensively during the 2009/10 and 2010/11 breeding seasons. Upon detection, each nest was visited every fourth day to monitor egg or nestling stages while minimizing disturbance, and tarsus length and body mass were measured when nestlings were nine days old. We randomly selected 113 of 264 discovered nests for environmental measurements (CH: 46 nests; NG: 67 nests), using a stratified approach with respect to the position of each nest relative to the fragment edge or interior. Next, we selected 99 random sites that had never contained a nest within a 25m radius since 2006 (CH: 38 sites; NG 61 sites; Fig. 2.1). During 2012, we measured 17 habitat characteristics (Table 2.1) that we considered relevant for predator avoidance or resource availability within a 10m radius of both nest and non-nest sites, by centring a circular plot (± 314m2) around each nest or the nearest tree for non-nesting sites (nest-patch scale sensu Benson et al., 2009; Crampton & Sedinger, 2011). In addition, we identified the vegetation composition in each plot. To do this, we summed the basal area of all mature tree specimens (¿ 5m high) for each individual species within each plot. Next, we conducted a regional (i.e. both fragments combined) Principal Components Analysis on the basal area of a subset of 20 tree species that are forest community indicators within our study area (Thijs, 2015). We retained the first three axes (Eigenvalue > 1), cumulatively explaining over 50% of the variation, and calculated the plot scores on each of these axes. Each axis represents a gradient between contrasting forest communities, for instance between pioneer and

Page 27 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

2

Figure 2.1: Map of the Taita Hills (SE Kenya), indicating the location of the indigenous forest fragments (black), the largest town (Wundanyi) and the 1500m a.s.l altitudinal zone (grey contour line). Nest sites (•) and random, non-nest sites (◦) are indicated for both studied populations from Ngangao forest fragment (NG, moderately disturbed) and Chawia (CH, heavily disturbed).

Page 28 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore late-successional forest (Table 2.1; see appendix and Table A.1). Detailed descriptions and average values (± SD) of each habitat variable are given in Table 2.1. The average values express the availability of a specific habitat variable to the population in the respective fragment. We calculated F-test statistic for each variable in function of fragment identity to detect significant differences in availability. From these 20 habitat variables we selected a set of 14 uncorrelated predictor variables to assess if, and to what extent, nest site 2 selection differed between both populations (Overview in Table 2.1; all Pearsons r < 0.5 in either population; Table A.2& A.3; Zuur et al., 2010).

2.3.3 Nest-site Selection

Prior to statistical analysis, all 14 predictors were standardized to a mean of zero and standard deviation of one to make model coefficients directly comparable. Because testing for differences in nest site selection strategies between the two populations would require modelling of 14 two-way interactions with variable ’fragment’, we instead opted to create separate statistical models for each fragment. We then compared parameter estimates for each of our predictors in each fragment. We built a full multiple logistic regression model with binomial distribution and logit-link (GLM; nest: 1 vs. non-nest: 0), including all 14 scaled predictors, and adopted an AIC-based multi-model inference approach to quantify the relative importance of each specific habitat feature for nest-site selection. We weighted and ranked all possible models based on a small-sample information criterion (AICc; Burnham & Anderson, 2002), and derived ’full’ average parameter estimates and 95% confidence intervals (’zero-method’ sensu Grueber et al., 2011) based on a reduced set of models with good empirical support (∆AICc 6 4; Burnham & Anderson, 2002). Models were run using the MuMIn package in R (Barton, 2016). We used Nagelkerkes pseudo R2 to quantify model explanatory power (Nagelkerke, 1991).

2.3.4 Reproductive correlates of nest-site selection

We used the averaged models obtained earlier to predict site-scores in each fragment. This score is the likelihood that, given its habitat characteristics, a site would be selected for nesting and is henceforth referred to as the Nest Site Probability Score (NSPS). We then quantified relationships between NSPS and reproductive success for 60 nests with known breeding outcome (CH: 19 nests; NG: 41 nests). Reproductive success was both measured as ’hatching success’ and ’fledging success’, which is the probability that at least one egg of a clutch hatches, or fledges, respectively. We do this because the effects of nest site selection may vary between stages of the breeding cycle depending on parental behaviour. For example, parents largely remain cryptic during incubation, but visit the nest frequently during nestling provisioning (Roper & Goldstein, 1997; Martin et al., 2000). We used GLMs with logit link functions and binomial error distributions to model both hatching

Page 29 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

Table 2.1: Habitat variables measured at (113) nest-sites and at (99) randomly selected non-nest sites of two Placid Greenbul populations from a heavily disturbed (CH) and moderately disturbed (NG) forest fragment. Out of 20 habitat variables, 14 uncorrelated variables are retained as nest-site predictors (marked by a filled circle; •). Average values ± SD over all nest and non-nest 2 sites combined indicate the availability and distribution of each habitat feature. Differences (p 6 0.05; F-test) in average between both forest fragments are noted in bold. When applicable, abbreviated names used in Fig. 1 are noted in brackets.

Variable Unit Description Values ± SD CH NG

Variables quantified at the center of the plot 1 Distance edge m Distance from plot centre to the nearest forest edge 123.35 ± 74.77 136.74 ± 73.92 (measured using ArcGIS 9.2, ESRI, Redlands, CA) 2 • Distance indige- m Distance from plot centre to the nearest indigenous 63.02 ± 43.58 89.69 ± 54.44 nous edge [Indi. forest edge (measured using ArcGIS 9.2) Edge] 3 • Concealment % Determined by placing the middle of a 0.5 × 0.5 m2 38.42 ± 18.97 33.66 ± 22.60 cover board consisting of 25 squares in front of the nest or between 1 and 1.5 m above the ground for non-nest sitesa. At 5 m distance, the percentage of the board obscured by vegetation was recorded in four cardinal directions and averaged Variables quantified at plot level 4 • Canopy closure % Average % closed canopy measured with a spherical 40.09 ± 27.75 78.10 ± 14.66 densitometer in the four cardinal directions 5 • Herbaceous % Percentage of the forest floor covered based on 68.27 ± 19.20 33.69 ± 17.70 cover [Herb. 40 systematic presence/absence reading at 1 m Cover] intervals from plot centre in each cardinal direction by using an ocular tube (diameter = 9 mm) 6 Litter cover % As above 30.06 ± 18.41 64.36 ± 16.78 7 Shrub cover % Percentage of the forest floor covered by the vertical 66.44 ± 14.05 58.70 ± 17.22 projection of shrub crowns 8 • Tree height [Tree m Average height of the three tallest trees, measured 25.09 ± 10.20 25.74 ± 8.04 hght.] with a Suunto inclinometer 9 • Nest substrates # Number of Dracaena steudneri, Chassalia sp., 22.15 ± 20.33 12.19 ± 13.27 [Nest subst.] Uvaria lucida, Landolphia buchananii (Dominant nest substrates during 2007 - 2008 see Table A.8) 10 • Dead wood m/plot Total volume of fallen and standing dead wood 0.60 ± 1.34 0.70 ± 1.05 V = π(DBH)2/4 with DBH the diameter at breast height for all fallen and standing, respectively, trees with a diameter > 12 cm. 11 Basal area m/ha Total basal area calculated from DBH of all the 163.45 ± 132.93 149.86 ± 135.40 mature trees (height > 5 m) in the plot 12 • Tree density #/ha Number of mature tree stems 632.64 ± 325.53 1113.35 ± 363.94 13 • Sapling density #/ha Number of sapling stems (height <5 m) 568.31 ± 614.48 759.32 ± 559.05 [Sap. density] 14 Tree species # Number of mature tree species 5.70 ± 2.22 9.61 ± 2.88 15 Sapling species # Number of sapling species 4.68 ± 2.46 5.36 ± 2.35 16 • Shannon Tree Shannon diversity index (H0) for mature trees 0.83 ± 0.45 1.29 ± 0.40 spec. species (> 5 m). 17 • Shannon Sapling Idem for sapling species (<5 m) 1.19 ± 0.53 1.31 ± 0.46 spec. [Shannon sap.] 18 • Vegetation com- First axis of a PCA (Principal Component Analysis) -0.070 ± 0.11 0.047 ± 0.10 position PC1 on species basal area data to assess gradients in [Veg. Comp. vegetation composition. This axis represents a PC1] gradient from typical pioneer species (-) to late successional species (+) 19 • Vegetation com- Second axis of the PCA. Gradient from interior (-) 0.010 ± 0.051 -0.0060 ± 0.093 position PC2 to edge and gap species (+) [Veg. Comp. PC2] 20 • Vegetation com- Third axis of the PCA. Gradient from late- -0.015 ± 0.043 0.011 ± 0.075 position PC3 successional species found at higher altitudes (-) [Veg. Comp. to species typical for late successional, but lower PC3] altitudinal patches (+)

Page 30 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore success and fledging success as a function of NSPS. In both analyses, reproductive success was modelled as a binary variable whereby nests either failed (i.e. no egg hatched or no nestling fledged) or succeeded (i.e. > 1 egg hatched or > 1 nestling fledged). Because the number of nests with information on reproductive success was limited, we could not use two separate, independent datasets for each analysis on nest-site selection and reproductive success. We however reran the nest-site selection analysis 10 times on a subsetted dataset, 2 whereby we each time randomly excluded half of the nests with information on reproductive success (30 nests, CH : 9 nests; NG : 21 nests). We then predicted site-scores, and tested for relationships with reproductive success, on the remaining nests. As results were identical (see appendix Table A.9), we consider our results unbiased with respect to non-independence of our datasets used for evaluating nest-site selection and reproductive success. Due to logistical difficulties, we only obtained information on nest-exposure (i.e. the timespan a nest is monitored) of a subset of nests. The outcome of an analysis on this subset, which account for the fact that survival probability depends on the interval length between two nest checks (daily survival rates sensu Shaffer, 2004), was similar to the results based on simple binary nest fate and are given in the appendix (Table A.4). This suggests that our models are unbiased with respect to nest-exposure. Finally, we applied a linear mixed model (LMM) to assess relationships between NSPS and nestling Scaled Mass Index (SMI). This index corrects body mass for variation in body size using a linear regression of log-body mass on log-tarsus length estimated by type-2 (standardized major axis) regression, and is a reliable proxy of organismal health and fitness (Peig & Green, 2010). Tarsus length was strongly correlated with body mass on a log-log scale (r = 0.88, p < 0.001), and we obtained a regression slope of 1.59 and average tarsus length of 24.01 mm for SMI calculation after excluding one outlier (i.e. |standardized residual| > 3). We thus calculated the SMI as body mass × (24.01/tarsus length)1.59 (Peig & Green, 2009, 2010). Because only a minority of nests successfully fledged, sample sizes for this model were reduced (CH: 7 nests, 12 nestlings; NG: 17, 30).

For each analysis on hatching success, fledging success, and nestling condition, we weighted and ranked three a priori models based on AICc. We defined the following models with success or SMI as dependent variable and (i) fragment, (ii) fragment + NSPS, and (iii) fragment × NSPS (i.e. the interaction between both variables) and its main effects as predictor variables. We added year as a categorical variable in all models to control for any factor that may affect reproductive success or SMI between years, and brood size in the SMI-models only, to control for the effect of brood size on SMI. We also added nest identity as random factor in the SMI-models, to account for non-independence of siblings. We used Nagelkerkes pseudo R2 (GLM) and marginal R2 (LMM) to quantify model explanatory power (Nagelkerke, 1991; Nakagawa & Schielzeth, 2013). LMM models are fitted by maximum likelihood (ML) with Satterthwaite approximated denominator degrees of freedom using the ’lme4’ package in R (Bates et al., 2015). Plots of fitted

Page 31 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

values against residuals indicated that normality or homoscedasticity assumptions were not violated. All analysis were performed in R 3.3.3 (R Core Team, 2017). Data are available from Figshare.

2 2.4 Results

Figure 2.2: Nest-site Selection (left panel). Placid Greenbul nests were more (+, right from dotted line), or less (-), likely to be found at specific habitat features in both a heavily disturbed (CH) and moderately disturbed (NG) forest fragment. Values (symbols) represent model-averaged parameter estimates and 95% CI (lines). Habitat values (right panel). Relative availability of all habitat in both fragments. Boxplots visualize the median, the first and third quartile; the whiskers extend to 1.5 × Inter-quartile range and values beyond this range are considered outliers and not plotted to improve visibility. Values standardized and centred (x - mean/SD), and differences (p 6 0.05; F-test) in availability (average value) between both forest fragments are denoted by ∗.

The two forest fragments (CH and NG) differed significantly for most of the 14 habitat variables (Table 2.1, Fig. 2.1); only concealment opportunities, tree height, the amount of dead wood, the sapling Shannon diversity and Vegetation Composition

Page 32 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

Table 2.2: Summary of model selection on a set of three a priori Generalized Linear models investigating effect of Nest Site Probability Score (NSPS) on hatching and fledging success (> 1 egg hatched or > 1 nestling fledged) of Placid Greenbul nests showing the coefficients with 95% CI on the logit scale.

Model Intercept Year a Fragment b NSPS NSPS × Fragment df R2 c AICc ∆AICc Fledging success 0.40 -1.43 -0.10 – – 1 3 0.15 80.1 0.00 (-0.71, 1.54) (-2.60, -0.35) (-1.29, 1.09) 2 0.84 -1.35 -0.047 -0.81 – 2 4 0.15 82.0 1.92 (-0.94, 2.72) (-2.54, -0.23) (-1.25, 1.17) (-3.41, 1.76) -0.36 -1.48 -1.84 1.29 -3.06 3 5 0.18 83.2 3.09 (-3.33, 2.48) (-2.74, -0.32) (-1.71, 5.61) (-3.27, 6.07) (-8.80, 2.37) Hatching success 0.45 -0.66 0.20 – – 1 3 0.04 86.8 0.00 (-0.61, 1.57) (-1.72, 0.37) (-0.92, 1.32) -0.10 -0.77 0.16 1.00 – 2 4 0.05 88.4 1.66 (-1.84, 1.66) (-1.89, 0.30) (-0.97, 1.28) (-1.45, 3.55) -1.13 -0.87 1.73 2.78 -2.52 3 5 0.07 89.9 3.15 (-4.26, 1.63) (-2.02, 0.23) (-1.70, 5.48) (-1.61, 7.86) (-8.25, 2.66) a reference category is breeding season 2009 - 2010, estimated category 2010 - 2011 b reference category fragment CH, estimated category fragment NG c Nagelkerkes pseudo R2

PC2 did not differ). In both forests, Placid Greenbul nest sites were significantly better concealed than randomly selected sites, and there was a non-significant tendency for nests to occur in areas with lower amounts of dead wood (strongest in CH, where zero was only marginally included in the 95% CI; Fig. 2.2, Table A.5). In NG, individuals nested at sites with significantly taller trees and showed a non-significant tendency to nest further from the indigenous forest edge and selected sites (Fig. 2.2, Table A.5). In CH, there was a non-significant tendency to nest in areas characterized by a higher abundance of plants typically used as nest substrates (Fig. 2.2, Table A.5). Our models explained jointly up to 50% (CH) and 35% (NG) of the variation in nest-site use (Nagelkerkes pseudo R2 of the most explanatory model in the ∆AICc 6 4 model subset; Table A.6& A.7). Overall, 58% (35 out of 60 nests) of our randomly selected Placid Greenbul nests failed before fledging, of which 74% (26 out of 35 nests) failed before hatching. Predation was the main cause of nest failure (88%), while the remaining brood failures were due to nest abandonment (i.e. containing cold eggs or dead nestlings) for unknown reason. Variation in NSPS (Nest Site Probability Score) did not explain variation in hatching or fledging success, also not in interaction with forest fragment (low AICc support and zero included in the 95% CI of the parameter estimates; Table 2.2). Nest success did not differ between both populations, but varied among years (Table 2.2). Our models explained up to 7% and 18% of variation in hatching or fledging success, respectively (Nagelkerkes pseudo R2 of the most explanatory model; Table 2.2). Likewise, variation in nestling condition (SMI) was not explained by variation in NSPS, nor did we detect an interaction between NSPS and forest fragment (Table 2.3). SMI only varied with brood size, but not among years or populations (Table 2.3) and our models explained up to 19% of the variation (marginal R2 values of the most explanatory model; Table 2.3).

Page 33 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

Table 2.3: Summary of model selection on a set of three a priori Linear Mixed Effect models investigating effect of Nest Site Probability Score (NSPS) on Placid Greenbul nestling scaled mass index (Nestling SMI) showing the coefficients with 95% CI.

Model Intercept Brood size Year a Fragment b NSPS NSPS Fragment df R2 c AICc ∆AICc Nestling SMI 20.57 -0.77 -0.36 0.48 1 – – 6 0.18 122.6 0.00 (19.22, 21.92) (-1.42, -0.11) (-1.12, 0.39) (-0.23, 1.20) 20.53 -0.77 -0.36 0.48 0.073 2 – 7 0.18 125.5 2.88 2 (19.03, 22.07) (-1.42, -0.11) (-1.12, 0.39) (-0.24, 1.20) (-1.29, 1.36) 20.45 -0.76 -0.37 0.58 0.20 -0.17 3 8 0.19 128.6 5.94 (18,40, 22.56) (-1.43, -0.09) (-1.16, 0.41) (-1.35, 2.49) (-2.42, 2.70) (-3.22, 2.91) a reference category is breeding season 2009 - 2010, estimated category 2010 - 2011 b reference category fragment CH, estimated category fragment NG c Marginal R2

2.5 Discussion

Nest site selection constitutes a central life-history event in tropical birds and may strongly affect individual fitness. Yet, compared to temperate species, few studies so far have assessed how and to what extent tropical birds cope with anthropogenic habitat disturbance through nest-site selection strategies. Here, based on two years of breeding monitoring of the Placid Greenbul, a typical understory insectivorous species of the East-African Taita Hills, we found that Placid Greenbuls from two populations inhabiting fragments differing in their degree of anthropogenic disturbance use nest sites that differ in habitat characteristics. Given the generally high predation pressure in the study area, we expected nest-site selection to be primarily driven by anti-predation strategies. However, we found no evidence that nests at sites more likely to be used for breeding experienced lower predation rates, and nor did these nests produce nestlings in better condition. As open-cup nesters generally benefit from greater nest concealment (Liebezeit & George, 2002; Remeˇs, 2005), Placid Greenbuls may be expected to prefer to nest at sites offering well vegetated and leafy cover, as in dense, shrubby areas. Such areas are typically patchily distributed in forests with an intact canopy layer, whereas (selective) logging that causes thinning of the canopy layer often results in a more uniform, dense understory scrub layer due to increased light penetration (Fredericksen & Mostacedo, 2000; Marsden et al., 2002). These structural changes are often accompanied by shifts in the species community, whereby some species increase in abundance at the cost of others (e.g. Sagar et al., 2003). In the most disturbed fragment (CH), species of understorey plant previously found to be typical nest substrates for the Placid Greenbul (notably Dracaena steudneri, Table 2.1, Table A.8) may attain high local densities under these disturbed conditions (represented by high average value in CH only; Table 2.1, Fig 2.2). Apart from offering a concealed nesting site, the high local density of these nest-substrate plants creates an area with many potential nest sites. Because many potential nest sites might reduce the search efficiency of predators, nesting at such sites may help reduce predation risk in this heavily disturbed fragment. Predation risk can also be reduced by avoiding areas with abundant dead wood, as woody debris is used by predatory small reptiles and

Page 34 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore mammals for denning and nesting (Harmon et al., 2004). Indeed, camera-trap recordings of rodents near nests and encounters of egg-predating snakes in our study area (Van de Loock & Bates, 2016) indicate that woody debris can act as shelter for greenbul nest predators. Together, the fact that in both study areas, the nest sites of this species are characterized by a greater concealment and a lower prevalence of dead wood is in line with our expectation that Placid Greenbuls may nest at sites likely to reduce predation 2 risk, irrespective of the overall degree of habitat degradation of the forest fragment they inhabit. The main difference in nest-site selection between our study populations is that greenbuls only favoured nesting at sites with taller trees and further away from the indigenous forest edge in the least disturbed fragment (NG; Fig. 2.2). Traditional silvicultural practices involve harvesting tall, valuable and easily exploitable trees and (re-)planting fast- and often tall-growing exotic species (Lamprecht, 1989). In heavily disturbed habitat such as fragment CH, tall trees might therefore represent both (for foresters invaluable) indigenous left-over trees as relicts from past, more pristine conditions (e.g. (Schlawin & Zahawi, 2008), as well as long-established exotic species. Indeed, tall stands of exotic Eucalyptus are more prevalent in fragment CH (Pellikka et al., 2009; Omoro et al., 2010). This suggests that tall trees might not be indicative of favourable greenbul nesting or foraging habitat in such heavily disturbed habitats (Gray et al., 2007; Johnson, 2007). In addition, another consequence of severe habitat modification is its effect on the spatial distribution of predator communities (Saunders et al., 1991; Chalfoun et al., 2002). For instance, forest-dependent nest predators of greenbuls may prefer the forest interior of small patches because of a higher habitat quality, lower human impact, or both (Carlson & Hartman, 2001; Spanhove et al., 2009). Under these conditions, greenbuls nesting in the centre of a remnant forest fragment might become prone to more nest failure because of predation (’inverse edge effect’ Lahti, 2001; Vetter et al., 2013), a pattern which has indeed been reported earlier for our study areas (Spanhove et al., 2014). This inverse edge effect implies that in fragment NG, where greenbul nests were more likely to be found further from the edge, breeding birds may potentially be exposed to a higher, not lower, predation risk. In contrast, because the vegetation structure of fragment CH is severely altered and resembles edge habitat throughout the fragment, greenbuls from fragment CH may no longer rely on such structural habitat variables. However, given that in our study area only two larger fragments can be compared, we cannot definitively rule out that observed differences in nest-site selection between both study populations are a consequence of a differing availability of certain habitat features between the forest fragments (Sih et al., 2011; Tuomainen & Candolin, 2011), rather than resulting from a plastic response to fragment-specific conditions. Although we found clear evidence that in both forest fragments, greenbul nests are not placed randomly with respect to available habitat variables, we did not find

Page 35 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore

evidence for any relationship with our fitness indicators (i.e. reproductive success and nestling condition). While we cannot dismiss the possibility that we failed to include potentially important habitat variables that might be used by Placid Greenbuls to select nesting sites (pseudo R2 of the nest-site selection models only ranged from 0.35 (NG) to 0.50 (CH)), we suggest two specific possible causes that may obscure the relationships we 2 hypothesized. First, nest predation is a highly stochastic event (e.g. (Wilson & Cooper, 1998; Githiru et al., 2005; Vigallon & Marzluff, 2005)), and larger sample sizes may be required to uncover more subtle relationships between nests selected and predation rates experienced. Second, it is likely that individuals may detect, and respond to, variation in perceived predation pressure (e.g. (Eggers et al., 2006)). This may be a contributing factor in the Taita Hills as well, as Spanhove et al. 2014 found that greenbul nest predation rates can vary in space and time. Hence, we require more detailed information on predator presence and phenological factors such as timing of nest initiation, for a more robust test on the adaptive value, if any, of greenbul nest-site selection.

More generally, a failure to detect strong relationships between nest-site selection and reproductive success is commonly reported in literature (see Chalfoun & Schmidt, 2012). Therefore, we here propose four mechanisms that should be taken into consideration by future studies of the breeding ecology of Placid Greenbuls and other tropical open-cup nesting passerines. First, parents may actively deter predators and thereby compensate for poor nest concealment, allowing successful breeding in predator-rich environments (Andersson et al., 1980; Weidinger, 2002; Remeˇs, 2005; Merrill et al., 2016). This may, in particular, be the case for cooperative breeders, where subordinates assist in protecting offspring against predators before and after fledging (Mumme, 1992a; Innes & Johnston, 1996; Riehl & Jara, 2009). In the Taita Hills, Placid Greenbuls display a facultative cooperative breeding strategy with up to five subordinates, and fledglings raised in larger groups have previously been shown to have a higher probability of post-fledging survival (Van de Loock et al., 2017). Second, as nest predation is highly stochastic, a higher reproductive success might not be achieved by selecting nest sites that reduce predation risk, but rather by reducing re-nesting intervals (Roper et al., 2010). This can be a particularly rewarding strategy in long-lived, tropical species with extended breeding seasons, as in our study system, and may hence explain why we observe up to four nesting attempts within one breeding season (Van de Loock, pers. obs.). Third, optimal nest site selection may be constrained by the fact that territories or home ranges are primarily selected to fulfil year-round nutritional needs and survival probability of the parents (or cooperative flocks) at larger spatial and temporal scales (hierarchical process ssensu (Wiens et al., 1987; Chalfoun & Martin, 2007; Orians & Wittenberger, 1991; Cornell & Donovan, 2010)). Fourth, variation in hunting strategies among different predator guilds can mask relationships between nest success and nest-site characteristics. For instance, nest sites may be optimized to match the hunting strategies of the predominant predator in one

Page 36 Flexible nest-site selection under anthropogenic habitat change in an Afrotropical understorey insectivore area, but be optimized differently to match different predators in another (Santisteban et al., 2002; Schmidt et al., 2006; Benson et al., 2010). In addition, high local densities of nests at preferred sites may facilitate nest predation as predators may learn cues and develop targeted hunting strategies (Pelech et al., 2010; Weidinger & Koˇcvara, 2010). In conclusion, while anthropogenic habitat change and disturbance can influence reproductive success of bird populations through several mechanisms, possible consequences 2 of such changes through impacts on nest-site selection are not yet fully understood (Robinson & Sherry, 2012; Ib´a˜nez-Alamo´ et al., 2015). Here, we report differences in nest-site selection between populations inhabiting forest fragments differing in the degree of anthropogenic disturbance, but conclude that more research is need to disentangle possible underlying drivers, as discussed above. In particular, we propose a better understanding of the spatio-temporal variation in predator communities would be a first important step to identify the underpinnings of nest-site selection strategies of this, and other tropical passerines (Lima, 2002).

2.6 Acknowledgements

We are grateful to P. Kafusi, A. Mwakulombe, L. Chovu and O. Mwakesi for fieldwork, H. Matheve for GIS mapping and L. Cousseau and J. Engler for feedback on earlier drafts of the manuscript. The comments of T. Spanhove, Jeremy Wilson and two anonymous reviewers greatly improved this manuscript. This work was approved by the National Commission for Science, Technology and Innovation of Kenya (NACOSTI/P/14/9325/3932) and was supported by FWO-grant G.0308.13N to LL and EM. Kenya Forest Service (KFS) kindly facilitated access to the forest fragments.

Page 37 African Goshawk (Accipiter tachiro) is a notorious predator of Placid Greenbul nestlings. Currently we know too little about the relative importance of this, or other predators within the greenbuls’ predator community. CHAPTER 3

3

Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

Dries Van de Loock1,2,3,∗ Laurence Cousseau1,∗ Mwangi Githiru4 Erik Matthysen2 Luc Lens1

Manuscript

∗ Contributed equally to this work 1 Terrestrial Ecology Unit (TEREC), Department of Biology, Ghent University, K. L. Ledeganckstraat 35, 9000 Gent, Belgium 2 Evolutionary Ecology Group (EVECO), Department of Biology, University of Antwerp, Campus Drie Eiken, Universiteit- splein 1, 2610 Wilrijk, Belgium 3 Ornithology Section, National Museums of Kenya, P.O. Box 40658-00100, Nairobi, Kenya 4 Wildlife Works, P.O. Box 310-80300, Voi, Kenya

Page 39 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

3.1 Abstract

Cooperative breeding can occur among related individuals, usually through delayed dispersal, among unrelated individuals, or the combination of both. Dispersal strategies of individuals are known to be influenced by environmental conditions, to vary among sexes and ages, and to depend on the relatedness with nearby conspecifics. Hence, spatiotemporal variation in individual dispersal strategies may cause variation in the genetic and demographic composition of groups and cooperative breeding behaviour. Such 3 variation in group composition and cooperative breeding may be particularly pronounced in anthropogenically modified landscapes, as human-induced habitat fragmentation and degradation may impose greater dispersal costs or alter the fitness consequences of delayed dispersal. Currently however, the evidence on the consequences of anthropogenic habitat change on group composition and cooperative breeding is mixed and appears to be very context dependent. We study spatial variation in delayed dispersal and group composition in a fragmented population of the Placid Greenbul (Phyllastrephus placidus), a facultative cooperative breeder from South-East Kenya. We found that up to 90% of the offspring delayed dispersal until the first subsequent breeding season, with a maximum of four breeding seasons (6% of all offspring). Up to 64% pairs bred in group together with subordinates and in 48% of these group-breeding pairs at least one, and up to three, of the subordinates are helpers (i.e., assist with providing food to the nestlings). Delayed dispersal, cooperative breeding and most group characteristics did not differ across seven forest fragments that vary in size, quality and degree of isolation. However, we found a significant difference in subordinate sex-ratio across fragments which may be due to sex-specific mortality and timing of delayed dispersal. Taken together, these results suggest that cooperative breeding is maintained in this fragmented Placid Greenbul population.

3.2 Introduction

Cooperative breeding occurs when individuals forego the opportunity to breed indepen- dently and opt to help raise other individuals’ offspring as non-breeding subordinates instead (Koenig & Dickinson, 2016; Komdeur et al., 2017a). These social groups are formed through delayed dispersal of offspring, through the immigration of unrelated individuals, or the combination of both (Riehl, 2013). Offspring may delay dispersal in response to ecological constraints (Hatchwell & Komdeur, 2000), such as when breeding opportunities are scarce or when dispersal costs are high, as well as due to the benefits experienced on the natal territory, such as access to resources (benefits of philopatry, Stacey & Ligon, 1991). Immigrants may join an unrelated breeding pair or family due to the direct benefits they obtain from group membership, such as access to mates or better survival prospects (Kingma et al., 2014; Kingma, 2017). These subordinate individuals then regularly help the

Page 40 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape breeding pair with parental duties, such as incubating, nestling provisioning or territory defense (Koenig & Dickinson, 2016). As the benefits of philopatry, direct benefits of group membership and ecological constraints may vary in both space and time, cooperative breeding can be viewed as a plastic response to local environmental conditions (Hatchwell & Komdeur, 2000; Baglione et al., 2006; Rubenstein, 2011). Several field experiments involving translocation of eggs or individuals support the hypothesis that variation in cooperative behaviour is due to phenotypic plasticity rather than genetically fixed strategies. For instance, Carrion Crow (Corvus corone corone) offspring hatched from eggs translocated from a non-cooperative colony from Italy, in 3 which offspring do not delay dispersal, to a cooperative colony from Spain in which offspring do, now delayed dispersal and remained in cooperative groups just as their foster colony (Baglione et al., 2003). Offspring delaying dispersal in the Spanish population benefit from access to resources on the natal territory, which is held year round in this, but not in the other, population (Baglione et al., 2005). Likewise, in a newly founded population of Seychelles Warblers (Acrocephalus sechellensis), the frequency of delayed dispersal and cooperative breeding increased with population size, as population growth caused complete occupation of high-quality territories (Komdeur, 1992; Komdeur et al., 1995). The idea that food abundance or access to resources influences the offspring’s decision to depart or stay, is further supported by experiments increasing territory quality through supplementary food. This indeed increased social group sizes and the incidence of cooperative breeding for some species (Baglione et al., 2006; Stacey & Ligon, 1991; Lin et al., 2006), but not for all. In particular, food supplemented Sociable Weaver offspring (Philetairus socius) showed the opposite pattern and dispersed more rapidly and started independent breeding sooner, with more breeding pairs and smaller cooperative groups as a consequence (Covas et al., 2004). Dispersal strategies are, in addition to environmental factors, also shaped by the individual’s age, sex and relatedness to other group members. In birds, females typically disperse both sooner and farther than males, due to inbreeding avoidance mechanisms or variation in benefits of local territory establishment (Greenwood, 1980; Perrin & Mazalov, 2000; Pusey, 1987; Eikenaar et al., 2008). Dispersal may also be induced at high local densities, such as in large groups, when kin competition decreases the inclusive fitness benefits among kin (Hamilton & May, 1977; Moore et al., 2006). Hence, variation in individual dispersal strategies may reflect variation in cooperative breeding, and the genetic and demographic composition of groups. Consider, for example, groups that are formed through delayed dispersal of offspring. As females usually disperse sooner than males, older subordinates in a group will, as a result, most likely be male (Cockburn, 1998; Komdeur et al., 2017b). Also, relatedness among group-members may decline with subordinate age as older birds may be immigrants in search for mating opportunities (Dierkes et al., 2005; Groenewoud et al., 2018).

Page 41 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

Variation in cooperative breeding and group composition may be particularly pronounced in highly heterogeneous landscapes, such as caused by anthropogenic activities. Anthropogenic habitat change is known to affect benefits and constraints of natal dispersal (Bonte et al., 2012; Bullock et al., 2018; Russell, 2001), and may, in turn, alter the fitness payoffs of helping behaviour (Russell, 2001; Banks et al., 2007; Harrisson et al., 2013). On the one hand, habitat degradation may cause offspring to disperse sooner, as the fitness benefits from delayed dispersal may be lowered on low quality habitat (Koenig et al., 1992; Kokko & Lundberg, 2001; Komdeur, 1992). On the other hand, habitat fragmentation 3 may impair natal dispersal and increase natal philopatry and helping behavior as the fragmentation process increases the costs to cross hostile matrix (Bonte et al., 2012; Temple et al., 2006; Russell, 2001). The effect of habitat fragmentation on group composition and helping behaviour appears to vary among species and with the severity of habitat change. In the Florida Scrub Jay (Aphelocoma coerulescen), for instance, group sizes were smaller with less helpers in fragmented habitat, likely due to higher mortality caused by the greater dispersal distance (Breininger, 1999; Thaxton & Hingtgen, 1996). A similar pattern of smaller groups in a fragmented landscape was also observed in Brown Treecreepers (Climacteris picumnus; Walters et al., 1999). However, in the latter species, fragment isolation completely disrupted immigration of unrelated females, which resulted in smaller and all male groups unable to breed (Walters et al., 1999; Cooper et al., 2002; Cooper & Walters, 2002). In contrast, forest fragmentation increased group sizes and both male and female helping behaviour in Long-tailed Tits (Aegithalos caudatus), as more females remained philopatric in isolated sites (Russell, 2001). As it stands, cooperative breeding has been suggested to buffer against demographic and environmental stochasticity caused by anthropogenic habitat change because of the pool of backup breeders (Walters et al., 2004). Yet, as demonstrated, it remains unclear how, and to what extent, anthropogenic habitat change causes variation in dispersal strategies, group composition and cooperative breeding, and how this breeding strategy may, in turn, buffer the population (Andr´en, 1994; Lampila et al., 2005; Fischer & Lindenmayer, 2007).

In this study, we assess spatial variation in delayed dispersal, group composition and cooperative breeding in a fragmented population of a facultative cooperative breeder from South-east Kenya (Taita Hills). The Placid Greenbul (Phyllastrephus placidus) is known to forage in small family groups (Fry & Keith, 2000), and was already noted as a likely cooperative breeder as early as the seventies (Grimes, 1976). Yet, cooperative breeding behaviour, with individuals helping the dominant breeding pair with providing food to the nestlings, was observed for the first time in this species in the Taita Hills in 2007 (first published in Callens, 2012). The indigenous cloud-forest of these hills is severely fragmented and degraded, and restricted to a set of isolated, variously sized and contrastingly degraded fragments. As expected for an understory insectivore that depends on intact indigenous forest for foraging and breeding (Bregman et al., 2014; Powell et al.,

Page 42 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

2015; Sekerciolu, 2011), the species shows genetic clustering between fragments, due to limited among fragment dispersal (Lens et al., 2002; Callens et al., 2011; Husemann et al., 2015; Aben et al., 2012, 2014), and suffered from environmental stress caused by habitat degradation (Lens et al., 1999). This evidence suggests that the constraints and benefits of delayed dispersal and cooperative breeding may vary across fragments. First, we use resighting data to construct individual dispersal trajectories, and use this to evaluate variation in timing of dispersal (degree of philopatry) across age, sex, and fragments. Second, we use a combination of genotyping and field techniques to evaluate variation in group composition and cooperative breeding. We compare the 3 presence, number and helping behaviour (i.e. whether or not subordinates provision nestlings) of non-breeding subordinates, as well as the relatedness, age composition and sex-ratio among these individuals across fragments. Formulating conclusive predictions is challenging because fragment size, degree of isolation and level of habitat degradation, may influence dispersal strategies and cooperative breeding differently, as noted in the introduction.

3.3 Methods

3.3.1 Study area and species

The Taita Hills (30°25’S, 38°20’E) are an isolated mountain block (max. altitude 2200 m.a.s.l.) surrounded by semi-deciduous lowland savannah plains. They represent the northern section of the Eastern-Arc Mountains, a mountain chain boasting high levels of endemism but which suffered from an estimated 95% forest loss during the last 200 years due to agricultural encroachment (Pellikka et al., 2009; Burgess et al., 2007a; Lovett & Wasser, 1993). At present, a dozen or so (depending on the authority and the adopted forest classification protocol) cloud-forest fragments remain, which are embedded within a matrix of agricultural fields and exotic plantations (Pellikka et al., 2009; Thijs, 2015; Aerts et al., 2011; Chege & Bytebier, 2005; Wilder et al., 1998). At least nine of these fragments host populations of the Placid Greenbul (Phyllastrephus placidus; formerly considered a subspecies of the Cabanis’s Greenbul P. cabanisi) of which ca. 75% are individually color-banded as a result of ongoing fieldwork since 1996. The breeding season of this common, medium-sized understory insectivore coincides with the onset of the short rainy season in November and lasts until March. As a facultative cooperative breeder, both pairs and groups may produce a clutch of two eggs (mean ± SD : 1.94 ± 0.39 eggs, range 1 - 3, n = 677 clutches) that hatch into nestlings in 15 to 17 days which fledge about 11 to 12 days later. In the study area, both fledging success and post-fledging survival increase with group size (Chapter4, Van de Loock et al., 2017). Both pairs and groups often re-nest after breeding failure, and occasionally after successful breeding.

Page 43 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

3.3.2 Data collection & Genotyping

Breeding Group Composition

During seven breeding seasons (2007-08, 08-09, 09-10, 12-13, 13-14, 14-15 and 15-16), a team of local field assistants mapped Placid Greenbul nests in seven forest fragments where greenbuls are known to breed: one large and moderately degraded fragment (Ngangao, NG, ca. 120 ha), another large, but heavily degraded fragment (Chawia, CH, ca. 80 ha), and five small and heavily degraded fragments (< 15 ha). From 07-08 onwards, we 3 identified the breeding pair, subordinates and helpers at each active nest. We did this by a combination of multiple observations, targeted mist-netting and video recordings (see Chapter4 for methodological details on observation- and trap sessions and video recording). The breeding pair was identified based on cloacal swelling (breeding males) and the presence of a brood patch or observed incubating behaviour (breeding females). All non-breeding individuals are considered subordinates. Helpers were defined as subordinates observed feeding the nestlings at least once during a minimum of four hours of continuous nest video recording (average number of observed feeding events per helper ± SD : 6 ± 4.43; range : 1 - 19, n = 59 helpers). A breeding pair with at least one subordinate is defined as a group-breeding pair, and is considered a cooperatively breeding pair when one or more of these subordinates feed the nestlings. Group composition did not change between successive nesting attempts, except for the addition of fledglings from earlier successful nests. This was confirmed by monitoring radio-tagged breeding females at regular intervals during the breeding season, which made it possible to detect changes in group composition within this period (Van de Loock et al., 2017). From this total set of observed pairs and groups, we only retained the first case per observed pair/group within a breeding season to avoid pseudo-replication in our analysis. The age of each subordinate at the time of nesting was determined based on the year of fledging (when ringed as nestling) or the estimated age when first captured. Indi- viduals ringed as nestlings were hence considered one year old during the first subsequent breeding season, two year old during the second breeding season, and so on. Subordinates were sexed molecularly using a set of sex-linked primers P2/P8 (Griffiths et al., 1998). Relatedness between breeding females and subordinates was based on genotyping at 12 microsatellite loci. The DNeasy Blood and Tissue kit were used to extract DNA from feathers and InstaGene matrix kit for blood samples, depending on the tissue that was collected. PCR were performed in 6 L reactions containing 2 L DNA, 2 L of primer mix and 2 L of Qiagen multiplex PCR Master Mix. The PCR products were analyzed on an ABI 3130XL Genetic Analyzer (Applied Biosystems), and genotypes were scored with GENEIOUS 7.0.5 (Kearse et al. 2012). See Husemann et al. 2015 and Cousseau et al. in prep for more details on the microsatellite markers. We used the exclusion method (Jones & Ardren, 2003) to identify first-order kin (son or daughter) of 124 subordinate-breeding

Page 44 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape female pairs. Analysis of known mother-offspring combinations showed only 8% (19 out of 236) cases of genetic mismatch at a single locus due to genotyping error (Cousseau et al. in prep). Thus, subordinates were considered first-order relatives of breeding females if they mismatched at fewer than two loci, and henceforth referred to as related subordinates.

Degree of philopatry

Prior to fledging, nestlings were colourbanded, and a blood sample taken to determine the sex using the above mentioned method (Griffiths et al., 1998). Out of 399 fledglings (excluding the final year of study), 54 (13,5%) were resighted at least once. For 15 of these 3 individuals we did not have information on the composition of the natal group and the associated group during resighting. As we cannot infer whether these individuals were philopatric or dispersed, they were therefore not considered further. Roughly half of all individuals (18 out of 37) were resighted only once with a maximum of 4 times. Of these individuals that were resighted only once, 11 were resighted when philopatric and 7 when dispersed. Of these individuals that were resighted more than once, 7 were resighted when philopatric only, 6 when dispersed only and 6 during both occasions.

3.3.3 Statistical Analysis

We evaluated spatial variation in group composition and cooperative breeding using seven response variables : presence of (i) subordinates and (ii) helpers, number of (iii) subordinates and (iv) helpers and within-group proportion of (v) related subordinates, (vi) helpers and (vii) male subordinates. For each response variable, we built a model that included fragment as main effect and year as random term. Year was modelled as discrete factor with seven levels (all breeding seasons since 07 - 08) or less as not all response variables could be collected during all breeding seasons (see below). Fragment was always modelled as discrete factor with three levels (fragment NG : large and moderately degraded, fragment CH : large and heavily degraded and small fragments : small and heavily degraded). We used Generalized Linear Mixed Models (GLMM) with log link function and Poisson error structure for models on number of subordinates and helpers, and GLMMs with logit link function and binomial error structure for the other five models. Based on a comparison between the sum of squared Pearson residuals and the residual degrees of freedom, we found no evidence for overdispersion. Note that sample sizes may vary between models due to variation in the stage at which nests were predated, due to practical constraints during breeding group characterisation and due to variation in the set of data collected during each breeding season (see Table 3.1). Helping behaviour was recorded for six breeding seasons (n = 122 pairs with information on number of helpers) and group size was recorded for four breeding seasons (n = 188 pairs with information on number of subordinates). Recording number of subordinates and helpers overlapped during

Page 45 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

three breeding seasons (n = 63 pairs with information on both number of subordinates and helpers). We did not test variation in the proportion of within-group male helpers between fragments as sample size were too low (n = 13 groups). We modelled philopatry as a binary variable (1 = philopatric, 0 = dispersed) using a Generalized Linear Mixed Model (GLMM, binomial error structure, logit link) with natal fragment, year since fledging and sex as predictor variables. Natal fragment is modelled as a categorical variable with the same three levels as the group composition models; year since fledging is modelled as a continuous variable. We did not run pair-wise 3 interactions among the three predictors because any fitted interaction created perfectly separated response variables in some of the categories. Due to the restricted number of individuals which were resighted both at their natal territory (philopatric), as well as when considered dispersed, we could not run the models including ID as random factor. We added year of fledging as a random factor to control for non-specific factors that may affect philopatry among different fledging-cohorts. All GLMM models were fitted by Maximum Likelihood. We calculated Type II Wald χ2-statistic to determine significance levels of main factors (function ’Anova’ in car package (Fox & Weisberg, 2011). We performed all statistical using R 3.4.1 software (R Core Team, 2017).

3.4 Results

3.4.1 Cooperative breeding and group composition

Over all fragments and years, 64% (121 out of 188) pairs bred in groups with one up to five subordinates (mean number of subordinates among group-breeding pairs ± SD : 1.5 ± 0.79, n = 121). In 48% (20 out of 42) of these group-breeding pairs one up to three subordinates are helpers (mean number of helpers ± SD : 1.35 ± 0.67, n = 20; Fig. 3.1). When not taking into account group-breeding behaviour, 40% (50 out of 122) of all pairs bred cooperatively with at least one helper. Subordinates included both related and unrelated individuals of either sex, but nestlings were fed by related subordinates only (Fig. 3.2). Among groups, on average 48% ± 46% SD of the subordinates in a group were males (range 0 - 100%, n = 43 groups). The age of subordinates varied from one to four years for both males and females, and the oldest male and female helper were four and two years, respectively (Fig. 3.2). Among groups, on average 75% ± 42% SD of the subordinates were related to the breeding female (range : 0-100%, n = 37 groups). More specifically, 73% of all groups had only related subordinates, 22% only unrelated, and 5% had a combination of both. Among groups, on average 40% ± 46% SD of the subordinates within a group helped with provisioning the nestlings (range 0-100%, n = 42 groups). Of these, on average 57% ± 45% SD were males (range 0-100%, n = 13 groups).

Page 46 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

3

Figure 3.1: Variation in Placid Greenbul group size and number of helpers. Group size two indicates pairs without subordinates, whereas pairs with at least one subordinate have group size > 3.

Figure 3.2: Variation in sex with age (left) and relatedness (right) among subordinates (top) and helpers (bottom). Note differently scaled y-axis between age and relatedness graphs. Age calculated in years (yr) since birth.

Page 47 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

Neither the presence or number of subordinates and helpers, nor the proportion of related or helping subordinates varied among fragments (all p > 0.05, Table 3.1). Only the within-group sex-ratio of subordinates varied significantly among fragments (p = 0.04 3.1). Groups were more male-biased in fragment NG, but more female-biased in fragment CH and the small fragments (Fig. 3.3).

3

Figure 3.3: Model prediction showing the proportion of male subordinates within groups in each fragment (average % ± SE).

3.4.2 Degree of philopatry

As expected, philopatry decreased with the number of years since fledging (Fig 3.4; Table 3.2). Almost all birds remained in the natal breeding group in the first year after fledging (predicted degree of philopatry = 89.9%), whereas almost all were dispersed in their fourth year after fledging (predicted degree of philopatry = 6.0%). Degree of philopatry did not vary across fragments, although birds tended to disperse sooner in the small fragments (Table 3.2). The predicted degree of philopatry at the first year after fledging in Ngangao, Chawia and the Small fragments was 99.9%, 99.9% and 69.8%, respectively (Fig. 3.4). Degree of philopatry was not significantly related to sex (Table 3.2).

3.5 Discussion

Within the fragmented and degraded cloud-forest of the Taita Hills, a majority of Placid Greenbul pairs (64%) breed in groups together with one or more subordinates. Only a subset of these subordinates are helpers (i.e., providing food to the nestlings), thus the proportion of cooperatively breeding pairs in this population is 40%. The majority of groups consist of subordinates related to the breeding female, one quarter of the groups contained unrelated subordinates only, and a minority a combination of both. Both male

Page 48 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

Table 3.1: Mixed model analysis of the effect of fragment on seven response variables related to group composition and cooperative breeding. Reference fragment for the fixed factor was Fragment Ngangao. Sample size (number of pairs/groups) for each model, and number of levels of random factor is shown in brackets. Significant p-values noted in bold.

Estimate ± SE χ2 (d.f.) p-value Presence of subordinates (n = 188) Fixed term intercept 0.51 ± 0.21 Fragment 0.74 (2) 0.69 Fragment Chawia 0.30 ± 0.37 Small fragments 0 ± 3.89 Random term Variance 3 Year (levels = 4) 0 Presence of helpers (122) Fixed term intercept -0.41 ± 0.29 Fragment 0.57 (2) 0.75 Fragment Chawia 0.16 ± 0.39 Small fragments -0.29 ± 0.62 Random term Year (6) 0 Number of subordinates (188) Fixed term intercept -0.13 ± 0.11 Fragment 2.31 (2) 0.32 Fragment Chawia 0.24 ± 0.17 Small fragments -0.0013 ± 0.20 Random term Year (4) 0.0022 Number of helpers (122) Fixed term intercept -0.78 ± 0.21 Fragment 0.057 (2) 0.97 Fragment Chawia 0.066 ± 0.28 Small fragments -0.014 ± 0.43 Random term Year (6) 0 Proportion related subordinates within groups (37) Fixed term intercept 1.34 ± 0.46 Fragment 0.43 (2) 0.81 Fragment Chawia -0.50 ± 0.83 Small fragments 0.043 ± 0.79 Random term Year (3) 0 Proportion helpers within groups (42) Fixed term intercept -0.31 ± 0.40 Fragment 1.12 (2) 0.57 Fragment Chawia -0.38 ± 0.57 Small fragments 0.31 ± 0.67 Random term Year (3) 0 Proportion male subordinates within groups (43) Fixed term intercept 0.61 ± 0.36 Fragment 6.31 (2) 0.043 Fragment Chawia -1.17 ± 0.72 Small fragments -1.70 ± 0.76 Random term Year (4) 0

Page 49 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

3

Figure 3.4: Model predictions showing the proportion (± SE) of male (round) and female (triangle) philopatric offspring for each year since fledging, in each fragment. Only 5 years since fledging are represented here to increase visibility.

Table 3.2: Mixed model analysis of degree of philopatry in a fragmented population of the Placid Greenbul. Reference category for fragment and offspring sex was Ngangao and female, respectively.

Fixed term Estimate (± SE) χ2 (d.f.) p-values Intercept 11.79 ± 5.14 Years since fledging -3.31 ± 1.26 6.94 (1) 0.0084 Fragment 4.29 (2) 0.12 Fragment Chawia -1.01 ± 1.71 Small fragments -7.09 ± 3.51 Offspring sex 0.27 (1) 0.60 Male -0.84 ± 1.61

Random term Variance Year of fledging 0

Page 50 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape and female subordinates engaged in helping, but only if they were related to the breeding female. Dispersal was typically delayed until the second year and this is reflected in the observed age composition of groups. Based on this dominant delayed dispersal strategy and group composition characteristics, the Placid Greenbul can be considered as a rather typical cooperatively breeding species (Riehl, 2013; Griesser et al., 2017; Kingma et al., 2016a; Koenig & Dickinson, 2016; Cockburn, 2003). In contrast to our expectations, neither the proportion of cooperatively breeding pairs, nor group composition differed between fragments or years, with the exception of the subordinate sex-ratio. This lack of variation corresponds with the lack of variation in 3 natal dispersal strategies across fragments. We have two possible explanations for these findings. Firstly, we hypothesize that the degree of fragment isolation is less severe than initially presumed based on the high levels of genetic differentiation (Callens et al., 2011), and does not, or only to a lesser extent, prevent among-fragment dispersal in our study area. Indeed, levels of genetic admixture have been shown to increase since 2000, possibly indicating increased gene flow over the past decade (Vangestel et al., 2013; Husemann et al., 2015). Likewise, out of 23 fledglings known to settle as breeders since 2006, three individuals (14%) did so in a different fragment than the one in which they were born (DVL, pers. observation). Secondly, it is possible that habitat degradation does not, or only to a lesser extent, influence the critical resources that are associated with variation in natal dispersal strategy (Banks et al., 2007). If the availability of resources is similar across both heavily degraded and moderately degraded fragments, then it is likely that the benefits of delayed dispersal do not vary among territories across fragments. To formally test this hypothesis, it would however be necessary to experimentally control for variation in resources through food supplementation (e.g. Baglione et al., 2006; Covas et al., 2004. Nonetheless, our results indicate that natal dispersal strategies, cooperative breeding and most group characteristics of the Placid Greenbul do not vary across a number of forest fragments that differ in size and degree of degradation. Notwithstanding the similarities across fragments, we detected variation in within- group subordinate sex-ratio. In our study area, groups were more female-biased in both small and fragment Chawia (CH), but more male-biased in fragment Ngangao (NG). We propose three possible explanations for this observed pattern : (i) In Chawia and the small fragments, the sex-ratio at fledging may be female-biased (Ewen et al., 2004; Komdeur, 2004); however this does not appear to be the case (Ngangao: 60% females (n = 28 nests), CH: 40% females (n = 25), small fragments: 70% females (n = 8)). (ii) Males may suffer higher mortality in Chawia and the small fragments, for instance due to differential access to food between sexes or parental manipulation (Donald, 2007; Clutton-Brock, 1985; Szekely et al., 2014). In turn, this would reduce the number of males and may causes a female-bias among subordinates. (iii) Males may delay dispersal longer in Ngangao, for instance due a higher natal territory quality (Komdeur, 1992). In turn, this would increase

Page 51 Cooperative breeding of an Afrotropical bird in a degraded and fragmented landscape

the relative proportion of male subordinates in Ngangao. As our dataset is currently too limited to formally test these hypotheses, future studies are necessary to test whether variation in subordinate sex-ratio across fragments is caused by different timing of delayed dispersal, is due to different survival probability of males and females, or is the result of a combination of both. Variation in sex-ratio of subordinates may affect the fitness benefits of cooperative breeding when the amount and quality of help differs between males and females (Lejeune et al., 2016; Brouwer et al., 2014; Legge, 2000a; Blackmore & Heinsohn, 2007). For example, 3 Red-winged Fairy-wren breeders (Malurus elegans) benefit only from female helpers as nestlings grew bigger and had higher post-fledging survival with female, but not with male helpers (Brouwer et al., 2014). This does not appear to be an issue in this species as reproductive success (i.e., successfully fledging > 1 offspring) was earlier shown to be similar across fragments (Chapter4). However, future work is necessary to study whether helper sex may affect other factors such as parental investment strategies, which, in turn, have also been shown to affect reproductive success (Kingma et al., 2010; Covas & Griesser, 2007; Khan & Walters, 2002; MacColl & Hatchwell, 2004). In conclusion, we here show that cooperative breeding and group composition, with the exception of subordinate sex-ratio, do not vary across a set of forest fragments that vary in size, degree of isolation and level of degradation. Although future work should focus on the causes and consequences of this sex-ratio variation, it currently appears that the cooperative breeding strategy of this species is maintained under anthropogenic habitat change. It however remains open whether the large number of subordinates may act as pool of potential breeding recruits that may buffer the population against stochastic fluctuations in environmental conditions and demographic processes(Walters et al., 2004). This may be particularly important for populations from the small and isolated fragments, where filling of breeding vacancies may hence not depend on timely immigration, but may happen by subordinates from nearby territories.

3.6 Acknowledgements

We are grateful to all local fieldworkers who assisted with nest searching and data collecting since 2006. We specifically like to thank T. Callens for collecting data during the first years of this study, V. Vandomme for sexing Greenbul samples, and all students who assisted in the field or screened video recordings on helper behaviour. This work was approved by the National Commission for Science, Technology and Innovation of Kenya (NACOSTI/P/14/9325/3932) and was supported by FWO-grant G.0308.13N to LL and EM. We thank the Kenya Forest Service (KFS) for granting access to the forest fragments, and the forest guards of the Taita-Taveta district in particular for on-site assistance.

Page 52 The Taita Hills in SE Kenya are isolated from other mountains in every direction and surrounded by semi-deciduous lowlands. The appearance of the landscape changes drastically between the dry and wet season. Photo’s taken in December 2013 (Upper) and January 2015 (Lower, by Sybryn Maes) from Lion’s Bluff.

CHAPTER 4

Maternal investment by a facultative cooperative breeder varies 4 with habitat degradation in a human-dominated landscape

Dries Van de Loock1,2,3 Liesbeth De Neve1 Laurence Cousseau1 Mwangi Githiru3,4 Erik Matthysen2 Luc Lens1

Manuscript

1 Terrestrial Ecology Unit (TEREC), Ghent University, K. L. Ledeganckstraat 35, 9000 Ghent, Belgium 2 Evolutionary Ecology Group (EVECO), Department of Biology, University of Antwerp, Campus Drie Eiken, Universiteit- splein 1, 2610 Wilrijk, Belgium 3 Ornithology Section, National Museums of Kenya, P.O. Box 40658-00100, Nairobi, Kenya 4 Wildlife Works, P.O. Box 310-80300, Voi, Kenya

Page 55 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape 4.1 Abstract

In cooperatively breeding species, subordinates provide extra care and/or protect a brood against predators. Breeding females of these species can apply multiple pre- and post- hatching strategies to optimize the trade-off between current and future reproduction. Such female behavioural plasticity may buffer natural variation in critical resources. For instance, investment in offspring can be increased under harsh conditions, or alternatively, decreased when times are benign to benefit the breeding female. To what extent human-induced environmental change may trigger plastic maternal strategies in cooperative breeders, however, remains unclear. Here, we assessed how maternal investment strategies and reproductive success in a facultative cooperative breeder vary with different levels of habitat degradation in a cloud forest archipelago in South-East Kenya. We compared pre- 4 and post-hatching maternal investment (i.e., variation in egg size and nestling provisioning, resp.) and reproductive success (i.e., fledging success and nestling body condition) in a population of Placid Greenbuls (Phyllastrephus placidus) from forest fragments that differ in size and degree of habitat degradation, and related this to variation in group size (all subordinates) and number of helpers (subordinates that provision food to nestlings). We found that egg investment varied across fragments with presence of subordinates, but was irrespective of the presence of helpers. Compared to pair-breeding females, females with subordinates laid larger eggs in the large and heavily degraded fragment, but smaller eggs in both the large and more intact fragment as well as in small and heavily degraded fragments. In contrast, investment in provisioning decreased with actual number of helpers, but was irrespective of habitat degradation. This suggests that maternal investment may be modulated by the degree of degradation and/or other (fragment-specific) environmental conditions but depends on the type of investment and the size and composition of the cooperative flock. While nestling condition was not affected by any of the tested variables, fledging success increased with group size. Exploring the long-term fitness consequences of plastic maternal strategies in both degraded and pristine landscapes may now be necessary to judge whether these can mitigate the negative consequences of habitat degradation.

4.2 Introduction

Understanding the evolution of maternal investment strategies is a central topic in life- history theory. A large number of empirical studies across a wide range of taxa have investigated putative trade-offs between current reproductive investment and future repro- ductive success (Roff, 2002; Stearns, 1992; Roff & Fairbairn, 2007; Bowers et al., 2012). In egg-laying species such as birds, breeding females may adjust their investment in eggs (pre-hatching) and/or nestlings (post-hatching) to current or expected environmental conditions, such as temperature, food availability and predation pressure (Cunningham &

Page 56 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

Russell, 2000; Horvathova et al., 2012; Lima, 2009; Yoon et al., 2017; Fontaine & Martin, 2006). Egg size is positively associated with early-life development of offspring (Krist, 2011) but involves a high production cost, which may compromise energetic investment in other life-history traits (Nager et al., 2000, 2001; Monaghan & Nager, 1997). Hence, there may be a trade-off between egg size and clutch size, which would result in large clutches of small eggs and vice versa (e.g. Williams, 2001). Yet, evidence is growing that females may also vary egg investment when clutch size variation is constraint, for instance among long-lived tropical passerines. In these species, breeding females may lay large eggs and give their young a head start under harsh environmental conditions (bigger is better hypothesis Kuijper & Johnstone, 2013; McGinley et al., 1987). However, when strong predation pressure reduces the likelihood of a successful nesting attempt, females may lay smaller eggs and save energy for future breeding attempts (Lima, 2009; Fontaine & Martin, 4 2006). Besides such variation in egg investment, ambient conditions and predator pressure are known to shape food provisioning strategies to nestlings as well (Yoon et al., 2017; Fontaine & Martin, 2006). For instance, in a safer environment with reduced predation pressure, females have been shown to increase nestling feeding rates to fledge better quality chicks (Fontaine & Martin, 2006).

Maternal investment strategies are also expected to vary based on the contri- butions of other individuals, of which cooperatively breeding species offer a particularly interesting case (Heinsohn, 2004; Hatchwell, 1999; Carranza et al., 2008; Savage et al., 2015; Cockburn, 2006). In these species, breeding pairs are assisted by subordinate individuals in vigilance against predators and/or food provisioning, and the joint effort by all group mem- bers ultimately determines reproductive success (McGowan & Woolfenden, 1989; Valencia et al., 2006; Mumme, 1992a; Koenig & Dickinson, 2016). These (expected) contributions by cooperative group members may interact with ambient conditions in shaping maternal investment strategies. Under favourable environmental conditions, breeding females that receive (or expect) help with provisioning food to their nestlings are predicted to reduce their own reproductive investment. This increases their survival, with no (or little) net cost to current reproductive outcome (’load-lightening hypothesis’; Crick, 1992). Such a strategy is typically predicted in long-lived species, which have multiple reproductive opportunities and for which life-time reproductive success is a direct function of longevity (Russell & Lummaa, 2009). Empirical evidence is growing that females effectively lay smaller eggs (e.g. Paquet et al., 2013; Russell et al., 2007; Canestrari et al., 2011) or reduce their provisioning rates (Meade et al., 2010; Li et al., 2015) in relation to the presence or number of helpers in cooperative groups (e.g. Legge, 2000b; Paquet et al., 2013). Under reduced food availability (e.g. due to drought), however, females are predicted to reduce the likelihood of brood starvation or delayed offspring development by maintaining or increasing their investment in the presence of helpers (’additive effect’, e.g. (Valencia et al., 2006; Liebl et al., 2016)). Along these lines, Superb Fairy-wren (Malurus cyaneus) females

Page 57 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

with helpers laid larger eggs when faced with long periods of drought (bigger is better strategy) but smaller eggs during more favourable conditions (load-lightening strategy) whereas egg size in pair breeders was independent of environmental conditions (Langmore et al., 2016). Maternal strategies have been shown to be shaped by both spatial (e.g. spatial clustering of food resources) and temporal (e.g. inter-annual fluctuations in rainfall) variation in natural environmental conditions (Langmore et al., 2016; Koenig et al., 2011; Koenig & Walters, 2015; Kuijper & Johnstone, 2013; Rubenstein, 2011). However, to the best of our knowledge, only one study so far studied how maternal investment strategies may be shaped by anthropogenic habitat change: female Rufous Treecreepers (Climacteris rufa) with helpers adopted a load-lightening strategy in large, intact forests, but maintained 4 constant provisioning rates in fragmented, degraded ones, likely as a result of depleted food resources (Luck, 2002). Apart from food resources, predator communities may be influenced by anthropogenic habitat change too, e.g. due to shifts in core-edge ratios in fragmented landscapes (Crooks & Soul´e, 1999). Generally, increased habitat fragmentation and degradation is associated with higher nest predation rates (Robinson et al., 1995; Tewksbury et al., 2006; Newmark & Stanley, 2011), which in turn may impose a strong direct selection on avian reproductive strategies (Lima, 2009; Fontaine & Martin, 2006; Eggers et al., 2006). Hence, maternal investment strategies may vary with the degree of nest predation and anti-predator behaviour of subordinate group members (Carranza et al., 2008). Along these lines, Iberian Magpie (Cyanopica cooki) females increased maternal investment when breeding cooperatively with helpers, as helpers reduced nest predation likely through improved vigilance and nest defence (Valencia et al., 2006; Carranza et al., 2008). Here we assess pre- and post-hatching maternal investment strategies under different levels of anthropogenic habitat degradation in a population of an Afrotropical, facultative cooperative breeder (Placid Greenbul, Phyllastrephus placidus). In the frag- mented Taita forest archipelago of south-east Kenya, this species breeds in pairs or in social groups that consist of one to five subordinates, of which zero to three assist with nestling provisioning and are referred to as helping subordinates or helpers (Van de Loock, unpubl. data). As a forest-dependent, understory insectivore with a long life-span and a small modal clutch size of two eggs, this species is predicted to be strongly affected by anthropogenic habitat change (Bregman et al., 2014; Powell et al., 2015). Indeed, levels of developmental stress were previously shown to correlate with increased levels of habitat degradation (Lens et al., 1999), whereas nest predation rates suffered from changes in edge-to-interior rates caused by the fragmentation process (Spanhove et al., 2014). Concurrently, Placid Greenbul females can be expected to vary investment strategies to overcome the negative consequences of anthropogenic habitat change. We study maternal investment strategies and reproductive success in seven study

Page 58 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape sites characterized by different levels of habitat degradation: a large and moderately degraded indigenous forest fragment (Ngangao, ca. 120 ha), another large but heavily degraded fragment (Chawia, ca. 80 ha), and five tiny (< 15 ha), heavily degraded fragments. First, we test to what extent maternal investment strategies vary with group size and/or number of helpers in relation to habitat degradation. We predict that cooperatively breeding females invest more in eggs and/or in nestling provisioning in more degraded habitat (i.e. bigger is better), but adopt a load-lightening strategy in more pristine habitat and hence invest less. Second, we test to what extent reproductive success and nestling body condition varies with group size and/or the number of helpers, in relation to habitat degradation. Given the high predation pressure (up to 70% initiated nests fail due to predation; Spanhove et al., 2014), we predict that reproductive success will improve in larger groups, whereas nestling condition will depend on the adopted maternal investment 4 strategy. Under a bigger is better scenario, we expect a higher nestling condition, whereas nestling condition is expected to remain constant under a load-lightening scenario.

4.3 Methods

4.3.1 Study area & species

The Taita Hills (30°25’S, 38°20’E, SE Kenya) are a verdant, fertile mountain block (max. altitude 2200 m a.s.l.) embedded within dry lowland savannah plains. They represent the northern most isolate of the Eastern-Arc Mountains, a mountain chain boasting high levels of endemism but which suffered from an estimated 95% forest loss during the last 200 years due to agricultural encroachment (Pellikka et al., 2009; Burgess et al., 2007b; Lovett & Wasser, 1993). At present, the indigenous cloud forest in the Taita Hills is severely fragmented into 13 remnant fragments. These are embedded within a matrix of agricultural fields and plantations of exotic crops and differ in size and degree of habitat degradation (Chege & Bytebier, 2005; Wilder et al., 1998; Aerts et al., 2011). Several of these forest fragments host variably-sized populations of the Placid Greenbul (Husemann et al., 2015), a common, medium-sized understory insectivore (Phyllastrephus placidus; formerly considered a subspecies of the Cabanis’s Greenbul P. cabanisi) of which ca. 75% are individually colour-banded as a result of ongoing fieldwork since 1996. Since 2007, nests have been consistently mapped for the full duration of the breeding season and have been further monitored (every fourth day) upon detection to determine their outcome. The breeding season starts with the onset of the short rainy season in November and lasts up to March. As a facultative cooperative breeder, pairs or small groups typically produce a clutch of two eggs (mean ± SE : 1.94 ± 0.015 eggs, range 1 - 3, n = 677 clutches) which are incubated by the breeding female for about 15 to 17 days and nestlings fledge about 11 to 12 days after hatching. Eggs are laid and incubated by the breeding female, but

Page 59 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

nestling provisioning and care are shared between the breeding pair and helpers. While only philopatric offspring from previous breeding seasons help with food provisioning, unrelated individuals may also join cooperative groups and likely assist with anti-predator tactics (Van de Loock & Cousseau, Unpubl. Data). In our population, 64% of the pairs breed in social groups with at least one subordinate, while 40% of pairs have at least one helper at the nest. Group size and composition do not vary significantly between the fragments (Chapter3).

4.3.2 Data collection & handling

During three consecutive breeding seasons (2012 - 13, 2013 - 14, and 2014 15), we 4 measured eggs, nestlings and provisioning rates, and assessed size and composition of groups associated with each of the located nests. The length and width of 405 eggs from 203 clutches laid by 78 identified breeding females (66% of all clutches) were measured with a digital calliper (± 0.01mm) upon clutch completion and by the same person (DVL). We used Hoyt’s formula (0.51 × length × width2) to calculate individual egg sizes. We used egg size rather than mass, as the former does not show a linear decrease over time and could be determined more accurately and irrespectively of timing of measurement (Krist, 2011). Nestlings were metal- and colour- banded and their body mass and tarsus length measured when about 9 days old (mean age ± SE: 8.84 ± 0.080, range 7 - 11, n = 137 nestlings). In those nests with offspring reaching fledging age and where no signs of nest predation were observed (i.e. destroyed or ragged nest conditions), we considered the number of fledglings equal to the number of nestlings observed at the last nest visit (mean age ± SE: 9.77 ± 0.12, range 8 - 12, n = 77 nests). For all other nests we recorded no fledglings. Nests were considered successful when at least one nestling fledged. We recorded group size, identified breeding and subordinate individuals, and quantified nestling provisioning rates and number of helpers through targeted mist-netting, multiple observations, and nest video recordings. We conducted focal observations both opportunistically when checking nests, as well as immediately after incubation initiation (i.e., warm eggs) by hiding less than 10m from the nest and using a playback with conspecific distress calls deployed for a maximum duration of 10 min to lure group members without attracting predators. The same approach was used to capture group members when nestlings were ca. 5 days old (mean age ± SE: 5.70 ± 0.15, range 3 - 8, n = 74 nests) or after nest depredation (n = 22 nests). New individuals were colour-banded and the breeding status of each group member was determined by the presence of brood patches or cloacal swellings. When nestlings were ca. 8 days old (mean age nestlings ± SE: 8.34 ± 0.13, range 6 - 10, n = 70 nests), we continuously video-recorded food provisioning from 7 a.m. to 1 p.m. using a HD camera (Sony Corp.) installed on a tri-pod, positioned about 1.5 m from the nest, protected and camouflaged by a waterproof casing and camo-coloured

Page 60 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape poncho and secured to a nearby tree by a metal wire. From these recordings we extracted maternal provisioning rates and average prey size, as well as the number of helpers. Prey length and width were determined relative to the birds’ length and multiplied to calculate prey size. We calculated hourly per capita provisioning rates to correct for effects of brood size on provisioning rates. Hourly per capita provisioning rates are not related to brood size (LMM with nestling age as covariate and female ID and year as random effects : Fbroodsize(n.d.f, d.d.f) = 3.38(1, 46.8), p = 0.78). Only video-recordings in which individuals could reliably be identified in more than 70% of the nest visits were used to calculate provisioning rates. The final dataset used for each analysis is reduced and varies due to several reasons (see Table 4.1). First, we were not able to extract all information from each nests due to logistic constraints, variation in the stage at which nests were detected, or due 4 to nest predation. For instance, egg- and clutch size are not recorded for nests detected in the nestling stage, while number of helpers could never be inferred in nests that were predated prior to video-recording. Second, we omitted cases from our data set if group size or number of helpers could not reliably be determined. Third, we excluded 9 nests from the analysis on reproductive success as they were abandoned for reasons other than natural predation and hence contained cold eggs or dead pulli (3 abandoned after adult ringing, 3 due to bad weather and 3 from unknown causes). Finally, rather than adding laying date as a covariate, we selected first breeding attempts of each group during each breeding season to reduce biases from seasonal changes in maternal investment strategies when re-nesting (e.g. Langmore et al., 2016) and to reduce model complexity. Based on nests with information on nest discovery stage (collected for all nests only during the 2014-2015 breeding season), nests were usually discovered before, or during, incubation (92%, 79 out of 85 nests). This was the case among both pairs (94%, 33 out of 35 nests) and groups (92%, 46 out of 50 nests).

4.3.3 Statistical analysis

We investigated variation in maternal investment and reproductive success (fledging success and nestling body condition) in relation to group size, number of helpers, and degree of habitat degradation. For all analyses, we defined group size (pair + all subordinates; range: 2 - 7 individuals) and number of helpers (range: 0 - 3 individuals) as continuous variables (group size 2 = 58 out of 178 nests, 3 = 73, 4 = 30, 5 = 12, 6 = 4 and 7 = 1 nest; number of helpers 0 = 27 nests out of 56, 1 = 22, 2 = 6, 3 = 1 nest). We ran two separate models because females may vary their investment depending on group size in general, or to the number of helpers specifically. All analyses involving number of helpers were conducted on group-breeding females. Fragments were classified in three ’types’ that combined fragment size and level of degradation and were modelled as a discrete factor

Page 61 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

Table 4.1: Overview of the number of nests (and eggs or nestlings) used for two sets of analysis involving group size or number of helpers on five response variables related to maternal investment and reproductive success. See text for more information on variation in sample sizes between analysis.

Group sizea Number of helpersa Maternal investment Egg size 77 (156 eggs) 29 (58 eggs) Provisioning rate 52 35 Reproductive success Fledging successb 101 38 Number of fledglingsb 89 35 Nestling mass 42 (71 nestlings) 28 (47 nestlings) a Number of helpers was obtained from fewer nests because many nests were predated prior to assessing the helping behaviour, which was only done late in the nestling stage 4 b differences in sample size between fledging success and number of fledglings models are due to unknown clutch sizes

with three levels: fragment Ngangao (large and moderately degraded), fragment Chawia (large and heavily degraded) and small fragments (all small and heavily degraded). When effects of group size or number of helpers were not significant, analyses were repeated with presence/absence of subordinates or helpers, respectively. When the effect of both group size and number of helpers was significant, we ran an additional analysis testing for the effect of non-helping subordinates.

Maternal Investment

We modelled variation in egg size and hourly per capita provisioning rate (log(feeds/h/nestling)) using Linear Mixed Models (LMM). For both response variables, we fitted two Linear Mixed Models (LMM) with either group size (77 and 52 nests, resp.) or number of helpers (29 and 35 nests, resp.), fragment type and the interaction as fixed effects. In the egg models we added clutch size and female tarsus length, to account for clutch effects and maternal body size. In the provisioning models we added nestling age (days since hatching (day 1), broods hatch synchronously) to account for age effects on provisioning rate. We also include mean prey size to detect possible trade-offs between maternal provisioning rates and prey size (e.g. Grieco, 2002). Breeding season and female ID were included as random effects to control, respectively, for non-specific factors that may affect reproductive success between years, and for non-independence of repeated observations of the same female over years. In the egg size model we also added nest ID to control for non-independence of eggs, which was nested in female ID because some females bred in multiple breeding seasons. We dropped female ID as random effect in the egg size/helper model as only 4 out of 25 females occurred twice in the dataset (out of a total of 29 nests). The same analysis on a subset of this dataset whereby we randomly selected one nesting event per breeding female yielded similar results (n = 25 nests). Hence, we consider our analysis on the full

Page 62 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape dataset unbiased with respect to non-independence of data from these 4 breeding females (supplementary material, Table B.2). In all egg size models we removed clutches from one breeding female where egg sizes were identified as outliers (egg size > Q3 + 1.5 × IQR).

Reproductive success & nestling condition

We modelled variation in fledging success (> 1 fledgling (1) vs. 0 fledglings) and number of fledglings (0, 1, 2 or 3) using Generalized Linear Mixed Models (GLMM) with binomial and Poisson distributions, respectively. For both response variables, we fitted two models with either group size (n = 101 and 89 nests, respectively) or number of helpers (n = 38 and 35 nests, respectively), fragment type and the interaction as fixed effects. Clutch size was added as an additional fixed effect in the number of fledglings models only. As random effects we included female ID and breeding season. Out of 38 nests with information on 4 number of helpers, only 1 nest failed and the model on fledging success with number of helpers is therefore not reported. The reason for this skewed success vs. failure ratio in this model stems from the fact that number of helpers is determined late in the nestling phase and thus not known for a majority of rapidly predated nests. Next, we modelled variation in nestling body mass using two LMM’s with group size (71 nestlings) or number of helpers (47 nestlings) and fragment type and the interaction as fixed factors. In addition, we added nestling tarsus length to account for variation in mass due to variation in body size, and number of nestlings and nestling age to account for other factors that may affect mass. As random factors, we added breeding season, female ID and nest ID (nested in Female ID) to control for non-independence of siblings. We reran both analyses with mean egg size per clutch as an additional covariate on a subset of nests where eggs were measured, to detect effects of egg size on nestling condition (Krist, 2011).

Model selection & analytical details

LMM models were fitted by Restricted Maximum Likelihood with Satterthwaite approx- imation to calculate denominator degrees of freedom. For each analysis, we calculated the Type III Sum of Squares F-statistic to determine significance levels of main effects and interactions. We used the function anova in lmerTest-package (Kuznetsova et al., 2016), which is implemented through the lme4-package (Bates et al., 2015). Visual in- spection of plots of fitted values against residuals did not reveal violation of normality and homoscedasticity assumptions. GLMM models were fitted by Maximum Likelihood. We calculated Type II Wald χ2-statistic to determine significance levels of main factors and interactions (function ’Anova’ in car-package; Fox & Weisberg, 2011). Based on a comparison between the sum of squared Pearson residuals and the residual degrees of freedom, we found no evidence for overdispersion. Only non-significant interactions (i.e.

Page 63 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

4

Figure 4.1: Raw data (small symbols) and ML model predictions (large symbols) showing the egg size for pair breeding (round symbols; group size = 2) and group-living (square symbols, group size > 2) Placid Greenbul females under different levels of habitat degradation (fragment type; Green = moderately degraded fragment Ngangao; Orange and Red = heavily degraded large fragment Chawia and small fragments, respectively). Predictions (± SE) were controlled for variables not represented in the graph (overall among-fragment mean of clutch size and female tarsus length) and error structure includes both fixed and random error terms.

between group size or number of helpers and fragment type) were omitted from the full model and results from this reduced model were reported. All models were run in R 3.3.3 (R Core Team 2017).

4.4 Results

4.4.1 Maternal Investment

Maternal egg investment varied between fragment types and with presence of subordinates (Group-living × Fragment type, p = 0.03, Table 4.2, Fig. 4.1), but not with group size (supplementary table B.1 for models on group size). In the presence of subordinates, females from Chawia laid on average 6.3% larger eggs than when breeding as a pair (percentage increase between model estimated egg sizes of pair and group-living females). In contrast, group-living females from both Ngangao and the small fragments laid on average 4.6% and 1.3% smaller eggs, respectively. Variation in egg size was not related to the number or presence of helpers, nor did it interact with fragment type (all p > 0.05, supplementary table B.2). Across all fragments, maternal food provisioning rates were inversely related with group size (Table 4.3). Provisioning rates did not differ between fragment type nor did

Page 64 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

4

Figure 4.2: Raw data (dots) and REML model predictions (lines) showing Placid Greenbul maternal provisioning rate in relation to number of helpers. Predictions are back-transformed from the original model which was calculated on a log-transformed response variable, and controlled for fragment type and the effect of other variables (overall among-fragment mean of number and age of nestlings, and mean prey size). Shaded area and whiskers reflect ± SE, which comprises both fixed and random error terms. Size of dots depict number of females (n).

the effect of group size differ among fragment types (Table 4.3). Females also provided less food to the nestlings when assisted by a larger number of helpers (Fig. 4.2, Table 4.3), independently of fragment type (Table 4.3). In contrast, maternal provisioning rates were not related to the number of non-helping subordinates (supplementary analysis where number of helpers were substituted by number of non-helping subordinates, Table B.3).

4.4.2 Reproductive success & nestling condition

The probability of at least one offspring successfully fledging increased significantly with group size but did not vary with fragment type, nor did these two factors significantly interact (Fig. 4.3, Table 4.4). Cooperative groups of five individuals were 36.2% more likely to have at least one offspring fledge compared to pair-only breeders (pairs: 46.9% vs. groups: 82.1%). Also, number of fledglings increased significantly with group size, but was not affected by fragment type or the two-factor interaction (Fig. 4.3; Table 4.4). Number of fledglings did not significantly vary in relation to the number of helpers, fragment type, or their interaction, nor when substituting number of helpers by presence of helpers (all p > 0.05, supplementary table B.4). Nestling mass was not related to fragment type, the presence or number of subordinates or helpers, or egg size (main effects and interactions: all p > 0.05, supplementary table B.5 and B.6).

Page 65 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

4

Figure 4.3: Raw data (dots) and REML model predictions (lines) showing the (upper graph) proportion of successful Placid Greenbul nests (i.e. at least one offspring fledged) and (lower graph) the number of fledged offspring for different group sizes. Shaded area and whiskers reflect ± SE. Predictions were controlled for fragment type and variables not represented in the graph (overall among-fragment mean of clutch size), and error structure comprises both fixed and random error terms. Size of dots depicts number of broods (n).

Page 66 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

Table 4.2: Mixed model analysis of female Placid Greenbul egg investment (cm 3) in relation to absence/presence of subordinates (group-living). Reference fragment type and social condition for the fixed factor and the interaction was Ngangao and pair-breeding.

Fixed term Estimate ± SE F (n.d.f, d.d.f) p-values Intercept 0.22 ± 1.42 Clutch size -0.059 ± 0.070 0.71 (1, 56.52) 0.40 Tarsus Length Female 0.12 ± 0.052 5.15 (1, 55.30) 0.027 Group-living -0.15 ± 0.0063 0.0040 (1, 64.19) 0.95 Fragment type 6.33 (2, 55.06) 0.0034 Fragment Chawia -0.41 ± 0.11 Small fragments -0.12 ± 0.11 Group-living × Fragment type 3.57 (2, 62.75) 0.034 Group-living × Chawia 0.33 ± 0.13 Group-living × Small 0.11 ± 0.12 4 Random term Variance Nest ID (Female ID) 0.0027 Female ID 0.040 Breeding Season 0 Residual variation 0.013

4.5 Discussion

We found evidence that group-living females in a heavily degraded forest fragment (Chawia) produced larger eggs than pair-breeding females. In contrast, females in a more intact fragment (Ngangao) produced smaller eggs when breeding as a group. Contrary to our expectations however, group-living females in the small and heavily degraded fragments also laid smaller eggs, though to a lesser extent than in Ngangao. Furthermore, in contrast with our expectations, females reduced their nestling provisioning rate in relation to group size and number of helpers across all fragments. Although this is a correlative study and experiments are needed to test causal relationships, these results suggest that maternal investment strategies in the Placid Greenbul (i) may be modulated by the level of habitat degradation but depend on which subordinate contributions are taken into account (i.e. whether or not helping with food provisioning), and (ii) depend on the type of maternal investment, i.e. pre-hatching (i.e. investing in egg size) or post-hatching (i.e. investing in nestling provisioning). Finally, our results may support the hypothesis that larger cooperative groups obtain a higher reproductive success, likely through reducing nest predation rates. Variation in nestling condition, however, was not explained by group size or number of helpers. Maternal load-lightening saves resources for future reproduction and should be favoured by long-lived cooperative breeders with many reproductive opportunities (Russell & Lummaa, 2009). Indeed, a comparative study on a taxonomically diverse set of 27 cooperative breeding species identified load-lightening as the prime strategy in 63% of these

Page 67 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

Table 4.3: Mixed model analysis of maternal per capita provisioning rates in relation to group size and number of helpers. Model fitted on log-transformed response variable (log per capita provisioning rate). Reference fragment type for the fixed factor and the interaction was Ngangao.

Fixed term Estimate ± SE F (n.d.f, d.d.f) p-values Group Size Intercept -0.72 ± 0.69 Nestling age 0.18 ± 0.071 6.12 (1,45.32) 0.017 4 Mean prey size 0.41 ± 0.49 0.73 (1,15.45) 0.41 Group size -0.21 ± 0.067 10.26 (1,42.52) 0.0026 Fragment type 0.073 (2,33.24) 0.93 Fragment Chawia 0.037 ± 0.17 Small fragments 0.080 ± 0.22 Group size × Fragment type * 1.20 (2, 38.97) 0.31 Group size × Chawia 0.20 ± 0.15 Group size × Small -0.031 ± 0.23

Number of Helpers Intercept -0.63 ± 1.07 Nestling age 0.12 ± 0.12 1.15(1, 26.26) 0.29 Mean prey size 0.00043 ± 0.56 0.00 (1, 10.15) 0.99 Number of helpers -0.30 ± 0.13 4.92 (1, 22.87) 0.037 Fragment type 0.26 (2, 25.05) 0.78 Fragment Chawia 0.060 ± 0.25 Small fragments -0.18 ± 0.34 Number of helpers × Fragment type * 1.08 (2, 24.71) 0.35 Number of helpers × Chawia 0.032 ± 0.29 Number of helpers × Small 0.91 ± 0.62 Random term Variance Group Size Female ID 0.00040 Breeding Season 0.023 Number of helpers Female ID 0.18 Breeding Season 0.026 Residual variation(Group size model) 0.28 Residual variation(Number of helpers model) 0.22 * Estimates and F-statistics and p-values of interaction obtained from full model

Page 68 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

Table 4.4: Mixed model analysis of Placid Greenbul fledging success (i.e. > 1 fledgling) and number of fledglings in relation to group size. Reference fragment type for the fixed factor and the interaction was Ngangao.

Fixed term Estimate ± SE χ2(d.f.) p-values Intercept -1.40 ± 0.71 Group size 0.55 ± 0.23 5.60 (1) 0.018 4 Fragment type 0.71 (2) 0.7 Fragment Chawia 0.054 ± 0.47 Small fragments 0.49 ± 0.59 Group size × Fragment type * 1.22 (2) 0.54 Group size × Chawia -0.020 ± 0.54 Group size × Small -0.61 ± 0.59 Fledging Success Random term Variance Female ID 0 Breeding Season 0 Fixed term Intercept -1.16 ± 0.81 Clutch size 0.25 ± 0.32 0.58 (1) 0.44 Group size 0.21 ± 0.098 4.51 (1) 0.034 Fragment type 2.10 (2) 0.35 Fragment Chawia -0.35 ± 0.27 Small fragments 0.059 ± 0.29 Group size × Fragment type * 0.40 (2) 0.82 Group size × Chawia -0.012 ± 0.22 Group size × Small -0.16 ± 0.26

Number of fledglings Random term Female ID 0 Breeding Season 0 * Estimates and χ2-statistics and p-values of interaction obtained from full model

Page 69 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

species (Hatchwell, 1999), and has been shown to result in higher survival probabilities and higher lifetime reproductive success in these, and many others (Covas et al., 2008; Khan & Walters, 2002; Li et al., 2015; Russell et al., 2007; Cockburn et al., 2008; Kingma et al., 2010; MacColl & Hatchwell, 2004). However, both past and current empirical studies typically focus on load-lightening strategies during nestling provisioning only (e.g. Covas et al., 2008; Meade et al., 2010; Hatchwell, 1999; Crick, 1992), while largely neglecting pre-hatching strategies. A recent meta-analysis on pre-hatching investment in cooperatively breeding vertebrates reports only 12 studies on 10 species (Dixit et al., 2017). Although several studies find that females in larger groups laid smaller eggs (e.g. Russell et al., 2007; Canestrari et al., 2011; Santos & Macedo, 2011; Paquet et al., 2013), other studies reported no, inverse, or context-dependent patterns in egg investment (e.g. Valencia 4 et al., 2017; Lejeune et al., 2016; Yang et al., 2016; Koenig et al., 2009; Russell et al., 2010; Langmore et al., 2016). Savage et al. 2015 predicted breeding females to reduce pre-hatching investment whenever helpers can compensate for initial loss in investment post-hatching, pending on the fact that they are able to predict subordinate contributions at the start of nesting. As with many cooperatively breeding birds, Placid Greenbul group composition remains constant throughout the breeding season (Van de Loock et al. 2017), and overall provisioning rates increase with number of helpers (DVL, unpublished data). This implies that helper contributions can be anticipated by females, and also suggests that these may compensate for the initial maternal investment loss. Taken together, these results confirm the idea that the Placid Greenbul, as a long-lived with extended breeding season and many reproductive opportunities, high levels of nest predation and low rates of nestling starvation, adopts a pre-and post-hatching load-lightening strategy. Future work should then focus on how Placid Greenbul females may gain fitness from load-lightening, for instance through increased survival or shorter re-nesting intervals (Russell et al., 2007; Woxvold & Magrath, 2005; Blackmore & Heinsohn, 2007).

An increase in maternal investment, i.e., bigger is better (Valencia et al., 2017; Langmore et al., 2016; Yang et al., 2016) can be expected to be favoured under harsh environmental conditions as it may increase the survival chances of offspring (Hatchwell, 1999; Carranza et al., 2008; Savage et al., 2015). In support of this hypothesis, females breeding in a large and heavily degraded fragment (Chawia) produced larger eggs when breeding in group. At the same time however, this is contradicted by the fact that females from small and heavily degraded fragments laid smaller, rather than larger, eggs when breeding in group. In turn, this variation in investment strategy suggests that we should interpret the influence of habitat degradation on maternal investment cautiously. It is, for instance, not yet possible to assess to what extent habitat degradation (as inferred from changes in forest structure and community, Aerts et al., 2011) parallels habitat quality, or whether there are other (local) factors that create a harsh and challenging environment. Likewise, it remains an open question what environmental or intrinsic factors trigger

Page 70 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape females to adopt different investment strategies both pre-and post-hatching, as observed in this population. Gaining a better understanding in maternal investment strategies may be possible by evaluating how the investment of her partner or helpers varies (which can be inferred from provisioning rates) and by further exploring factors that may shape habitat quality. Optimal maternal investment strategies are a function of the collective contribu- tions by all subordinate group members, including help with food provisioning, predator defence and other types of social behaviour (Cockburn, 1998; Hatchwell, 1999; Carranza et al., 2008). Yet, disentangling the relative impact of each contribution is deemed chal- lenging, at least in part because each subordinate may vary its contribution depending on the relative contributions by the other subordinates (Baglione et al., 2010), and because one type of contribution does not automatically exclude another. For instance, food provi- 4 sioning may simultaneously contribute to chick growth as well as chick survival through increased nest attentiveness (McGowan & Woolfenden, 1989; Hunter, 1987), although this may also attract predators (Lloyd et al., 2009). Nonetheless, the relation between egg investment and presence of subordinates, but not of helpers per se, suggests that breeding females tune their egg investment to the benefits obtained from group-living. We thereby hypothesize that larger groups are better in detecting and deterring predators, which increases the a priori probability of fledging (this study) and hence influences initial brood investment (Taborsky et al., 2007; Carranza et al., 2008). In contrast, maternal provisioning rates are primarily related to the number of helpers, suggesting that breeding females tune their post-hatching care to the level of help with food provisioning. Offspring fitness is generally assumed to depend to a great extent on the adopted maternal strategy, whereby a load-lightening strategy is predicted to result in equally fit offspring, whereas greater investment (bigger is better) in more fit offspring (additive effect; Russell et al., 2007; Valencia et al., 2006). Here, we found that nestling body condition did not vary with fragment quality nor with group size or number of helpers, all factors which interactively affect maternal strategies. How can we then explain that greater egg investment by group-living females in fragment Chawia did not result in more heavy nestlings? First, the small proportion of pair-raised broods that successfully raise young in this population may have masked statistically significant differences in nestling body condition with group-raised broods. Second, the adopted load-lightening during nestling provisioning may have further masked the effect of egg size on nestling condition (Savage et al., 2015; M¨aenp¨a¨a& Smiseth, 2017). In conclusion, our study corroborates the growing evidence that maternal strate- gies in avian cooperative breeders are more complex and flexible than previously thought, and may be influenced by both environmental conditions and group composition (Langmore et al., 2016; Royle et al., 2014). While past work highlighted that flexible maternal strate- gies, and cooperative breeding more broadly, may act as a buffer to maintain reproductive

Page 71 Maternal investment by a facultative cooperative breeder varies with habitat degradation in a human-dominated landscape

success under unpredictable climatic conditions (Shen et al., 2017; Rubenstein, 2011), we argue that such mechanisms remain largely unexplored with respect to human-induced habitat change. Although we here present empirical evidence for flexible maternal strate- gies in a fragmented and degraded landscape, we should now explore the long-term fitness consequences of these strategies in both degraded and pristine landscapes to judge whether these can actually mitigate the negative consequences of habitat degradation.

4.6 Acknowledgements

We are grateful to P. Kafusi, A. Mwakulombe, L. Chovu and O. Mwakesi for fieldwork and to D. Strubbe for conceptual, S. Maes for textual and L. Hertzog for statistical 4 feedback. We thank V. Vandomme for sexing greenbul samples, B. Apfelbeck and M. L. Vandegehuchte for constructive comments on earlier drafts, and all students who screened video-recordings. This work was approved by the National Commission for Science, Technology and Innovation of Kenya (NACOSTI/P/14/9325/3932) and was supported by FWO-grant G.0308.13N to LL and EM. Kenya Forest Service (KFS) kindly facilitated access to the forest fragments.

Page 72 Oliver is surprised to discover a previously unknown large chunk of forest on a mountain slope, and wonders how it remained overlooked during all these years (Upper photo, January 2014). We named it Susu West, and it is home to at least four greenbul groups. Two months later, Oliver discovers another two tiny indigenous pockets a bit farther north. These were well hidden within a Eucalypthus plantation and sheltered at least two greenbul groups. We named them, somewhat uninspiring, Susu North (lower left photo, March 2014). Forest disturbance is still ongoing in the Taita Hills, and particularly in the Susu area. In Susu East, just weeks after a Greenbul group successfully raised nestlings, the local community cleared the undergrowth for agriculture (lower right photo, January 2015). The interior of Ngangao forest. Dense, humid, sometimes surprisingly cold, but very beautiful. Some stands of Dracaena steudneri are visible in foreground of the left picture and are preferred substrates for nesting. CHAPTER 5

Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird 5

Dries Van de Loock1,2,3 Diederik Strubbe1 Liesbeth De Neve1 Mwangi Githiru4 Erik Matthysen2 Luc Lens1

Slightly modified from: Van de Loock, D., Strubbe, D., De Neve, L., Githiru, M., Matthysen, E. and Lens, L. 2017. Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird. Ecology and Evolution 7: 3489-3493.

1 Terrestrial Ecology Unit (TEREC), Department of Biology, Ghent University, K. L. Ledeganckstraat 35, 9000 Gent, Belgium 2 Evolutionary Ecology Group (EVECO), Department of Biology, University of Antwerp, Campus Drie Eiken, Universiteit- splein 1, 2610 Wilrijk, Belgium 3 Ornithology Section, National Museums of Kenya, P.O. Box 40658-00100, Nairobi, Kenya 4 Wildlife Works, P.O. Box 310-80300, Voi, Kenya

Page 75 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

5.1 Abstract

For avian group living to be evolutionary stable, multiple fitness benefits are expected. Yet, the difficulty of tracking fledglings, and thus estimating their survival rates, limits our knowledge on how such benefits may manifest postfledging. We radio-tagged breeding females of the Afrotropical cooperatively breeding Placid greenbul (Phyllastrephus placidus) during nesting. Tracking these females after fledging permitted us to locate juvenile birds, their parents, and any helpers present and to build individual fledgling resighting datasets without incurring mortality costs or causing premature fledging due to handling or transmitter effects. A Bayesian framework was used to infer age-specific mortality rates in relation to group size, fledging date, maternal condition, and nestling condition. Postfledging survival was positively related to group size, with fledglings raised in groups with four helpers showing nearly 30% higher survival until independence compared with pair-only offspring, independent of fledging date, maternal condition or nestling condition. Our results demonstrate the importance of studying the early dependency period just 5 after fledging when assessing presumed benefits of cooperative breeding. While studying small, mobile organisms after they leave the nest remains highly challenging, we argue that the telemetric approach proposed here may be a broadly applicable method to obtain unbiased estimates of postfledging survival.

5.2 Introduction

While it has generally been acknowledged that avian group living has to confer multiple benefits to be evolutionary stable (Dickinson & Hatchwell, 2004), the difficulty of following birds after fledging (reviewed by Cox et al., 2014) limits our knowledge on postfledging benefits. Yet, offspring of cooperative breeders typically receive extended care after fledging, suggesting that helpers can contribute substantially to a breeders fitness during this phase too (Langen, 2000). Quantifying postfledging survival in group-living species can hence result in more accurate measures of reproductive success and provide new insights into the ecology and evolution of sociality (Hatchwell et al., 2004; Preston et al., 2016). However, studies addressing possible effects of cooperative breeding on postfledg- ing survival have yielded mixed results (see Appendix C.1 and Table C.1), possibly due to methodological and/or taxonomic heterogeneity. For instance, few studies quantified survival during early dependency, when avian mortality rates are assumed to be highest (e.g. Sankamethawee et al., 2009), and only some have carried out this in Afrotropical forest birds, where cooperative breeding is common (Jetz & Rubenstein, 2011). Most studies found that helpers appear to have a neutral effect on postfledging survival. Yet, the influence of helpers on postfledging juvenile survival may vary temporally. For example, helpers may positively influence survival in the first days after fledging but not later on

Page 76 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

(e.g. Williams & Hale, 2006; Mumme et al., 2015). If postfledging survival is only assessed at a longer time span, the impact of cooperative breeding on juvenile survival may be misunderstood (e.g. Covas et al., 2011). Here, we use an innovative radio- tracking approach to obtain post-fledging survival estimates of an Afrotropical cooperatively breeding bird and relate these to group size. Although radio-tagging nestlings may appear most straightforward, both premature fledging due to handling and detrimental transmitter effects can bias survival estimates (Mattsson et al., 2006; Streby et al., 2013). To overcome this, we tagged breeding females prior to fledging and hence were able to track fledglings without the associated costs to juveniles. Age-specific postfledging mortality rates were obtained by fitting parametric survival functions to resighting data using a Bayesian framework (Colchero et al., 2012).

5.3 Material and Methods 5 5.3.1 Study system

Between November 2014 and April 2015, resighting data were collected for 40 juveniles fledged from 25 Placid greenbul (Phyllastrephus placidus) nests. Nests were detected by experienced field assistants in five remnant cloud forest patches within the Taita Hills, southeast Kenya (1500 m A.S.L., 30°25’S, 38°20’E). The Taita Hills represent the northernmost range of the Eastern Arc Mountains biodiversity hotspot and are isolated from similar highland areas by low-altitudinal savannah (900 m A.S.L.) in every direction (Lovett & Wasser, 1993; Myers et al., 2000). The need for agricultural land strongly pressured the indigenous cloud forest cover since precolonial times, and current distribution resembles an archipelago imbedded within small agroforestry fields and surrounded by exotic plantations (Pellikka et al., 2009). The Placid greenbul is a medium- sized insectivorous passerine of montane cloud forests understory. They build open cup-shaped nests at the forks of saplings, shrubs or climbers that resemble trapped leaf debris. Clutches (2-3 eggs) are incubated by the female, hatch after about 15 days, and fledge 11 (10-13) days later. In Taita Hills, stable territories with ill-defined borders are defended by pairs or cooperatively breeding groups consisting of a socially monogamous breeding pair and up to four helpers (Van de Loock, D. unpubl. data). Breeding coincides predominantly with the onset of the short rains in November and can last until March. Predation pressure is high with up to 70% of initiated nests failing completely (Spanhove et al., 2014). Importantly, an intensive color-banding effort is ongoing since 1996 through standard-effort ringing and nest monitoring, with at least 70% of the population being individually recognizable at all times (Lens, L. unpubl. data).

Page 77 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

5.3.2 Resighting data

To relocate fledglings, we attached lightweight VHF transmitters (Pip tag, Biotrack Ltd., Wareham, U.K.; <4% body mass) to breeding females during nesting, using a leg-loop harness (Rappole & Tipton, 1990). There were no indications that maternal nest attentiveness and mobility were affected by tagging, nor did any of the females disappear or die during the study. To distinguish between individuals, nestlings were banded with a metal ring and a unique combination of color rings prior to fledging. Each mother was relocated within 6 days after fledging of her brood and every subsequent fifth or sixth day until 55 days post-fledging, when fledglings are considered nutritionally independent (pers. obs. Van de Loock, D). Upon location of the tagged female, known offspring were intensively searched for over a 45-min search session. Flocks forage as coherent groups, and known flock members could nearly always be observed within this timeframe (Van de Loock, D. pers. obs.), hence suggesting that repeatedly undetected offspring was almost certainly dead. In addition, capture-recapture data since 2007 revealed no juvenile dispersal within 5 2 months after fledging (Lens, L. unpubl. data). Therefore, disappearance due to dispersal instead of mortality and therefore false-negative observations (i.e., offspring considered deceased while still alive) were highly unlikely.

5.3.3 Statistical modeling

Predictor variables

We considered four variables as predictors of postfledging survival. (i) Group size (GS). GS was determined through multiple focal observations around the nest at several occasions pre- and posthatching, combined with targeted mistnet traps with tape lure after hatching (average nestling age at targeted mistnet traps: 6.3 days, range 4-10, n = 37) and video recording during feeding of nestling (average nestling age at recording: 8 days, range 6-10, n = 28). The tape lure track consisted of recorded distress calls of conspecifics, intermitted with silence periods, and was deployed for a maximum duration of 10 min. Allofeeding behavior was recorded from 7 a.m.-1 p.m. continuously using a HD video camera set up at approximately 1.5 m from the nest. P. placidus individuals are known to be very sedentary and are often found in the same small area for long consecutive periods (Fry & Keith, 2000). Long-term observations of groups with known composition in our study area confirmed that both their size and composition rarely change during pre- and postfledging periods (Van de Loock, D. unpubl. data). Group size is the combined sum of the breeding pair and all helpers and was distributed as follows: 2: 6 broods; 3: 8 broods; 4: 6 broods; 5: 4 broods; and 6: 1 brood. When a breeding pair is thus not aided by helpers, the GS is 2, while in a GS of 6, four helpers aid the breeding pair. For a subset of fledglings (n = 28), we could reliably determine the number

Page 78 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird of allofeeders (AF) as well. Numbers of allofeeders were distributed as follows 1: 2 broods; 2: 11 broods; 3: 3 broods; and 4: 1 broods. As with GS, AF is the combined total of the breeding pair and the helpers. Only the male or female breeder provided the nestling with care when AF was 1, while both breeders received aid from 2 helpers when AF was 4. GS was related to AF (r = .46; p = .01; n = 28). (ii) Scaled Mass Index (SMI). Nestlings were measured and weighed at an average age of 9 days (range 7-11 days), and SMI values were calculated that accounted for age and variation in body size (following Peig & Green, 2009). SMI was not related to GS, AF or brood size (all p > .5). (iii) Fledging date (FLD). This was calculated as the number of days since the earliest 2 fledged nest. FLD was not related to GS (GLM with Poisson errors, χ(26) = 543.16; p = .34). (iv) Maternal condition (MC). This was inferred from daily feather growth rates of the breeding female, quantified by the width of five consecutive growth bars averaged over the left and right second outermost rectrix (ptilochronology sensu Grubb, 2006) and reflecting 5 individual and environmental quality during the period of feather growth (Grubb, 2006). To correct for variation in body size, residuals from an ordinary least squares regression of average bar width against tarsus were used (Vangestel et al., 2010). MC was not related to GS (LM, F4,19 = 0.45; p = .77).

Bayesian survival trajectory analysis

Time-specific mortality rates were estimated using the Bayesian Survival Trajectory Analysis (BaSTA) package in R (Colchero et al., 2012). This package is specifically designed for right- censored data (i.e., deaths unknown) and allows for obscured timing of death resulting from our indirect telemetry approach. First, the most appropriate mortality function was chosen based on the lowest deviance information criterion (DIC), after comparing all possible functions (Gompertz, Exponential, Logistic, and Weibull) and shapes (Simple, Weibull, Bathtub) (Millar, 2009). Second, the most appropriate model was used to infer the relationship between our four predictor variables and postfledging mortality. Forest patch ID in which the nest was located was added as a covariate in each model (Colchero et al., 2012). Models were initially run with all four predictor variables included (i.e., group size, scaled mass index, fledging date, maternal condition). Non- relevant variables (i.e., 95% confidence interval of the model estimate gamma (γ) includes 0) were removed in a stepwise manner until only relevant models remained (minimal adequate model, MAM; Klein & Moeschberger, 2003). Reported parameter values were derived from the MAM for the relevant variables, and parameter values of nonrelevant variables were obtained by forcing the variable into the MAM. Models were run 100 times, and outputs were averaged to obtain stable estimates. Finally, as the BaSTA analytical framework does not accommodate random effects, we bootstrapped the final model 100

Page 79 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

times while only selecting one fledgling per nest. Mortality functions were optimized using a Markov Chain Monte Carlo (MCMC) simulation procedure using four parallel simulations with 20000 iterations, 2001 burn-in periods, and 150 interval sampling each. Models converged appropriately, and serial autocorrelation or choice of priors did not affect the model estimation. All statistical analyses were run in R 3.2.2 (R Core Team, 2017).

5.4 Results

5

Figure 5.1: Estimates of (A) daily mortality rates during postfledging dependency for juveniles raised by pairs (group size = 2) or by the largest recorded group (group size = 6, occurred once in our dataset), and (B) survival probability until independence in relation to group size. Shaded area and whiskers reflect 95% CI in (A) and (B), respectively.

Overall, fledglings had an estimated 45.2% survival probability up to nutritional independence (95%CI 15.7% - 70.1%). The postfledging mortality rate was best explained by a Weibull function with a Makeham shape (∆ DIC second best model: 13.9) and decreased over time. Fledglings in larger groups and with a better body condition had lower mortality rates (Table 5.1, Figure 5.1a). In the largest recorded group size (pair with four helpers, which occurred once in our dataset), fledglings had an estimated 29.3% higher survival up to independence compared with individuals raised by a breeding pair only (Figure 5.1b), while the heaviest fledglings had an estimated 8.2% higher survival compared with the lightest ones. Both factors did not significantly interact, and mortality was not affected by fledging date or maternal condition (Table 1). In the subset of nests for which the number of allofeeders could be reliably quantified, the latter failed to predict postfledging survival when substituting GS in the MAM (average posterior: γ AF = −0.19,

Page 80 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

95%CI −0.74 - 0.34), while SMI still did (average posterior: γ SMI = −0.20, 95%CI −0.28 - −0.12). Finally, the bootstrapping procedure rerunning the MAM randomly selecting one fledgling per nest produced similar estimates, so we consider our analysis unbiased with respect to non-independence of fledglings.

Table 5.1: Posterior estimates of gamma (γ) for each covariate and interaction. (*) Relevant predictor of fledgling mortality; sign of estimate indicates direction of relationship

Lower Upper Estimate SE estimate 95% CI 95% CI Cooperative group size -0.41 0.17 -0.74 -0.092 * Scaled mass index -0.14 0.038 -0.21 -0.061 * Ordinal fledging day -0.016 0.0086 -0.033 0.00084 Maternal condition -0.18 0.17 -0.51 0.13 Group size × scaled mass index -0.15 0.61 -1.28 1.08

5 5.5 Discussion

By estimating daily mortality rates of fledglings based on telemetry of mothers during dependency, we show that group size had a positive effect on postfledging survival, independently of nestling condition. The few studies that quantified helper effects on postfledging survival yielded mixed results (Table C.1). However, these studies strongly varied in timing over which postfledging survival was quantified, from 2 weeks (e.g., Mumme et al., 2015) to 1 year (e.g., Covas et al., 2011). As selection pressures can vary temporally, presence of helpers may enhance fledging survival soon after fledging but not later on (e.g., Mumme et al., 2015), for example, due to local competition between helpers and fledglings (Covas et al., 2011). For example, the number of juvenile Brown jays ( morio) surviving to 30 days after fledging was strongly correlated with group size, while the number of offspring still alive after 1 year showed only a weak relationship (Williams & Hale, 2006). Ignoring temporal dynamics of helper contributions to fledgling survival can thus obscure subtle mechanisms through which cooperative breeding impacts on species life-history traits. Apart from duration over which survival rates are being assessed, helper effects may also be confounded by territory quality, as better territories often hold higher-quality breeding pairs, larger groups or both (Eguchi et al., 2002). While such intertwined effects are best disentangled by removal or cross-foster experiments (Brown et al., 1982), the latter potentially disrupt social relationships within the group (Mumme, 1992b) and may be too intrusive for tropical species with a slow life history and high disturbance sensitivity. Nonetheless, our correlative approach shows group benefits independently of maternal condition.

Page 81 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

How exactly helpers contribute to the reduced mortality in P. placidus fledglings currently remains unknown. Both antipredator behavior and alloparental care have been suggested as important, nonexclusive direct mechanisms (Mcgowan & Woolfenden, 1990; Ridley & Raihani, 2007). In addition, helpers may also contribute to the learning of foraging skills, although empirical studies on this mechanism remain scarce (Heinsohn, 1991; ?). Based on our ecological knowledge of P. placidus, we hypothesize that antipredator behavioral strategies may underlie the positive effect of helpers on postfledging survival observed in our study population. Indeed, while nest-predation rates were already shown to be high in P. placidus (Spanhove et al., 2014), a recent pilot study in which transmitters were placed on P. placidus nestlings showed that over 90% of postfledging mortality could be attributed to predation as well (Van de Loock, D. unpubl. data). The latter often involved African Goshawks (Accipiter tachiro), an aerial predator also specialized on bird eggs and nestlings (Spanhove et al., 2014). As overall survival until independence (45.2%) is low in P. placidus compared with other tropical species (Lloyd & Martin, 2016; 5 Sankamethawee et al., 2009), antipredator strategies hence are an obvious avenue for increasing fitness.

We acknowledge our data cannot disentangle the relative contributions of an- tipredator behavior versus alloparental care. Quantifying the role of alloparental care can be made through relating postfledging survival to feeding rates or the number of allofeeders in case not all group members allofeed the young (Ridley & Raihani, 2007; Mcgowan & Woolfenden, 1990; Brouwer et al., 2012). For P. placidus, survival models run on a subset of data for which the number of allofeeders at the nest was known showed that fledgling survival was not correlated with number of allofeeders. Data on postfledging feeding events and on the identity of the allofeeder, however, remain scarce for P. placidus. The fact that we did not find any relationship should therefore be interpreted cautiously, as we currently cannot exclude the possibility that alloparental care contributes to postfledging survival as well.

In conclusion, we demonstrate that variation in group size in cooperative breeding species can substantially affect juvenile survival and thereby reproductive success. We agree with Heinsohn 1992 that full breeding cycle research in cooperative breeding ecology is crucial, yet remains rare due to logistic difficulties when studying evasive, mobile organisms during the early postfledging phase (Marra et al., 2015). We believe that the telemetric approach employed here may help to overcome some of these challenges and may ultimately lead to a better mechanistic understanding of relationships between cooperative breeding and postfledging survival.

Page 82 Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird

5.6 Acknowledgements

We thank the Kenyan government and the Kenya Forest Service for permitting research in the Taita Hills and A. Ramos, A. Mwakulomba, P. Kafusi, L. Chovu, and O. Mwakio for fieldwork. This work was approved by the National Commission for Science, Technology and Innovation of Kenya (NACOSTI/P/14/9325/3932) and was supported by FWO- grant G.0308.13N to LDN, EM, and LL. We are grateful for the comments of one anonymous reviewer, which greatly improved our final manuscript.

5

Page 83

CHAPTER 6

General discussion

6

In this thesis, I investigated if, and how, life-history strategies of a facultative cooperative breeder vary in a severely modified and anthropogenically degraded landscape. To do this, I evaluated five different aspects associated with reproduction in a chronological order : nest-site selection (Chapter2 ), natal dispersal and composition of breeding groups (Chapter3 ), maternal investment strategies (Chapter4 ), reproductive success (Chapter4 ), and the post-fledging period (Chapter5 ).

I integrate and elaborate further on the results of these chapters in this general discussion, which is separated into two distinctive parts. In a first part, I explore the fitness payoffs of cooperative breeding for both subordinates and breeders, and discuss the ecological drivers of this breeding strategy. In a second part, I first look at the impact of anthropogenic habitat change on this species and how this may affect, or interact with, cooperative breeding. I critically evaluate these findings and propose some important considerations for future research. Then I discuss in a somewhat more speculative manner whether, and to what extent, cooperative breeding may modulate the impact of, and ultimately buffer the species against habitat change. In particular, I address whether cooperative breeding may moderate nest-site selection strategies, buffer fluctuations in population dynamics and reduce variation in reproductive success. I finish this discussion with the main conclusions that can be drawn from this thesis.

Page 87 General discussion

6.1 Fitness payoffs and ecological drivers of coopera- tive breeding

6.1.1 Subordinates benefit from cooperating

In the Taita Hills, most fledglings (90%) delay dispersal and remain on the natal territory until the subsequent breeding season, with a maximum of four breeding seasons (6% of all fledglings; Chapter3). Such dispersal strategy facilitates the formation of family groups consisting of retained offspring and parents, as is observed in the majority of groups (75%). Whereas these family groups are in some rare cases also joined by unrelated immigrants, about a quarter of all groups were composed of unrelated immigrants only, and did not contain retained offspring (Chapter3). All non-breeding individuals within a group are considered subordinate to the breeding pair, and collectively referred to as subordinates. Helping behaviour (i.e., providing food to the nestlings; Chapter4) was documented for roughly 50% of related subordinates, but never observed for unrelated immigrants. These findings suggest that delayed dispersal is a common strategy in our study species, while helping behaviour and (prolonged) group-membership of both related and unrelated individuals are more variable. In the following paragraphs I speculate on the likely causes 6 and underlying mechanisms of the observed strategies by evaluating the possible costs and benefits of delayed dispersal, helping and immigrating into an unrelated group.

The natal territory is a safe haven for retained offspring

In accordance with the ’benefits-of-philopatry’ hypothesis, it appears that Placid Greenbul offspring benefit from delayed dispersal on the natal territory, rather than to leave and wait elsewhere (Stacey & Koenig, 1990; Kokko & Ekman, 2002; Griesser et al., 2006; Kingma et al., 2011, 2014). In particular, I hypothesize that the access to resources within a familiar and safe environment may improve the survival probability of offspring and subsequently offer better prospects on future reproduction. Although my data do currently not permit to quantify the relative fitness increase from philopatric behaviour, two anecdotal observations illustrate that greenbuls may use the natal territory as a ’safe haven’ from where to explore the forest and prospect for breeding vacancies or available mates (Kokko & Ekman, 2002; Kingma et al., 2016b; Jungwirth et al., 2015). First, at least one two year old helper which was trapped far off its natal territory just after the breeding season was later encountered back in its natal territory. The fact that this bird was trapped 300m away from its natal territory, spanning several sites where other groups were known to breed, illustrates that individuals may cross territorial boundaries and also illustrates the range that may be covered during prospecting. Second, several philopatric two year old birds which were radio-tagged to record movements and behaviour before and during the breeding season were occasionally observed solo away from their natal territory

Page 88 General discussion whereas they re-joined their natal flock on later occasions (DVL, Unpubl. Data). Individuals that delay dispersal may also inherit, or settle near the edge of the natal territory (territory budding and inheritance; Woolfenden & Fitzpatrick, 1978; Kokko & Ekman, 2002; Komdeur & Edelaar, 2001). In spite of the overall short-distance dispersal with many individuals settling in their natal fragment (Vangestel et al., 2013), no offspring is recorded as breeder in, or adjacent to, their natal territory with a possible exception among individuals from the small fragments (DVL Unpubl. Data). Out of 23 individuals, 20 (86%) dispersed and settled within their natal fragment. The mean dispersal distance for these 20 individuals is as follows (mean ± SD (n)): Ngangao, 678±450m (7); Chawia, 401±391m (11) and 43.3±18m (2) in the small fragments. The shortest distances are noted among males : 190m in Ngangao, 101m in Chawia and 30m in the small fragments. All females (n = 3) born in the small fragments dispersed to another fragment. It is hence likely that, due to inbreeding avoidance mechanisms (Kokko & Ekman, 2002), territory budding and inheritance represent a rarely used strategy to obtain a breeding position in our population. Notwithstanding the benefits associated with philopatry, individuals may also pay the costs of, among other things, greater competition for local resources (Dickinson & Hatchwell, 2004; Sorato et al., 2016). For example, removal of food resources caused birds to disperse sooner in Western Bluebirds due to increased competition among group 6 members, which lowered the net benefits of philopatry (Dickinson & McGowan, 2005). In order to fully understand the benefits of philopatry in the Placid Greenbul, it is hence necessary to compare its costs and benefits among alternative strategies such as leaving to float through the population (Ridley et al., 2008; Kingma et al., 2016a). In general however, the high incidence of delayed dispersal illustrates the fact that the net fitness benefits is presumably higher than any associated costs, particularly during the first year after fledging.

Helping behaviour is explained by kin-selection, but depends on environmental and social conditions

In chapter3 I show that all helpers are related to the breeding female, and thus care for their younger siblings. Given that reproductive success is correlated with group size (Chapter4&5), it appears that subordinates accrue indirect fitness benefits through helping. Hence, kin selection offers an unambiguous and powerful explanation for helping behaviour in the Placid Greenbul (Green et al., 2016). However, the fact that not all related subordinates help (Chapter3) indicates that various proximate factors additionally determine helping behaviour. Given the high incidence of extra-pair paternity in our study species (48% of offspring sired by males from outside the cooperative group; Cousseau et al. in prep), it is possible that helping behaviour is modulated by the actual genetic relatedness between

Page 89 General discussion

subordinates and the young in the nest. However, this strongly depends on the currently untested assumption that subordinates possess a kin-discriminating mechanism that allows them to discriminate between full sibs and half sibs (e.g., fathered by extra pair individuals; Griffin & West, 2003; Cornwallis et al., 2009). Kin recognition is widespread in birds but typically learned by association (Riehl & Stern, 2015). Indeed, past research on kin recognition mechanisms documented that male breeders are generally unable to detect extra-pair young sired by other males, while siblings fail to discriminate between related and fostered birds (Kempenaers & Sheldon, 1996; Griffin et al., 2013). Overall, as studies offer mixed support for kin discrimination without prior association (Breed, 2014, but see e.g. Petrie et al., 1999), it is unlikely that this mechanism would explain variation in helping behaviour in this population. Another alternative explanation may be that helpers specialize in different tasks, and non-helping subordinates which refrain from nestling provisioning thus carry out a larger share of other duties such as territory defence (task partitioning; Arnold et al., 2005; Le Vin et al., 2011; Bruintjes & Taborsky, 2011; English et al., 2010). Few studies however have explored such task-partitioning among subordinates in cooperatively breeding birds (e.g., Arnold et al., 2005), and it is currently unclear whether this represents a 6 common strategy among these species. I therefore suggest that future work should aim to quantify the contribution of helpers to other tasks than provisioning behaviour. Measuring individual contributions to predator vigilance and defence may, for instance, be possible by experimentally triggering a response using acoustic and visual predator stimuli. Non-helping subordinates that do not regularly provision the young, may also be ’insurance’ helpers that may compensate for sudden changes in provisioning effort of others and may buffer unexpected and unfavourable circumstance during breeding (lazy helper hypothesis; Baglione et al., 2010). Because helping behaviour is potentially costly, subordinates should refrain from doing so if the fitness gains are only marginal (Heinsohn & Legge, 1999). Under unfavourable conditions (e.g., limited resources), or when the provisioning effort of others declines from whatever cause (e.g., injury or mortality of a group member), fewer young are likely to fledge, which will lower the indirect fitness gains. To compensate for such negative effects, helpers are hypothesized to contribute more and in this way prevent a strong reduction of reproductive success, in order to maintain their personal indirect fitness gains. In this regard, helpers may be important to reduce variance in reproductive success (Baglione et al., 2010). It is necessary to focus on individual variation in helping behaviour, both within as well as across nesting attempts, to answer this hypothesis. One possible experimental approach may be to handicap helpers (e.g. through wing-clipping) and infer whether other, initially non-helping subordinates take over the helping behaviour of their impaired group mates. Finally, helping can also be seen as a form of rent-paying that is enforced by the breeders to offset the costs of subordinate presence (the pay-to-stay hypothesis; Gaston,

Page 90 General discussion

1978; Kokko & Ekman, 2002; Hamilton & Taborsky, 2005). Given such scenario, parental enforcement of helping behaviour can be hypothesized to co-vary with the costs of having retained offspring. Hence, retained offspring on low-quality territories with fewer resources and many competitive neighbouring groups may be forced to contribute more to the reproductive attempt. However, parental enforcement of helping behaviour is not well- supported among avian cooperative societies (but see e.g., Koenig & Walters, 2011; Mulder & Langmore, 1993) and it may be equally plausible that subordinates leave a low quality territory, where fitness benefits of helping are expected to be lower (Komdeur, 1992). Overall, I conclude that kin selection is likely to be the principle underlying mechanism that determines whether individuals will engage in helping behaviour in our study population, while the amount and type of help is likely shaped by local environmental and social conditions.

Unrelated immigrants may gain benefits from group living and potential future access to mates

In contrast to cooperation among kin, through which the contributors may accumulate a variety of both indirect and direct fitness benefits (as discussed earlier), unrelated immigrants can only benefit directly from group-membership (Kingma, 2017). I hypothesize 6 that, in the Placid Greenbul, immigrant birds may join groups for two reasons. First, group-living immigrants may have higher survival probability than solitary floaters as a consequence of group-living (Ridley et al., 2008). Comparing the survival probability between a group-living immigrant and solitary floater is necessary to test this effect, but is in our study population complicated by possible transient group-membership of immigrants (year-to-year changes) and gaps in resighting history. Second, immigrants may gain access to a mate or breeding position and as such increase their reproductive success (Groenewoud et al., 2018). As none of the extra-pair offspring were sired by birds from within the group (Cousseau et al. in prep), unrelated subordinates are not expected to engage in extra-pair mating. However, they may pair with retained offspring and settle on a nearby territory, or wait until one of the dominant breeders dies.

6.1.2 Breeding cooperatively benefits breeders

Subordinates increase nestling and fledging survival

This thesis corroborates the well-supported assumption that subordinate presence may boost the breeding pairs’ reproductive success through direct mechanisms affecting current reproductive attempt, as well as through possible indirect mechanisms affecting future reproductive attempts (Hatchwell, 1999; Kingma et al., 2010). In chapter4 and chapter5, I demonstrated that Placid Greenbul nestlings are more likely to fledge (Chapter4), and have a higher survival probability once fledged, when reared by larger cooperative groups

Page 91 General discussion

(Chapter5). A positive relationship between number of subordinates and reproductive success was thus detected both before and after fledging. While in both cases the total number of subordinates was a stronger predictor of survival than the number of actual helpers, it should be noted that the analysis on number of helpers was conducted on a smaller dataset. Nonetheless, and despite being only a correlative study, this pattern may suggest that mechanisms other than food provisioning are equally, or even more important in securing offspring survival and increasing reproductive success. These may, for instance, involve improved predator vigilance in larger groups and the contribution to anti-predator behaviour by subordinates, which is coherent with both the low nestling starvation rates and high nest predation rates (as argued previously in the discussion of chapters4&5). Larger groups may however also attract more predators due to being more conspicuous (Krause & Godin, 1995). Yet, it is generally acknowledged that such disadvantage is ample overcome by, among other things, the greater vigilance and better communal defence in larger groups (see Davies et al., 2012 for a full overview of advantages of group living). Frequent observations of conspicuous and noisy greenbul groups in the vicinity of predators may indeed indicate that groups engage in group defence (DVL, Pers. Obs). Nonetheless, it would be necessary to experimentally control for predation risk to fully test this hypothesis. One possible approach would be to wire-cage nests of pairs, 6 cooperative and non-cooperative groups, and compare fledging success between these three groups. In spite of some smaller, but only rarely recorded nest predators (i.e., snakes and rodents) that may enter these wire-cages, such a barrier will successfully obstruct the large and most dominant Placid Greenbul nest predator (African Goshawk Accipiter tachiro; Van de Loock & Bates, 2016 and DVL, Pers. Obs). Although the greenbuls are highly resilient against disturbance from nest visits, mist-net traps and video-recordings with respect to nest attentiveness (DVL, pers. Obs.), it remains to be tested whether breeders and subordinates would also willingly enter the wire cages.

It is important to note that the effect of subordinates on reproductive success and post-fledging survival may be influenced by confounding factors such as the quality of the territory and condition of the breeders (e.g., Legge, 2000b). For instance, pairs on higher quality territories may live in larger groups, simply because they produce more potential recruits for their group. In such scenario, a greater reproductive success may be a more direct consequence of breeding in higher quality territories than from breeding in larger groups. Although these confounding factors are ideally controlled for experimentally (Boland et al., 1997; Brown et al., 1982), they can also be taken into account correlatively using a metric of territory or breeder quality (e.g., Komdeur, 1992), or statistically through pairwise, or within subject comparison (e.g., Legge, 2000b). In the latter case, this means that reproductive success is compared among groups that vary in size and composition across several breeding seasons, but remain on the same territory. I used two of the aforementioned approaches to control for these confounding factors in this thesis (Chapters

Page 92 General discussion

4&5) and therefore conclude that my results on reproductive success and post-fledging survival do not appear to be biased with respect to territory or breeder quality. Nonetheless, I would like to highlight the value of a quantitative measure of habitat quality to be broadly applicable in future analysis (see below for some suggestions on possible metrics).

Nestling mass and clutch size are unaffected by cooperative breeding

In contrast to several other cooperatively breeding species (see Kingma et al. 2010 for review), two direct mechanisms through which subordinates and helpers may increase reproductive success are likely less important in our study system. First, as demonstrated and argued in Chapter4 , the presence of more helpers does not increase nestling condition. This is likely due to a compensatory (load-lightening) response of the breeding pair masking the contributions of helpers. Such load-lightening strategy is expected to affect future reproductive success of the breeders and will be discussed in the next paragraph. Second, laying small clutches is a typical life-history strategy of tropical passerines because clutch size variation is constrained by strong predation pressure (Martin, 2015). Indeed, as demonstrated and discussed in Chapter4, variation in clutch size is extremely small in the Placid Greenbul, and is likely not influenced by the presence of subordinates or helpers. 6 Cooperative breeding may boost future reproductive success

In Chapter4, I demonstrate that breeding females may adopt a load-lightening strategy in response to nestling provisioning behaviour of helpers. Such strategy typically enhances future reproductive success, and hence illustrates that breeding pairs benefit also indirectly from helper presence. In the following paragraphs, I elaborate further on two earlier mentioned drivers of enhanced future reproductive success in the Placid Greenbul : re- nesting and female survival (see discussion chapter4). In the short term, pairs may benefit from cooperative breeding by being able to re-nest faster after success or failure, initiate more nesting attempts and ultimately raise more fledglings within one breeding season (Woxvold & Magrath, 2005; Blackmore & Heinsohn, 2007; Rowley & Russell, 1990; Russell & Rowley, 1988). An important mechanism that could contribute to such short-term benefit is reproductive division of labour, with helpers becoming the primary care-givers of young after fledging (Langen & Vehrencamp, 1999). This allows breeders to initiate a re-nesting attempt much faster than would otherwise be possible when fledged young require care for prolonged periods (Ridley & Raihani, 2008), as is the case for the Placid Greenbul (see Chapter5). Anecdotal observations of incubating breeding females with, at that time, young that depend on adults for food corroborates this hypothesis. Future work is now necessary to quantify the variation in timing and frequency of re-nesting strategies among pair and group-breeding Placid Greenbuls.

Page 93 General discussion

In cooperatively breeding species, females generally experience higher survival rates over multiple years after reducing her reproductive investment during nesting (e.g., Li et al., 2015; Russell et al., 2007; Kingma et al., 2010). Hence, helpers may indirectly increase the reproductive success over a longer period, as females are more likely to survive and breed during future breeding seasons. An effect of load-lightening on female survival may, however, be difficult to detect in our population. First, individuals may remain undetected during a breeding season when their reproductive attempt fails at an early stage or when their active nest remains undiscovered (see also Robinson, 2018 on the challenges that are associated with surveying tropical birds). Second, the Placid Greenbul is a long-lived species that may alternate between breeding as a pair or a cooperative group in different breeding seasons. Out of 38 females observed during more than one breeding season, 24% (n = 9) bred as a pair in one season and as cooperative group in another or vice versa. Due to this variation in cooperative behaviour, and the longevity of the species, any positive influence of helpers on survival may be well masked and subtle to detect. Finally, breeders in larger groups may have lower mortality risk due to the non-specific benefits of group-living, irrespective of the adopted load-lightening strategies. These issues could, to a large extent, be resolved by applying a multi-state encounter analysis that accounts for variation in cooperative behaviour and detectability over multiple breeding 6 seasons.

Is breeding cooperatively costly for the breeding pair?

In spite of the suite of benefits that breeding pairs obtain from cooperating with sub- ordinates and helpers, accepting these extra individuals on their territory can be costly (Dickinson & Hatchwell, 2004). Costs of subordinate and helper presence can, for example, accrue if these individuals parasitize reproduction (Komdeur, 1994; Richardson et al., 2001) or compete with breeders for food or other resources (Gaston, 1978; Brouwer et al., 2006). In some extreme cases, these extra individuals may even reduce, rather than increase, reproductive success (Komdeur, 1994; Legge, 2000a). In our population, subordinates do not appear to parasitize reproduction as extra-pair young were never sired by the immigrants but always by extra-group individuals (Cousseau et al. in prep). Hence, accepting an extra individual on the territory is not per se associated with the cost of a higher likelihood of extra-pair copulations. Costs accrued from competition for food or other resources may give rise to density-dependent mortality (e.g. Brouwer et al., 2006). As such analysis is yet to be conducted (see previous paragraph on challenges and suggestions for survival analysis in this study population), it remains unclear to what extent, and in which way, extra individuals are costly to the breeding pair. Nonetheless and irrespective of the type of costs that are inflicted on the breeding pair, I hypothesize that such costs will be outweighed by the benefits of helping behaviour (see previous section) and passive group augmentation. The passive group augmentation

Page 94 General discussion hypothesis postulates that individuals benefit from by-products associated with the mere presence of other group members (Bergm¨uller et al., 2007). These benefits do not involve costly investment between specific group members, such as nestling provisioning, and are automatically shared by all members (Clutton-Brock, 2002; Bergm¨uller et al., 2007). For instance, mortality risk may be reduced under high predation pressure due to dilution effects and increased vigilance (Sorato et al., 2015). These effects are evident during the post-fledging period where survival probability of fledglings are positively related to group size (Chapter5). In this regard, passive group augmentation may also explain why unrelated and unhelping individuals are accepted in the group.

6.1.3 Does habitat change promote cooperative breeding?

Whereas it appears that Placid Greenbul subordinates attain benefits of delayed dispersal, through helping behaviour and from group membership (as discussed above), it is unlikely that the accumulated fitness from these action outweigh the benefits of breeding inde- pendently (MacColl & Hatchwell, 2004). So, why do individuals then forgo independent breeding to help others? One likely explanation is that independent breeding is constrained, and coopera- 6 tive breeding hence promoted, under particular ecological conditions such as the restricted availability of localized or limited resources (Emlen, 1982). These conditions are often associated with habitat saturation and shortages of available mates and may be caused by high dispersal costs in conjunction with a slow-life history strategy (Russell & Hatchwell, 2001; Hatchwell, 2009). I hypothesize that habitat loss, fragmentation and degradation influence the dispersal costs and the availability of high quality territories and mates, and hence drastically reduce the independent breeding opportunities. Accordingly, cooperative breeding in the Placid Greenbul may be viewed as a plastic response to anthropogenic habitat change (see discussion chapter3). Both experimental and correlative studies on a suite of cooperatively breeding species support the fact that delayed dispersal and helping are plastic responses rather than genetically fixed strategies (see introduction of chapter3 for more elaborate information). Nonetheless, it is currently not yet possible to elucidate the relative contribution of habitat change among other drivers of cooperative breeding in this population. For instance, high predator pressure or slow rates of high quality territory turn-over may also trigger cooperative breeding through benefits of cooperation rather than constraints on independent breeding (Shen et al., 2017). The Placid Greenbul appears to be a social species foraging in flocks throughout its global range (Fry & Keith, 2000), including several areas in Kenya and Tanzania (see Borghesio & Laiolo, 2004; Dinesen, 1997 and Newmark B., Amakobe B., Borghesio L. & Dinesen L. pers. communication). At present however, a formal assessment of the cooperative breeding strategy of this species has not been

Page 95 General discussion

conducted in other areas than the Taita Hills. Only by comparing cooperative behaviour among areas with different environmental conditions can we resolve whether cooperative breeding occurs due to local stimuli of anthropogenic habitat change, or whether it is due to other factors.

6.2 Effect of anthropogenic habitat change

Earlier work detected a strong impact of anthropogenic habitat change on the Placid Greenbul, including its fragmented genetic population structure (Callens et al., 2011), an association between developmental stress and degree of habitat degradation (Lens et al., 1999), spatiotemporal variation in nest predation risk (Spanhove et al., 2014) and variation in occurrence of the species across the landscape (Lens et al., 2002). In spite of these consequences, comparisons of the genetic and demographic population structure over time revealed a recent increase (1996 - 2010) in population size and among-fragment connectivity and also reported stable adult survival rates (yearly apparent survival Φ > 0.69; Husemann et al., 2015). Though such evidence may suggest that Placid Greenbuls are able to cope with habitat fragmentation and degradation, I here evaluate further whether habitat 6 change may affect, or interact with, the cooperative breeding strategy of the species. I show in chapter3&4 that the natal dispersal strategy, incidence of cooperation, most group characteristics, maternal provisioning rates and reproductive success do not vary between forest fragments. Only within-group sex-ratio of subordinates (Chapter3) and egg investment of cooperatively breeding females (Chapter4) varies across fragments (See the respective chapters for a more elaborate discussion on the possible causes and consequences of the similarities and differences across fragments for these traits). Given that most of the tested traits and processes associated with cooperative breeding do not vary across the forest fragments, it may now be tempting to conclude that anthropogenic habitat change does not interfere with the cooperative breeding strategy of this species. However, such conclusion may be premature for three reasons. (i) No large and intact forest fragment could be included in the set of forest fragments under study. Without such pristine fragment as a ’baseline’ reference for comparison, it is not possible to judge whether the cooperative breeding strategy of this species is affected by habitat change or not (see discussion Chapter3). (ii) In spite of the overall variation in degree of degradation, both large fragments (Ngangao and Chawia) represent unique and non- replicated forest conditions. Hence, disentangling the actual influence of forest degradation among other forest specific environmental conditions may be difficult and problematic. (iii) The categories of habitat degradation at fragment level represent generalizations of the expected conditions at territory level. While such generalisations are valuable to detect general patterns, they fall short of detecting variation within fragments. Yet, the conditions that pairs or groups experience at their territory may overlap across fragments.

Page 96 General discussion

In this regard, an important question to be raised is what defines habitat quality for a Placid Greenbul. Possible resolutions for these three concerns are discussed in the next two sections.

6.2.1 Increasing spatial replication

Absence of reference and replicated forest conditions may be resolved in the future by enlarging the suit of forest fragments under study. Within the global range of the Placid Greenbul (see Fig. 1.1), the forests of the Taita Hills can arguably be considered among the more severely degraded areas (Burgess et al., 2007b). Whereas the Mbololo forest fragment is the largest and least degraded fragment in the region, and may constitute a first noteworthy candidate to enlarge the range of forest fragments under study, this fragment is still part of the Taita Hills’ degraded and fragmented ecosystem and only covers 220ha (Aerts et al., 2011). I therefore consider it highly relevant to include other forested areas outside the Taita Hills area in future studies when addressing the influence of habitat change on cooperative breeding strategies. Other forests of the Eastern Arc Mountains in Tanzania may offer such conditions, of which the East Usambara section may be particularly interesting. These mountains support seven forest fragments ranging in size from 0.2 to 521 ha and an adjacent large block of continuous forest of 7571 ha, and 6 have been subject to ongoing avian research for more than two decades (Newmark, 2002, 2006; Newmark & Stanley, 2011; Newmark et al., 2017).

6.2.2 The importance of a reliable metric of habitat quality

Another approach to improve our understanding of the effect of habitat change, is to gain a better understanding of habitat quality at the fragment, and smaller (i.e., territory) scales. At present, we consider habitat quality, as experienced by birds, to correspond with fragment-specific levels of forest degradation, measured by changes in vegetation structure (Thijs, 2015; Aerts et al., 2011; Wilder et al., 1998). This assumption is based on previously detected relationships between levels of habitat degradation and developmental stress (Lens et al., 1999), as well as other metrics of body condition in this, and other forest restricted bird species (e.g., Grubb, 2006). While this provides a rough generalisation that is valuable for detecting broad patterns, it neglects within-fragment variation in habitat conditions and may not completely correspond to what birds experience as habitat quality. It is hence necessary to increase our understanding of what defines good, or bad quality habitat, and be able to measure this at a smaller scale than at fragment level. Measures of avian habitat quality can be divided broadly into (i) direct measurements of the environment (e.g., the abundance of predators, resources and nesting opportunities) (ii) fitness-related, or indirect, measurements (e.g., individual measures of body condition and measures of population viability; Johnson, 2007; Minias, 2015. In the following paragraph,

Page 97 General discussion

I will discuss those measures that may be most appropriate for our study system based on the species’ ecology and logistical implementation, whereas a complete and extensive overview can be found in Johnson’s review on avian habitat quality (Johnson, 2007).

Direct measurements of habitat quality

Direct measurements of the environment have been shown to yield highly appropriate and valuable metrics (e.g., Komdeur, 1992), but are based on our assumptions on what birds value in their environment (Johnson, 2007). As an insectivore that experiences high nest predation rates, greenbuls may for instance value habitat that is associated with many safe nesting opportunities, low predator abundance, high arthropod abundance, or a combination of these factors. Arthropod abundance may be relatively straightforward to quantify using a variety of field techniques, whereas predation risk may be more difficult to comprehend. Appropriate parameters that may influence predation risk include the degree of nest concealment, which is positively associated with nest-site choice and inversely related to predation risk (Chapter2 and Spanhove et al., 2014), the amount of dead wood in the area, which is negatively associated with nest-site choice as it may shelter predators (Chapter2) and the position of a territory relative to the fragment edge, which may also 6 affect predation risk (Spanhove et al., 2014).

Indirect measurements of habitat quality

Morphological metrics (e.g., deviations from bilateral symmetry), analysis of growth bars on feathers (ptilochronology) and measurements of hormone levels or blood metabolites (e.g., Glucocorticoid levels and oxidative stress) have all been considered as indirect indicators of habitat quality (Johnson, 2007). While these metrics have demonstrated a clear relationship with habitat quality in several species and contexts (e.g., Lens et al., 1999; Anci˜aes& Marini, 2000), they occasionally fail to support similar relationship in others (e.g., Le Tortorec et al., 2012; Elderbrock et al., 2012; Helle et al., 2011). Although these metrics should thus be interpreted cautiously, they may be highly valuable proxies to detect small scale variation in habitat quality. Habitat quality may correlate with the amount of environmental stress suffered during ontogeny. This can be inferred from levels of fluctuating asymmetry, which are random deviations from left-right symmetry in bilateral traits and a proxy for developmental stability (Parsons, 1992). A positive relation between habitat degradation and levels of fluctuating asymmetry in forest birds has earlier been demonstrated in our study area (Lens et al., 1999), as well as in other degraded areas (e.g., Anci˜aes & Marini, 2000), but not in all (e.g., Helle et al., 2011). Habitat quality may also depend on food abundance and availability, which influences individual nutrient uptake. Nutrient uptake can be inferred from the rate of daily feather growth and is visualised in growth bars, which are series of

Page 98 General discussion alternating bands depicting 24h growth (Grubb, 1989, 2006). The relationship between growth bar width and habitat quality has been shown in some studies (Grubb & Yosef, 1994; Carlson, 1998; Vangestel et al., 2010), but not all (Le Tortorec et al., 2012), and may be obscured by other factors such as individual condition, age or genotype (Saino et al., 2012; Matysiokov´a& Remeˇs, 2010; Gienapp & Meril¨a, 2010; Elderbrock et al., 2012). In our study area, growth bar width has earlier been shown to correlate with fragment size in the similarly forest dependent White-starred Robin Pogonocichla stellata (Grubb, 2006), which may be indicative of reduced food availability in smaller fragments. Also endocrine techniques may be useful to infer habitat quality. Among these, glucocorticoid (CORT or ”stress” hormone) levels are physiological indices of individual condition that are known to correspond to habitat quality (reviewed by Walker et al., 2005; Wikelski & Cooke, 2006). Because CORT is part of the physiological response to environmental and social challenges, high baseline CORT levels are considered indicative for a lowered body condition and habitat quality (e.g., Marra & Holberton, 1998; Jenni- Eiermann et al., 2008; Williams et al., 2008; Crule et al., 2017). However, these relationships are not always considered to be directly causal as CORT levels may covary with other factors (e.g., stage during nesting; Bonier et al., 2009). The oxidative status of an organism may constitute another interesting proxy of habitat quality (e.g., Van de Crommenacker et al., 2011). This status is defined by the oxidant-antioxidant balance, which depicts the rate 6 between the production of harmful oxidants and the neutralizing antioxidants. Oxidants are reactive oxygen species that are generated as a by-product of energy consumption, whereas antioxidants can be extracted from food or produced metabolically to neutralize these oxidants (Balaban et al., 2005; Monaghan et al., 2009). Increased oxidant levels are generally associated with greater energy expenditure, for instance when individuals work harder to secure scarce resources or are faced with greater predator pressure (Costantini, 2008; Pinya et al., 2016; Janssens & Stoks, 2013). Hence, changes in oxidative status can be expected when food becomes more scarce, or predation pressure increases, which in turn, may point towards lower quality habitat.

6.2.3 Cooperative breeding may moderate effects of anthropogenic habitat change

Though it may be difficult to judge the extent to which habitat change interferes with cooperative breeding in this species, some findings do indicate that this breeding strategy may moderate, and even buffer, consequences of habitat change.

Cooperative breeding may moderate nest-site selection strategies

Nest predation is the primary cause of nest failure in Placid Greenbuls in our study area, and with up to 70% predation strongly drives variation in reproductive success. One

Page 99 General discussion

possible mechanism through which birds can reduce nest predation rates is by selecting safe nest-sites that are less prone to predation. In chapter2 I show that the set of habitat features associated with nest-site selection varies between the two largest and differently degraded forest fragments (Ngangao and Chawia). However, neither in Ngangao, nor in Chawia did nest-site selection correlate with reproductive success, which may question the adaptive value of nest-site selection strategies for reducing nest predation. As I argue in chapter2, it is however possible that nest-site selection strategies may be moderated by other factors, such as individual behaviour against predators (Eggers et al., 2008; Remeˇs, 2005). Unfortunately, however, information on the associated breeding pair or cooperative group was not collected at that time, which does not allow a formal comparison between nest-site selection strategies of pairs and groups. In spite of this, the fact that larger groups have higher reproductive success in all fragments, and hence irrespective of differences in nest-site selection strategies between fragments (chapter4), supports the idea that cooperative breeding may moderate nest-site selection strategies through group anti-predator behaviour. Along these lines, predation risk has been proposed as an important selective force in the evolution of cooperative breeding among birds (Doerr & Doerr, 2006) and fish (Heg et al., 2004), and even been demonstrated to shape group composition and complexity (Groenewoud et al., 2016). Future work should not only focus 6 on identifying, and quantifying, the dominant predators in the study area (as is argued in discussion chapter2), but should also focus on whether, and how, pairs and cooperatively breeding groups may differ in their choice of nesting sites. Such information may offer important clues on the relative influence of nest-site selection, group-living and predator communities on predation risk in the Taita Hills.

Cooperative breeding may buffer against demographic and environmental stochasticity and small population effects

Cooperative breeding may buffer the negative consequences of environmental or demo- graphic stochasticity, as the surplus of subordinates may serve as a pool of reserve breeders (Walters et al., 2004). This hypothesis however depends on two assumptions. Firstly, subordinates should readily take over empty territories as soon as they arise. For example, empty Seychelles Warbler territories resulting from experimental breeder removal were, indeed, immediately claimed by subordinates from adjacent territories (Eikenaar et al., 2008). Also, removal experiments on a sympatric, albeit non-cooperatively breeding species from the Taita Hills (White-starred Robin) yielded similar results whereby floaters, rather than subordinates, rapidly settled on recently cleared territories (Githiru & Lens, 2006). Secondly, density of breeders and territories should rapidly be restored when population- wide perturbation cause severe loss of breeders. For example, territories of the Galapagos Mockingbird (Mimus parvulus) that became available when the breeders deceased after sequential climatically harsh years, where quickly filled by birds residing on other territories

Page 100 General discussion

(Curry & Grant, 1989). Similarly, when a disease killed 45% of the Florida Scrub-Jay breeders, all vacant territories were filled within two years by subordinates (Woolfenden & Fitzpatrick, 1985). The large number of readily available subordinates (based on low Ne/Nc ratios; Husemann et al., 2015 and high incidence of group living; Chapter3), indicate that there is a surplus of replacement breeders that may buffer demographic and environmental stochasticity. This mechanism may be particularly relevant in the smaller fragments, where filling up breeding vacancies would otherwise only be possible by immigrants. Yet, breeder removal experiments, or severe natural perturbations are now necessary to reveal whether these subordinates may indeed rapidly claim newly available territories. Small populations may suffer from the accumulation of deleterious alleles due to inbreeding (Keller & Waller, 2002). Cooperatively breeding species are predicted to be more vulnerable to inbreeding due to both delayed and short-distance dispersal, which may result in close-kin neighbourhoods (Koenig & Haydock, 2004; Nichols, 2017). Accordingly, it has been suggested that cooperatively breeding birds have evolved mechanisms to overcome the negative effects of inbreeding, such as sex-biased dispersal (Greenwood, 1980; Pusey, 1987; Koenig & Haydock, 2004; but see Riehl & Stern, 2015). Although habitat fragmentation may alter dispersal strategies and thus cause inbreeding, this does not appear to be the case in our study population as genetic analysis registered dispersal and 6 gene flow among forest fragments (Husemann et al., 2015; Vangestel et al., 2013). Hence, our population is not likely suffering from (severe) inbreeding.

Cooperative breeding may buffer against habitat degradation

As cooperatively breeding birds occur more frequently in unpredictable environments where annual rainfall is low and highly variable through time (Jetz & Rubenstein, 2011), it has been suggested that this breeding strategy may serve to buffer against harsh environmental conditions (Shen et al., 2017). Accordingly, the ’hard life hypothesis’ predicts that cooperative breeding may allow successful reproduction when it would otherwise be impossible to raise young without the assistance of helpers (Koenig et al., 2011). In a number of studies on species from highly unpredictable environments, cooperative breeding only affected reproductive success during harsh years, which lends support to this hypothesis (Ebensperger et al., 2014; Can´ario et al., 2004; Covas et al., 2008). However, in other species from similarly unpredictable environments, reproductive success of cooperative breeders is more consistent across both harsh and benign years (Rubenstein, 2007; Legge, 2000b). These inconsistencies may, in part, be due to variation in breeder investment strategies, which may subsequently mask variation in reproductive benefits across variable years. Breeders with helpers may decrease, maintain, or increase their investment in response to environmental conditions (Guindre-Parker & Rubenstein, 2018; Langmore et al., 2016). For instance, cooperatively breeding Superb Fairy-wren females

Page 101 General discussion

laid larger eggs under harsher conditions, but smaller eggs under more benign conditions (Langmore et al., 2016). Similarly, male Superb starlings (Lamprotornis superbus) with helpers reduce provisioning rates only in benign conditions, but maintained investment levels during harsh years (Guindre-Parker & Rubenstein, 2018). Given this flexibility in investment strategies, one can argue that breeders may buffer variation in reproductive success resulting from harsh environmental conditions by a matching response strategy. To what extent can we relate environmental variation caused by climatic variability with human-induced environmental variation? And can we hence argue that anthropogenic habitat change may create harsh conditions that can be buffered by flexible investment strategies in our study species? Typical forest dependent bird species, such as the Placid Greenbul, are adapted to stable climatological conditions of the forest understory. Hence, when habitat fragmentation and degradation cause abiotic changes at forest edge and interior, birds may experience harsher conditions due to changes in predator pressure, the availability of food resources and changes in micro-climatic (Chalfoun et al., 2002). A study on Rufous Treecreepers (Climacteris rufus) revealed that breeding females from fragmented habitat maintained their investment when helpers were present (additive strategy), while females from continuous habitat reduced investment (load-lightening strategy; Luck, 2002). Loss in reproductive success was mitigated by the female’s additive strategy in 6 the fragmented habitat, which supports the hypothesis that investment strategies of cooperative breeders may buffer the negative effects of habitat degradation (Luck, 2002). Along these lines, I showed that breeding females vary their egg investment according to variation in degree of habitat degradation (chapter4). And may hence constitute a strategy to buffer variation in reproductive success among degraded landscapes. It is however necessary to compare life-time reproductive success across both pair and cooperatively breeding females from differently degraded forests to test this hypothesis. Nonetheless, I hypothesize that such mechanism is currently under evaluated in literature and the importance thereof hence under appreciated.

6.3 General conclusions

This thesis revealed new information regarding the cooperative breeding strategy of the Placid Greenbul under anthropogenic habitat change, of which the main conclusions are summarized below.

1. The majority of fledglings delay dispersal and remain in the natal group for at least one, and up to four years. About half of the retained individuals contribute to nestling provisioning and as such accrue indirect fitness benefits. Delayed dispersal appears to be promoted by the combined benefits and constraints of philopatry and dispersal, respectively, and further mediated by kin selection benefits. The relative

Page 102 General discussion

importance of each driver requires further investigation.

2. Cooperative breeding benefits breeders both during and after nesting. In general, breeding in larger groups is associated with a lower probability of nest predation, and higher survival of fledglings. This is likely due to increased vigilance and anti- predator strategies associated with increasing group size, though this should yet be confirmed experimentally. More specifically, females may adjust their investment in eggs and nestlings in relation to subordinates, helpers and local environmental conditions. This may, depending on which strategy the female choses, benefit the brood, herself or a combination of both. Elucidating these benefits requires more elaborated studies on condition and survival of nestlings, and on re-nesting strategies and survival of females.

3. In general, there is no evidence that anthropogenic habitat change interferes with cooperative breeding, as few investigated relationships varied across different forest fragments. Yet, information from populations in more pristine forest conditions, as well as a better understanding of habitat quality, both among and within forest frag- ments, are now needed to confirm these observations. Then, further (experimental) investigation is required to confirm if, and to what extent, this breeding strategy 6 may actually buffer populations against consequences of habitat change, as some observations suggest this may indeed be the case.

Page 103 Olliver, Adam, Peter and Laurence, a.k.a Team Fantastic. (December 2015) CHAPTER 7

Bibliography

Aben, J., Adriaensen, F., Thijs, K. W., Pellikka, P., Siljander, M., Lens, L. & Matthysen, E. (2012). Effects of matrix composition and configuration on forest bird movements in a fragmented Afromontane biodiversity hot spot. Conservation, 15(6):658–668.

Aben, J., Bocedi, G., Palmer, S. C. F., Pellikka, P., Strubbe, D., Hallmann, C., Travis, J. M. J., Lens, L., Matthysen, E. & Wilson, J. (2016). The importance of realistic dispersal models in conservation planning: application of a novel modelling platform to evaluate management scenarios in an Afrotropical biodiversity hotspot. Journal of Applied Ecology, 53(4):1055–1065. 7 Aben, J., Strubbe, D., Adriaensen, F., Palmer, S. C. F., Travis, J. M. J., Lens, L. & Matthysen, E. (2014). Simple individual-based models effectively represent Afrotropical forest bird movement in complex landscapes. Journal of Applied Ecology, 51(3):693–702.

Aerts, R., Thijs, K. W., Lehouck, V., Beentje, H., Bytebier, B., Matthysen, E., Gulinck, H., Lens, L. & Muys, B. (2011). Woody plant communities of isolated Afromontane cloud forests in Taita Hills, Kenya. Plant Ecology, 212(4):639–649.

Anci˜aes,M. & Marini, M. (2000). The effects of fragmentation on fluctuating asymmetry in passerine birds of Brazilian tropical forests. Journal of Applied Ecology, 37(6):1013–1028.

Andersson, M., Wiklund, C. G. & Rundgren, H. (1980). Parental defence of offspring: A model and an example. Animal Behaviour, 28(2):536–542.

Andr´en,H. (1994). Effects of Habitat Fragmentation on Birds and Mammals in Landscapes With Different Proportions of Suitable Habitat - a Review. Oikos, 71(3):355–366.

Arnold, K., Goldizen, A. & Owens, I. (2005). Division of labour within cooperatively breeding groups. Behaviour, 142(11):1577–1590.

Baglione, V., Canestrari, D., Chiarati, E., Vera, R. & Marcos, J. M. (2010). Lazy group members are substitute helpers in carrion crows. Proceedings. Biological sciences / The Royal Society, 277(1698):3275– 82.

Baglione, V., Canestrari, D., Marcos, J. M. & Ekman, J. (2003). Kin selection in cooperative alliances of carrion crows. Science, 300(5627):1947–1949.

Page 105 Bibliography

Baglione, V., Canestrari, D., Marcos, J. M. & Ekman, J. (2006). Experimentally increased food resources in the natal territory promote offspring philopatry and helping in cooperatively breeding carrion crows. Proceedings. Biological sciences / The Royal Society, 273(1593):1529–1535.

Baglione, V., Canestrari, D., Marcos, J. M., Griesser, M. & Ekman, J. (2002). History, environment and social behaviour: experimentally induced cooperative breeding in the carrion crow. Proceedings. Biological sciences / The Royal Society, 269(1497):1247–51.

Baglione, V., Marcos, J. M., Canestrari, D., Griesser, M., Andreotti, G., Bardini, C. & Bogliani, G. (2005). Does year-round territoriality rather than habitat saturation explain delayed natal dispersal and cooperative breeding in the carrion crow? Journal of Animal Ecology, 74(5):842–851.

Baguette, M., Blanchet, S., Legrand, D., Stevens, V. M. & Turlure, C. (2013). Individual dispersal, landscape connectivity and ecological networks. Biological Reviews, 88(2):310–326.

Balaban, R. S., Nemoto, S. & Finkel, T. (2005). Mitochondria, oxidants, and aging. Cell, 120(4):483–495.

Banks, S. C., Piggott, M. P., Stow, A. J. & Taylor, A. C. (2007). Sex and sociality in a disconnected world: a review of the impacts of habitat fragmentation on animal social interactionsThis review is one of a series dealing with some aspects of the impact of habitat fragmentation on animals and plants. This series is . Canadian Journal of Zoology, 85(10):1065–1079.

Banks-Leite, C., Ewers, R. M. & Metzger, J. P. (2010). Edge effects as the principal cause of area effects on birds in fragmented secondary forest. Oikos, 119(6):918–926.

Barlow, J., Lennox, G. D., Ferreira, J., Berenguer, E., Lees, A. C., Nally, R. M., Thomson, J. R., Ferraz, S. F. D. B., Louzada, J., Oliveira, V. H. F., Parry, L., Ribeiro De Castro Solar, R., Vieira, I. C., Araga˜o, L. E., Begotti, R. A., Braga, R. F., Cardoso, T. M., Jr, R. C. D. O., Souza, C. M., Moura, N. G., Nunes, S. S., Siqueira, J. V., Pardini, R., Silveira, J. M., Vaz-De-Mello, F. Z., Veiga, R. C. S., Venturieri, A. & Gardner, T. A. (2016). Anthropogenic disturbance in tropical forests can double biodiversity loss from deforestation. Nature, 535(7610):144–147. 7 Barton, K. (2016). MuMIn: Multi-Model Inference. R package version 1.15.6. https://CRAN.R- project.org/package=MuMIn.

Bates, D., M¨achler, M., Bolker, B. & Walker, S. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1).

Bennun, L., Dranzoa, C. & Pomeroy, D. (1996). The forest birds of Kenya and Uganda. Journal of East African Natural History, 85(1):23–48.

Benson, T. J., Anich, N. M., Brown, J. D. & Bednarz, J. C. (2009). Swainson’s Warbler Nest-Site Selection in Eastern Arkansas. The Condor, 111(4):694–705.

Benson, T. J., Brown, J. D. & Bednarz, J. C. (2010). Identifying predators clarifies predictors of nest success in a temperate passerine. Journal of Animal Ecology, 79(1):225–234.

Bergm¨uller, R., Johnstone, R. A., Russell, A. F. & Bshary, R. (2007). Integrating cooperative breeding into theoretical concepts of cooperation. Behavioural processes, 76(2):61–72.

Blackmore, C. J. & Heinsohn, R. (2007). Reproductive success and helper effects in the cooperatively breeding grey-crowned babbler. Journal of Zoology, 273(3):326–332.

Boivin, N. L., Zeder, M. A., Fuller, D. Q., Crowther, A., Larson, G., Erlandson, J. M., Denham, T. & Petraglia, M. D. (2016). Ecological consequences of human niche construction: Examining long-term anthropogenic shaping of global species distributions. Proceedings of the National Academy of Sciences, 113(23):6388–6396.

Boland, C. R. J., Heinsohn, R. & Cockburn, A. (1997). Experimental manipulation of brood reduction and parental care in cooperatively breeding white-winged . Journal of Animal Ecology, 66(5):683– 691.

Page 106 Bibliography

Bonier, F., Moore, I. T., Martin, P. R. & Robertson, R. J. (2009). The relationship between fitness and baseline glucocorticoids in a passerine bird. General and Comparative Endocrinology, 163(1-2):208–213.

Bonte, D., Van Dyck, H., Bullock, J. M., Coulon, A., Delgado, M., Gibbs, M., Lehouck, V., Matthysen, E., Mustin, K., Saastamoinen, M., Schtickzelle, N., Stevens, V. M., Vandewoestijne, S., Baguette, M., Barton, K., Benton, T. G., Chaput-Bardy, A., Clobert, J., Dytham, C., Hovestadt, T., Meier, C. M., Palmer, S. C., Turlure, C. & Travis, J. M. (2012). Costs of dispersal. Biological Reviews, 87(2):290–312.

Borghesio, L. & Laiolo, P. (2004). Seasonal foraging ecology in a forest avifauna of northern Kenya. Journal of Tropical Ecology, 20(2):145–155.

Borgmann, K. L. & Conway, C. J. (2015). The Nest-concealment hypothesis: new insights from a comparative analysis. Wilson Journal of Ornithology, 4(4):646–660.

Bowers, E. K., Sakaluk, S. K. & Thompson, C. F. (2012). Experimentally increased egg production constrains future reproduction of female house wrens. Animal Behaviour, 83(2):495–500.

Bowman, R. & Woolfenden, G. E. (2002). Nest site selection by Florida Scrub-Jays in natural and human-modified habitats. The Wilson Bulletin, 114(1):128–135.

Breed, M. D. (2014). Kin and nestmate recognition: The influence of W. D. Hamilton on 50years of research. Animal Behaviour, 92:271–279.

Bregman, T. P., Lees, A. C., MacGregor, H. E. A., Darski, B., de Moura, N. G., Aleixo, A., Barlow, J. & Tobias, J. A. (2016). Using avian functional traits to assess the impact of land-cover change on ecosystem processes linked to resilience in tropical forests. Proceedings of the Royal Society B: Biological Sciences, 283(1844).

Bregman, T. P., Sekercioglu, C. H. & Tobias, J. A. (2014). Global patterns and predictors of bird species responses to forest fragmentation: Implications for ecosystem function and conservation. Biological Conservation, 169:372–383. Breininger, D. R. (1999). Florida Scrub-Jay demography and dispersal in a fragmented landscape. The 7 Auk, 116(2):520–527.

Brooks, T. M., Mittermeier, R. a., Mittermeier, C. G., Da Fonseca, G. a. B., Rylands, A. B., Konstant, W. R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G. & Hilton-Taylor, C. (2002). Habitat Loss and Extinction in the Hotspots of Biodiversity. Conservation Biology, 16(4):909–923.

Brouwer, L., Richardson, D. S., Eikenaar, C. & Komdeur, J. (2006). The role of group size and environmental factors on survival in a cooperatively breeding tropical passerine. Journal of Animal Ecology, 75(6):1321–1329.

Brouwer, L., Richardson, D. S. & Komdeur, J. (2012). Helpers at the nest improve late-life offspring performance: evidence from a long-term study and a cross-foster experiment. PloS one, 7(4):e33167.

Brouwer, L., Van de Pol, M. & Cockburn, A. (2014). The role of social environment on parental care: Offspring benefit more from the presence of female than male helpers. Journal of Animal Ecology, 83(2):491–503.

Brown, J. H. & Kodric-Brown, A. (1977). Turnover Rates in Insular Biogeography: Effect of Immigration on Extinction. Ecology, 58(2):445–449.

Brown, J. L. (1987). Helping and communal breeding in birds. Princeton University Press, Princeton, NJ.

Brown, J. L., Brown, E. R., Brown, S. D. & Dow, D. D. (1982). Helpers: effects of experimental removal on reproductive success. Science, 215(4531):421–422.

Browning, L. E., Patrick, S. C., Rollins, L. a., Griffith, S. C. & Russell, a. F. (2012). Kin selection, not group augmentation, predicts helping in an obligate cooperatively breeding bird. Proceedings of the Royal Society B: Biological Sciences, 279(1743):3861–3869.

Page 107 Bibliography

Bruintjes, R. & Taborsky, M. (2011). Size-dependent task specialization in a cooperative cichlid in response to experimental variation of demand. Animal Behaviour, 81(2):387–394.

Bullock, J. M., Bonte, D., Pufal, G., da Silva Carvalho, C., Chapman, D. S., Garc´ıa,C., Garc´ıa,D., Matthysen, E. & Delgado, M. M. (2018). Human-Mediated Dispersal and the Rewiring of Spatial Networks. Trends in Ecology & Evolution, xx:1–13.

Burgess, N. D., Balmford, A., Cordeiro, N. J., Fjelds??, J., K??per, W., Rahbek, C., Sanderson, E. W., Scharlemann, J. P. W., Sommer, J. H. & Williams, P. H. (2007a). Correlations among species distributions, human density and human infrastructure across the high biodiversity tropical mountains of Africa. Biological Conservation, 134(2):164–177.

Burgess, N. D., Butynski, T. M., Cordeiro, N. J., Doggart, N. H., Fjelds˚a,J., Howell, K. M., Kilahama, F. B., Loader, S. P., Lovett, J. C., Mbilinyi, B., Menegon, M., Moyer, D. C., Nashanda, E., Perkin, A., Rovero, F., Stanley, W. T. & Stuart, S. N. (2007b). The biological importance of the Eastern Arc Mountains of Tanzania and Kenya. Biological Conservation, 134(2):209–231.

Burivalova, Z., ekerciolu, C¸ . H. & Koh, L. P. (2014). Thresholds of logging intensity to maintain tropical forest biodiversity. Current Biology, 24(16):1893–1898.

Burnham, K. P. & Anderson, D. R. (2002). Model Selection and Multimodel Inference. Springer-Verlag New York, Inc., New York, 2nd edition.

Bytebier, B. (2001). Taita Hills Biodiversity Project Report. Technical report, National Museums of Kenya, Nairaboi.

Caffrey, C. (2000). Correlates of reproductive success in cooperatively breeding western American crows: if helpers help, it’s not by much. The Condor, 102(August):333–341.

Callens, T. (2012). Genetic and demographic signatures of population fragmentation in a cooperatively breeding bird from south-east Kenya. Ph.D. thesis, Ghent University, Ghent, Belgium.

Callens, T., Galbusera, P., Matthysen, E., Durand, E. Y., Githiru, M., Huyghe, J. R. & Lens, L. (2011). 7 Genetic signature of population fragmentation varies with mobility in seven bird species of a fragmented Kenyan cloud forest. Molecular Ecology, 20(9):1829–1844.

Can´ario,F., Matos, S. & Soler, M. (2004). Environmental Constraints and Cooperative Breeding in the Azure-Winged Magpie. The Condor, 106(3):608–617.

Canestrari, D., Marcos, J. M. & Baglione, V. (2011). Helpers at the nest compensate for reduced maternal investment in egg size in Carrion Crows. Journal of evolutionary biology, 24(9):1870–8.

Carlson, A. (1998). Territory Quality and Feather Growth in the White-Backed Woodpecker Dendrocopos leucotos. Journal of Avian Biology, 29(2):205–207.

Carlson, A. & Hartman, G. (2001). Tropical forest fragmentation and nest predation an experimental study in an Eastern Arc montane forest , Tanzania. Biodiversity and Conservation, 10(7):1077–1085.

Carranza, J., Polo, V., Valencia, J., Mateos, C. & de la Cruz, C. (2008). How should breeders react when aided by helpers? Animal Behaviour, 75(4):1535–1542.

Chalfoun, A. D. & Martin, T. E. (2007). Assessments of habitat preferences and quality depend on spatial scale and metrics of fitness. Journal of Applied Ecology, 44(5):983–992.

Chalfoun, A. D. & Martin, T. E. (2009). Habitat structure mediates predation risk for sedentary prey: Experimental tests of alternative hypotheses. Journal of Animal Ecology, 78(3):497–503.

Chalfoun, A. D. & Schmidt, K. A. (2012). Adaptive Breeding-Habitat Selection : Is It for the Birds ? The Auk, 129(4):589–599.

Chalfoun, A. D., Thompson, F. R. & Ratnaswamy, M. J. (2002). Nest predators and fragmentation: A review and meta-analysis. Conservation Biology, 16(2):306–318.

Page 108 Bibliography

Chege, J. & Bytebier, B. (2005). Vegetation Structure of Four Small Forest Fragments in Taita Hills, Kenya. Journal of East African Natural History, 94(1):231–234.

Clutton-Brock, T. (2002). Breeding together: kin selection and mutualism in cooperative vertebrates. Science (New York, N.Y.), 296(5565):69–72.

Clutton-Brock, T. (2009). Cooperation between non-kin in animal societies. Nature, 462(7269):51–57.

Clutton-Brock, T. H. (1985). Sex ratio variation in birds. Ibis, 128(3):317–329.

Cockburn, A. (1998). Evolution of helping behavior in cooperatively breeding birds. Annual Review of Ecology and Systematics, 29(1):141–177.

Cockburn, A. (2003). Cooperative breeding in oscine passerines: does sociality inhibit speciation? Proceedings. Biological sciences / The Royal Society, 270(1530):2207–2214.

Cockburn, A. (2006). Prevalence of different modes of parental care in birds. Proceedings of the Royal Society B: Biological Sciences, 273(1592):1375–1383.

Cockburn, A., Sims, R. A., Osmond, H. L., Green, D. J., Double, M. C. & Mulder, R. A. (2008). Can we measure the benefits of help in cooperatively breeding birds: The case of superb fairy-wrens Malurus cyaneus? Journal of Animal Ecology, 77(3):430–438.

Colchero, F., Jones, O. R. & Rebke, M. (2012). BaSTA: An R package for Bayesian estimation of age-specific survival from incomplete mark-recapture/recovery data with covariates. Methods in Ecology and Evolution, 3(3):466–470.

Conway, C. J. & Martin, T. E. (2000). Evolution of passerine incubation behavior: influence of food, temperature, and nest predation. Evolution, 54(2):670–685.

Cooper, C. B. & Walters, J. R. (2002). Experimental evidence of disrupted dispersal causing decline of an Australian passerine in fragmented habitat. Conservation Biology, 16(2):471–478. 7 Cooper, C. B., Walters, J. R. & Ford, H. A. (2002). Effects of remnant size and connectivity on the response of Brown Treecreepers to habitat fragmentation. Emu, 102:249–256.

Cornell, K. L. & Donovan, T. M. (2010). Scale-Dependent Mechanisms of Habitat Selection for a Migratory Passerine: An Experimental Approach. The Auk, 127(4):899–908.

Cornwallis, C. K., West, S. a. & Griffin, a. S. (2009). Routes to indirect fitness in cooperatively breeding vertebrates: Kin discrimination and limited dispersal. Journal of Evolutionary Biology, 22(12):2445– 2457.

Costantini, D. (2008). Oxidative stress in ecology and evolution: Lessons from avian studies. Ecology Letters, 11(11):1238–1251.

Covas, R., Deville, A. S., Doutrelant, C., Spottiswoode, C. N. & Gr´egoire,A. (2011). The effect of helpers on the postfledging period in a cooperatively breeding bird, the Sociable Weaver. Animal Behaviour, 81(1):121–126.

Covas, R., Doutrelant, C. & du Plessis, M. A. (2004). Experimental evidence of a link between breeding conditions and the decision to breed or to help in a colonial cooperative bird. Proceedings of the Royal Society B: Biological Sciences, 271(1541):827–832.

Covas, R., du Plessis, M. a. & Doutrelant, C. (2008). Helpers in colonial cooperatively breeding sociable weavers Philetairus socius contribute to buffer the effects of adverse breeding conditions. Behavioral Ecology and Sociobiology, 63(1):103–112.

Covas, R. & Griesser, M. (2007). Life history and the evolution of family living in birds. Proceedings of the Royal Society B: Biological Sciences, 274(1616):1349–1357.

Page 109 Bibliography

Cox, W. A., Thompson, F. R., Cox, A. S. & Faaborg, J. (2014). Post-fledging survival in passerine birds and the value of post-fledging studies to conservation. Journal of Wildlife Management, 78(2):183–193.

Crampton, L. H. & Sedinger, J. S. (2011). Nest-habitat selection by the Phainopepla: Congruence across spatial scales but not habitat types. Condor, 113(1):209–222.

Crick, H. P. (1992). Load-lightening in cooperatively breeding birds and the cost of reproduction. Ibis.

Crooks, K. & Soul´e,M. (1999). Mesopredator release and avifaunal extinctions in a fragmented system. Nature, 400:563–566.

Crule, D., Krama, T., Krams, R., Elferts, D., Kaasik, A., Rantala, M. J., Mierauskas, P., Luoto, S. & Krams, I. A. (2017). Habitat quality affects stress responses and survival in a bird wintering under extremely low ambient temperatures. Die Naturwissenschaften, 104(11-12):99.

Cunningham, E. J. A. & Russell, A. F. (2000). Egg investment is influenced by male attractiveness in the mallard. Nature, 404(March):74–77.

Curry, R. L. & Grant, P. R. (1989). Demography of the cooperatively breeding Galapagos Mockingbird, Nesomimus parvulus, in a climatically variable environment. Journal of Animal Ecology, 58(2):441–463.

Darwin, C. (1859). On the origin of species by means of natural selection. John Murray, London.

Davies, N. B. (1985). Cooperation and conflict among dunnocks, Prunella modularis, in a variable mating system. Animal Behaviour, 33(2):628–648.

Davies, N. B., Krebs, J. R. & West, S. A. (2012). An Introduction to Behavioural Ecology. Wiley-Blackwell, Chichester, UK, 4th edition.

DeCesare, N. J., Hebblewhite, M., Bradley, M., Hervieux, D., Neufeld, L. & Musiani, M. (2014). Linking habitat selection and predation risk to spatial variation in survival. Journal of Animal Ecology, 83(2):343–352.

7 Deikumah, J. P., Mcalpine, C. A. & Maron, M. (2014). Biogeographical and Taxonomic Biases in Tropical Forest Fragmentation Research. Conservation Biology, 28(6):1522–1531.

Di Marco, M., Chapman, S., Althor, G., Kearney, S., Besancon, C., Butt, N., Maina, J. M., Possingham, H. P., Rogalla von Bieberstein, K., Venter, O. & Watson, J. E. (2017). Changing trends and persisting biases in three decades of conservation science. Global Ecology and Conservation, 10:32–42.

Dickinson, J. L. & Hatchwell, B. J. (2004). Fitness consequences of helping. In W. D. Koenig & J. L. Dickinson, editors, Ecology and evolution of cooperative breeding in birds, chapter 4, pp. 48 – 66. Cambridge University Press, Cambridge.

Dickinson, J. L., Koenig, W. D. & Pitelka, F. A. (1996). Fitness consequences of helping behavior in the Western bluebird. Behavioral Ecology, 7(2):168–177.

Dickinson, J. L. & McGowan, A. (2005). Winter resource wealth drives delayed dispersal and family-group living in western bluebirds. Proceedings of the Royal Society B: Biological Sciences, 272(1579):2423–2428.

Didham, R. K. & Lawton, J. H. (1999). Edge Structure Determines the Magnitude of Changes in Microclimate and Vegetation Structure in Tropical Forest Fragments. Biotropica, 31(1):17.

Dierkes, P., Heg, D., Taborsky, M., Skubic, E. & Achmann, R. (2005). Genetic relatedness in groups is sex-specific and declines with age of helpers in a cooperatively breeding cichlid. Ecology Letters, 8(9):968–975.

Dinesen, L. (1997). Succeshistorier blandt afrikanske regnskovsfugle. Naturens verden, 181.

Dixit, T., English, S. & Lukas, D. (2017). The relationship between egg size and helper number in cooperative breeders: a meta-analysis across species. PeerJ, 5:e4028.

Page 110 Bibliography

Doerr, E. D. & Doerr, V. a. J. (2006). Comparative demography of treecreepers: evaluating hypotheses for the evolution and maintenance of cooperative breeding. Animal Behaviour, 72(1):147–159.

Donald, P. F. (2007). Adult sex ratios in wild bird populations. Ibis, 149(4):671–692.

Duval, E. H. (2013). Does cooperation increase helpers’ later success as breeders? A test of the skills hypothesis in the cooperatively displaying lance-tailed manakin. The Journal of animal ecology, 82(4):884–93.

Ebensperger, L. A., Villegas, A.,´ Abades, S. & Hayes, L. D. (2014). Mean ecological conditions modulate the effects of group living and communal rearing on offspring production and survival. Behavioral Ecology, 25(4):862–870.

Eggers, S., Griesser, M. & Ekman, J. (2008). Predator-induced reductions in nest visitation rates are modified by forest cover and food availability. Behavioral Ecology, 19(5):1056–1062.

Eggers, S., Griesser, M., Nystrand, M. & Ekman, J. (2006). Predation risk induces changes in nest-site selection and clutch size in the Siberian jay. Proceedings of the Royal Society B-Biological Sciences, 273(1587):701–6.

Eguchi, K., Yamagishi, S., Asai, S., Nagata, H. & Hino, T. (2002). Helping does not enhance reproductive success of cooperatively breeding rufous vanga in Madagascar. Journal of Animal Ecology, 71(1):123–130.

Eikenaar, C., Richardson, D. S., Brouwer, L., Bristol, R., Komdeur, J., Group, A. E., Studies, E., Tech, V. & Kingdom, U. (2008). Experimental evaluation of sex differences in territory acquisition in a cooperatively breeding bird. Behavioral Ecology, (October).

Ekman, J. (2006). Family living among birds.

Elderbrock, E. K., Kern, M. D. & Lynn, S. E. (2012). Problems with Using Ptilochronology to Measure the Growth and Nutritional Status of Nestling Bluebirds. The Condor, 114(4):823–830.

Emlen, S. T. (1982). The Evolution of Helping. I. An Ecological Constraints Model. The American 7 Naturalist, 119(1):29–39.

Emlen, S. T. & Wrege, P. H. (1991). Breeding Biology of White-Fronted Bee-Eaters at Nakuru: The Influence of Helpers on Breeder Fitness. The Journal of Animal Ecology, 60(1):309–326.

English, S., Nakagawa, S. & Clutton-Brock, T. H. (2010). Consistent individual differences in cooperative behaviour in meerkats (Suricata suricatta). Journal of Evolutionary Biology, 23(8):1597–1604.

Ewen, J. G., Cassey, P. & Møller, A. P. (2004). Facultative primary sex ratio variation: A lack of evidence in birds? Proceedings of the Royal Society B: Biological Sciences, 271(1545):1277–1282.

Ewers, R. M. & Didham, R. K. (2006). Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews of the Cambridge Philosophical Society, 81(1):117–142.

Fischer, J. & Lindenmayer, D. B. (2007). Landscape modification and habitat fragmentation: a synthesis. Global Ecology and Biogeography, 16(3):265–280.

Fontaine, J. J. & Martin, T. E. (2006). Parent birds assess nest predation risk and adjust their reproductive strategies. Ecology Letters, 9(4):428–434.

Forstmeier, W. & Weiss, I. (2004). Adaptive plasticity in nest-site selection in response to changing predation risk. Oikos, 104(3):487.

Fox, J. & Weisberg, S. (2011). An {R} Companion to Applied Regression, Second Edition. Thousand CA, Sage.

Frankham, R. (2005). Genetics and extinction.

Page 111 Bibliography

Fredericksen, T. S. & Mostacedo, B. (2000). Regeneration of timber species following selection logging in a Bolivian tropical dry forest. Forest Ecology and Management, 131(1-3):47–55.

Fry, C. H. & Keith, S. (2000). The Birds of Africa, Vol VI : Picathartes to Oxpeckers. Academic Press, London.

Futuyma, D. J. & Moreno, G. (1988). The Evolution of Ecological Specialization. Annual Review of Ecology and Systematics, 19(1):207–233.

Gaston, A. J. (1978). The Evolution of Group Territorial Behavior and Cooperative Breeding. The American Naturalist, 112(988):1091–1100.

Gaston, K. J., Blackburn, T. M. & Goldewijk, K. K. (2003). Habitat conversion and global avian biodiversity loss. Proceedings of the Royal Society B: Biological Sciences, 270(1521):1293–1300.

Gienapp, P. & Meril¨a,J. (2010). Genetic and environmental effects on a condition-dependent trait: feather growth in Siberian jays. Journal of Evolutionary Biology, 23(4):715–723.

Githiru, M. & Lens, L. (2006). Demography of an Afrotropical passerine in a highly fragmented landscape. Animal Conservation, 9(1):21–27.

Githiru, M., Lens, L. & Cresswell, W. (2005). Nest predation in a fragmented Afrotropical forest: Evidence from natural and artificial nests. Biological Conservation, 123(2):189–196.

Gray, M. A., Baldauf, S. L., Mayhew, P. J. & Hill, J. K. (2007). The response of avian feeding guilds to tropical forest disturbance. Conservation Biology, 21(1):133–141.

Green, J. P., Freckleton, R. P. & Hatchwell, B. J. (2016). Variation in helper effort among cooperatively breeding bird species is consistent with Hamilton’s rule. Nature Communications, 7:12663.

Greenwood, P. J. (1980). Mating systems, philopatry and dispersal in birds and mammals. Animal 7 Behaviour, 28(4):1140–1162. Grieco, F. (2002). How Different Provisioning Strategies Result in Equal Rates of Food Delivery: An Experimental Study of Blue Tits Parus caeruleus. Journal of Avian Biology, 33(33):331–341.

Griesser, M., Drobniak, S. M., Nakagawa, S. & Botero, C. A. (2017). Family living sets the stage for cooperative breeding and ecological resilience in birds. PLoS Biology, (621):1–17.

Griesser, M., Nystrand, M. & Ekman, J. (2006). Reduced mortality selects for family cohesion in a social species. Proceedings of the Royal Society B: Biological Sciences, 273(1596):1881–1886.

Griffin, A. S., Alonzo, S. H. & Cornwallis, C. K. (2013). Why Do Cuckolded Males Provide Paternal Care? PLoS Biology, 11(3).

Griffin, A. S. & West, S. a. (2003). Kin discrimination and the benefit of helping in cooperatively breeding vertebrates. Science, 302(5645):634–6.

Griffiths, R., Double, M., Orr, K. & Dawson, R. (1998). A DNA test to sex most birds. Molecular ecology.

Grimes, L. G. (1976). The Occurrence Of Cooperative Breeding Behaviour In African Birds. Ostrich, 47(1):1–15.

Groenewoud, F., Frommen, J. G., Josi, D., Tanaka, H., Jungwirth, A. & Taborsky, M. (2016). Predation risk drives social complexity in cooperative breeders. Proceedings of the National Academy of Sciences, 113(15):4104–4109.

Groenewoud, F., Kingma, S. A., Hammers, M., Dugdale, H. L., Burke, T., Richardson, D. S. & Komdeur, J. (2018). Subordinate females in the cooperatively breeding Seychelles warbler obtain direct benefits by joining unrelated groups. Journal of Animal Ecology, 87(5):1251–1263.

Page 112 Bibliography

Grubb, T. (1989). Ptilochronology - feather growth bars as indicators of nutritional status. AUK, 106(2):314–320.

Grubb, T. C. (2006). Ptilochronology - feather time and the biology of birds. Oxford University Press, Oxford.

Grubb, T. C. & Yosef, R. (1994). Habitat-Specific Nutritional Condition in Loggerhead Shrikes (Lanius ludovicianus): Evidence from Ptilochronology. The Auk, 111(3):756–759.

Grueber, C. E., Nakagawa, S., Laws, R. J. & Jamieson, I. G. (2011). Multimodel inference in ecology and evolution: Challenges and solutions. Journal of Evolutionary Biology, 24(4):699–711.

Guindre-Parker, S. & Rubenstein, D. R. (2018). Multiple benefits of alloparental care in a fluctuating environment. Royal Society Open Science, 5(2):172406.

Haddad, N. M., Brudvig, L. A., Clobert, J., Davies, K. F., Gonzalez, A., Holt, R. D., Lovejoy, T. E., Sexton, J. O., Austin, M. P., Collins, C. D., Cook, W. M., Damschen, E. I., Ewers, R. M., Foster, B. L., Jenkins, C. N., King, A. J., Laurance, W. F., Levey, D. J., Margules, C. R., Melbourne, B. A., Nicholls, A. O., Orrock, J. L., Song, D.-X. & Townshend, J. R. (2015). Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances, 1(2):e1500052–e1500052.

Hale, R. & Swearer, S. E. (2016). Ecological traps: current evidence and future directions. Proceedings of the Royal Society B-Biological Sciences, 283(1824).

Hamilton, I. M. & Taborsky, M. (2005). Unrelated helpers will not fully compensate for costs imposed on breeders when they pay to stay. Proceedings of the Royal Society of London B: Biological Sciences, 272(1561):445–454.

Hamilton, W. D. (1963). The Evolution of Altruistic Behavior. The American Naturalist, 97(896):354–356.

Hamilton, W. D. & May, R. M. (1977). Dispersal in stable habitats. Nature, 269(5629):578–581.

Harmon, M. E., Franklin, J. F., Swanson, F. J., Sollins, P., Gregory, S. V., Lattin, J. D., Anderson, N. H., Cline, S. P., Aumen, N. G., Sedell, J. R., Lienkaemper, G. W., Cromack, K. & Cummins, K. W. 7 (2004). Ecology of Coarse Woody Debris in Temperate Ecosystems. Advances in Ecological Research, 34(03):59–234.

Harrisson, K. A., Pavlova, A., Amos, J. N., Takeuchi, N., Lill, A., Radford, J. Q. & Sunnucks, P. (2013). Disrupted fine-scale population processes in fragmented landscapes despite large-scale genetic connectivity for a widespread and common cooperative breeder: The superb fairy-wren (Malurus cyaneus). Journal of Animal Ecology, 82(2):322–333.

Hart, L. A., Downs, C. T. & Brown, M. (2016). Sitting in the sun: Nest microhabitat affects incubation temperatures in seabirds. Journal of Thermal Biology, 60:149–154.

Hatchwell, B. & Komdeur, J. (2000). Ecological constraints, life history traits and the evolution of cooperative breeding. Animal Behaviour, 59:1079–1086.

Hatchwell, B. J. (1999). Investment strategies of breeders in avian cooperative breeding systems. The American Naturalist, 154(2):205–219.

Hatchwell, B. J. (2009). The evolution of cooperative breeding in birds: kinship, dispersal and life history. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1533):3217–27.

Hatchwell, B. J., Gullett, P. R. & Adams, M. J. (2014). Helping in cooperatively breeding long-tailed tits: a test of Hamilton ’ s rule. Philosophical Transactions of the Royal Society B: Biological Sciences, 369(March):20130565.

Hatchwell, B. J., Russell, A. F., MacColl, A. D. C., Ross, D. J., Fowlie, M. K. & McGowan, A. (2004). Helpers increase long-term but not short-term productivity in cooperatively breeding Long-tailed Tits. Behavioral Ecology, 15(1):1–10.

Page 113 Bibliography

Heg, D., Bachar, Z., Brouwer, L. & Taborsky, M. (2004). Predation risk is an ecological constraint for helper dispersal in a cooperatively breeding cichlid. Proceedings of the Royal Society B: Biological Sciences, 271(1555):2367–2374.

Heinsohn, R. & Legge, S. (1999). The cost of helping. Trends in ecology & evolution, 14(2):53–57.

Heinsohn, R. G. (1991). Slow learning of foraging skills and extended parental care in cooperatively breeding White-Winged Choughs. The American Naturalist, 137(6):864.

Heinsohn, R. G. (1992). Cooperative enhancement of reproductive success in White-Winged Choughs. Evolutionary Ecology, 6(2):97–114.

Heinsohn, R. G. (2004). Parental care, load-lightening, and costs. In W. D. Koenig & J. L. Dickinson, editors, Ecology and evolution of cooperative breeding in birds, chapter Chapter 4, pp. 67 – 80. Cambridge University Press, Cambridge.

Helle, S., Huhta, E., Suorsa, P. & Hakkarainen, H. (2011). Fluctuating asymmetry as a biomarker of habitat fragmentation in an area-sensitive passerine, the Eurasian treecreeper (Certhia familiaris). Ecological Indicators, 11(3):861–867.

Horvathova, T., Nakagawa, S. & Uller, T. (2012). Strategic female reproductive investment in response to male attractiveness in birds. Proceedings of the Royal Society B: Biological Sciences, 279(1726):163–170.

Hunter, L. A. (1987). Cooperative breeding in purple gallinules: the role of helpers in feeding chicks. Behavioral Ecology and Sociobiology, 20(3):171–177.

Husemann, M., Cousseau, L., Callens, T., Matthysen, E., Vangestel, C., Hallmann, C. & Lens, L. (2015). Post-fragmentation population structure in a cooperative breeding Afrotropical cloud forest bird: emergence of a source-sink population network. Molecular Ecology, 24:1172–1187.

Ib´a˜nez-Alamo,´ J. D., Magrath, R. D., Oteyza, J. C., Chalfoun, A. D., Haff, T. M., Schmidt, K. A., 7 Thomson, R. L. & Martin, T. E. (2015). Nest predation research: recent findings and future perspectives. Journal of Ornithology, 156(1):247–262.

Innes, K. E. & Johnston, R. E. (1996). Cooperative breeding in the white-throated magpie-jay. How do auxiliaries influence nesting success? Animal Behaviour, 51(3):519–533.

Janiga, M. & Viˇsovsk´a,Z. (2004). The placement and the size and shape of nests of ring ouzel (Turdus torquatus) reflect the environemental conditions at different altitude. Oecologia Montana, pp. 11–16.

Janssens, L. & Stoks, R. (2013). Predation risk causes oxidative damage in prey. Biology Letters, 9(4).

Jenni-Eiermann, S., Glaus, E., Gr¨uebler,M., Schwabl, H. & Jenni, L. (2008). Glucocorticoid response to food availability in breeding barn swallows (Hirundo rustica). General and Comparative Endocrinology, 155(3):558–565.

Jetz, W. & Rubenstein, D. R. (2011). Environmental uncertainty and the global biogeography of cooperative breeding in birds. Current Biology, 21(1):72–78.

Johnson, M. D. (2007). Measuring habitat quality: a review. The Condor, 109(3):489–504.

Johnston, J., Peach, W., Gregory, R. & White, S. (1997). Survival Rates of Tropical and Temperate Passerines: A Trinidadian Perspective. The American Naturalist, 150(6):771–789.

Jones, A. G. & Ardren, W. R. (2003). Methods of parentage analysis in natural populations. Molecular Ecology, 12(10):2511–2523.

Jungwirth, A., Walker, J. & Taborsky, M. (2015). Prospecting precedes dispersal and increases survival chances in cooperatively breeding cichlids. Animal Behaviour, 106:107–114.

Page 114 Bibliography

Keinath, D. A., Doak, D. F., Hodges, K. E., Prugh, L. R., Fagan, W., Sekercioglu, C. H., Buchart, S. H. & Kauffman, M. (2017). A global analysis of traits predicting species sensitivity to habitat fragmentation. Global Ecology and Biogeography, 26(1):115–127.

Keller, L. F. & Waller, D. M. (2002). Inbreeding effects in wild populations.

Kempenaers, B. & Sheldon, B. C. (1996). Why do male birds not discriminate between their own and extra-pair offspring? Animal Behaviour, 51(5):1165–1173.

Khan, M. Z. & Walters, J. R. (2002). Effects of helpers on breeder survival in the red-cockaded woodpecker (Picoides borealis). Behavioral Ecology and Sociobiology, 51(4):336–344.

Khimoun, A., Eraud, C., Ollivier, A., Arnoux, E., Rocheteau, V., Bely, M., Lefol, E., Delpuech, M., Carpentier, M. L., Leblond, G., Levesque, A., Charbonnel, A., Faivre, B. & Garnier, S. (2016). Habitat specialization predicts genetic response to fragmentation in tropical birds. Molecular Ecology, 25(16):3831–3844.

Kingma, S. A. (2017). Direct benefits explain interspecific variation in helping behaviour among coopera- tively breeding birds. Nature Communications, 8(1):1094.

Kingma, S. A., Bebbington, K., Hammers, M., Richardson, D. S. & Komdeur, J. (2016a). Delayed dispersal and the costs and benefits of different routes to independent breeding in a cooperatively breeding bird. Evolution, 70(11):2595–2610.

Kingma, S. A., Hall, M. L., Arriero, E. & Peters, A. (2010). Multiple benefits of cooperative breeding in purplecrowned fairy-wrens: A consequence of fidelity? Journal of Animal Ecology, 79(4):757–768.

Kingma, S. A., Hall, M. L. & Peters, A. (2011). Multiple benefits drive helping behavior in a cooperatively breeding bird: an integrated analysis. The American naturalist, 177(4):486–495.

Kingma, S. A., Komdeur, J., Hammers, M. & Richardson, D. S. (2016b). The cost of prospecting for dispersal opportunities in a social bird. Biology Letters, 12(6):20160316. 7 Kingma, S. A., Santema, P., Taborsky, M. & Komdeur, J. (2014). Group augmentation and the evolution of cooperation. Trends in Ecology and Evolution, 29(8):476–484.

Kinnaird, M. F. & Grant, P. R. (1982). Cooperative breeding by the Galapagos Mockingbird, Nesomimus parvulus. Behavioral Ecology and Sociobiology, 10(1):65–73.

Klein, J. & Moeschberger, M. (2003). Survival analysis. Techniques for censored and truncated data. Springer, New York, USA, 2nd edn edition.

Koenig, W. D. & Dickinson, J. L., editors (2016). Cooperative Breeding in Vertebrates : studies of ecology, evolution and behavior. Cambridge University Press, Cambridge.

Koenig, W. D. & Haydock, J. (2004). Incest and incest avoidance. In W. D. Koenig & J. L. Dickinson, editors, Ecology and evolution of cooperative breeding in birds, pp. 142–156.

Koenig, W. D., Pitelka, F. A., Carmen, W. J., Mumme, R. L. & Stanback, M. T. (1992). The Evolution of Delayed Dispersal in Cooperative Breeders. The Quarterly Review of Biology, 67(2):111–150.

Koenig, W. D. & Walters, E. L. (2011). Age-related provisioning behaviour in the cooperatively breeding acorn woodpecker: Testing the skills and the pay-to-stay hypotheses. Animal Behaviour, 82(3):437–444.

Koenig, W. D. & Walters, E. L. (2015). Temporal variability and cooperative breeding: testing the bet-hedging hypothesis in the acorn woodpecker. Proceedings of the Royal Society B, 282.

Koenig, W. D., Walters, E. L. & Haydock, J. (2009). Helpers and egg investment in the cooperatively breeding acorn woodpecker: Testing the concealed helper effects hypothesis. Behavioral Ecology and Sociobiology, 63(11):1659–1665.

Page 115 Bibliography

Koenig, W. D., Walters, E. L. & Haydock, J. (2011). Variable helper effects, ecological conditions, and the evolution of cooperative breeding in the acorn woodpecker. The American naturalist, 178(2):145–158.

Kokko, H. & Ekman, J. (2002). Delayed dispersal as a route to breeding: territorial inheritance, safe havens, and ecological constraints. The American naturalist, 160(4):468–84.

Kokko, H., Johnstone, R. a. & T. H., C.-B. (2001). The evolution of cooperative breeding through group augmentation. Proceedings of the Royal Society B: Biological Sciences, 268(1463):187–196.

Kokko, H. & Lundberg, P. (2001). Dispersal, Migration, and Offspring Retention in Saturated Habitats. The American Naturalist, 157(2):188–202.

Komdeur, J. (1992). Importance of habitat saturation and territory quality for evolution of cooperative breeding in the Seychelles warbler. Nature, 358(6386):493–495.

Komdeur, J. (1994). Experimental evidence for helping and hindering by previous offspring in the cooperative-breeding Seychelles warbler Acrocephalus sechellensis. Behavioral Ecology and Sociobiology, 34(3):175–186.

Komdeur, J. (2004). Sex-ratio manipulation. In Ecology and evolution of cooperative breeding in birds. Cambridge University Press, Cambridge.

Komdeur, J. & Edelaar, P. (2001). Male Seychelles warblers use territory budding to maximize lifetime fitness in a saturated environment. Behavioral Ecology, 12(6):706–715.

Komdeur, J., Huffstadt, A., Prast, W., Castle, G., Mileto, R. & Wattel, J. (1995). Transfer experiments of Seychelles warblers to new islands: changes in dispersal and helping behaviour. Animal Behaviour, 49(3):695–708.

Komdeur, J., Richardson, D. S., Hammers, M., Eikenaar, C., Brouwer, L. & Kingma, S. (2017a). The Evolution of Cooperative Breeding in Vertebrates. (April).

Komdeur, J., Sz´ekely, T., Long, X. & Kingma, S. A. (2017b). Adult sex ratios and their implications for 7 cooperative breeding in birds. Philosophical Transactions of the Royal Society B: Biological Sciences, 372(1729):20160322.

Krause, J. & Godin, J. G. J. (1995). Predator preferences for attacking particular prey group sizes: Consequences for predator hunting success and prey predation risk. Animal Behaviour, 50(2):465–473.

Krist, M. (2011). Egg size and offspring quality: a meta-analysis in birds. Biological reviews of the Cambridge Philosophical Society, 86(3):692–716.

Kuijper, B. & Johnstone, R. A. (2013). How should parents adjust the size of their young in response to local environmental cues? Journal of Evolutionary Biology, 26(7):1488–1498.

Kuznetsova, A., Brockhoff, P. B. & Christensen, R. H. B. (2016). lmerTest: Tests in Linear Mixed Effects Models. R package version 2.0-33. https://CRAN.R-project.org/package=lmerTest.

Lahti, D. C. (2001). The ”edge effect on nest predation” hypothesis after twenty years. Biological Conservation, 99(3):365–374.

Lamanna, J. A., Hemenway, A. B., Boccadori, V. & Martin, T. E. (2015). Bird species turnover is related to changing predation risk along a vegetation gradient. Ecology, 96(6):1670–1680.

Lampila, P., M¨onkk¨onen,M. & Desrochers, A. (2005). Demographic responses by birds to forest fragmentation. Conservation Biology, 19(5):1537–1546.

Lamprecht, H. (1989). Silviculture in the Tropics: tropical forest ecosystems and their species-possibilities and methods of their long-term utilization. Deutsche Gesellschaft Fur Technische Zusammenarbeit (GTZ) GmbH, Eschborn.

Lande, R. (1988). Genetics and demography in biological conservation. Science (New York, N.Y.).

Page 116 Bibliography

Lande, R. (1993). Risks of Population Extinction from Demographic and Environmental Stochasticity and Random Catastrophes. The American Naturalist, 142(6):911–927.

Langen, T. A. (1996). The mating system of the White-throated Magpie-jay Calocitta formosa and Greenwood’s hypothesis for sex-biased dispersal. Ibis, 138(3):506–513.

Langen, T. A. (2000). Prolonged offspring dependence and cooperative breeding in birds. Behavioral Ecology, 11(4):367–377.

Langen, T. A. & Vehrencamp, S. L. (1999). How White-Throated Magpie-Jay helpers contribute during breeding. The Auk, 116(1):131–140.

Langmore, N. E., Bailey, L. D., Heinsohn, R. G., Russell, A. F. & Kilner, R. M. (2016). Egg size investment in superb fairy-wrens: helper effects are modulated by climate. Proceedings of the Royal Society B: Biological Sciences, 283(1843):20161875.

Latif, Q. S., Heath, S. K. & Rotenberry, J. T. (2012). How avian nest site selection responds to predation risk: Testing an ’adaptive peak hypothesis’. Journal of Animal Ecology, 81(1):127–138.

Laurance, W. F., Carolina Useche, D., Rendeiro, J., Kalka, M., Bradshaw, C. J. a., Sloan, S. P., Laurance, S. G., Campbell, M., Abernethy, K., Alvarez, P., Arroyo-Rodriguez, V., Ashton, P., Ben´ıtez-Malvido, J., Blom, A., Bobo, K. S., Cannon, C. H., Cao, M., Carroll, R., Chapman, C., Coates, R., Cords, M., Danielsen, F., De Dijn, B., Dinerstein, E., Donnelly, M. a., Edwards, D., Edwards, F., Farwig, N., Fashing, P., Forget, P.-M., Foster, M., Gale, G., Harris, D., Harrison, R., Hart, J., Karpanty, S., John Kress, W., Krishnaswamy, J., Logsdon, W., Lovett, J., Magnusson, W., Maisels, F., Marshall, A. R., McClearn, D., Mudappa, D., Nielsen, M. R., Pearson, R., Pitman, N., van der Ploeg, J., Plumptre, A., Poulsen, J., Quesada, M., Rainey, H., Robinson, D., Roetgers, C., Rovero, F., Scatena, F., Schulze, C., Sheil, D., Struhsaker, T., Terborgh, J., Thomas, D., Timm, R., Nicolas Urbina-Cardona, J., Vasudevan, K., Joseph Wright, S., Carlos Arias-G., J., Arroyo, L., Ashton, M., Auzel, P., Babaasa, D., Babweteera, F., Baker, P., Banki, O., Bass, M., Bila-Isia, I., Blake, S., Brockelman, W., Brokaw, N., Br¨uhl,C. a., Bunyavejchewin, S., Chao, J.-T., Chave, J., Chellam, R., Clark, C. J., Clavijo, J., Congdon, R., Corlett, R., Dattaraja, H. S., Dave, C., Davies, G., de Mello Beisiegel, B., de Nazar´ePaes da Silva, R., Di Fiore, A., Diesmos, A., Dirzo, R., Doran-Sheehy, D., Eaton, M., Emmons, L., Estrada, A., Ewango, 7 C., Fedigan, L., Feer, F., Fruth, B., Giacalone Willis, J., Goodale, U., Goodman, S., Guix, J. C., Guthiga, P., Haber, W., Hamer, K., Herbinger, I., Hill, J., Huang, Z., Fang Sun, I., Ickes, K., Itoh, A., Ivanauskas, N., Jackes, B., Janovec, J., Janzen, D., Jiangming, M., Jin, C., Jones, T., Justiniano, H., Kalko, E., Kasangaki, A., Killeen, T., King, H.-b., Klop, E., Knott, C., Kon´e,I., Kudavidanage, E., Lahoz da Silva Ribeiro, J., Lattke, J., Laval, R., Lawton, R., Leal, M., Leighton, M., Lentino, M., Leonel, C., Lindsell, J., Ling-Ling, L., Eduard Linsenmair, K., Losos, E., Lugo, A., Lwanga, J., Mack, A. L., Martins, M., Scott McGraw, W., McNab, R., Montag, L., Myers Thompson, J., Nabe-Nielsen, J., Nakagawa, M., Nepal, S., Norconk, M., Novotny, V., O’Donnell, S., Opiang, M., Ouboter, P., Parker, K., Parthasarathy, N., Pisciotta, K., Prawiradilaga, D., Pringle, C., Rajathurai, S., Reichard, U., Reinartz, G., Renton, K., Reynolds, G., Reynolds, V., Riley, E., R¨odel,M.-O., Rothman, J., Round, P., Sakai, S., Sanaiotti, T., Savini, T., Schaab, G., Seidensticker, J., Siaka, A., Silman, M. R., Smith, T. B., de Almeida, S. S., Sodhi, N., Stanford, C., Stewart, K., Stokes, E., Stoner, K. E., Sukumar, R., Surbeck, M., Tobler, M., Tscharntke, T., Turkalo, A., Umapathy, G., van Weerd, M., Vega Rivera, J., Venkataraman, M., Venn, L., Verea, C., Volkmer de Castilho, C., Waltert, M., Wang, B., Watts, D., Weber, W., West, P., Whitacre, D., Whitney, K., Wilkie, D., Williams, S., Wright, D. D., Wright, P., Xiankai, L., Yonzon, P. & Zamzani, F. (2012). Averting biodiversity collapse in tropical forest protected areas. Nature, 489(7415):290–294.

Laurance, W. F., Goosem, M. & Laurance, S. G. (2009). Impacts of roads and linear clearings on tropical forests. Trends in Ecology and Evolution, 24(12):659–669.

Le Tortorec, E., Helle, S., Suorsa, P., Sirki¨a,P., Huhta, E., Nivala, V. & Hakkarainen, H. (2012). Feather growth bars as a biomarker of habitat fragmentation in the Eurasian treecreeper. Ecological Indicators, 15(1):72–75.

Le Vin, A., Mable, B., Taborsky, M., Heg, D. & Arnold, K. (2011). Individual variation in helping in a cooperative breeder: relatedness versus behavioural type. Animal Behaviour, 82(3):467–477.

Page 117 Bibliography

Leedale, A. E., Sharp, S. P., Simeoni, M., Robinson, E. J. H. & Hatchwell, B. J. (2018). Fine-scale genetic structure and helping decisions in a cooperatively breeding bird. Molecular Ecology, pp. 0–1.

Lees, A. C. & Peres, C. A. (2009). Gap-crossing movements predict species occupancy in Amazonian forest fragments. Oikos, 118(2):280–290.

Legendre, P. & Gallagher, E. D. (2001). Ecologically meaningful transformations for ordination of species data. Oecologia, 129(2):271–280.

Legendre, P. & Legendre, L. (2012). Numerical Ecology. Elsevier, Oxford, UK, 3th editio edition.

Legge, S. (2000a). Helper contributions in the cooperatively breeding laughing kookaburra: feeding young is no laughing matter. Animal Behaviour, 59(5):1009–1018.

Legge, S. (2000b). The effect of helpers on reproductive success in the laughing kookaburra. Journal of Animal Ecology, 69:714–724.

Lejeune, L., van de Pol, M., Cockburn, A., Louter, M. & Brouwer, L. (2016). Male and female helper effects on maternal investment and adult survival in red-winged fairy-wrens. Behavioral Ecology.

Lens, L., Dongen, S. V., Wilder, C. M., Brooks, T. M. & Matthysen, E. (1999). Fluctuating asymmetry increases with habitat disturbance in seven bird species of a fragmented afrotropical forest. Proceedings of the Royal Society B-Biological Sciences, 266(February):1241.

Lens, L., Van Dongen, S., Norris, K., Githiru, M. & Matthysen, E. (2002). Avian persistence in fragmented rainforest. Science, 298(5596):1236–1238.

Leung, B., Forbes, M. & Houle, D. (2000). Fluctuating Asymmetry as a Bioindicator of Stress: Comparing Efficacy of Analyses Involving Multiple Traits. The American naturalist, 155(1):101–115.

Lewis, D. M. (1982). Cooperative breeding in a population of White-browed Sparrow Weavers Plocepasser mahali. Ibis, 124:511–522.

7 Li, Y., Li, S., Guo, C., Zhang, G., Zhou, Y. & Lu, X. (2015). Nest helpers improve parental survival but not offspring production in a high-elevation passerine , the Ground Tit Pseudopodoces humilis. Ibis, 157:567–574.

Liebezeit, J. R. & George, T. L. (2002). Nest Predators , Nest-Site Selection , and Nesting Success of the Dusky Flycatcher in a Managed Ponderosa Forest Nest Predators , Nest-Site Selection , and Nesting Success of the Dusky Flycatcher in a Managed. The Condor, 104(3):507–517.

Liebl, A., Nomano, F., Browning, L. & Russell, A. (2016). Experimental evidence for fully additive care among male carers in the cooperatively breeding chestnut-crowned babbler. Animal Behaviour, 115:47–53.

Lima, S. L. (2002). Putting predators back into behavioral predatorprey interactions. Trends in Ecology & Evolution, 17(2):70–75.

Lima, S. L. (2009). Predators and the breeding bird: Behavioral and reproductive flexibility under the risk of predation. Biological Reviews, 84(3):485–513.

Lin, Y. K., Keane, B., Isenhour, A. & Solomon, N. G. (2006). Effects of Patch Quality on Dispersal and Social Organization of Prairie Voles: an Experimental Approach. Journal of Mammalogy, 87(3):446–453.

Lindenmayer, D. B. & Fischer, J. (2007). Tackling the habitat fragmentation panchreston. Trends in Ecology and Evolution, 22(3):127–132.

Lloyd, P., Andrew Taylor, W., Du Plessis, M. A. & Martin, T. E. (2009). Females increase reproductive investment in response to helper-mediated improvements in allo-feeding, nest survival, nestling provi- sioning and post-fledging survival in the Karoo scrub-robin Cercotrichas coryphaeus. Journal of Avian Biology, 40(4):400–411.

Page 118 Bibliography

Lloyd, P. & Martin, T. E. (2016). Fledgling survival increases with development time and adult survival across north and south temperate zones. Ibis, 158(1):135–143.

Losier, C. L., Couturier, S., St-Laurent, M. H., Drapeau, P., Dussault, C., Rudolph, T., Brodeur, V., Merkle, J. A. & Fortin, D. (2015). Adjustments in habitat selection to changing availability induce fitness costs for a threatened ungulate. Journal of Applied Ecology, 52(2):496–504.

Lovett, J. C. & Wasser, S. K., editors (1993). Biogeography and Ecology of the Rain forests of Eastern Africa. Cambridge University Press, Cambridge.

Luck, G. W. (2002). The Parental Investment Strategy of an Avian Cooperative Breeder Differs between a Fragmented and an Unfragmented Landscape. American Naturalist, 160(6):809–814.

MacColl, A. D. C. & Hatchwell, B. J. (2002). Temporal Variation in Fitness Payoffs Promotes Cooperative Breeding in LongTailed Tits Aegithalos caudatus. The American Naturalist, 160(2):186–194.

MacColl, A. D. C. & Hatchwell, B. J. (2004). Determinants of lifetime fitness in a cooperative breeder, the long-tailed tit Aegithalos caudatus. Journal of Animal Ecology, 73(6):1137–1148.

M¨aenp¨a¨a,M. & Smiseth, P. (2017). Egg size, begging behaviour and offspring fitness in Nicrophorus vespilloides. Animal Behaviour, 134(November):201–208.

Maga˜na,M., Alonso, J. C., Mart´ın,C. A., Bautista, L. M. & Mart´ın,B. (2010). Nest-site selection by Great Bustards Otis tarda suggests a trade-off between concealment and visibility. Ibis, 152(1):77–89.

Magrath, R. D. & Yezerinac, S. M. (1997). Facultative helping does not influence reproductive success or survival in cooperatively breeding white-browed scrubwrens. Journal of Animal Ecology, 66(5):658–670.

Marra, P. P., Cohen, E. B., Loss, S. R., Rutter, J. E. & Tonra, C. M. (2015). A call for full annual cycle research in animal ecology. Biology Letters, 11(8):20150552.

Marra, P. P. & Holberton, R. L. (1998). Corticosterone levels as indicators of habitat quality: effects of habitat segregation in a migratory bird during the non-breeding season. Oecologia, 116(1-2):284–292. 7

Marsden, S. J., Fielding, A. H., Mead, C. & Hussin, M. Z. (2002). A technique for measuring the density and complexity of understorey vegetation in tropical forests. Forest Ecology and Management, 165(1-3):117–123.

Martin, T. E. (1993). Nest Predation Among Vegetation Layers and Habitat Types : Revising the Dogmas. The American naturalist, 141(6):897–913.

Martin, T. E. (1998). Are Microhabitat Preferences of Coexisting Species under Selection and Adaptive? Ecology, 79(2):656.

Martin, T. E. (2015). Age-related mortality explains life history strategies of tropical and temperate . Science, 349(6251):966–970.

Martin, T. E. & Roper, J. J. (1988). Nest Predation and Nest-Site Selection of a Western Population of the Hermit Thrush. The Condor, 90(1):51–57.

Martin, T. E., Scott, J. & Menge, C. (2000). Nest predation increases with parental activity: separating nest site and parental activity effects. Proceedings of the Royal Society B-Biological Sciences, 267(1459):2287– 2293.

Marzluff, J. M. & Balda, R. P. (1990). Pinyon Jays: making the best of a bad situation by helping. In P. B. Stacey & W. D. Koenig, editors, Cooperative Breeding in Birds : Long term studies of Ecology and Behaviour. Cambridge University Press, Cambridge.

Mattsson, B. J., Meyers, J. M. & Cooper, R. J. (2006). Detrimental impacts of radiotransmitters on juvenile Louisiana Waterthrushes. Journal of Field Ornithology, 77(2):173–177.

Page 119 Bibliography

Matysiokov´a,B. & Remeˇs,V. (2010). Assessing the usefulness of ptilochronology in the study of melanin- and carotenoid-based ornaments in the Great Tit Parus major. Ibis, 152(2):397–401.

Maxwell, S. L., Fuller, R. A., Brooks, T. M. & Watson, J. E. M. (2016). The ravages of guns, nets and bulldozers. Nature, 536:146–145.

McGinley, M. A., Temme, D. H. & Geber, M. A. (1987). Parental Investment in Offspring in Variable Environments: Theoretical and Empirical Considerations. The American Naturalist, 130(3):370–398.

McGowan, A., Hatchwell, B. J. & Woodburn, R. J. W. (2003). The effect of helping behaviour on the survival of juvenile and adult long-tailed tits Aegithalos caudatus. Journal of Animal Ecology, 72(3):491–499.

McGowan, K. J. & Woolfenden, G. E. (1989). A sentinel system in the Florida scrub jay. Animal Behaviour, 37:1000–1006.

Mcgowan, K. J. & Woolfenden, G. E. (1990). Contributions to fledgling feeding in the Florida Scrub Jay. Journal of Animal Ecology, 59(2):691–707.

McGraw, J. B. & Caswell, H. (1996). Estimation of Individual Fitness from Life-History Data. The American Naturalist, 147(1):47–64.

Meade, J., Nam, K. B., Beckerman, A. P. & Hatchwell, B. J. (2010). Consequences of ’load-lightening’ for future indirect fitness gains by helpers in a cooperatively breeding bird. Journal of Animal Ecology, 79(3):529–537.

Merrill, L., Baehl, E. M., Ripple, K. E., Benson, T. J. & Koenig, W. (2016). Do Birds Alter Levels of Parental Investment Based on Nest-Site Habitat Features? Ethology, 122(11):859–868.

Millar, R. B. (2009). Comparison of hierarchical Bayesian models for overdispersed count data using DIC and Bayes’ factors. Biometrics, 65(3):962–969.

7 Miller, D. A., Grand, J. B., Fondell, T. F. & Anthony, R. M. (2007). Optimizing Nest Survival and Female Survival : Consequences of Nest Site Selection for Canada Geese. The Condor, 109(4):769 – 780.

Minias, P. (2015). The use of haemoglobin concentrations to assess physiological condition in birds: A review. Conservation Physiology, 3(1):1–15.

Monaghan, P., Metcalfe, N. B. & Torres, R. (2009). Oxidative stress as a mediator of life history trade-offs: Mechanisms, measurements and interpretation. Ecology Letters, 12(1):75–92.

Monaghan, P. & Nager, R. G. (1997). Why don’t birds lay more eggs? Trends in Ecology and Evolution.

Moore, J. C., Loggenberg, A. & Greeff, J. M. (2006). Kin competition promotes dispersal in a male pollinating fig wasp. Biology Letters, 2(1):17–19.

Moore, R. P., Robinson, W. D., Lovette, I. J. & Robinson, T. R. (2008). Experimental evidence for extreme dispersal limitation in tropical forest birds. Ecology Letters, 11(9):960–968.

Morris, D. W. (1989). Density-dependent habitat selection: Testing the theory with fitness data. Evolu- tionary Ecology, 3(1):80–94.

Mulder, R. A. & Langmore, N. E. (1993). Dominant males punish helpers for temporary defection in superb fairy-wrens. Animal Behaviour, 45(4):830–833.

Mumme, R. L. (1992a). Do helpers increase reproductive success - an experimental-analysis in the Florida scrub jay. Behavioral Ecology and Sociobiology, 31(5):319–328.

Mumme, R. L. (1992b). Do helpers increase reproductive success? An experimental analysis in the Florida scrub jay. Behavioral Ecology and Sociobiology, 31(5):319–328.

Page 120 Bibliography

Mumme, R. L., Bowman, R., Pruett, M. S. & Fitzpatrick, J. W. (2015). Natal territory size, group size, and body mass affect lifetime fitness in the cooperatively breeding Florida Scrub-Jay. The Auk, 132(3):634–646.

Murcia, C. (1995). Edge effects in fragmented forests: implications for conservation. Trends in Ecology & Evolution, 10(2):58–62.

Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. & Kent, J. (2000). Biodiversity hotspots for conservation priorities. Nature, 403(6772):853–8.

Nagelkerke, N. J. D. (1991). A note on a general definition of the coefficient of determination. Biometrika, 78(3):691–692.

Nager, R. G., Monaghan, P. & Houston, D. C. (2000). Within-clutch trade-offs between the number and quality of eggs: Experimental manipulations in gulls. Ecology, 81(5):1339–1350.

Nager, R. G., Monaghan, P. & Houston, D. C. (2001). The cost of egg production: increased egg production reduces future fitness in gulls. Journal of Avian Biology, 32(2):159–166.

Nakagawa, S. & Schielzeth, H. (2013). A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods in Ecology and Evolution, 4(2):133–142.

Nelson-Flower, M. J., Wiley, E. M., Flower, T. P. & Ridley, A. R. (2018). Individual dispersal delays in a cooperative breeder: Ecological constraints, the benefits of philopatry and the social queue for dominance. Journal of Animal Ecology, (August 2017):1–12.

Newbold, T., Scharlemann, J. P. W., Butchart, S. H. M., Sekercioglu, C¸ . H., Alkemade, R., Booth, H. & Purves, D. W. (2013). Ecological traits affect the response of tropical for- est bird species to land-use intensity. Proceedings of the Royal Society B: Biological Sciences, 280(1750):dx.doi.org/10.1098/rspb.2012.2131. Newmark, W. D. (2002). Conserving biodiversity in East African forests : a study of the eastern arc 7 mountains. Springer-Verlag Berlin Heidelberg, Berlin.

Newmark, W. D. (2006). A 16-year study of forest disturbance and understory bird community structure and composition in Tanzania. Conservation Biology, 20(1):122–134.

Newmark, W. D., Jenkins, C. N., Pimm, S. L., McNeally, P. B. & Halley, J. M. (2017). Targeted habitat restoration can reduce extinction rates in fragmented forests. Proceedings of the National Academy of Sciences, 114(36):9635–9640.

Newmark, W. D. & Stanley, T. R. (2011). Habitat fragmentation reduces nest survival in an Afrotropical bird community in a biodiversity hotspot. Proceedings of the National Academy of Sciences of the United States of America.

Nichols, H. J. (2017). The causes and consequences of inbreeding avoidance and tolerance in cooperatively breeding vertebrates. Journal of Zoology, 303(1):1–14.

Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin, P. R., O’Hara, R., Simpson, G. L., Solymos, P., Stevens, M. H. H., Szoecs, E. & Wagner, H. (2017). vegan: Community Ecology Package. R package version 2.4-2. https://CRAN.R-project.org/package=vegan.

Oli, M. K. (2003). Hamilton goes empirical: estimation of inclusive fitness from life-history data. Proceedings of the Royal Society B: Biological Sciences, 270(1512):307–311.

Olson, D. M., Dinerstein, E., Wikramanayake, E. D., Burgess, N. D., Powell, G. V. N., Underwood, E. C., D’amico, J. a., Itoua, I., Strand, H. E., Morrison, J. C., Loucks, C. J., Allnutt, T. F., Ricketts, T. H., Kura, Y., Lamoreux, J. F., Wettengel, W. W., Hedao, P. & Kassem, K. R. (2001). Terrestrial ecoregions of the world: a new map of life on earth. BioScience, 51(11):933.

Page 121 Bibliography

Omoro, L. M. A., Pellikka, P. K. E. & Rogers, P. C. (2010). Tree species diversity, richness, and similarity between exotic and indigenous forests in the cloud forests of Eastern Arc Mountains, Taita Hills, Kenya. Journal of Forestry Research, 21(3):255–264.

Orians, G. & Wittenberger, J. F. (1991). Spatial and temporal scales in habitat selection. The American Naturalist, 137(s1):S29–S49.

Paquet, M., Covas, R., Chastel, O., Parenteau, C. & Doutrelant, C. (2013). Maternal Effects in Relation to Helper Presence in the Cooperatively Breeding Sociable Weaver. PLoS ONE, 8(3).

Parmesan, C. & Yohe, G. (2002). A globally coherent fingerprint of climate change impacts across natural systems. Nature(London), 421:37–42.

Parsons, P. A. (1992). Fluctuating asymmetry: A biological monitor of environmental and genomic stress. Heredity, 68(4):361–364.

Peig, J. & Green, A. J. (2009). New perspectives for estimating body condition from mass/length data: the scaled mass index as an alternative method. Oikos, 118(12):1883–1891.

Peig, J. & Green, A. J. (2010). The paradigm of body condition: a critical reappraisal of current methods based on mass and length. Functional Ecology, 24(6):1323–1332.

Pelech, S. A., Smith, J. N. M. & Boutin, S. (2010). A predator’s perspective of nest predation: Predation by red squirrels is learned, not incidental. Oikos, 119(5):841–851.

Pellikka, P. K., L¨otj¨onen,M., Siljander, M. & Lens, L. (2009). Airborne remote sensing of spatiotemporal change (19552004) in indigenous and exotic forest cover in the Taita Hills, Kenya. International Journal of Applied Earth Observation and Geoinformation, 11(4):221–232.

Perrin, N. & Mazalov, V. (2000). Local Competition, Inbreeding, and the Evolution of SexBiased Dispersal. 7 The American Naturalist, 155(1):116–127. Petrie, M., Krupa, A. & Bucke, T. (1999). Peacocks lek with relatives even in the absence of social and environmental cues. Nature, 401(6749):155–157.

Pfeifer, M., Lefebvre, V., Peres, C. A., Banks-Leite, C., Wearn, O. R., Marsh, C. J., Butchart, S. H., Arroyo-Rodr´ıguez,V., Barlow, J., Cerezo, A., Cisneros, L., D’Cruze, N., Faria, D., Hadley, A., Harris, S. M., Klingbeil, B. T., Kormann, U., Lens, L., Medina-Rangel, G. F., Morante-Filho, J. C., Olivier, P., Peters, S. L., Pidgeon, A., Ribeiro, D. B., Scherber, C., Schneider-Maunoury, L., Struebig, M., Urbina-Cardona, N., Watling, J. I., Willig, M. R., Wood, E. M. & Ewers, R. M. (2017). Creation of forest edges has a global impact on forest vertebrates. Nature, 551(7679):187–191.

Pimm, S. L. & Raven, P. (2000). Biodiversity. Extinction by numbers. Nature, 403(6772):843–845.

Pinya, S., Tejada, S., Cap´o,X. & Sureda, A. (2016). Invasive predator snake induces oxidative stress responses in insular amphibian species. Science of the Total Environment, 566-567:57–62.

Potapov, P., Hansen, M. C., Laestadius, L., Turubanova, S., Yaroshenko, A., Thies, C., Smith, W., Zhuravleva, I., Komarova, A., Minnemeyer, S. & Esipova, E. (2017). The last frontiers of wilderness: Tracking loss of intactforest landscapes from 2000 to 2013. Science Advances, 3:e1600821.

Powell, L. L., Cordeiro, N. J. & Stratford, J. A. (2015). Ecology and conservation of avian insectivores of the rainforest understory: A pantropical perspective. Biological Conservation, 188(January):1–10.

Preston, S. A., Briskie, J. V. & Hatchwell, B. J. (2016). Adult helpers increase the recruitment of closely related offspring in the cooperatively breeding rifleman. Behavioral Ecology, 27(6):1617–1626.

Preston, S. A. J., Briskie, J. V., Burke, T. & Hatchwell, B. J. (2013). Genetic analysis reveals diverse kin-directed routes to helping in the rifleman Acanthisitta chloris. Molecular ecology, 22(19):5027–39.

Page 122 Bibliography

Prugh, L. R., Hodges, K. E., Sinclair, A. R. E. & Brashares, J. S. (2008). Effect of habitat area and isolation on fragmented animal populations. Proceedings of the National Academy of Sciences, 105(52):20770–20775.

Pusey, A. E. (1987). Sex-biased dispersal and inbreeding avoidance in birds and mammals. Trends in Ecology and Evolution, 2(10):295–299.

R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.

Rappole, J. H. & Tipton, A. R. (1990). New harness design for attachment of radio transmitters to small passerines. Journal of Field Ornithology, 62(3):335–337.

Remeˇs,V. (2005). Nest concealment and parental behaviour interact in affecting nest survival in the blackcap (Sylvia atricapilla): An experimental evaluation of the parental compensation hypothesis. Behavioral Ecology and Sociobiology, 58(3):326–332.

Reyer, H.-U. (1984). Investment and relatedness: A cost/benefit analysis of breeding and helping in the pied kingfisher (Ceryle rudis). Animal Behaviour, 32(4):1163–1178.

Richardson, D. S., Burke, T. & Komdeur, J. (2002). Direct benefits and the evolution of female-biased cooperative breeding in Seychelles warblers. Evolution, 56(11):2313–2321.

Richardson, D. S., Jury, F. L., Blaakmeer, K., Komdeur, J. & Burke, T. (2001). Parentage assignment and extra-group paternity in a cooperative breeder: The Seychelles warbler (Acrocephalus sechellensis). Molecular Ecology, 10(9):2263–2273.

Ricklefs, R. E. (1969). An analysis of nesting mortality in birds. Smithsonian Contributions to Zoology, 9(9):1–48.

Ridley, A. R. & Raihani, N. J. (2007). Variable postfledging care in a cooperative bird: Causes and consequences. Behavioral Ecology, 18(6):994–1000. 7 Ridley, A. R. & Raihani, N. J. (2008). Task partitioning increases reproductive output in a cooperative bird. Behavioral Ecology, 19(6):1136–1142.

Ridley, A. R., Raihani, N. J. & Nelson-Flower, M. J. (2008). The cost of being alone: The fate of floaters in a population of cooperatively breeding pied babblers Turdoides bicolor. Journal of Avian Biology, 39(4):389–392.

Riehl, C. (2011). Living with strangers: direct benefits favour non-kin cooperation in a communally nesting bird. Proceedings of the Royal Society B: Biological Sciences, 278(1712):1728–1735.

Riehl, C. (2013). Evolutionary routes to non-kin cooperative breeding in birds. Proceedings. Biological sciences / The Royal Society, 280(1772):20132245.

Riehl, C. & Jara, L. (2009). Natural history and reproductive biology of the communally breeding greater ani (Crotophaga major) at Gat´un Lake, Panama. The Wilson Journal of Ornithology, 121(4):679–687.

Riehl, C. & Stern, C. A. (2015). How cooperatively breeding birds identify relatives and avoid incest: New insights into dispersal and kin recognition. BioEssays, 37(12):1303–1308.

Robertson, B. A. & Hutto, R. L. (2006). A framework for understanding ecological traps and an evaluation of existing evidence. Ecology, 87(5):1075–1085.

Robinson, S. K., Thompson, F. R., Donovan, T. M., Whitehead, D. R. & Faaborg, J. (1995). Regional forest fragmentation and the nesting success of migratory birds. Science (New York, N.Y.), 267(5206):1987–90.

Robinson, W. D. (2018). Surveying tropical birds is much harder than you think: a primer of best practices. Biotropica, 0(0):1–4.

Page 123 Bibliography

Robinson, W. D. & Sherry, T. W. (2012). Mechanisms of avian population decline and species loss in tropical forest fragments. Journal of Ornithology, 153(SUPPL. 1):141–152.

Rodrigues, A. M. M. & Kokko, H. (2016). Models of social evolution: can we do better to predict who helps whom to achieve what’? Philosophical Transactions of the Royal Society B: Biological Sciences, 371(1687):20150088.

Roff, D. (2002). Life History Evolution. Sinauer Associates, Sunderland (Massachusetts, USA).

Roff, D. A. & Fairbairn, D. J. (2007). The evolution of trade-offs: Where are we? Journal of Evolutionary Biology, 20(2):433–447.

Rold´an,M., Mart´ın-G´alvez, D., Rodr´ıguez,J. & Soler, M. (2013). Breeding biology and fledgling survival in a Carrion Crow Corvus corone population of Southern Spain: a comparison of group and pair breeder. Acta Ornithologica, 48(2):221–235.

Rollins, L. A., Browning, L. E., Holleley, C. E., Savage, J. L., Russell, A. F. & Griffith, S. C. (2012). Building genetic networks using relatedness information: A novel approach for the estimation of dispersal and characterization of group structure in social animals. Molecular Ecology, 21(7):1727–1740.

Roper, J. J. (2005). Try and try again: Nest predation favors persistence in a neotropical bird. Ornitologia Neotropical, 16:253–262.

Roper, J. J. & Goldstein, R. R. (1997). A test of the Skutch hypothesis: Does activity at nests increase nest predation risk? Journal of Avian Biology, 28(2):111–116.

Roper, J. J., Sullivan, K. A. & Ricklefs, R. E. (2010). Avoid nest predation when predation rates are low, and other lessons: testing the tropical-temperate nest predation paradigm. Oikos, 119(4):719–729.

Rowley, I. C. R. & Russell, E. (1990). Splendid Fairy-wrens: demonstrating the importance of longevity. In P. B. Stacey & W. D. Koenig, editors, Cooperative Breeding in Birds : Long term studies of Ecology and Behaviour, pp. 1 – 30. Cambridge University Press, Cambridge.

7 Royle, N. J., Russell, A. F. & Wilson, A. J. (2014). The evolution of flexible parenting. Science, 345(6198):776–781.

Rubenstein, D. R. (2007). Territory quality drives intraspecific patterns of extrapair paternity. Behavioral Ecology, 18(6):1058–1064.

Rubenstein, D. R. (2011). Spatiotemporal environmental variation, risk aversion, and the evolution of cooperative breeding as a bet-hedging strategy. PNAS, 108:10816–10822.

Rubenstein, D. R. & Lovette, I. J. (2007). Temporal Environmental Variability Drives the Evolution of Cooperative Breeding in Birds. Current Biology, 17(16):1414–1419.

Russell, A. F. (2001). Dispersal costs set the scene for helping in an atypical avian cooperative breeder. Proceedings of the Royal Society of London Series B-Biological Sciences, 268(1462):95–99.

Russell, A. F. & Hatchwell, B. J. (2001). Experimental evidence for kin-biased helping in a cooperatively breeding vertebrate. Proceedings. Biological sciences / The Royal Society, 268(1481):2169–74.

Russell, A. F., Langmore, N. E., Cockburn, A., Astheimer, L. B. & Kilner, R. M. (2007). Reduced egg investment can conceal helper effects in cooperatively breeding birds. Science (New York, N.Y.), 317(5840):941–944.

Russell, A. F. & Lummaa, V. (2009). Maternal effects in cooperative breeders: from hymenopterans to humans. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 364(1520):1143–67.

Russell, A. F., Portelli, D. J., Russell, D. J. F. & Barclay, H. (2010). Breeding ecology of the Chestnut- crowned Babbler: A cooperative breeder in the desert. Emu, 110(4):324–331.

Page 124 Bibliography

Russell, E. & Rowley, I. (1988). Helper contributions to reproductive success in the Splendid Fairy-wren (Malurus splendens). Behavioral Ecology and Sociobiology, 22(2):131–140.

Russell, E. M., Brown, R. J. & Brown, M. N. (2004). Life history of the White-breasted robin, Eopsaltria georgiana (Petroicidae), in south-western Australia. Australian Journal of Zoology, 52(2):111–145.

Sagar, R., Raghubanshi, A. S. & Singh, J. S. (2003). Tree species composition, dispersion and diversity along a disturbance gradient in a dry tropical forest region of India. Forest Ecology and Management, 186(1-3):61–71.

Saino, N., Romano, M., Caprioli, M., Ambrosini, R., Rubolini, D., Scandolara, C. & Romano, A. (2012). A ptilochronological study of carry-over effects of conditions during wintering on breeding performance in the barn swallow Hirundo rustica. Journal of Avian Biology, 43(6):513–524.

Sanderson, E. W., Jaiteh, M., Levy, M. a., Redford, K. H., Wannebo, A. V. & Woolmer, G. (2002). The Human Footprint and the Last of the Wild. BioScience, 52(10):891.

Sankamethawee, W., Gale, G. A. & Hardesty, B. D. (2009). Post-Fledgling survival of the cooperatively breeding Puff-Throated (Alophoixus pallidus). The Condor, 111(4):675–683.

Santisteban, L., Sieving, K. E. & Avery, M. L. (2002). Use of sensory cues by fish crows Corvus ossifragus preying on artificial bird nests. Journal of Avian Biology, 33(3):245–252.

Santos, E. S. A. & Macedo, R. H. (2011). Load lightening in southern lapwings: Group-Living mothers lay smaller eggs than pair-living mothers. Ethology, 117(6):547–555.

Saunders, D. A., Hobbs, R. J. & Margules, C. R. (1991). Biological Consequences of Ecosystem Fragmentation: A Review. Conservation Biology, 5(1):18–32.

Savage, J. L., Russell, A. F. & Johnstone, R. A. (2015). Maternal allocation in cooperative breeders: should mothers match or compensate for expected helper contributions? Animal Behaviour, 102:189–197.

Schaefer, V. H. (1976). Geographic Variation in the Placement and Structure of Oriole Nests. The Condor, 78(4):443–448. 7

Schlawin, J. R. & Zahawi, R. (2008). ’Nucleating’ succession in recovering neotropical wet forests: The legacy of remnant trees. Journal of Vegetation Science, 19(4):485–492.

Schmidt, K. A., Ostfeld, R. S. & Smyth, K. N. (2006). Spatial heterogeneity in predator activity, nest survivorship, and nest-site selection in two forest thrushes. Oecologia, 148(1):22–29.

Sekerciolu, C. H. (2011). Functional Extinctions of Bird pollinators cause plant declines. Science, 331(February).

Sekerciolu, C. H., Ehrlich, P. R., Daily, G. C., Aygen, D., Goehring, D. M., Sandi, R. F. R. F., Sekercioglu, C. H., Ehrlich, P. R., Daily, G. C., Aygen, D., Goehring, D. M. & Sandi, R. F. R. F. (2002). Disappearance of insectivorous birds from tropical forest fragments. Proceedings of the National Academy of Sciences of the United States of America, 99(1):263–267.

Shaffer, T. L. (2004). A unified approach to analyzing nest success. Auk, 121(2):526–540.

Shen, S.-f., Emlen, S. T., Koenig, W. D. & Rubenstein, D. R. (2017). The Ecology of Cooperative Breeding Behavior. Ecology Letters.

Sherley, G. H. (1990). Co-operative breeding in Riflemen (Acanthissitta chloris) benefits to parents , offspring and helpers. Behaviour, 112(1):1–22.

Sih, A., Ferrari, M. C. O. & Harris, D. J. (2011). Evolution and behavioural responses to human-induced rapid environmental change. Evolutionary Applications, 4(2):367–387.

Sodhi, N. S., Liow, L. & Bazzaz, F. (2004). Avian Extinctions from Tropical and Subtropical Forests. Annual Review of Ecology, Evolution, and Systematics, 35(1):323–345.

Page 125 Bibliography

Sorato, E., Griffith, S. C., Russell, A. F. & Dantzer, B. (2016). The price of associating with breeders in the cooperatively breeding chestnut-crowned babbler: foraging constraints, survival and sociality. Journal of Animal Ecology, 85(5):1340–1351.

Sorato, E., Gullett, P. R., Creasey, M. J., Griffith, S. C. & Russell, A. F. (2015). Plastic territoriality in group-living chestnut-crowned babblers: roles of resource value, holding potential and predation risk. Animal Behaviour, 101:155–168.

Spanhove, T., Callens, T., Hallmann, C. A., Pellikka, P. & Lens, L. (2014). Nest predation in Afrotropical forest fragments shaped by inverse edge effects, timing of nest initiation and vegetation structure. Journal of Ornithology, 155(2):411–420.

Spanhove, T., Lehouck, V. & Lens, L. (2009). Inverse edge effect on nest predation in a Kenyan forest fragment: an experimental case study. Bird Conservation International, 19(04):367.

Spielman, D., Brook, B. W. & Frankham, R. (2004). Most species are not driven to extinction before genetic factors impact them. Proceedings of the National Academy of Sciences, 101(42):15261–15264.

Stacey, P. (1992). Environmental Variation and the Persistence of Small Populations. Ecological Applica- tions, 2(1):18–29.

Stacey, P. B. & Koenig, W. D. (1990). Splendid Fairy-wrens : demonstrating the importance of longevity. In P. B. Stacey & W. D. Koenig, editors, Cooperative Breeding in Birds : Long term studies of Ecology and Behaviour, chapter 1, p. 636. Cambridge University Press, Cambridge.

Stacey, P. B. & Ligon, J. D. (1987). Territory Quality and Dispersal Options in the Acorn Woodpecker, and a Challenge to the Habitat-Saturation Model of Cooperative Breeding. The American Naturalist, 130(5):654–676.

Stacey, P. B. & Ligon, J. D. (1991). The benefits-of-philopatry hypothesis for the evolution of cooperative breeding: variation in territory quality and group size effects. The American Naturalist, 137(6):831–846.

7 Stearns, S. C. (1992). The Evolution of Life-Histories. Oxford University Press, Oxford, UK.

Stratford, J. A. & Robinson, W. D. (2005). Gulliver travels to the fragmented tropics: Geographic variation in mechanisms of avian extinction. Frontiers in Ecology and the Environment, 3(2):91–98.

Streby, H. M., Peterson, S. M., Lehman, J. A., Kramer, G. R., Iknayan, K. J. & Andersen, D. E. (2013). The effects of force-fledging and premature fledging on the survival of nestling songbirds. Ibis, 155(3):616–620.

Stutchbury, B. J. M. & Morton, E. S. (2001). Behavioral Ecology of Tropical Birds. Academic Press, London/San Diego.

Sydeman et al., W. J. (1988). Annual reproductive yield in the cooperative Pygmy Nuthatch (Sitta pygmaea). Auk, 105(1):70–77.

Szekely, T., Liker, A., Freckleton, R. P., Fichtel, C. & Kappeler, P. M. (2014). Sex-biased survival predicts adult sex ratio variation in wild birds. Proceedings of the Royal Society B: Biological Sciences, 281(1788):20140342–20140342.

Taborsky, B., Skubic, E. & Bruintjes, R. (2007). Mothers adjust egg size to helper number in a cooperatively breeding cichlid. Behavioral Ecology, 18(4):652–657.

Temple, H. J., Hoffman, J. I. & Amos, W. (2006). Dispersal, philopatry and intergroup relatedness: Fine-scale genetic structure in the white-breasted thrasher, Ramphocinclus brachyurus. Molecular Ecology, 15(11):3449–3458.

Tewksbury, J. J., Garner, L., Garner, S., Lloyd, J. D., Saab, V. & Martin, T. E. (2006). Tests of landscape influence: Nest predation and brood parasitism in fragmented ecosystems. Ecology, 87(3):759–768.

Page 126 Bibliography

Thaxton, J. E. & Hingtgen, T. M. (1996). Effects of suburbanization and habitat fragmentation on Florida Scrub-Jay dispersal. Florida Field Naturalist, 24(2):25–37.

Thijs, K. W. (2015). Tree community dynamics and ecosystem function in a tropical landscape under deforestation pressure. Phd thesis, KU Leuven, Belgium.

Thomas, C. D., Thomas, C. D., Cameron, A., Cameron, A., Green, R. E., Green, R. E., Bakkenes, M., Bakkenes, M., Beaumont, L. J., Beaumont, L. J., Collingham, Y. C., Collingham, Y. C., Erasmus, B. F. N., Erasmus, B. F. N., De Siqueira, M. F., De Siqueira, M. F., Grainger, A., Grainger, A., Hannah, L., Hannah, L., Hughes, L., Hughes, L., Huntley, B., Huntley, B., Van Jaarsveld, A. S., Van Jaarsveld, A. S., Midgley, G. F., Midgley, G. F., Miles, L., Miles, L., Ortega-Huerta, M. a., Ortega-Huerta, M. a., Peterson, a. T., Peterson, a. T., Phillips, O. L., Phillips, O. L., Williams, S. E. & Williams, S. E. (2004). Extinction risk from climate change. Nature, 427(6970):145–8.

Tuomainen, U. & Candolin, U. (2011). Behavioural responses to human-induced environmental change. Biological Reviews, 86(3):640–657.

Valencia, J., Cruz, C. & Gonzalez, B. (2003). Flexible Helping Behaviour in the Azure-Winged Magpie. Ethology, 109(7):545–558.

Valencia, J., De La Cruz, C., Carranza, J. & Mateos, C. (2006). Parents increase their parental effort when aided by helpers in a cooperatively breeding bird. Animal Behaviour, 71(5):1021–1028.

Valencia, J., Mateos, C., de la Cruz, C. & Carranza, J. (2017). Maternal allocation in eggs when counting on helpers in a cooperatively breeding bird. Journal of Avian Biology, 48(4):536–543.

Van de Crommenacker, J., Komdeur, J., Burke, T. & Richardson, D. S. (2011). Spatio-temporal variation in territory quality and oxidative status: A natural experiment in the Seychelles warbler (Acrocephalus sechellensis). Journal of Animal Ecology, 80(3):668–680.

Van de Loock, D. & Bates, M. F. (2016). Dasypeltis atra Sternfeld, 1913 - Montane Egg-eater, Diet and Distribution. African Herp News, 63:16–18. 7 Van de Loock, D., Strubbe, D., De Neve, L., Githiru, M., Matthysen, E. & Lens, L. (2017). Cooperative breeding shapes post-fledging survival in an Afrotropical forest bird. Ecology and Evolution, 7(10):3489– 3493.

Vangestel, C., Braeckman, B. P., Matheve, H. & Lens, L. (2010). Constraints on home range behaviour affect nutritional condition in urban House Sparrows (Passer domesticus). Biological Journal of the Linnean Society, 101(1):41–50.

Vangestel, C., Callens, T., Vandomme, V. & Lens, L. (2013). Sex-biased dispersal at different geographical scales in a cooperative breeder from fragmented rainforest. PloS one, 8(8):e71624.

Venter, O., Sanderson, E. W., Magrach, A., Allan, J. R., Beher, J., Jones, K. R., Possingham, H. P., Laurance, W. F., Wood, P., Fekete, B. M., Levy, M. A. & Watson, J. E. M. (2016). Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. Nature Communications, 7:1–11.

Vetter, D., R¨ucker, G. & Storch, I. (2013). A meta-analysis of tropical forest edge effects on bird nest predation risk: Edge effects in avian nest predation. Biological Conservation, 159:382–395.

Vigallon, S. M. & Marzluff, J. M. (2005). Is nest predation by Steller’s Jays ( stelleri) incidental or the result of a specialized search strategy? Auk, 122(1):36–49.

Walker, B. G., Boersma, P. D. & Wingfield, J. C. (2005). Field endocrinology and conservation biology. Integrative and Comparative Biology, 45(1):12–18.

Walters, J. R., Ford, H. A. & Cooper, C. B. (1999). The ecological basis of sensitivity of brown treecreepers to habitat fragmentation: A preliminary assessment. Biological Conservation, 90(1):13–20.

Page 127 Bibliography

Walters, J. R., Pasinelli, G., Cooper, C. B., Schiegg, K. & Daniels, S. J. (2004). Conservation Biology. In W. D. Koenig & J. L. Dickinson, editors, Ecology and evolution of cooperative breeding in birds. Cambridge University Press, Cambridge.

Waters, C. N., Zalasiewicz, J., Summerhayes, C., Barnosky, A. D., Poirier, C., Galuszka, A., Cearreta, A., Edgeworth, M., Ellis, E. C., Ellis, M., Jeandel, C., Leinfelder, R., McNeill, J. R., Richter, D. D. B., Steffen, W., Syvitski, J., Vidas, D., Wagreich, M., Williams, M., Zhisheng, A., Grinevald, J., Odada, E., Oreskes, N. & Wolfe, A. P. (2016). The Anthropocene is functionally and stratigraphically distinct from the Holocene. Science, 351(6269).

Watling, J. I., Nowakowski, A. J., Donnelly, M. A. & Orrock, J. L. (2011). Meta-analysis reveals the importance of matrix composition for animals in fragmented habitat. Global Ecology and Biogeography, 20(2):209–217.

Watson, J. E., Whittaker, R. J. & Dawson, T. P. (2004). Habitat structure and proximity to forest edge affect the abundance and distribution of forest-dependent birds in tropical coastal forests of southeastern Madagascar. Biological Conservation, 120(3):315–331.

Weidinger, K. (2002). Interactive effects of concealment, parental behaviour and predators on the survival of open passerine nests. Journal of Animal Ecology, 71(3):424–437.

Weidinger, K. & Koˇcvara, R. (2010). Repeatability of nest predation in passerines depends on predator species and time scale. Oikos, 119(1):138–146.

Wiens, J. A., Rotenberry, J. T. & Horne, B. V. (1987). Habitat Occupancy Patterns of North American Shrubsteppe Birds: The Effects of Spatial Scale. Oikos, 48(2):132.

Wikelski, M. & Cooke, S. J. (2006). Conservation physiology. Trends in Ecology and Evolution, 21(1):38–46.

Wikelski, M., Spinney, L., Schelsky, W., Scheuerlein, A. & Gwinner, E. (2003). Slow pace of life in tropical sedentary birds: a common-garden experiment on four stonechat populations from different latitudes. 7 Proceedings. Biological sciences / The Royal Society, 270(1531):2383–2388. Wilcove, D. S. (1985). Nest predation in forest tracks and the decline of migratory songbirds. Ecology, 66(4):1211–1214.

Wilder, C., Brooks, T. & Lens, L. (1998). Vegetation Structure and Composition of the Taita Hills Forests. Journal of East African Natural History, 87(1):181–187.

Williams, C. T., Kitaysky, A. S., Kettle, A. B. & Buck, C. L. (2008). Corticosterone levels of tufted puffins vary with breeding stage, body condition index, and reproductive performance. General and Comparative Endocrinology, 158(1):29–35.

Williams, D. A. & Hale, A. M. (2006). Helper Effects on Offspring Production in Cooperatively Breeding Brown Jays (Cyanocorax Morio). The Auk, 123(3):847.

Williams, T. D. (2001). Experimental manipulation of female reproduction reveals an intraspecific egg size clutch size trade-off. Proceedings of the Royal Society B: Biological Sciences, 268(1465):423–428.

Wilson, R. R. & Cooper, R. J. (1998). Acadian Flycatcher Nest Placement: Does Placement Influence Reproductive Success? The Condor, 100(4):673–679.

Woodroffe, R. & Ginsberg, J. R. (1998). Edge effects and the extinction of populations inside protected areas. Science, 280(5372):2126–2128.

Woolfenden, G. E. & Fitzpatrick, J. W. (1978). The Inheritance of Territory in Group-Breeding Birds. BioScience, 28(2):104–108.

Woolfenden, G. E. & Fitzpatrick, J. W. (1985). The Florida Scrub Jay. Princeton University Press.

Page 128 Bibliography

Woolfenden, G. E. & Fitzpatrick, J. W. (1990). Florida Scrub jays: a synopsis after 18 years of study. In P. B. Stacey & W. D. Koenig, editors, Cooperative Breeding in Birds : Long term studies of Ecology and Behaviour, chapter 8, pp. 239 – 266. Cambridge University Press.

Woxvold, I. A. & Magrath, M. J. L. (2005). Helping enhances multiple components of reproductive success in the cooperatively breeding apostlebird. Journal of Animal Ecology, 74(6):1039–1050.

Wright, J., McDonald, P. G., Marvelde, L., Kazem, A. J. N., Bishop, C. M., te Marvelde, L., Kazem, A. J. N. & Bishop, C. M. (2010). Helping effort increases with relatedness in bell miners, but ’unrelated’ helpers of both sexes still provide substantial care. Proceedings. Biological sciences / The Royal Society, 277(1680):437–45.

Yang, N., Moermond, T. C., Lloyd, H., Xu, Y., Dou, L., Zhang, K., Yue, B. & Ran, J. (2016). Effects of supplementary feeding on the breeding ecology of the buff-throated partridge in a Tibetan sacred site, China. PLoS ONE, 11(1):1–21.

Yoon, J., Sofaer, H. R., Sillett, T. S., Morrison, S. A. & Ghalambor, C. K. (2017). The relationship between female brooding and male nestling provisioning: does climate underlie geographic variation in sex roles? Journal of Avian Biology, 48(2):220–228.

Yuan, A. H.-w., Liu, M. & Shen, S.-f. (2004). Joint nesting in Taiwan yuhinas: a rare passerine case. The Condor, 106(November 2003):862–872.

Zack, S. (1990). Coupling Delayed Breeding with Shortdistance Dispersal in Cooperatively Breeding Birds. Ethology, 86(4):265–286.

Zuur, A. F., Ieno, E. N. & Elphick, C. S. (2010). A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution, 1(1):3–14.

7

Page 129

A

Appendix to Chapter 2

A.1 Vegetation Composition

Tree community can be an important determinant in nest-site selection and can be quantified using the basal areas of individual species within each plot. To assess vegetation gradients and determine which vegetation composition is important for nest-site selection, we performed a Principle Component Analysis (PCA) using the Vegan package in R (Oksanen et al., 2017). Species distributions are not uniform throughout both fragments (Chawia (CH), a heavily disturbed forest fragment and Ngangao (NG), a moderately disturbed forest fragment) and resulting basal areas strongly skewed. To still be able to A use an Euclidean-based ordination method, we first performed a Hellinger transformation on the absolute basal areas (Legendre & Gallagher, 2001). The Hellinger transformation is defined as the square root of the relative basal area, calculated as the basal area of the species on the total basal area of that plot (Legendre & Gallagher, 2001). Then, we only selected the ’tree community indicator species’ (sensu Thijs, 2015) as these attributed to the majority of the plots’ total basal area measurements and hereby restrict the number of species and resulting ordination axis. We conducted an unconstrained Principle Component Analyses on the global dataset (comprising both fragments) and calculated site scores at each axis (Legendre & Legendre, 2012). We retained the first three axis principle components (PCs), which cumulatively explained over 50% of the total variation and performed Spearman correlations between the PC scores and the abundance (i.e. basal area) of each species in each plot to assess individual species contributions (Table S2). The first axis is represents a gradient from pioneer vegetation to typical late successional tree species (community VI to I, respectively sensu Thijs, 2015). Sites with high scores along the first axis will thus be presented by

Page 131 Appendix to Chapter 2

more late successional tree species. The second axis associates with a gradient from late successional species to species typical for edges and clearings (community I to III sensu Thijs, 2015). The third axis represents a gradient from Cola greenwayi, a late successional tree seen at higher altitudes, and Phoenix reclinata, a species associated with edge and gaps (community I and III sensu Thijs, 2015) to late successional trees associated with lower altitudes (community II sensu Thijs, 2015).

A.2 Daily Survival Rates

In addition to evaluating the relationship between Nest Site Probability Score (NSPS) and nest success as the simple binary outcome of success/fail, we also used logistic-exposure models (Shaffer, 2004). These are designed for flexible nest predation based on daily survival rates and account for the fact that survival probability depends on the interval length between two nest checks. Sample size for this analysis is reduced to nests from the 2009-2010 breeding season only (CH : 9 nests; NG : 20 nests), due to unfortunate data management during the subsequent 2010-2011 breeding season which resulted in loss of detailed nest check information for that season. The models are Generalised Linear Mixed Models (GLMM) with binomial distribution and a logistic-exposure link function defined as ln[Θ1/t/(1 − Θ1/t)], with Θ the daily survival rate and t the interval length (in days) between two nest checks (Shaffer, 2004). More specifically, success (1)/fail (0) of a specific nest check interval is modelled against NSPS and fragment, while the interval length is accounted for in the link function. Nest identity is added as a random term to account for variation in nest monitoring length. Rapidly predated nests might only be represented in the dataset by one nest check interval, while nests monitored for a longer period of A time will be visited multiple times and represented by a number of nest check intervals. We adopted a similar AICc-based ranking and model selection approach on three a priori defined models including : (i) fragment, (ii) fragment + NSPS, (iii) fragment × NSPS as predictor variables. GLMM models were fitted using ’lme4’ package (Bates et al., 2015), and ranked using ’MuMIn’ package in R (Barton, 2016). Model ranking and parameter estimates are presented in table S3 and adhere to the results from the analysis on the larger, two-season dataset using simple binary nest outcome rather than daily survival rates. Based on relative difference in AICc values and 95% confidence intervals of the parameter estimates, there is no strong support for a relationship between daily survival rates and NPS, or for an interaction between the latter and fragment.

Page 132 Appendix to Chapter 2

Table A.1: The species loadings of the five most influential species of each axis, calculated by performing Spearman correlations between the PC scores and the basal area of each species in each plot.

PC1 PC2 PC3 Proportion variation explained 32.4% 14.6% 9.8% Species Tabernaemontana stapfiana -0.976273681 Cola greenwayi 0.459357343 -0.666351498 -0.467128126 Craibia sp. 0.402859490 -0.306952911 Phoenix reclinata 0.320648190 0.798205851 -0.353839260 Millettia oblata ssp. teitensis 0.295853395 Teclea nobilis (syn : Vepris nobilis) -0.218327324 Leptonychia usambarensis 0.169942297 Strombosia scheffleri 0.495630817 A Psychotria sp. 0.429333464 Newtonia buchananii 0.422624714

Page 133

Appendix to Chapter 2 eeaincmoiinPC3 composition Vegetation

1 eeaincmoiinPC2 composition Vegetation

1 -0,23 eeaincmoiinPC1 composition Vegetation

1 0,48 0,29

0,03 -0,14 0,14 hno aln spec. Sapling Shannon 1 0,03 -0,12 0,08

0,86 -0,12 0,23 0,1 0,25 hno respec. Tree Shannon 1 -0,03 0,39 0,02 0,46

0,59 -0,02 0,37 -0,1 -0,1 0,02 aln species Sapling 1 0,02

0,69 respecies Tree

1 -0,09 aln density Sapling

1 -0,02 redensity Tree

1 0,16 0,45 0,05 0,24 -0,03 0,06 0,21 0,04 aa area Basal

1 0,35 0,27 -0,18 0,18 -0,41 0,14 -0,11 0,03 -0,2 edwood Dead 1 0,07 0,05 0 -0,13 0,21 -0,09 0,25 -0,06 -0,15 0,08

Chawia etsubstrates Nest

1 0 -0,07 -0,1 -0,24 0,14 -0,19 0,1 -0,18 -0,12 -0,1 0,11 reheight Tree 1 -0,09 -0,11 0,23 0,05 0,18 0 0,14 0,04 0,16 0,01 -0,08 0,18

-0,03 0,31 -0,05 0,08 -0,27 -0,16 -0,05 -0,2 -0,16 -0,13 -0,19 0,11 -0,13 hu cover Shrub 1 0,01 0,19 0 0,06 -0,25 0,11 -0,12 0,08 -0,2 0,04 -0,29 -0,02 -0,07

0,53 -0,15 0,27 -0,18 0,05 0,05 -0,24 0,23 -0,3 0,22 -0,05 0,2 -0,19 -0,16 0,23 0.50) are noted in bold. cover Litter 1 0,1 -0,26 0,21 -0,04 -0,03 0,3 -0,25 0,31 -0,24 0,08 -0,23 0,24 0,18 -0,21 >

-0,92

A cover Herbaceous 1 aoyclosure Canopy

1 -0,03 0,03 0,1 -0,05 0 -0,03 0,07 -0,19 0,02 -0,18 -0,01 -0,03 0 0,1 -0,07 -0,04 Concealment

1 0,04 -0,09 0,03

0,07 -0,18 0 0,01 0,11 -0,16 0,01 0,06 -0,02 -0,1 -0,23 -0,22 -0,23 -0,19 -0,17 -0,12 0,18 -0,21 itneidgnu edge indigenous Distance 1 0,09 -0,27 -0,06 0,03 0,12 -0,23 -0,04 0 0,01 0 -0,07 0,03 -0,14 -0,07 -0,1 -0,01 0,34 -0,22

0,57 itneedge Distance 1 Pair-wise correlations between 20 habitat variables in one population (out of two) from a heavily degraded (Chawia) aforest fragment in Taita Hills, Kenya. Highly correlated variable pairs (Pearson’s r Distance edge Distance indigenous edge Concealment Canopy closure Herbaceous cover Litter cover Shrub cover Tree height Nest substrates Dead wood Basal area Tree density Sapling density Tree species Sapling species Shannon Tree spec. Shannon Sapling spec. Vegetation composition PC1 Vegetation composition PC2 Vegetation composition PC3 Table A.2:

Page 134

Appendix to Chapter 2 eeaincmoiinPC3 composition Vegetation

1 eeaincmoiinPC2 composition Vegetation

1 0,06 eeaincmoiinPC1 composition Vegetation 1 -0,08 -0,27

-0,25 0,05 0,33 hno aln spec. Sapling Shannon 1 -0,22 0,11 0,3

0,21 -0,18 0,04 0,35 0,84 -0,19 0,13 0,09 -0,22 hno respec. Tree Shannon 1 0,27 -0,3 0,02 0,44

0,13 0,13 -0,04 -0,08 0,11 0,61 -0,58 aln species Sapling 1 0,31

0,54 respecies Tree

1 0,2 aln density Sapling

1 0,04 redensity Tree

1 0,09 0,37 0 0,02 -0,06 -0,09 0,03 0,13 aa area Basal

1 0,46 -0,03 -0,19 -0,15 edwood Dead 1 -0,14 0 -0,13 0,12 0,1 0,09 0,16 -0,07 0,1 0,17

Ngangao etsubstrates Nest

1 0,03 -0,11 -0,02 -0,14 0,2 -0,05 0,2 0,01 -0,17 0,09 0,03 reheight Tree

1 0,21 -0,02 0,07 0,1 0,02 0,04 0,02 0,05 0 -0,08 0,02 0,1 0.50) are noted in bold. cover Shrub

1 -0,04 0,17 0,19 -0,17 -0,02 0,12 0,06 0,01 0,06 -0,03 -0,09 -0,03 0,19

> itrcover Litter

1 -0,11 0,08 -0,09 -0,05 0,15 0,24 -0,05 0,07 -0,2 -0,07 -0,15 0,32 -0,29 -0,09

ebcoscover Herbaceous A 1 -0,97 0,1 -0,07 0,08 0,02 -0,13 -0,21 0,12 -0,08 0,21 0,06 0,13 -0,29 0,25 0,05 aoyclosure Canopy

1 0,11 -0,14 0,21 -0,21 -0,06 0,06 -0,03 -0,07 -0,09 0 0,01 -0,09 0,04 0,05 0,12 -0,03 Concealment

1 0,09 0 0,01 0,46 0,05 0,19 0,13 -0,09 0 0,03 0,12 0,05 0,06 0,01 -0,1 0,05 0,12

0,08 -0,06 0,04 -0,06 -0,1 0,15 0,2 0,19 0 0,14 -0,03 0,18 0,25 0,09 0,33 -0,28 -0,03 0,29 itneidgnu edge indigenous Distance 1 0,07 -0,09 0,08 -0,08 0,01 0,25 0,23 0,09 -0,03 0,18 0,07 0,08 0,25 0,08 0,25 -0,31 -0,08 0,25

0,54 itneedge Distance 1 Pair-wise correlations between 20 habitat variables in one population (out of two) from a moderately degraded (Ngangao) forest fragment in Taita Hills, Kenya. Highly correlated variable pairs (Pearson’s r Distance edge Distance indigenous edge Concealment Canopy closure Herbaceous cover Litter cover Shrub cover Tree height Nest substrates Dead wood Basal area Tree density Sapling density Tree species Sapling species Shannon Tree spec. Shannon Sapling spec. Vegetation composition PC1 Vegetation composition PC2 Vegetation composition PC3 Table A.3:

Page 135 Appendix to Chapter 2

Table A.4: Summary of model selection on a set of three a priori models investigating effect of NSPS on nest success of Placid greenbul nests showing the coefficients with 95% CI on the logistic-exposure scale.

Model Intercept Fragment a NSPS NSPS × Fragment df AICc ∆AICc Fledging Success 1 4.00 (2.64, 6.06) 0.37 (-1.53, 2.46) – – 3 87.9 0.00 2 3.59 (1.19, 6.55) 0.26 (-1.72, 2.40) 0.82 (-3.38, 5.26) – 4 89.8 1.89 3 3.87 (-0.0069, 8.40) -0.18 (-5.49, 5.22) 0.26 (-7.28, 8.45) 0.83 (-8.61,10.16) 5 91.9 3.96 a reference category Chawia, estimated category Ngangao

Table A.5: AICc based model averages (and 95% CI) of 14 predictor habitat variables in nest-site selection. Global model is run on two populations from two forest fragments that differ in level of forest disturbance (Chawia : heavily disturbed; Ngangao : moderately disturbed), all models ∆AICc 6 4 included. Chawia Ngangao Variables Estimate 95% CI Estimate 95% CI (Intercept) 0.12 -0.77, 1.02 0.33 -0.26, 0.92 A Distance indigenous edge 0.02 -0.29, 0.34 0.42 -0.09, 0.93 Concealment 1.09 0.19, 1.99 0.86 0.36, 1.36 Canopy closure -0.01 -0.21, 0.19 -0.04 -0.40, 0.32 Herbaceous cover 0.01 -0.26, 0.28 0.02 -0.23, 0.27 Tree height 0.00 -0.14, 0.14 0.65 0.11, 1.19 Nest substrates 0.74 -0.02, 1.49 0.01 -0.20, 0.23 Dead wood -0.60 -1.29, 0.09 -0.30 -0.88, 0.27 Tree density 0.01 -0.24, 0.25 -0.18 -0.73, 037 Sapling density 0.00 -0.15, 0.16 0.03 -0.20, 0.25 Shannon Tree spec. 0.10 -0.38, 0.59 -0.06 -0.40, 0.28 Shannon Sapling spec. 0.24 -0.39, 0.87 -0.01 -0.19, 0.16 Vegetation composition PC1 -0.24 -1.13, 0.65 0.03 -0.22,0.27 Vegetation composition PC2 -0.83 -2.33, 0.67 -0.08 -0.38, 0.23 Vegetation composition PC3 0.58 -0.85, 2.02 0.00 -0.12, 0.11

Page 136 Appendix to Chapter 2 2 R AICc ∆ 7 86.44 0.11 0.48 7 86.72 0.39 0.48 8 88.02 1.707 0.49 88.33 2.00 0.46 6 86.71 0.386 0.45 87.237 0.90 87.34 0.45 1.018 88.00 0.47 1.677 0.49 88.19 1.86 0.46 9 89.68 3.35 0.50 df AICc and the lowest support (i.e. highest 2 R 4; CH : 87 models; NG : 132) were used 6 AICc ∆ 2). Models with the highest 6 AICc ∆ Chawia

A are reported. 2 R Concealment + Nest substrates + Dead wood + Shannon Sapling spec.Concealment + + Vegetation Nest composition substrates PC1PC3 + + Dead wood +Concealment Vegetation + composition Nest PC2 substrates + + Vegetation Dead composition wood +Concealment Shannon + Sapling Nest spec. substratesPC3 + + Dead Vegetation wood composition + PC2Concealment Vegetation + + composition Nest PC1 substrates + +composition Dead Vegetation PC2 wood composition + Shannon Tree spec. + ShannonConcealment Sapling + spec. Nest + substrates Vegetation +composition Dead PC1 wood + + Shannon VegetationConcealment composition Tree spec. + PC3 Nest + substrates Shannon Sapling + spec. Dead wood + Vegetation +Concealment Shannon + Sapling Nest spec. substratesPC2 + + + Dead Vegetation Vegetation wood composition composition + PC1Distance PC3 Vegetation + to composition PC1 Edge + + Vegetation Concealment composition + Nest substrates + DeadConcealment + wood Nest + substrates + Vegetationcomposition Dead composition PC1 wood + + PC2 Shannon Vegetation + composition Tree spec. PC2 + + Shannon Vegetation Sapling composition spec. PC3 + Vegetation Vegetation composition PC3 Vegetation composition PC3 Vegetation composition PC2 + Vegetation composition PC3 Vegetation composition PC3 Overview of multiple logistic regression models comparing environmental characteristics of 113 Placid greenbul nests with 99 random, non-nest 34 Concealment + Nest substrates + Dead wood + Vegetation composition PC26 5 86.66 0.33 0.42 12 Concealment + Nest substrates + Dead wood + Shannon Sapling spec. + Vegetation composition PC2 6 86.33 0.00 0.46 5 7 89 Concealment + Nest substrates + Dead wood + Shannon Tree spec. + Vegetation composition PC2 6 87.45 1.12 0.44 11 69 10 12 87 Concealment + Dead wood + Shannon Tree spec. + Vegetation composition PC2 5 90.32 3.99 0.38 AICc) are also reported. Nagelkerke’s pseudo Model Model variables sites in in oneto (out calculate of model-averages, two) forest but fragment here that only is the heavily top-models degraded are (Chawia). reported All ( models on a subset ( ∆ Table A.6:

Page 137 Appendix to Chapter 2 2 R AICc ∆ and the lowest support (i.e. 2 7 139.48 1.03 0.35 7 140.00 1.557 0.34 140.29 1.84 0.34 8 141.02 2.577 142.44 0.35 3.99 0.32 R df AICc 4; CH : 87 models; NG : 132) were 6 AICc ∆ 2). Models with the highest 6 AICc ∆ Ngangao are reported. 2 A R Distance indigenous edgecomposition + PC2 Concealment + Tree height + Dead wood + Tree density + Vegetation Distance indigenous edge +spec. Concealment + Tree height + Dead wood + Tree density + ShannonDistance Tree indigenous edgecomposition + PC1 Concealment + Tree height + Dead wood + Tree density + Vegetation Distance indigenous edge +spec. Concealment + + Tree Vegetation height composition + PC2 Dead wood +Distance Tree indigenous density edge + + ShannonSapling Concealment Sapling spec. + Tree height + Tree density + Sapling density + Shannon AICc) are also reported. Nagelkerke’s pseudo Overview of multiple logistic regression models comparing environmental characteristics of 113 Placid greenbul nests with 99 random, non-nest ∆ 12 Distance3 indigenous edge + Distance Concealment4 indigenous + edge Tree height + Distance + Concealment indigenous Dead + edge wood Tree height + + Concealment5 Dead + wood Tree height + +6 Tree Dead density wood + Distance7 Vegetation indigenous composition edge PC2 + Distance Concealment8 indigenous + edge Tree height + Distance + Concealment indigenous Dead + edge wood Tree 6 height + + + Concealment9 Shannon Tree + Tree density Tree spec. height 139.21 Distance indigenous 0.76 edge + Concealment + Tree height 0.33 + 6 Dead wood + Vegetation 138.57 composition PC1 6 0.12 139.86 6 5 0.33 1.41 140.04 138.45 0.32 1.59 0.00 0.32 0.32 5 139.94 1.49 0.30 4 139.96 1.51 0.28 1011 Distance indigenous edge + Concealment + Canopy closure + Tree height + Dead wood 6 140.16 1.71 0.32 1213 Distance indigenous14 edge + Distance Concealment indigenous15 + edge Tree height + Distance + Concealment indigenous16 Vegetation + composition edge Tree height + PC2 Distance + Concealment indigenous Dead + edge wood Tree height + + Distance + Concealment Tree indigenous30 Dead + density edge wood Canopy + + closure Sapling + Concealment + density Sapling + Tree density height Tree height + + Dead Tree wood 7 density + + Tree Vegetation density composition 140.33 PC2 7 1.88 6 140.36 5 140.36 0.34 1.91 140.31 1.91 6 0.34 1.86 140.36 0.32 0.30 1.91 0.32 132 Model Model variables sites in one (outused of to two) calculate forest model-averages, fragmenthighest but that here is only moderately the degraded top-models (Ngangao). are All reported models ( on a subset ( Table A.7:

Page 138 Appendix to Chapter 2

Table A.8: Nest substrates used by Placid greenbul during the 2007-2008 breeding season in Taita Hills, Kenya.

Species Family Relative importance (%) Dracaena steudneri Engl. Asparagaceae 0.27 Chassalia sp. Rubiaceae 0.11 Uvaria lucida Bojer ex Benth. Annonaceae 0.09 Culcasia falcifolia Engl. Araceae 0.07 Landolphia buchananii (Hallier f.) Stapf Apocynaceae 0.06 Fern sp. 0.05 Xymalos monospora (Harv.) Baill. Monimiaceae 0.04 Blotiella stipitata (Alston) Faden Dennstaedtiaceae 0.03 Keetia gueinzii (Sond.) Bridson Rubiaceae 0.03 Oxyanthus speciosus DC. Rubiaceae 0.03 Piper capense L.f. Piperaceae 0.03 Strombosia scheffleri Engl. Olacaceae 0.03 Dasylepis integra Warb. Achariaceae 0.02 Vepris sp. Rutaceae 0.02 Agelaea pentagyna (Lam.) Baill. Connaraceae 0.02 Garcinia volkensii Engl. Clusiaceae 0.02 Justicia pseudorungia Lindau Acanthaceae 0.02 Phoenix reclinata Jacq. Arecaceae 0.01 A Syzygium sp. Myrtaceae 0.01 Turraea holstii Grke Meliaceae 0.01 Pouteria adolfi-friedericii (Engl.) A. Meeuse Sapotaceae <0.01 Camellia sinensis (L.) Kuntze Theaceae <0.01 Psychotria pseudoplatyphylla E.M.A.Petit Rubiaceae <0.01 Rauvolfia mannii Stapf Apocynaceae <0.01

Page 139 Appendix to Chapter 2

Table A.9: Summary of model selection on a set of three a priori Generalized Linear Models investigating the effect of Nest Site Probability Score (NSPS) on fledging success (> 1 nestling fledged). Models were run on 10 different subsets of randomly selected nests (n = 30 nests) that were excluded from the nest-site selection models (n = 83 nests and 99 random points). Model III also contained the main effects of Year and Frag (Fragment). Model I was always the best model, or ∆AICc 6 2.

Model I Fledging success ∼ Year + Frag Model II ... ∼ Year + Frag + NSPS Model III ... ∼ Year + Frag × NSPS df logLik AICc ∆AICc Subset 1 I 3 -19.119 45.2 0.00 II 4 -18.015 45.6 0.47 III 5 -18.009 48.5 3.36 Subset 2 I 3 -18.904 44.8 0.00 II 4 -18.855 47.4 2.61 III 5 -18.724 50.1 5.59 Subset 3 I 3 -16.647 40.2 0.00 II 4 -16.618 42.8 2.62 III 5 -16.411 45.3 5.11 Subset 4 III 5 -8.059 28.7 0.00 I 3 -11.596 30.2 1.42 II 4 -11.485 32.6 3.91 Subset 5 I 3 -18.030 43.0 0.00 II 4 -18.030 45.7 2.71 III 5 -17.969 48.5 5.53 Subset 6 I 3 -15.933 38.8 0.00 A II 4 -15.904 41.4 2.62 III 5 -15.438 43.4 4.59 Subset 7 I 3 -18.856 44.6 0.00 II 4 -18.119 45.8 1.20 III 5 -17.138 46.8 2.14 Subset 8 I 3 -16.890 40.7 0.00 II 4 -16.853 43.3 2.60 III 5 -15.537 43.6 2.87 Subset 9 I 3 -18.409 43.7 0.00 II 4 -18.304 46.2 2.47 III 5 -17.275 47.1 3.31 Subset 10 II 4 -17.061 43.7 0.00 I 3 -18.664 44.3 0.53 III 5 -16.977 46.5 2.73

Page 140 B

Appendix to Chapter 4

Table B.1: Mixed model analysis of female Placid Greenbul egg investment (egg size, cm3) in relation to group size. Reference fragment type for the fixed factor and the interaction was Ngangao.

Fixed term Estimate (± SE) F (n.d.f, d.d.f) p-values Intercept 0.81 ± 1.42 Clutch size -0.065 ± 0.075 0.74 (1, 62.62) 0.39 Tarsus Length Female 0.094 ± 0.053 3.12 (1, 57.24) 0.083 Group size 0.0072 ± 0.022 0.11 (1, 58.13) 0.75 Fragment type 4.16 (2, 55.33) 0.021 Fragment Chawia -0.20 ± 0.070 Small fragments -0.061 ± 0.074 Group size × Fragment type * 0.72 (2, 66.99) 0.49 Group size × Chawia 0.046 ± 0.051 Group size × Small 0.059 ± 0.057

Random term Variance Nest ID 0.0052 Female ID 0.038 B Breeding Season 0.00061 Residual variation 0.013 * Estimates and F-statistics and p-values of interaction obtained from full model

Page 141 Appendix to Chapter 4

Table B.2: Mixed model analysis of female Placid Greenbul egg investment (egg size, cm 3) in relation to number and presence/absence of helpers. Reference fragment type and social condition for the fixed factor and the interaction was Ngangao and absence of helpers. Note that female ID is not added as a random factor in the model due to model convergence issues. We therefore also report p-values of the same model run on a subset of the data whereby we randomly selected one nesting event per breeding female.

Fixed term Estimate (± SE) F (n.d.f, d.d.f) p-values p-values complete dataset complete dataset complete subset dataset dataset Number of helpers Intercept 0.56 ± 1.67 Clutch size -0.15 ± 0.15 1.09 (1, 23.67) 0.31 0.36 Tarsus Length Female 0.11 ± 0.063 3.12 (1, 22.64) 0.091 0.19 Number of helpers -0.041 ± 0.049 0.72 (1, 22.70) 0.41 0.21 Fragment type 1.59 (2, 22.56) 0.23 0.42 Fragment Chawia -0.17 ± 0.096 Small fragments -0.10 ± 0.099 Number of helpers × Fragment type * 1.43 (2, 20.91) 0.26 0.26 Number of helpers × Chawia -0.20 ± 0.15 Number of helpers × Small 0.091 ± 0.12

Presence/absence helpers Intercept 0.42 ± 1.64 Clutch size -0.13 ± 0.14 0.85 (1, 24.09) 0.37 0.49 Tarsus Length Female 0.12 ± 0.061 3.52 (1, 22.60) 0.07 0.19 Presence/absence helpers -0.095 ± 0.072 1.75 (1, 22.56) 0.20 0.18 Fragment type 1.48 (2, 22.52) 0.25 0.38 Fragment Chawia -0.16 ± 0.095 Small fragments -0.11 ± 0.097 Presence/absence helpers × Fragment type * 0.22 (2, 20.57) 0.81 0.60 Presence/absence helpers × Chawia 0.041 ± 0.19 Presence/absence helpers × Small 0.13 ± 0.19 Random term Variance Number of helpers B Nest ID 0.031 Breeding Season 0 Presence/absence helpers Nest ID 0.029 Breeding Season 0 Residual variation - Number of helpers 0.012 Residual variation - Presence/absence helpers 0.012 * Estimates and F-statistics and p-values of interaction obtained from full model

Page 142 Appendix to Chapter 4

Table B.3: Mixed model analysis of maternal per capita provisioning rates in relation to number of non-helping subordinates (i.e., subordinates that do not provision the nestlings). Model fitted on log-transformed response variable (log per capita provisioning rate). Reference fragment type for the fixed factor and the interaction was Ngangao.

Fixed term Estimate (± SE) F (n.d.f, d.d.f) p-values Intercept -1.11 ± 0.73 Nestling age 0.15 ± 0.077 3.91 (1,45.99) 0.054 Mean prey size 0.41 ± 0.47 0.75 (1,10.51) 0.40 Number of non-helping subordinates -0.086 ± 0.078 1.20 (1,44.65) 0.28 Fragment type 0.15 (2,35.72) 0.86 Fragment Chawia -0.031 ± 0.19 Small fragments 0.10 ± 0.24

Random term Variance Female ID 0.038 Breeding Season 0.00067 Residual variation 0.011

Table B.4: Mixed model analysis of Placid Greenbul number of fledglings in relation to number of helpers and presence/absence of helpers. Reference fragment type and social condition for the fixed factor and the interaction was Ngangao and absence of helpers.

Fixed term Estimate (± SE) χ2 (d.f.) p-values Number of helpers Intercept -0.51 ± 1.07 Clutch size 0.49 ± 0.50 0.97 (1) 0.42 Number of helpers 0.037 ± 0.17 0.05 (1) 0.83 Fragment type 0.44 (2) 0.80 Fragment Chawia -0.17 ± 0.32 Small fragments 0.085 ± 0.39 Number of helpers × Fragment type * 0.14 (2) 0.93 Number of helpers × Chawia -0.098 ± 0.42 Number of helpers × Small 0.11 ± 0.49

Presence/absence helpers Intercept -0.50 ± 1.07 Clutch size 0.51 ± 0.51 0.98 (1) 0.32 B Presence/absence helpers -0.034 ± 0.28 0.015 (1) 0.90 Fragment type 0.45 (2) 0.80 Fragment Chawia -0.17 ± 0.32 Small fragments 0.094 ± 0.39 Presence/absence helpers × Fragment type * 0.26 (2) 0.88 Presence/absence helpers × Chawia 0.078 ± 0.62 Presence/absence helpers × Small 0.41 ± 0.80 Random term Variance Number of helpers Mother ID 0 Breeding Season 0 Presence/absence helpers Mother ID 0 Breeding Season 0 * Estimates and χ2-statistics and p-values of interaction obtained from full model

Page 143 Appendix to Chapter 4

Table B.5: Mixed model analysis of Placid Greenbul nestling body condition in relation to group size and presence/absence of subordinate (group-living). Reference fragment type and social condition for the fixed factor and the interaction was Ngangao and pair-breeding.

Fixed term Estimate (± SE) F (n.d.f, d.d.f) p-values Group size Intercept -8.89 ± 4.48 Number of nestlings -0.25 ± 0.59 0.17 (1, 29.16) 0.68 Nestling age 0.14 ± 0.35 0.15 (1, 46.85) 0.70 Nestling tarsus length 1.02 ± 0.19 27.21 (1, 57.05) <0.0001 Mean clutch egg size 0.0015 ± 0.00093 2.47 (1, 27.59) 0.13 Group size -0.30 ± 0.19 2.42 (1, 32.58) 0.13 Fragment type 1.60 (2, 26.31) 0.22 Chawia -0.26 ± 0.69 Small -1.09 ± 0.62 Group size × Fragment identity a 0.21 (2, 30.52) 0.81 Group size × Chawia 0.31 ± 0.49 Group size × Small 0.16 ± 0.50

Presence/absence subordinates b Group-living -0.72 ± 0.48 2.23 (1, 42.35) 0.14 Fragment type 1.55 (2, 28.96) 0.23 Chawia -0.42 ± 0.66 Small -1.11 ± 0.63 Group-living × Fragment identity a 0.11 (2, 31.97) 0.90 Group-living × Chawia 0.87 ± 1.87 Group-living × Small 0.15 ± 1.18 Random term Variance Group size Nest (Female ID) 0 Female ID 1.46 Breeding Season 0 B Presence/absence subordinates Nest (female ID) 0 Female ID 14.75 Breeding Season 0 Residual variation group size model 0.75 Residual variation presence subordinates model 0.71 a Estimates and F-statistics and p-values of interaction obtained from full model b Intercept and covariables not reported for presence/absence subordinate model

Page 144 Appendix to Chapter 4

Table B.6: Mixed model analysis of Placid Greenbul nestling body condition in relation to number of helpers and presence/absence of helpers. Reference fragment type and social condition for the fixed factor and the interaction was Ngangao and absence of helpers.

Fixed term Estimate (± SE) F (n.d.f, d.d.f) p-values Number of helpers Intercept -12.12 ± 4.71 Number of nestlings -0.43 ± 0.54 0.63 (1, 21.48) 0.44 Nestling age -0.13 ± 0.39 0.12 (1, 37.04) 0.74 Nestling tarsus length 1.21 ± 0.22 29.18 (1, 38.06) <0.0001 Mean clutch egg size 0.0016 ± 0.0011 2.17 (1, 15.61) 0.16 Number of helpers 0.16 ± 0.26 0.40 (1, 38.79) 0.53 Fragment identity 0.59 (2, 16.94) 0.56 Chawia -0.22 ± 0.72 Small -0.92 ± 0.85 Number of helpers × Fragment identity a 0.94 (2, 18.87) 0.41 Number of helpers × Chawia 0.78 ± 0.85 Number of helpers × Small 1.02 ± 0.93

Presence/absence helpers b Presence/absence helpers 0.52 ± 0.46 1.30 (1, 32.63) 0.26 Fragment identity 0.48 (2, 16.07) 0.63 Chawia -0.14 ± 0.70 Small -0.80 ± 0.84 Presence/absence helpers × Fragment identity a 1.63 (2, 17.27) 0.23 Presence/absence helpers × Chawia 1.23 ± 1.13 Presence/absence helpers × Small 2.25 ± 1.40 Random term Variance Number of helpers Nest (Female ID) 0 Female ID 1.38 Breeding Season 0 Presence/absence helpers B Nest (Female ID) 0 Female ID 12.67 Breeding Season 0.27 Residual variation number of helpers model 0.64 Residual variation presence of helpers model 0.64 a Estimates and F-statistics and p-values of interaction obtained from full model b Intercept and covariables not reported for presence/absence subordinate model

Page 145 Appendix to Chapter 4

B

Page 146 C

Appendix to Chapter 5

C.1 Literature Review

A literature review on the effects of helpers on post-fledging survival in altricial cooperatively breeding birds yielded a limited number of studies. Three approaches were used: (i) a systematic search in Web of Science using a combination of specific keywords (post-fledging, post-fledgling, postfledging, juvenile survival, or fledgling survival and cooperative bird). (ii) a systematic search for specific cooperative breeding species listed in review papers (Kingma et al., 2010 complemented with additional species noted in the appendix of Li et al., 2015 and in Cockburn, 1998; Hatchwell, 1999; Legge, 2000a; Khan & Walters, 2002) using Web of Science (keywords : Latin name and English name in combination with keywords fledgling, juvenile, postfledging or post-fledging). When multiple studies have been conducted on the same species, all are incorporated if methodology or outcome differs. Otherwise the one with the highest sample size is retained.

C

Page 147 Appendix to Chapter 5 , 2009 et al. , 2011 , 2003 , 1996 , 2012 , 2015 , 2015 , 2016 , 2013 , 2013 , 2004 , 2004 , 2011 , 2009 et al. et al. et al. et al. et al. et al. et al. et al. et al. , 2015 et al. et al. et al. et al. et al. Ridley & Raihani , 2008 Covas Russell Russell Dickinson Sydeman et al. , 1988 Russell &, Rowley 1988 Emlen &, Wrege 1991 Lewis , 1982 Langen &, Vehrencamp 1999 Woxvold & Magrath , 2005 Kinnaird &, Grant 1982 Rold´an Sankamethawee Heinsohn , 1992 Heinsohn , 1992 Caffrey , 2000 Caffrey , 2000 Canestrari Rold´an Williams & Hale , 2006 Lloyd Brouwer McGowan Sherley , 1990 Mumme , 1992a Mumme Mumme Reference )* Williams & , Hale 2006 Cont Cat / POS / POS * / POS * Ridley & Raihani , 2007 Ridley & Raihani , 2007 e (POS * * Li Preston Cat Cont Cat Cat Cat Cat Cat Cont Cont Cat Cont Cat Cat Cat Cont Cont Cat Cat Cont Cat Cont Cat Cont Cont Cont Cat Cont Cont Cat Cat Cat Cont Cat NEG NEU NEU POS Helper effect on survival d Y early W eekly Y early Y early : number of helpers analysed continuously or categorical. When effect CAT Statistical Method and c CONT Detection Method b Methodology 49 (ND) O, MT T NEU 90 (ND) O T NEU 365 O, MT T NEU ± 365 RT T NEU 36560± 45 (ND) O O RT T T T NEU NEU NEU 365 T NEU 365365 O O T T NEU POS 1010 - 70 (ND) * O, MT O, MT T T NEU POS Follow-up period (days) a ) AT (7) 180 O, MT T NEU ) AA (4) 30 * O, MT T NEU ) NT (2) 180 (ND) O T NEU ) NT (13) 365 O, MT T NEU ) AT (7) 180 (ND) O T NEU ) AT (1) 365 + O, MT CMR ) AT (12) 21 O T NEU ) NA (5) 60 (ND) O T POS ) AA (12) ) IM (2) 56 (ND) O CMR ) AA (13) 365 O T NEU , 2001 )) ) PA (13) 56 (ND) O T NEU ) NA (12) 14 O T NEU ) AA (12) 365 O T NEU et al. ) PA (4) 365 + O CMR ) PA (10) 365 MT T POS ) AT (13) 365 + T CMR ) NA (5) 365 O, MT T NEU ) AA (8) 365 O, MT T NEU Plocepasser mahali ) NA (12) 365 O T NEU ) NT (1) 30 O T POS ) AA (4) 365 O T NEU C ) PA (12) Calocitta formosa Merops bullockoides Nesomimus parvulus Turdoides bicolor Corcorax melanorhamphos Eopsaltria georgiana Alophoixus pallidus Cercotrichas coryphoeus Acrocephalus sechellensis Malurus splendens Aphelocoma coerulescens Turdoides squamiceps Sialia sialis Sitta pygmaea Corvus brachyrhynchos (Philetairus socius Aegithalos caudatus Corvus corone Struthidea cinerea Pseudopodoces humilis Cyanocorax morio Acanthisitta chloris Overview of key methods and results of case-studies on effects of helpers on post-fledging survival in altricial cooperatively breeding birds. Realm and (biome) sensuSince (( Olson fledging, or notedO if : otherwise. targeted Recalculated observation to or days study using population 7 censusing; days MT in : a mist-net week, traps; 30 RT in : a radio-telemetry month and 365 in a year. ND : period coincides with nutritional dependency T : effect ofrecapture/resighting helpers analysis on incorporates post-fledging temporal survivalPOS, variability assessed NEU (timeframe using & for temporal NEG parameterof invariable : estimation statistical number positive, in methods of neutral subscript) (ex. allofeeders orsee on regression negative references survival analysis, effect for differs GLM(M), of study-specific this contingency helpers is details tables); on indicated CMR post-fledging using : survival. a capture-mark forward slash ( / ). a b c d e * Sociable Weaver Southern Pied Babbler ( White-breasted Robin ( Western Bluebird ( Pygmy Nuthatch ( White-browed Sparrow-weaver ( White-throated Magpie-jay ( Apostlebirds ( Galapagos Mockingbird ( Splendid Fairywen ( White-fronted Bee-eater ( White-winged ( American Crow ( Carrion Crow ( Puff-throated Bulbul ( Karoo Scrub Robin ( Brown Jay ( Seychelles Warbler ( Ground Tit ( Rifleman ( Long-tailed Tit ( Arabian Babblers ( SpeciesFlorida Scrub Jay ( Ecoregion Table C.1:

Page 148