Ecological opportunity leads to the emergence of an alternative behavioural phenotype in a tropical bird

Ecological opportunity leads to the emergence of an alternative behavioural phenotype in a tropical bird

Janeene M. Touchton

* and Martin Wikelski

1Department of Ecology and Evolutionary Biology, 106A Guyot Hall, Princeton University, Princeton, NJ 08544-2016, USA;²Max-Planck-Institut f€ur Ornithologie, Am Obstberg 1, D-78315, Radolfzell, Germany;³Smithsonian Tropical Research Institute, Apartado Postal 0843-03092 Panama, Repu´blica de Panama; and ⁴Department of Biology, Universit€at Konstanz, Postfach M633, D-78457, Konstanz, Germany

Summary

1. Loss of a dominant competitor can open ecological opportunities. Ecological opportunities are considered prerequisites for adaptive radiations. Nonetheless, initiation of diversification in response to ecological opportunity is seldom observed, so we know little about the stages by which behavioural variation either increases or coalesces into distinct phenotypes.

2. Here, a natural experiment showed that in a tropical island’s guild of army ant-following birds, a new behavioural phenotype emerged in subordinate spotted antbirds (Hylophylax naevioides) after the socially dominant ocellated antbird (Phaenostictus mcleannani) died out.

3. Individuals with this behavioural phenotype are less territorial; instead, they roam in search of ant swarms where they feed in locations from which dominant competitors formerly excluded them. Roaming individuals fledge more young than territorial individuals.

4. We conclude that ecological opportunity arising from species loss may enhance the success of alternative behavioural phenotypes and can favour further intraspecific diversification in life-history traits in surviving species.

Key-words: alternative behavioural phenotypes, ant-following birds, behavioural diversifica- tion, competitive release, ecological opportunity,Hylophylax naevioides, individual differences, territorial breakdown

Introduction

Local extinction of a species can alter the costs and benefits of resource acquisition for its fellow community members. For competitors, such an extinction could open an ecological opportunity a surplus of unexploited resources long considered a necessary condition for adaptive radiation (Simpson 1953; Schluter 2000; Losos 2010; Yoder et al.2010). By relaxing stabilizing selection, ecological opportunity presumably leads to ecological release, diversification and ultimately speciation (Yoder et al.2010). Yet, we know little about how exploiting eco logical opportunity can lead to increased trait variation and subsequent evolutionary diversification (Losos 2010;

Yoder et al.2010).Most studies of phenotypic divergence resulting from ecological opportunity rely on clades that have already diversified (e.g. Parent & Crespi 2009; Martin

& Wainwright 2011; Setiadiet al.2011; Priceet al.2014).

The importance of intraspecific trait variation particu larly behavioural in ecological and evolutionary processes is becoming increasingly clear (Bolnick et al.

2011; Sih et al. 2012). Additionally, the notion that behaviour takes a leading role in adaptive responses to environmental change is more and more widely accepted (West

Eberhard 2003; Duckworth 2008; Zuket al.2014). Evolu tionary models have shown how behavioural changes could lead to associated evolutionary responses in mor phological traits, for example as adoption of arboreal for aging behaviour in Columbiformes led to changes in tarsal and tail length (Lapiedraet al. 2013). Additionally, a growing body of studies document rapid evolutionary changes following behavioural shifts in invasive species or in response to invasive species [e.g. flatwing morphs lack ing song producing apparati in invasive Pacific field crickets (Teleogryllus oceanicus) (Zuk, Rotenberry & Ting hitella 2006) and adaptive anti predatory defences from fence lizards (Sceloporus undulates) following invasion from predatory fire ants (Solenopsis invicta) (Langkilde 2009)]. Similarly, behavioural shifts may result in adaptive changes in species facing new opportunities as a result of the loss of natural competitors. Individuals that express different behaviours may undergo differing selective pressures when either intra or interspecific competition

*Correspondence author: E mail: touchtonj@gmail.com

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-0-292225

changes (Grant & Price 1981; Calsbeeket al. 2004), lead ing to subsequent evolutionary diversification. The study of individual variation in behaviour during early responses of populations to ecological opportunity follow ing species extinction may thus help reveal how popula tions begin diversifying in the face of new opportunity.

Where pre extinction baseline data are available, local extinctions provide natural experiments on how individu als respond to ecological opportunity. An exceptional nat ural experiment on how ecological opportunity opened by a competitor’s extinction allows behavioural diversifica tion is provided by the guild of ant following birds on an island in central Panama. This guild originally included a dominant competitor, 50 g ocellated antbirds (Phaenostic tus mcleannani), a subdominant competitor, 30 g bicol oured antbirds (Gymnopithys leucaspis), and a subordinate species, 16 g spotted antbirds (Hylophylax naevioides). An 11 year study by E. O. Willis provides data on the popu lation density, behaviour, and social hierarchy of spotted, bicoloured and ocellated antbirds on Barro Colorado Island Panama (Willis & Oniki 1978), 50 years after its isolation by Gatun Lake, which was created to form the Panama Canal’s central segment. Dominant ocellated ant birds were common during the study’s first 5 years and declined afterwards towards extinction (Willis 1967, 1972). Population, density and behavioural data are also available from a nearby mainland site (8 km distant) where ocellated antbirds are still common (Styrsky 2003;

Styrsky, Brawn & Robinson 2005; Touchton & Smith 2011). The minimum distance from Barro Colorado Island to the mainland is 200 m, much further than ant birds can fly over continuous water (Moore et al. 2008).

Our study population on Barro Colorado is therefore effectively isolated, as immigrants from the mainland can not reach it. Thus, we can compare findings from (i) a locale that was studied in detail before ocellated antbirds disappeared (historic competitor present), (ii) the same locale nowadays, after ocellated antbirds disappeared (current competitor absent) and (iii) a nearby control locale where ocellated antbirds are common (current com petitor present), to assess how the extinction of a domi nant competitor affects individuals of surviving species.

In locales where all three species coexist, ocellated and bicoloured antbirds feed exclusively at army ant swarms.

The larger ocellated antbirds feed at the army ant swarm’s richest sites, from which they exclude antbirds of other species. Spotted antbirds only attend army ant swarms when a swarm enters their territory (Willis & Oniki 1978;

Zimmer & Isler 2003): even then, more dominant obligate antbirds restrict spotted antbirds to inferior swarm edges.

These three ant following species accordingly depend on army ant swarms to different extents, representing two very different types of space use. We define these types of space use here as (i) roaming ant following and (ii) terri torial foraging. Roaming ant following is adopted by obli gate ant followers such as ocellated and bicoloured antbirds: they require large home ranges and relaxed terri

toriality to forage at the scattered, nomadic army ant col oniesEciton burchellii andLabidus praedator,which flush the insects they depend on. Pairs and family groups of these obligate ant followers forage in subordinate posi tions on neighbouring territories (Willis 1973; Chaves Campos et al. 2009). Territorial foraging, on the other hand, occurs in the facultative ant following spotted ant birds. These individuals normally maintain and defend much smaller all purpose territories that they seldom leave (Willis 1972; Styrsky 2003). Spotted antbirds follow army ants only when the ants enter their territories, because territorial neighbours normally exclude both mated pairs and young individuals from their territories (Willis 1967, 1972). When army ant swarms are not pass ing through their territories, spotted antbirds will forage within their territorial borders either alone, with their mates or offspring, or on the periphery of mixed species foraging flocks comprised of antwrens and their allies (Willis 1972). Obligate ant following is thought to have evolved from facultative ant followers that ventured beyond normally well defended territorial borders to for age at ant swarms (Brumfieldet al.2007; Chaves Campos et al.2009).

Before ocellated antbirds disappeared, the behaviour of Barro Colorado’s bicoloured and spotted antbirds resem bled their behaviour in nearby sites where ocellated ant birds are still present (Willis 1972; Styrsky 2003). After ocellated antbirds disappeared from Barro Colorado, bicol oured antbird numbers did not increase, though they had access to the richer foraging locations at ant swarms for merly defended by ocellated antbirds, which suggests that their numbers are limited by other factors (Touchton &

Smith 2011). On the other hand, according to various cen suses, only decades after the ocellated antbirds disappeared spotted antbird density more than doubled, as did their numbers at army ant swarms suggesting a shift in territo riality for this species occurred (Touchton & Smith 2011).

These findings indicate that only the spotted antbird popu lation exploited and grew from the ecological opportunity opened by the ocellated antbird’s extinction on Barro Colo rado. Yet, the question remained: What behavioural shifts by spotted antbirds were required to exploit this opportu nity and did responses vary between individuals?

This circumstance offered a rare chance to learn how ecological opportunity allows the expression of novel behaviours that might promote behavioural dimorphism.

Moreover, interspecific competition is thought to be ele vated in tropical bird communities (Jankowski et al.

2012), making these communities a good testing ground for the study of behavioural responses to competitive release and one where findings are likely to be general.

Given the ecological release of spotted antbirds on Barro Colorado (hereafter referred to as the competitor absent site) after their dominant competitor disappeared, we doc umented the change in resource use among spotted antbirds and asked whether this release allowed (i) indi vidual differences in the exploitation of that ecological 1042

opportunity and (ii) differences in fitness resulting from contrasting resource use.

Materials and methods

STUDY SITES AND ANIMALS

We studied spotted antbirds during the breeding season (April October) in 2007 and 2008 at an isolated competitor absent site, Barro Colorado Island, Panama, and a nearby competitor pres ent site in Soberania National Park, Panama. Information about these study sites and study plots can be found in Touchton and Smith (Touchton & Smith 2011). Our competitor absent site was in the area where Willis worked in the 1960s (Willis 1967, 1972).

Spotted antbirds are suboscine passerines that range through out Central America, Northern Columbia and Western Ecuador.

Monogamous pairs form long term pairbonds. Though they only breed during the rainy season, all purpose territories are defended year round. Both males and females contribute equally to terri tory defence and parental care (Willis 1972; Styrsky 2003).

DATA COLLECTION

To record behaviours and home ranges of spotted antbirds at both competitor present and competitor absent sites, we attached 0· 75 g transmitters (Holohil Systems Ltd., Carp, Ontario, Canada) using figure 8 harnesses to eleven male spotted antbirds at the competitor present site and 26 males at the competitor absent site. Because both male and female spotted antbirds often forage together in pairs, and are similarly territorial (Willis 1972), we focused our sampling effort on males for logistical reasons.

We were only able to recapture and obtain sufficient data from three males during both years of our study. Though these males showed consistent behaviours between years (see Results), we

only present data from the first year of study for these three males for our analyses.

For 3 6 weeks, we followed radiotagged males from a distance of 10 20 m every day during pre assigned hour long observa tiona) periods all times of day, using a SIKA receiver and Y AGI antennae (Biotrack Ltd., Dorset, UK). During hour long obser vational periods, we recorded (i) whether individuals were present at an army ant swarm upon first encounter, and (ii) the total time spent actively foraging or engaging in aggressive acts {aggressive vocalizations, charges, posturing, displacements or submissive retreats) during one continuous 5 min observation. Additionally, in the competitor absent site in 2007, we made 9 10 continuous hour long observations for 12 males where we similarly recorded aggressive acts as above. We used these data to calculate the total proportion of time individuals were engaged in aggressive activity (aggressive vocalizations, charges, aggressive posturing, aggressive displacing or attacks) towards another individual. To augment our sample sizes, we observed foraging during continuous 5 min observations as above in non radiotagged but individuaUy banded male and female spotted antbirds (n 109) [when refer ring to the ringed birds] at army ant swarms to estimate variation in the proportion of time spent foraging while at ant swarms between the competitor present and competitor absent sites.

Following behavioural observations, in each observational per iod we used GPS devices {Garmin 60CSx; Garmin International, Inc., Olathe, KS, USA) to record the locations where individuals were first I oca ted to de tinea te home ranges. For each radio tracked individual, we used 20 60 GPS points to delineate home ranges. Home range sizes were then estimated with 95% fixed kernel estimators with least squares eross validation in Biotas (BiotasTM 2005), although Fig. I depicts 100% minimum convex polygons for ease of visual interpretation. To compare current day home ranges to historical ranges in Fig. I, we delineated ranges based on Willis's sketches {Willis 1972) of the centre of ranges and descriptions of borders with respect to neigh houri ng

{a) 10

8 6 4 2 0

{c) Competitor-absent site ^{e) Competitor-present site

a:

(b)

350

g.

;:: 200

~

a:

I

0

--.--

1 I

g

Absent Present Competitor

{d)_

(l) c:

(/)

~ a.

2007 2008

~

~ ~

(f)

c

(l) (/)

a. ~

....

g

Qi a.

E 0 (.)

1998-2001

Fig. 1. Territorial breakdown in spotted ant birds following ecological opportunity. (a) home range sizes (ha) and (b) total percentage of home range overlap for radiotracked male spotted ant birds in the current competitor absent (n 26) and competitor present {11 ll) sites (95% kernel estimates). Range maps {100% minimum convex polygons) of these same males in current day (c) competitor absent and (e) competitor present sites. Darker areas represent overlapping ranges: the darker the shade, the greater number of overlapping ranges. Light grey tines depict trails, while asterisks represent residential non radiotagged males' nest locations. Historical ranges in the (d) competitor absent site, while ocellated antbirds were still present in 1964 and (f) competitor present site in 1998 2001 (redrawn from home range sketches by Willis (1972) (see Materials and methods) and Styrsky (2003).

pairs. Proportion of home range overlap was estimated by sum ming each overlapping proportion of an individual’s home range with all other portions of overlapping ranges by other individu als, amounting to>100% overlap in some cases.

To distinguish roamers from territorial individuals, we created a foraging style index score by multiplying the proportion of an indi vidual’s total estimated time spent attending army ant swarms by their total calculated range size (ha). We used this index to capture both elements necessary for roaming behaviour: large undefended home ranges, and increased time spent seeking out and foraging at army ant swarms. Though increased home range appears to be associated with increased time spent at army ant swarms (see Results), which suggest that time spent foraging at army ants alone would be sufficient to determine roaming behaviour, we included home range in our index to ensure we distinguish roamers from territorial individuals that may have profited disproportionately during our study duration from army ant colonies that remained in their territories longer than usual. We classified an individual with a score over 1 as a roamer, a clear biologically relevant cut off in the distribution of both populations. ‘Roaming’ individuals routinely foraged at ant swarms over 30% of their overall time, an amount which, judging by army ant movement and density, is not feasible in a territory of traditional size (3 4 ha).

As a measure of fitness, we quantified the reproductive success of roaming spotted antbirds in the competitor absent site, vs. terri torial conspecifics at both competitor absent and competitor pres ent sites. We monitored the total number of fledged offspring for all focal individuals throughout the 2008 breeding season (April November). Nests were monitored every 3 4 days scheduled around expected transition dates (egg laying, incubating, hatching and fledgling). Nest owners were confirmed during behavioural observations during nest monitoring (nearly all individuals in our study areas were individually banded), and nestlings were uniquely banded 1 2 days prior to fledging. Only those individuals we were able to monitor for the entire breeding season were included in our analysis.

s t a t i s t i c a l a n a l y s e s

To analyse the proportion of time radiotagged male individuals spent attending ant swarms in both study sites, we used theLME4 package (Bates, Maechler & Bolker 2011) to run a generalized linear mixed model (GLMM) with a binomial error distribution.

We included presence/absence at ant swarm attendance as a bin ary response variable (nobs 929) with site (competitor absent vs.

competitor present) as a fixed variable and set individual ID as a random variable to account for repeated sampling. To analyse the proportion of time all observed males and females spent for aging while at ant swarms we similarly ran a GLMM with forag ing/not foraging as a binary response variable (nobs 612), site (competitor absent vs. competitor present) as a fixed variable, and set individual ID as a random variable to account for repeated sampling. We included sex as a variable in our final model even though it was insignificant (P>005) because it slightly improved the fit of our model (AIC 372 vs. 374).

Reproductive output of individuals was analysed with a general linear model (GLM) with Poisson error distributions. We included site (competitor absent vs. competitor present) and for aging strategy (roaming or territorial) as fixed variables and num ber of offspring as a response variable. A Poisson error distribution was used to account for the integer valued variable

‘number of offspring fledged’.

Results

e v i d e n c e o f e x p l o i t a t i o n o f e c o l o g i c a l o p p o r t u n i t y

By radiotracking reproductively active spotted antbird males at the competitor absent site, we discovered that the mean proportion of time birds spent following ant swarms (018 002 SE;n 26) was greater than that of counterparts at the competitor present site (007002 SE;n 11) (GLMM binomial:v²₁ 81,P 0005). Only some spotted antbirds, however, apparently exploited the newly released resources at ant swarms. At the competi tor absent site, 12 individuals spent more than 25% of their time at ant swarms (half of which spent more than 40% of their time at ant swarms), while nine individuals spent <20% of their time at ant swarms. On the other hand, only one individual in the competitor present site was able to spend up to 25% of its time at ant swarms, with all other individuals spending<20% of their time at ant swarms.

Further, behavioural observations showed that, presum ably because competitor absent spotted antbird males and females spend less time avoiding or fleeing competitors, they spend 33% more of their time foraging while at ant swarms (mean proportion foraging 078 001 SE;

n 43 males, 39 females) than do their competitor pres ent conspecifics (mean proportion foraging 059003 SE; n 28 males, 27 females) (GLMM binomial:

v²1 218,P<000001).

e x p l o i t a t i o n o f e c o l o g i c a l o p p o r t u n i t y t h r o u g h t h e e m e r g e n c e o f a n a l t e r n a t i v e b e h a v i o u r a l p h e n o t y p e

Monitoring spotted antbird adults in the competitor absent site revealed that some adopted an alternative space use tactic: roaming ant following. Yet, spotted antbird mean home range size does not differ between competitor absent and competitor present sites (Fig. 1a;

two sample ttest; t35 13, P 02) and is similar to historical reports (Willis 1972). While 30% of male individuals in the competitor absent site now have home ranges (5 9 ha) that are larger than in the 1960s prior to the loss of ocellated antbirds (3 4 ha), or to those in the competitor present site (2 4 ha), 10% of individuals in the competitor absent site have smaller home ranges (1 ha) than previously reported. Greater home range size by some individuals appears to have led to far greater home range overlap between competitor absent individu als compared to competitor present individuals (Fig. 1).

In the competitor absent site, spotted antbirds have partially or completely overlapping ranges with up to nine other radiotagged individuals, amounting to a cumulative overlap of 340% (see Materials and methods). By con trast, range overlap by competitor present individuals never exceeds 30% and only occurs along territorial 1044

borders (competitor absent vs. competitor present home range overlap: two sample t test; t₃₅ 6·5, P < 0·00001).

In the competitor absent site, home ranges are larger for spotted antbirds that spend more time following army ant swarms (Fig. 2; Pearson's r 0·5, n 26, P 0·005).

Based on our foraging style index (see Materials and methods), in a spotted antbird population that was uni formly territorial 40 years ago, over 40% of its males are now roamers (Fig. 3). The spotted antbirds that now

10

a;-

Q) 6

Cl c:

~

Q)

E 4

0 I

2

.. . ^. .

0·0 0·1 0·2 0·3 0·4 0·5 Proportion time at ant swarms

Fig. 2. Positive relationship between home range size (ha) and proportion of total time spent attending army ant swarms of radiotracked competitor absent males (Pearson's r 0·5, n 26, p 0·005).

0·25 0·20 0·15 0·10 0·05

>.

(.) 0·00 c:

Q) :J r:::r

~

0·20 0·15 0·10 0·05 0·00

Competitor -present

+T R+

Competitor-absent

+T.R+

I I I I I

rnm1~ ^DOD

0 2 3 4 5

Foraging style index

Fig. 3. The distribution of radiotracked male spotted antbirds in the competitor present site (top panel; n II) and competitor absent site (lower panel; n 26) based on foraging style index (the product of an individual's home range size (ha) and proportion of total Lime spent attending ant swarms). We consider individuals with a foraging style index score of less than one territorial forag ers (1) and greater than one roaming ant followers (R).

roam, and specialize more on foraging at army ant swarms, engaged in fewer aggressive acts (see Materials and methods) per proportion time than territorial individ uals (Pearson's r 0·63, n 12, P 0·027). Though we have few individuals we were able to obtain sufficient data from in both years of our study, they were consistent in foraging strategy during both years of the study (Fig. 4).

DIFFERENTIAL REPRODUCTIVE SUCCESS OF ALTERNATIVE BEHAVIOURAL PHENOTYPES

Roaming spotted antbirds evidently experience a consider able fitness payoff (Fig. 5). Roaming individuals fledged

2

~ 1·5

"C

.5

Ol c:

·~

&

t R

---

~ T

o~--~---~---

2007 2008

Fig. 4. Foraging style index (the product of an individual's home range size (ha) and proportion of total time spent attending ant swarms] of three males (each depicted by a different symbol) observed both in 2007 and 2008. Horizontal dashed tine separates territorial foragers (T) and roaming ant followers (R). Individuals were consistent in foraging style between both years.

3

0

Roaming

Competitor-absent 0 Competitor-present •

Territorial

Fig. 5. Reproductive output of males with different space use strategies. Mean ± I SEM number of ftedglings produced by roaming individuals and by territorial spotted antbirds during the 2008 breeding season. Light grey bars depict individuals from the competitor absent site (n 5 roaming, 9 territorial), while dark grey bars depict individuals from the competitor present site (n 0 roaming, II territorial).

twice as many offspring as territorial individuals from either the competitor absent or competitor present sites (GLM poisson: foraging strategy;F1,23 490;P 0037), while the number of fledged young of territorial individu als from competitor absent and competitor present sites did not differ (GLM poisson: Site; F_1,23 000;

P 0983). Overall levels of nest failure (nearly all due to predation) of spotted antbirds were similar between the sites and therefore do not contribute to the above differ ences [Mayfield method for % nest failure (Mayfield 1961) competitor absent:n 69, 79%; competitor present:

n 22, 71%].

Discussion

Here, we document a case where fragmentation facilitated the emergence of an alternative behavioural phenotype by leading to the sustained loss of a dominant competitor, thereby releasing new resources to a competitively subor dinate species. A population that was uniformly territorial 40 years ago now includes roamers that forage more at nomadic army ant swarms, exploiting resources released by the disappearance of ocellated antbirds. Only some spotted antbirds exploit these released resources; others have remained strictly territorial. Though we lack data on lifetime reproductive success and fledgling survival, roam ing spotted antbirds evidently experience a dramatic fit ness payoff; they fledge twice the number of offspring as their territorial counterparts in either their own competi tor absent population or the nearby competitor present population with ocellated antbirds.

Previous data collected on foraging resources suggest that the behavioural differences we observed in spotted antbirds are indeed a result of the loss of ocellated ant birds. First, resource abundance (density of ant swarms and insects flushed by ant swarms) was remarkably simi lar at the time of this study between the competitor absent and competitor present sites (Touchton & Smith 2011). Second, the army ant density documented at the time of this study (Touchton & Smith 2011) is similar to previous reports in the competitor absent site from multi ple censuses that have occurred over time since 1949 to the present day, prior to (Schneirla 1949, 1956; Willis 1967), and following the demise (Franks 1982; Touchton

& Smith 2011) of ocellated antbirds. Yet, though recent differences in foraging resources are not observed, differ ences in prey density or behaviour could have occurred between 1970 and the present. And, while E. burchellii army ant colonies may have a buffering effect for ant fol lowing birds in times of drought or heavy rains, for exam ple by increased or decreased swarming intensity based on their own food requirements, we still cannot rule out the possibility that in between census years army ant density itself could have fluctuated as a result of adverse or bene ficial conditions. Nor can we rule out other possible changes that may have fluctuated over the years since the loss of ocellated antbirds, subsequently affecting spotted

antbird behaviour, or nest predation, for example. Ideally, these issues would be addressed by a larger sample size of closed populations that had experienced competitive release with which to draw conclusions. One of the rea sons this study is unique, however, is the very fact that such populations are difficult to identify, particularly in the face of behavioural diversification with supporting his torical data. Nevertheless, were ocellated antbirds still present on Barro Colorado, it is highly unlikely that spot ted antbirds that adopted a roaming foraging strategy could persist, since ocellated antbirds excluded spotted antbirds from foraging locations at army ant swarms on Barro Colorado in the past when present and currently in the competitor present site (Willis 1972; Touchton &

Smith 2011). Hence, we conclude that the territorial breakdown and emergence of alternative foraging strate gies we observe in spotted antbirds here is a response at least in part to the ecological opportunity at ant swarms that was opened by the disappearance of ocellated ant birds.

Roaming a new behaviour for spotted antbirds was evidently the way in which individuals were able to exploit this ecological opportunity and involves different space use and foraging tactics. Although space use now varies continuously among individuals from strict territo riality to extreme roaming in the competitor absent site, there appears to be a maximum territory size (at approxi mately 4 ha, the average territory and range size reported historically on Barro Colorado (Willis 1972) and in the competitor present site, (Styrsky 2003)) that spotted ant birds can defend effectively. Tropical passerines com monly hold all purpose, year round territories with stable boundaries (Morton, Derrickson & Stutchbury 2000).

Before ocellated antbirds disappeared from Barro Colo rado in the 1970s, year round defence of all purpose terri tories allowed spotted antbirds exclusive use of nest sites and sufficient foraging resources to provision themselves and their fledglings (Willis 1972). On mainland sites with ocellated antbirds, spotted antbirds still defend such terri tories (Styrsky 2003). On sites without ocellated antbirds, spotted antbirds probably cannot both substantially increase their degree of roaming and successfully maintain and defend a territory. Thus far, relaxation of territorial ity in bird species that are normally strictly territorial has been reported only for individuals in inferior condition or habitat quality who intrude or pass through conspecific territories to forage (Fort & Otter 2004; Mazerolle &

Hobson 2004; Sol et al. 2005). Here, exploitation of new opportunities by genuinely competitive individuals also appears to lead to relaxed territoriality.

Specific traits of individual spotted antbirds may have enabled them to adopt new space use and foraging tactics in response to ecological opportunity, subsequently increasing behavioural variation among individuals. Spot ted antbirds that now roam, focusing more on foraging at army ant swarms, are evidently less aggressive than terri torial individuals. Perhaps, reduced aggression in these 1046

individuals was one of the attributes that led to the suc cess of the rise of the roaming phenotype. Aggressive ter ritorial owners engaged in aggressive disputes with owners from neighbouring territories conceivably might not have engaged as much with these more passive individuals particularly if they were foraging at an army ant swarm.

Similarly, the most successful bird invaders of new urban environments and opportunities in New South Wales are less aggressive than less successful invaders (Sol, Bartom eus & Griffin 2011). Levels of aggression have been shown to have a genetic component in other passerine bird spe cies (van Oers et al. 2005; Duckworth & Badyaev 2007).

Thus, in the face of ecological opportunity, individual dif ferences in levels of aggression could predispose certain spotted antbird individuals to become roamers.

Yet, other studies have found that ecological opportu nity alone may be insufficient to explain emergence of alternative phenotypes and increased trait variation. In one of the only other experimental tests of ecological release, removal of one competitor, trout (Oncorhynchus clarki), resulted in among individual diet expansion in stickleback (Gasterosteus aculeatus) individuals, but removing another competitor, sculpin (Cottus asper), had no such effect (Bolnick et al.2010). Studies on spadefoot tadpoles (Martin & Pfennig 2010) andZenaidadoves (Sol et al. 2005) indicate that both intraspecific competition and ecological opportunity are critical for changes in diet breadth among individuals and the establishment of resource polymorphisms extreme alternative phenotypes.

Similarly, Winkelmann and others (Winkelmann et al.

2014) conclude that competition played an important role in speciation and in maintenance of ecomorphs in cichlid fish in Lake Tanganyika, while Bailey and others (Bailey et al. 2013) found the extent of diversification in experi mentally altered Pseudomonas microcosms to be deter mined by the strength of resource competition by conspecifics, in relation to the degree of ecological oppor tunity provided by levels of interspecific competitors.

Likewise, in the spotted antbirds here, intraspecific com petition alone was evidently not sufficient to cause some individuals to roam. Willis (1972) reports a high degree of competition for territories, yet he observed no roaming spotted antbirds. Though intraspecific competition for ter ritories may have facilitated the success of roaming in spotted antbirds, individuals with roaming behaviour had to appear in the population for this ecological opportu nity to be exploited, an event that apparently took dec ades to occur. Thus, whether or not the roaming phenotype became successful in the face of new ecological opportunity through increased density, changes in behav iour such as aggression, or both, remains uncertain.

Such behavioural changes by individuals in response to environmental change can either inhibit or drive evolu tionary change (Duckworth 2008). Plasticity in behaviour al traits, for example, is generally thought to impede evolutionary change by dampening or preventing diver gent selection. Nonetheless, there is growing evidence that

adaptive plasticity can, in some cases, promote canaliza tion of genetic traits during adaptive evolution (West Eb erhard 2003). For example, marine sticklebacks exhibit diet induced plasticity that often coalesces into dimor phism in freshwater populations following changes in het erospecific competition (Wund et al. 2008), and spade foot tadpoles in allopatry demonstrate similar experimen tally initiated niche shifts that are analogous to canalized shifts present in sympatric populations (Pfennig, Rice &

Martin 2006), which are considered to be derived from the ancestral allopatric population (Rice, Leichty & Pfen nig 2009). Evidently, when considering whether or not phenotypic plasticity will play a role in divergence in spe cies traits, it is important to assess the degree of plasticity individuals can express. Highly plastic traits may not become genetically differentiated because divergent selec tion is too weak (Price, Qvarnstrom & Irwin 2003). The evidence we have with the spotted antbirds studied here suggests there may be little behavioural plasticity: we observed consistency in roaming vs. territoriality by the three individuals we were able to follow for 2 years.

Though, it is possible that there are ontogenetic shifts that occur from roaming to territorial or vice versa even after individuals become reproductive.

Our observations that roaming spotted antbirds now fledge twice as many offspring as their territorial counter parts could have evolutionary implications if pairs mate assortatively with respect to space use and foraging tactics and there is cultural inheritance from parent to offspring of those same space use and foraging tactics. Though con tentious, in some cases, a resource use dimorphism can facilitate speciation, while in others, it may inhibit further evolutionary diversification (Smith & Skulason 1996;

West Eberhard 1998, 2003, 2005; Pfennig & McGee 2010). Currently, we do not know whether spotted ant bird pairs mate assortatively with respect to the alterna tive behavioural phenotypes studied here, but there is reason to suggest they might. In spotted antbirds, success ful defence of territories requires both male and female participation (Willis 1972). Additionally, there appears to be ample opportunity to learn a parent’s foraging and space use tactic. We have observed roaming parents bring their fledglings near the edges of army ant swarms while they forage, and territorial parents hide their fledglings in a well protected area of their territory while foraging.

Imprinting by young on their parent’s behaviour can increase the likelihood of following similar foraging strat egies and assortative mating (see Dukas 2013).

In conclusion, novel individual variation appeared to be essential for the behavioural adaptations we observed in the face of new ecological opportunity. Recent experi mental and theoretical work have revealed the importance of including individual differences in predictions of popu lation responses to environmental change (Bolnick et al.

2011). Moreover, it is increasingly clear that individual variation in behaviour can initiate new pathways for ecological and evolutionary change (Duckworth 2008;

Pfennig et al. 2010; Bolnick et al. 2011; Sih et al. 2012).

Behavioural and physiological traits are highly responsive to rapid environmental change and can influence the strength and direction of selection pressures on other aspects of the phenotype, such as morphological traits (West Eberhard 2003; Duckworth 2008). Only certain populations may harbour sufficient individual variation in behavioural plasticity to exploit a newly opened opportu nity (Simpson 1953; Schluter 2000; Bolnick et al. 2010;

Losos 2010; Yoder et al. 2010), perhaps explaining why only certain clades undergo adaptive radiations (Losos 2010; Yoder et al. 2010). Nevertheless, increases in trait variation after the loss of a dominant competitor may be more common than we realize and deserve further study as progressive loss of species will make ecological pro cesses increasingly dependent on how surviving species respond to alterations created by these extinctions.

Acknowledgements

We thank L. Caro, I. Diaz, H. Arias, M. Rodriguez, D. Garcia, D. Marti- nez, E. Hurme, M. Smith, J. Moxley, R. Zambrano and C. Ziegler for field assistance. M Hau, H. Horn, E. Leigh Jr., J. Tobias, MJ. West- Eberhard, C. Ziegler and four anonymous reviewers provided helpful dis- cussion and clarification of the manuscript. This research was conducted under research permits by Autoridad Nacional del Ambiente (SE/A-61- 09), Panama and The Smithsonian Tropical Research Institute’s Animal Care and Use Committee (07-06-24-08). Logistical and financial support was provided by Princeton University, The Smithsonian Tropical Research Institute, The Max Planck Institute for Ornithology at Radolfzell, Ger- many, and Konstanz University, Germany.

References

Bailey, S.F., Dettman, J.R., Rainey, P.B. & Kassen, R. (2013) Competi- tion both drives and impedes diversification in a model adaptive radia- tion.Proceedings of the Royal Society of London. Series B, Biological Sciences,280, 20131253.

Bates, D., Maechler, M. & Bolker, B. (2011) lme4: Linear mixed-effects models using S4 classes.

BiotasTM. (2005) Ecological Software Solutions LLC.

Bolnick, D.I., Ingram, T., Stutz, W.E., Snowberg, L.K., Lau, O.L. &

Paull, J.S. (2010) Ecological release from interspecific competition leads to decoupled changes in population and individual niche width.Pro- ceedings of the Royal Society of London. Series B, Biological Sciences, 277, 1789 1797.

Bolnick, D.I., Amarasekare, P., Araujo, M.S., B€urger, R., Levine, J.M., Novak, M. et al. (2011) Why intraspecific trait variation matters in community ecology.Trends in Ecology & Evolution,26, 183 192.

Brumfield, R.T., Tello, J.G., Cheviron, Z.A., Carling, M.D., Crochet, N. & Rosenberg, K.V. (2007) Phylogenetic conservatism and antiq- uity of a tropical specialization: army-ant-following in the typical antbirds (Thamnophilidae).Molecular Phylogenetics and Evolution,45, 1 13.

Calsbeek, R., Alonzo, S., Zamudio, K. & Sinervo, B. (2004) Sexual selec- tion and alternative mating behaviours generate demographic stochastic- ity in small populations.Behavioral Ecology,133, 1023 1030.

Chaves-Campos, J., Araya-Ajoy, Y.-M., Lizana-Moreno, C.A. & Rabe- nold, K.N. (2009) The effect of local dominance and reciprocal toler- ance on feeding aggregations of ocellated antbirds.Proceedings of the Royal Society of London. Series B, Biological Sciences,276, 3995 4001.

Duckworth, R.A. (2008) The role of behavior in evolution: a search for mechanism.Evolutionary Ecology,23, 513 531.

Duckworth, R.A. & Badyaev, A.V. (2007) Coupling of dispersal and aggression facilitates the rapid range expansion of a passerine bird.Pro- ceedings of the National Academy of Sciences of the United States of America,104, 15017 15022.

Dukas, R. (2013) Effects of learning on evolution: robustness, innovation and speciation.Animal Behaviour,85, 1023 1030.

Fort, K. & Otter, K. (2004) Territorial breakdown of black-capped chicka- dees,Poecile atricapillus, in disturbed habitats? Animal Behaviour,68, 407 415.

Franks, N. (1982) A new method for censusing animal populations the number ofEciton burchelliarmy ant colonies on Barro-Colorado Island, Panama.Oecologia,52, 266 268.

Grant, P. & Price, T. (1981) Population variation in continuously varying traits as an ecological genetics problem.American Zoology,21, 795 811.

Jankowski, J.E., Graham, C.H., Parra, J.L. & Robinson, S.K. (2012) The role of competition in structuring tropical bird communities.Ornitologıa Neotropical,23, 115 124.

Langkilde, T. (2009) Invasive fire ants alter behavior and morphology of native lizards.Ecology,90, 208 217.

Lapiedra, O., Sol, D., Carranza, S. & Beaulieu, J.M. (2013) Behavioural changes and the adaptive diversification of pigeons and doves.Proceed- ings of the Royal Society of London. Series B, Biological Sciences,280, 20122893.

Losos, J.B. (2010) Adaptive radiation, ecological opportunity, and evolu- tionary determinism.The American Naturalist,175, 623 639.

Martin, R.A. & Pfennig, D.W. (2010) Field and experimental evidence that competition and ecological opportunity promote resource polymor- phism.Biological Journal of the Linnean Society,100, 73 88.

Martin, C.H. & Wainwright, P.C. (2011) Trophic novelty is linked to exceptional rates of morphological diversification in two adaptive radia- tions of Cyprinodon pupfish.Evolution,65, 2197 2212.

Mayfield, H. (1961) Nesting success calculated from exposure.The Wilson Bulletin,73, 7.

Mazerolle, D.F. & Hobson, K.A. (2004) Territory size and overlap in male Ovenbirds: contrasting a fragmented and contiguous boreal forest.

Canadian Journal of Zoology,82, 1774 1781.

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, 960 968.

Morton, E., Derrickson, K. & Stutchbury, B. (2000) Territory switching behavior in a sedentary tropical passerine, the dusky antbird (Cercomac- ra tyrannina).Behavioral Ecology,11, 648 653.

van Oers, K., de Jong, G., van Noordwijk, A., Kempenaers, B. & Drent, P. (2005) Contribution of genetics to the study of animal personalities:

a review of case studies.Behaviour,142, 1185 1206.

Parent, C.E. & Crespi, B.J. (2009) Ecological opportunity in adaptive radi- ation of Galapagos endemic land snails.The American Naturalist,174, 898 905.

Pfennig, D.W. & McGee, M. (2010) Resource polyphenism increases spe- cies richness: a test of the hypothesis.Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences,365, 577 591.

Pfennig, D., Rice, A. & Martin, R. (2006) Ecological opportunity and phe- notypic plasticity interact to promote character displacement and species coexistence.Ecology,87, 769 779.

Pfennig, D.W., Wund, M.A., Snell-Rood, E.C., Cruickshank, T., Schlich- ting, C.D. & Moczek, A.P. (2010) Phenotypic plasticity’s impacts on diversification and speciation.Trends in Ecology & Evolution,25, 459 467.

Price, T., Qvarnstrom, A. & Irwin, D. (2003) The role of phenotypic plas- ticity in driving genetic evolution.Proceedings of the Royal Society of London. Series B, Biological Sciences,270, 1433 1440.

Price, S.L., Powell, S., Kronauer, D.J.C., Tran, L.A.P., Pierce, N.E. &

Wayne, R.K. (2014) Renewed diversification is associated with new eco- logical opportunity in the Neotropical turtle ants.Journal of Evolution- ary Biology,27, 242 258.

Rice, A.M., Leichty, A.R. & Pfennig, D.W. (2009) Parallel evolution and ecological selection: replicated character displacement in spadefoot toads.Proceedings of the Royal Society of London. Series B, Biological Sciences,276, 4189 4196.

Schluter, D. (2000)The Ecology of Adaptive Radiation. Oxford University Press, New York, USA.

Schneirla, T. (1949) Army-ant life and behavior under dry-season condi- tions. 3. The course of reproduction and colony behavior.Bull Amer Mus Nat Hist,94, 7 81.

1048

Schneirla, T.C. (1956) A preliminary survey of colony division and related processes in two species of terrestrial army ants.Insectes Sociaux,3, 49 69.

Setiadi, M.I., McGuire, J.A., Brown, R.M., Zubairi, M., Iskandar, D.T., Andayani, N.et al.(2011) Adaptive radiation and ecological opportu- nity in Sulawesi and Philippine fanged frog (Limnonectes) communities.

The American Naturalist,178, 221 240.

Sih, A., Cote, J., Evans, M., Fogarty, S. & Pruitt, J. (2012) Ecological implications of behavioural syndromes.Ecology Letters,15, 278 289.

Simpson, G.G. (1953)The Major Features of Evolution. Columbia Univer- sity Press, New York, NY.

Smith, T. & Skulason, S. (1996) Evolutionary significance of resource polymorphisms in fishes, amphibians, and birds.Annual Review of Ecol- ogy and Systematics,27, 111 133.

Sol, D., Bartomeus, I. & Griffin, A.S. (2011) The paradox of invasion in birds: competitive superiority or ecological opportunism? Oecologia, 169, 553 564.

Sol, D., Elie, M., Marcoux, M., Chrostovsky, E., Porcher, C. & Lefebvre, L. (2005) Ecological mechanisms of a resource polymorphism in Zena- ida Doves of Barbados.Ecology,86, 2397 2407.

Styrsky, J. (2003) Life-History Evolution and Population Dynamics of a Neotropical Forest Bird (Hylophylax naevioides). University of Illinois, Urbana, IL, USA.

Styrsky, J., Brawn, J. & Robinson, S. (2005) Juvenile mortality increases with clutch size in a neotropical bird.Ecology,86, 3238 3244.

Touchton, J.M. & Smith, J.N.M. (2011) Species loss, delayed numerical responses, and functional compensation in an antbird guild.Ecology, 92, 1126 1136.

Touchton, J.M. & Wikelski, M. (2014) Data from: Ecological opportu- nity leads to the emergence of an alternative behavioural phenotype in a tropical bird. Movebank Data Repository, Doi: 10.5441/001/1.

f7j8vt43.

West-Eberhard, M. (1998) Evolution in the light of developmental and cell biology, and vice versa.Proceedings of the National Academy of Sciences of the United States of America,95, 8417 8419.

West-Eberhard, M.J. (2003) Developmental Plasticity and Evolution.

Oxford University Press, New York, USA.

West-Eberhard, M.J. (2005) Developmental plasticity and the origin of species differences.Proceedings of the National Academy of Sciences of the United States of America,102, 6543 6549.

Willis, E. (1967) The behavior of bicolored antbirds.University of Califor- nia Publications in Zoology,79, 1 132.

Willis, E.O. (1972) The behavior of spotted antbirds.Ornithological Mono- graphs,10, 162.

Willis, E. (1973) The behavior of ocellated antbirds.Smithsonian Contribu- tions to Zoology,144, 1 57.

Willis, E. & Oniki, Y. (1978) Birds and army ants.Annual Review of Ecol- ogy and Systematics,9, 243 263.

Winkelmann, K., Genner, M.J., Takahashi, T. & R€uber, L. (2014) Compe- tition-driven speciation in cichlid fish.Nature Communications,5, 1 8.

Wund, M.A., Baker, J.A., Clancy, B., Golub, J.L. & Foster, S.A. (2008) A test of the “flexible stem” model of evolution: ancestral plasticity, genetic accommodation, and morphological divergence in the threespine stickleback radiation.The American Naturalist,172, 449 462.

Yoder, J.B., Clancey, E., Des Roches, S., Eastman, J.M., Gentry, L., God- soe, W.et al.(2010) Ecological opportunity and the origin of adaptive radiations.Journal of Evolutionary Biology,23, 1581 1596.

Zimmer, K. & Isler, M. (2003) Family Thamnophilidae (typical antbirds).

Handbook of The Birds of the World, Vol. 8 (eds J. del Hoyo, A. Elliott

& D. Christie), pp. 448 681. Lynx Edicions, Barcelona, Spain.

Zuk, M., Rotenberry, J.T. & Tinghitella, R.M. (2006) Silent night: adap- tive disappearance of a sexual signal in a parasitized population of field crickets.Biology Letters,2, 521 524.

Zuk, M., Bastiaans, E., Langkilde, T. & Swanger, E. (2014) The role of behaviour in the establishment of novel traits. Animal Behaviour,92, 333 344.