Social Behaviours in Solitary Bees: Interactions Among Individuals

Social Behaviours in Solitary Bees: Interactions Among Individuals inXeralictus bicuspidariae Snelling (Hymenoptera: Halictidae: Rophitinae)

MIRIAM H. RICHARDS* AND LAURENCE PACKER

(MHR) Department of Biological Sciences, Brock University, 500 Glenridge Ave., St. Catharines,ON, Canada L2S 3A1

(LP) Department of Biology, York University, 4700 Keele St., Toronto, ON, Canada M3J 1P3

Social Behaviours in Solitary Bees: Interactions Among Individuals inXeralictus bicuspidariae Snelling (Hymenoptera: Halictidae: Rophitinae)

MIRIAM H. RICHARDS* AND LAURENCE PACKER

(MHR) Department of Biological Sciences, Brock University, 500 Glenridge Ave., St. Catharines,ON, Canada L2S 3A1

(LP) Department of Biology, York University, 4700 Keele St., Toronto, ON, Canada M3J 1P3

__________________________________________________________________________________________________________________________________________________________

Abstract.—Understanding behavioural interactions among ancestrally solitary species is key tounderstanding the evolutionary origins of group living and cooperation. Previously, Packer (2006)showed that circle tube arenas can be used to evaluate the social status of species for which nestingdata are unavailable. We used circle tube arenas to study the behaviour among 30 female dyads ofthe solitary halictid bee, Xeralictus bicuspidariae Snelling, a member of the subfamily Rophitinae, allmembers of which are ancestrally solitary. Overall, 75.2% of frontal encounters resulted inavoidance, 20.7% in aggression, and 4.1% in a successful pass, values which are similar to thosepreviously observed in solitary halictids. Although passing events, which are interpreted ascooperative behaviour, were rare, they were significantly correlated with bees’ rates of approachand avoidance, and also with differences between dyad members in rates of ovarian development.Rates of aggression were not correlated with physical traits of females or with other behaviours. Wecompare the circle tube behaviour of X. bicuspidariae to previously studied solitary and socialhalictids, and provide statistical support for this method of assessing social status.

__________________________________________________________________________________________________________________________________________________________

The origin of eusociality is one of themajor events in the evolutionary history oflife (Szathmary and Maynard Smith 1995),yet our understanding of what transpiresduring transitions to sociality remainspoor. One reason is the great age at whichmost solitary to eusocial transitions tookplace – over 100 million years ago fortermites, ants, and vespid wasps (Wenzel1990; Martinez-Delclos and Martinell1995), perhaps somewhat less in the bees(Michener and Grimaldi 1988), and around20 million years ago in the three mainlineages of eusocial Halictinae (Brady et al.2006). The great age of these social lineagesmeans that to investigate the evolutionaryorigins of sociality, we must often usecomparative methods based on detailedknowledge of the behaviour of extantspecies. However, a second reason for ourincomplete understanding of the origins of

eusociality stems from our poor knowl-edge of solitary species, whose behaviouris most likely to represent the ancestralforms from which sociality evolved.

Sweat bees (Halictidae: Halictini andAugochlorini) are the most socially vari-able group of animals on earth, includingspecies that run the gamut from obligatelysolitary to obligately social, with socialityvarying from communal to semisocial andeusocial forms (Schwarz et al. 2007). Thereare even examples of intraspecific socialpolymorphism, in which solitary or socialbehaviour is expressed within or amongpopulations, often in response to variabil-ity in environmental conditions (Schwarzet al. 2007). The ecological processes thatshape the social behaviour of modernhalictines are often considered to be analo-gous to those that shaped the evolution ofmajor social transitions in the subfamily asa whole, including at least three origins ofeusociality and multiple reversions to*Author for correspondence: [email protected]

J. HYM. RES.Vol. 19(1), 2010, pp. 66–76

solitary behaviour (Danforth et al. 2003;Danforth et al. 2008).

Recent behavioural studies provide in-triguing evidence that behavioural transi-tions from solitary to social behaviour mayoccur easily and rapidly. Jeanson et al.(2005) observed that in forced associationsof solitary L. (Chilalictus) NDA-1 or ofcommunal L. hemichalceum, both domi-nance interactions and division of labourarose as natural outcomes of normalsolitary behavioural patterns expressed inthe context of novel, social environments,the effect being stronger in the solitary thanin the communal species. A similar phe-nomenon occurred in experiments onsolitary Ceratina carpenter bees (Apidae,Xylocopinae) when females were forced tonest in social associations (Sakagami andMaeta 1977). This suggests that at the verybeginning of evolutionary transitions tocaste-based sociality, ‘emergent’ social be-haviour by solitary bees could provide thebehavioural substrate upon which naturalselection acts, before the evolution of caste-based sociality. This fascinating possibilitymakes understanding the behaviour ofobligately solitary bees all the more critical.

Although behavioural interactions amongindividuals of social and socially poly-morphic sweat bees have been studied indetail on numerous occasions (Breed et al.1978; Buckle 1984; McConnell-Garner andKukuk 1997; Wcislo 1997; Soucy 2002),solitary species have received less attention.Consequently, we know little about thepotential for social interactions amongindividuals of solitarily nesting species,and how naturally occurring variation inindividual behaviour might impinge uponthe development of sociality remains ob-scure. Of course, one problem with study-ing the behaviour of solitary bees is a dearthof opportunities for observing known in-dividuals under natural conditions at suffi-cient frequency to permit detailed analysis.Fortunately, a recent comparative studysuggests that the circle tube arena, a circleof clear plastic tubing in which bees are

forced to interact (Breed et al. 1978), is oneroute to obtain sufficient behavioural dataon interactions among individuals of soli-tary species (Packer 2006).

In this paper we analyse the results of adetailed study of interactions among in-dividuals in a solitary species of the beefamily Halictidae. Xeralictus bicuspidariaeSnelling is a member of the halictidsubfamily Rophitinae with several advan-tages as a study organism. First, thephylogenetic position of rophitine beessuggests that their solitary behaviour isancestral, i.e. there is no evidence that therehas been any sociality in the evolutionaryhistory of the entire subfamily (Danforth etal. 2008; Patiny et al. 2008), so any potentialfor social interactions that might be in-duced experimentally, is part of its solitaryground plan. Second, this bee exhibitsconsiderable variation in colour of themetasoma of females (Snelling and Stage1995) such that pairs can easily be chosento permit individual recognition withoutthe intervention of artificially marking thebees (marking has been shown to influenceinteractions among individuals; Packer2005). Third, it is a large bee, facilitatingobservations of behaviours.

METHODS

Xeralictus bicuspidariae was studied atDome Rock Road, La Paz County, Arizona,USA, in April 2005. Female bees werecollected from flowers of Mentzelia (Loasa-ceae) and retained in microcentrifuge tubesfor no more than 30 minutes beforebehavioural observations commenced. Thisduration between capture and observationwas maintained to reduce the effect ofcaptivity-induced physiological changesupon behaviour (Pabalan et al. 2000). Twobees were then placed simultaneously in aclean, plastic circle tube of internal diam-eter 7 mm and length 20 cm. Simultaneousentry precludes ownership effects (Wcislo1997), and this tube diameter was sufficientfor the two individuals to pass one anotherand to turn around (Packer 2005), but

VOLUME 19, NUMBER 1, 2010 67

narrow enough that one bee could block anattempted pass by the other. Observationslasted for fifteen minutes, a time periodsufficient for differences in behaviourbetween individuals within a pair to bedetected, and took place outdoors in theshade. The metasomal colour of femalesvaried from entirely brick red to entirelydark brown (Snelling and Stage 1995);pairs were set up with one red and onedark female that could be easily differen-tiated by the observer without beingartificially marked. Several experimentswere terminated when discrimination be-tween the individuals was found to bemore difficult than expected.

An approach was taken to have occurredwhen individuals came within a distanceof one body length of each other (Kukuk1992; Packer 2005). Both frontal (head tohead) and front-to-back (head to tail)encounters were assessed and their out-comes classified into categories: ap-proaches, aggression, avoidance andpasses. Aggressive interactions includednudges, lunges and C-postures. Interac-tions that resulted in avoidance arose whenone individual moved away from a sta-tionary individual or they both movedaway from each other. A pass was scoredwhen the two bees manoeuvred to permitone to move past the other, or they bothmoved past one another simultaneously.All behavioural observations were carriedout by LP, and are therefore directlycomparable to the data presented in Packer(2006). For more detailed descriptions ofindividual behaviours, see Batra (1966).

All bees were measured and assessed forrelative age and reproductive condition asfollows. Head width was measured as thegreatest distance across the compoundeyes; this was the greatest diameter of thehead in dorsal view. Relative wear wasassessed from mandibles, scored from 0(unworn) to 6 (worn to the base of thesubapical mandibular tooth), and fromwings (the total number of nicks alongthe margin of the left forewing).

Reproductive status was estimated basedupon dissection of the metasoma. Thespermatheca was inspected for the presenceof sperm and ovarian development wasassessed by estimating the size of oocytes ineach ovariole relative to the size of a fullydeveloped oocyte, and summing the result-ing proportions across all six ovarioles. Asexpected for a solitary bee during nestprovisioning, all females had mated andso matedness was not considered further.

Statistical analysis.—In circle tubes, thebehaviour of each member of a pair isaffected by the behaviour of the secondmember of the pair. This creates a problemof statistical non-independence betweenmembers of each dyad. A second problemis variation in behavioural rates amongpairs – some pairs are very active and somedo almost nothing. A common approachhas been to standardize focal behaviors bythe encounter rate, which in effect meansall the behaviours are analyzed as ratios, soinformation related to absolute frequencyis lost and the statistical problems ofanalysing ratios are gained. To addressthese issues, we present an approachsomewhat different than in previous circletube studies. First, when behavioural pat-terns of individuals are considered, weanalyse only one individual per dyad (redbees or dark bees), which avoids inflatingthe number of degrees of freedom in eachmeasurement. Second, when properties ofdyads are considered, we analyse bothbehavioural frequencies and physical traitsin terms of differences between each bee ina pair. Correlations between trait differ-ences can be interpreted in the same wayas correlations between the traits them-selves. For instance, a negative correlationbetween head width difference and weardifference would indicate that larger beestended to be less worn. All differencesbetween pair members were calculated as(value for red bee) – (value for dark bee),except for head width (HW) difference,which was calculated as (red HW – darkHW) / average HW.

68 JOURNAL OF HYMENOPTERA RESEARCH: FESTSCHRIFT HONORING ROY SNELLING

All variables, including differences, werechecked for normality using the array oftests in SAS 9.1 (PROC UNIVARIATE);since several variables were non-normallydistributed, we mainly used non-para-metric statistical methods. Additionallywe used principal components analysis(PCA) to further explore and confirmrelationships among physical and behav-ioural variables in X. bicuspidariae. Initially,the PCA was based on eight variables(entered as untransformed differences be-tween females in mandibular wear, wingwear, head width, total ovarian score,approach frequency, aggressive frequency,avoidance frequency, and pass frequency).However, since Kaiser’s Measure of Sam-pling Adequacy (MSA) with all eightvariables had a value of only 0.467, thevariable with the lowest communalitymeasure (head width) was dropped fromthe PCA. With the remaining seven vari-ables, MSA50.63, which exceeds the 0.6criterion. We present both factor loadingscores (the degree to which each variableinfluences the inferred factors) and com-munality estimates (a reliability scorewhich estimates the proportion of variancein each variable that is jointly explained byall three factors).

Packer (2006) argued that the socialstatus of halictine bees can be accuratelyassessed using circle tube assays of fe-males, even in the absence of nesting data.Solitary bees should be characterized byhigh levels of avoidance behaviour, com-munal bees by high levels of cooperativebehaviour (passing) and low levels ofaggression, and semisocial and eusocial

bees by low levels of cooperation andhigh levels of both aggression and avoid-ance. We used discriminant functionsanalysis (DFA) to assess how accuratelyX. bicuspidariae and 21 other species (refer-ences in Packer 2006) can be categorized assolitary, communal, or semi and eusocial,based on the percentages of avoidance,aggression, and passing behaviours incircle tubes.

RESULTS

Circle tube assays.—Physical traits of the60 females used in 30 circle tube trials arepresented in Table 1. There were no sig-nificant correlations among body size,degree of wear, and degree of ovariandevelopment within individuals used inthe behavioural tests, although degree ofmandibular wear was positively correlatedwith degree of wing wear (Pearson correla-tion coefficient, r50.55, n548, p,0.0001).All females had at least one J-size,developing oocyte, and 25 of 52 (48%)dissected females contained a full-sizeoocyte, ready to lay. The mean differencebetween red and dark females in eachdyad for each of these characteristics,was zero (Table 1), so bee colouration hadno significance other than providing aconvenient identification tool for the ob-server.

The frequencies of each of the fourclasses of behaviour per dyad and perfemale are given in Table 2. The mostfrequent behaviours were approaches(32.5 per dyad) and avoidance (25.1 perdyad), followed by aggressive behaviours(6.7 per dyad) and passing (1.2 per dyad).

Table 1. Physical characteristics of adult female Xeralictus bicuspidariae used in circle tube experiments.Signed rank tests were used to compare the physical characteristics of red and dark females in each dyad; non-significant (n.s.) results indicate that overall, red and dark females were equivalent.

Variable Mean SD Range Signed rank test

Head width (mm) (n556) 7.4 0.24 7.0–8.0 S 5 222.5, n.s.Mandibular wear (n556) 2.1 1.2 0–5 S 5 31.5, n.s.Wing wear (n548) 8.0 5.1 0–20 S 5 4.0, n.s.Total ovarian score (n555) 1.7 0.4 0.5–2.6 S 5 8.0, n.s.


Overall, 75.2% of frontal encounters re-sulted in avoidance, 20.7% in aggression,and 4.1% in a successful pass.

Aggressive acts were observed in all 30pairs, and by 52 of the 60 (87%) individualsassayed. Withdrawals were also observedin all 30 pairs; only 1 bee of 60 (2%) did notdisplay a unilateral withdrawal, but shedid take part in a mutual (bilateral) with-drawal. Passing or cooperative acts wererare, being observed in only 22 of 30 (73%)pairs. Of a total of 35 passes, 20 (57%) werebilateral (both bees moved past each other)and 15 (43%) were unilateral (1 bee movedpast the other bee).

Based on behavioural frequency differ-ences (red bee – dark bee), three behav-iours, approach, avoid, and pass werefound to be mutually positively correlated(i.e. the bee that did one behaviour morefrequently also did the other behaviourmore frequently; Spearman rank correla-tions: approach vs. avoid, r50.798, n529,p,0.0001, approach vs. pass: r50.577,n529, p50.001; avoid vs. pass: r50.460,n530, p50.010), but none was correlatedwith the frequency of aggression (aggress

vs. approach: r520.224, n529, n.s.; aggressvs. pass: r520.106, n530, n.s.; aggress vs.avoid: r520.193, n530, n.s.).

Differences between bees with respect tohead width, wing wear, and mandibularwear were not significantly correlated withdifferences in behavioural frequency forany of the behaviours. Differences in totalovarian score did correlate positively withthe rates of approach and pass, althoughnot with either avoidance or aggressivefrequencies (Table 3, Fig. 1). In otherwords, the female with greater ovariandevelopment was almost significantlylikely to approach and was significantlymore likely to pass than the female withlesser ovarian development.

A principal components analysis (PCA)further describes behavioural and physicalvariation among female interactants incircle tubes. As outlined in the Methods,the PCA (Table 4) included all variablesexcept head width, which contributed littleto understanding variation among thedyads. Three factors had eigenvalues .

1.0 and were retained, explaining 77.9% ofthe variation among dyads. Factor 1 was

Table 2. Behavioural frequencies for approach, avoid, pass, and aggressive behaviours. Note that‘aggression’ includes C-postures, biting, and pushing. Since the behavioural rates of each member of a dyadare non-independent, only one bee per dyad is used to provide an estimate of behavioural frequencies perindividual. Measurements of mutual behaviour refer to simultaneous performance of that behaviour by bothmembers of a dyad. Sample size N530 dyads, except where otherwise noted.

Behaviour Rate Mean SD Range

Frontal encounters Dyad total 33.0 8.3 13–46Approach Dyad total 1 32.5 9.1 11–52

Red female 1 4.6 4.8 0–22Dark female 7.6 7.4 0–35Mutual 20.3 10.6 0–44

Avoid Dyad total 25.1 7.5 10–37Red female 9.3 4.3 3–19Dark female 9.7 4.3 0–20Mutual 6.1 4.0 0–13

Pass Dyad total 1.2 1.1 0–4Red female 0.3 0.5 0–2Dark female 0.2 0.6 0–3Mutual 0.7 0.9 0–3

Aggression Dyad total 6.7 5.0 1–19Red female 3.8 3.8 0–16Dark female 2.9 3.0 0–11

1 N529


most influenced by non-aggressive behav-iour and ovarian development, reflectingthe previously noted positive associationbetween ovarian development and ap-proach and passing frequencies. Factor 2was influenced mainly by mandibular andwing wear; thus Factor 2 describes varia-

bility in wear differences among dyads,and so does not reflect behavioural varia-tion. Factor 3 was influenced mainly byaggression. PCA based only on the fourbehavioural frequency differences, re-sulted in two factors that together ex-plained 79.7% of the variation among

Fig. 1. Influence of ovarian development (OD, horizontal axes) on different behaviours (vertical axes), scoredin terms of the differences between individuals (red-dark). Positive values on the horizontal and vertical axesindicate a greater value for the red bee, whereas negative values indicate a greater value for the dark bee. Topleft: OD vs. aggression. Top right: OD vs. avoidance. Bottom left: OD vs. approach frequency. Bottom right: ODvs. cooperation (pass).

Table 3. Influence of female physical status on behaviour. Spearman rank correlations were based ondifferences in both female traits and differences in behaviour frequencies (red bee – dark bee). Positivecorrelations indicate that the bee with the greater trait value exhibited the behaviour more frequently.

Physical trait of females

Behaviour (N5number of dyads)

Approach Avoid Cooperate Aggression

Head width 0.197 (27) 0.018 (28) 20.054 (28) 0.292 (28)Wing wear 0.020 (20) 20.123 (21) 20.092 (21) 0.077 (21)Mandibular wear 20.007 (27) 20.100 (28) 20.068 (28) 0.215 (28)Total ovarian score 0.451 (26), p50.062 0.250 (27) 0.435 (27), p50.023 20.132 (27)


dyads. Factor 1, which explained 55.0% ofthe behavioural variation among dyads,was strongly influenced by approaches(loading score 0.910), avoidance (0.866),and passes (0.752). Factor 2, which ex-plained 24.6% of the variation, wasstrongly influenced by aggression (loadingscore 0.959).

Comparison of X. bicuspidariae with otherhalictines.—Fig. 2 compares X. bicuspidariaeto 21 other species, in terms of the propor-tion of avoidance, aggressive, and coopera-tive (passing) behaviours observed in circletube assays. It most closely resemblesPenapis toroi, another solitary rophitine.Discriminant functions analysis (DFA)based on four putative categories (solitary,communal, semisocial, and eusocial) per-fectly assigned solitary and communalspecies, but failed to distinguish betweenthe latter two, assigning 1 eusocial speciesto the semisocial category and 1 semisocialspecies to the eusocial category. DFA basedon three putative categories (solitary, com-munal, and caste-based social) reassignedeach species into the category presented inFig. 2. Moreover, when Caenohalictus pygo-sinuatum was categorized as communal(Michener et al. 1979), then DFA assignedit to the solitary group (as suggested byPacker 2006). The success of the DFAapproach is based on significant differ-ences among solitary, communal, and

caste-based social bees in the proportionsof aggressive behaviour (ANOVA,F550.32, df52,19, p,0.0001) and avoid-ance behaviour (ANOVA, F515.15,df52,19, p,0.0001), as well as significantlymore frequent passing behaviour in com-munal species, as compared to both soli-tary and social species (ANOVA, F562.55,df52,19, p,0.0001).

DISCUSSION

Solitary behaviour of X. bicuspidariae.—InX. bicuspidariae, differences between circletube interactants in head width, wingwear, and mandibular wear were notassociated with differences in behaviour,suggesting that neither body size nor wear(and possibly age) structured interactionsamong adult females. Differences in ovar-ian development (OD) also did not predictdifferences in either aggression or avoid-ance, but were associated with rates ofapproach and pass behaviours, these beingexhibited more frequently by the bee withgreater ovarian development. Why wouldhigh OD females be more likely to ap-proach and especially, to pass? One possi-bility is that the closer a female is to layingan egg, the more active she is likely to be.Under natural circumstances, a femalehalictine getting ready to lay an egg shouldbe spending considerable time readying abrood cell and provisioning it. In a circle

Table 4. Principal components analysis describing variation among dyads based on differences betweeninteractants in both physical and behavioural traits. Three factors were retained with eigenvalues . 1,explaining a cumulative total of 77.9% of the variation among dyads. Relatively strong factor loading scores(.0.6) are indicated in boldface. Communality estimates describe the proportion of variance in each trait that isjointly explained by Factors 1, 2 and 3. Kaiser’s overall Measure of Sampling Adequacy (MSA) was 0.6312.

Trait (difference between females) Factor 1 Factor 2 Factor 3 Communality estimate

Mandibular wear 0.0124 0.9064 20.0567 0.4962Wing wear 20.1391 0.7738 20.5193 0.4465Ovarian score 0.6217 0.3534 0.2441 0.7520Approaches 0.9017 0.0752 20.1593 0.6582Avoidance 0.8311 20.0742 20.2158 0.7008Pass 0.8116 20.0438 0.2910 0.7450Aggression 20.1848 0.5085 0.7388 0.5381Eigenvalue 2.602 1.817 1.035Variance explained 37.2% 26.0% 14.8%


tube, heightened activity rates make itmore likely that a bee will approach thesecond bee, and then perhaps continueright past her. In other words, bees withhigher ovarian development may be moremotivated to remain active, which in acircle tube would result in higher rates ofapproaching and passing.

The lack of correlation between ovarianstatus and aggression might seem surpris-ing, but is consistent with observations inother species, including the solitary halic-tine, L. (Ctenonomia) NDA-1, and thecommunal species, L. (Chilalictus) platyceph-alum, in which ovarian status was notassociated with aggression (McConnell-Garner and Kukuk 1997). In contrast, inLasioglossum figueresi the female with largerovaries was often first to be aggressive, and

the bee with smaller ovaries was often firstto withdraw (Wcislo 1997). Even in obli-gately eusocial species like Halictus ligatus,dominance, aggression, and defensive be-haviours are most likely and most severewhen the two members of a dyad bothhave relatively high ovarian development,for example, when a queen is paired with aworker with highly developed ovaries(Pabalan et al., 2000). This suggests thatcorrelations between OD and aggressionmight have more to do with reactions tothe threat of egg replacement than withdominance behaviour per se. In solitarybees, we should not then expect to see acorrelation between ovarian developmentand aggressive behaviour (or withdrawalbehaviour) except perhaps in species withhigh rates of intraspecific egg parasitism.

Fig. 2. Comparison of circle tube behaviour of X. bicuspidariae with literature values (partially redrawn fromPacker 2006, which also contains a complete list of references). Solitary bee species are characterized by highrates of avoidance (withdrawals), communal bees by high rates of cooperation or tolerance (passing), andsemisocial and eusocial species by high rates of aggression, coupled with very low rates of passing. Differencesin behavioural profiles of bees tested in circle tubes versus vertical tubes (Jeanson et al. 2005) are evident. Generarepresented (top to bottom) are Xeralictus (X.), Penapis (Pe.), Thrincohalictus (T.), Caenohalictus (Ca.), Lasioglossum

(L.), Corynura (Co.), Halictus (H.), Ruizantheda (R.), and Pseudagapostemon (Ps.).


Our experiments uncovered consider-able variability among individuals. Ran-dom pairings of individuals with differentbehavioural tendencies (personalities)must then have contributed to variationamong dyads, creating behavioural scenar-ios that bear a marked resemblance tothose expressed by social bees. Indeed, thiswould seem to be the basis of the phenom-enon of ‘emergent’ sociality as describedby Jeanson et al. (2005). However, the useof the term ‘emergent’ to describe forcedsocial interactions among solitary bees issomewhat problematic, even in thosewhich like X. bicuspidariae are ancestrallyand monomorphically solitary. This isbecause solitary bees may also experiencesocial interactions that insect sociobiolo-gists do not usually categorize as ‘social’,such as interactions between foragers onflowers, between nest residents in densenesting aggregations, between nest resi-dents and would-be nest usurpers, orbetween residents and egg kleptoparasites.Moreover, group living may occur at verylow frequencies in some solitary specieswithout extensive nest observations, as hasrecently been found for several species ofthe apid genus, Ceratina (Rehan et al. 2009).In other words, many solitary bees, bothancestrally solitary and ancestrally social,may have considerable scope for intraspe-cific social behaviour, even if they rarely ornever nest in multifemale groups. Thevariability in behavioural syndromes ofsolitary halictines (Fig. 2) suggests thateventually it may be possible to detectdifferences among obligately solitary, so-cially polymorphic, and reversed solitaryspecies, especially based on the frequencyof avoidance and aggession.

Behavioural changes in social transitions.—One caveat to the use of artificial arenas forobserving bee behaviour is that the fre-quencies of circle tube behaviours may ormay not represent the frequencies of sameor similar behaviours in natural settings.Indeed, there are obvious differences inbehavioural frequencies assessed using

horizontal circle tubes versus vertical lin-ear tubes (Fig. 2), implying that majordifferences in behavioural frequencies areproduced by different experimental meth-odologies. Nevertheless, the interspecificconsistency of behavioural syndromes ob-served in circle tube assays of solitary,communal, and caste-based social speciesis striking and statistically supportable,suggesting that when circle tube assaysare used consistently, they uncover funda-mental differences in behaviour amongsolitary, communal, and semisocial andeusocial species. These differences, if notthe behavioural frequencies themselves,can be used to infer general behaviouraltendencies in bees of different social levels.

In halictids, the ancestral trait of intoler-ance is suggested by high rates of avoid-ance in solitary bees such as X. bicuspidariaeand another solitary rophitine species,Penapis toroi, in which avoidance behav-iours comprise about 75% of encounters(Fig. 2). Transitions to communal versuscaste-based social behaviour may be quitedifferent. Circle tube assays imply thatsolitary-communal transitions involve sig-nificant decreases in both aggression andavoidance, whereas transitions to caste-based eusociality involve a significantincrease in aggression, coupled with adecrease in avoidance. To the extent thatpasses represent cooperative interactions,solitary-communal transitions would ap-pear to involve huge increases in coopera-tion whereas transitions to caste-basedeusociality involve little change or perhapseven a decrease in cooperative behaviour.It will be important in future studies ofboth solitary and social halictines, to assessthe degree of behavioural concordancebetween natural versus artificial contextswhenever possible, so that we can actuallyunderstand how representative circle tubebehaviour is for those species for whichnesting data are unobtainable.

Given that one of the most outstandingfeatures of the eusocial insects is theirfrequent and sophisticated cooperative


behaviour, the hypothesis that transitionsto caste-based sociality should involvedecreases in cooperation coupled withincreases in aggression seems contradic-tory. However, semisocial and eusocialhalictines not only interact with manymore individuals than solitary bees do,but they must also cope with dominance-subordinance relationships, many of whichare regulated by aggressive behaviour(Kukuk and May 1991; Pabalan et al.2000). Semisocial and eusocial bees mustbe able to exercise both tolerance andaggression with the same individuals.Although aggressive behaviours by soli-tary and caste-based social bees in circletubes may appear to be similar, a majordifference in natural settings is that ag-gressive behaviour by the latter is likelymodulated by nest-mate recognition, suchthat encounters with non-nestmates willlikely provoke aggression, whereas en-counters with nestmates may engenderaggression, tolerance, or cooperation (Pesoand Richards 2010), depending on theimmediate behavioural context.

Transitions to social behaviour, especiallyto caste-based sociality, are rarer in halic-tines than reversals to solitary behav-iour (Danforth et al. 2003). Recent evidencesuggests that reversals to solitary behav-iour do not necessarily retrace the originalevolutionary steps that led to sociality. Forinstance, reversed-solitary Lasioglossumhave retained the social nesting character-istic of constructing brood cells close to themain burrow, facilitating both maternalinspection and care of the cells (Plateaux-Quenu 2008), and the potential for socialinteractions among newly emerged, adultbrood. Thus reversed solitary bees mayhave lost caste-based sociality, but mayhave retained the context-dependent abilityto discriminate nestmates from non-nest-mates. Circle tube comparisons of ances-trally solitary species like X. bicuspidariae,and reversed-solitary species like L. figuer-esi, may help to illuminate and distinguishthe evolutionary sequences involved in

forward and reverse social transitions,especially where these involve the expres-sion of context-dependent behaviour.

ACKNOWLEDGMENTS

This paper, documenting behavioural interactions

among females of a species that was described by Roy

Snelling, is dedicated to the memory of this prodi-

gious taxonomist. We thank Rob Paxton for helpful

and perceptive comments on the manuscript and

Chris Starr for several felicitous suggestions for

improving the text. This research was funded by

Natural Sciences and Engineering Research Council of

Canada Discovery Grants to both authors.

LITERATURE CITED

Batra, S. W. 1966. The life cycle and social behavior of

the primitively social bee Lasioglossum zephyrum

(Hymenoptera: Halictidae). University of Kansas

Science Bulletin 46: 359–423.

Brady, S. G., S. Sipes, A. Pearson, and B. N. Danforth.

2006. Recent and simultaneous origins of eusoci-

ality in halictid bees. Proceedings of the Royal

Society B-Biological Sciences 273: 1643–1649.

Breed, M. D., J. M. Silverman, and W. J. Bell. 1978.

Agonistic behavior, social interactions, and be-

havioral specialization in a primitively eusocial

bee. Insectes Sociaux 25: 351–364.

Buckle, G. R. 1984. A second look at queen-forager

interactions in the primitively eusocial halictid,

Lasioglossum zephyrum. Journal of the Kansas

Entomological Society 57: 1–6.

Danforth, B. N., L. Conway, and S. Ji. 2003. Phylogeny

of eusocial Lasioglossum reveals multiple losses of

eusociality within a primitively eusocial clade of

bees (Hymenoptera: Halictidae). Systematic Biol-

ogy 52: 23–36.

———, C. Eardley, L. Packer, K. Walker, A. Pauly,

and F. J. Randrianambinintsoa. 2008. Phylogeny

of Halictidae with an emphasis on endemic

African Halictinae. Apidologie 39: 86–101.

Jeanson, R., P. F. Kukuk, and J. H. Fewell JH. 2005.

Emergence of division of labour in halictine bees:

contributions of social interactions and behav-

ioural variance. Animal Behaviour 70: 1183–1193.

Kukuk, P. F. and B. May. 1991. Colony dynamics in a

primitively eusocial halictine bee Lasioglossum

(Dialictus) zephyrum (Hymenoptera: Halictidae).

Insectes Sociaux 38: 171–189.

———. 1992. Social interactions and familiarity in a

communal halictine bee Lasioglossum (Chilalictus)

hemichalceum. Ethology 91: 291–300.

Martinez-Delclos, X. and J. Martinell. 1995. The oldest

known record of social insects. Journal of Paleon-

tology 69: 595–599.


McConnell-Garner, J. and P. F. Kukuk. 1997. Behav-ioral interactions of two solitary, halictine beeswith comparisons among solitary, communal andeusocial species. Ethology 103: 19–32.

Michener, C. D., M. D. Breed, and W. J. Bell. 1979.Seasonal cycles, nests, and social behavior ofsome Colombian halictine bees (Hymenoptera,Apoidea). Revista de Biologıa Tropical 27: 13–34.

——— and D. Grimaldi. 1988. The oldest fossil bee:Apoid history, evolutionary stasis, and antiquityof social behavior. Proceedings of the National

Academy of Sciences of the USA 85: 6424–6426.Pabalan, N., K. G. Davey, and L. Packer. 2000.

Escalation of aggressive interactions duringstaged encounters in Halictus ligatus Say (Hyme-noptera: Halictidae), with a comparison of circletube behaviors with other halictine species.Journal of Insect Behavior 13: 627–650.

Packer, L. 2005. The influence of marking upon beebehaviour in circle tube experiments with amethodological comparison among studies. In-

sectes Sociaux 52: 139–146.———. 2006. Use of artificial arenas to predict the

social organisation of halictine bees: Data forfourteen species from Chile. Insectes Sociaux 53:307–315.

Patiny, S., D. Michez, and B. N. Danforth. 2008.Phylogenetic relationships and host-plant evolu-tion within the basal clade of Halictidae (Hyme-noptera, Apoidea). Cladistics 24: 255–269.

Peso, M. and M. H. Richards. 2010. Knowing who’swho: Nestmate recognition in the facultativelysocial carpenter bee, Xylocopa virginica. Animal

Behaviour in press.

Plateaux-Quenu, C. 2008. Subsociality in halictinebees. Insectes Sociaux 55: 335–346.

Rehan, S., M. H. Richards, and M. P. Schwarz. 2009.Evidence of social nesting in the Ceratina ofBorneo. Journal of the Kansas Entomological Society

82: 194–209.Sakagami, S. F. and Y. Maeta. 1977. Some presumably

presocial habits of Japanese Ceratina bees, withnotes on various social types in Hymenoptera.Insectes Sociaux 24: 319–343.

Schwarz, M. P., M. H. Richards, and B. N. Danforth.2007. Changing paradigms in insect social evolu-tion: Insights from halictine and allodapine bees.Annual Review of Entomology 52: 127–150.

Snelling, R. R. and G. I. Stage. 1995. Systematics andbiology of the bee genus Xeralictus (Hymenop-tera: Halictidae, Rophitinae). Contributions in

Science, Natural History Museum of Los Angeles

County (451): 1–17.Soucy, S. L. 2002. Nesting biology and socially

polymorphic behavior of the sweat bee Halic-

tus rubicundus (Hymenoptera: Halictidae). Annals

of the Entomological Society of America 95: 57–65.

Szathmary, E. and J. Maynard Smith. 1995. Themajor evolutionary transitions. Nature 374: 227–232.

Wcislo, W. T. 1997. Social interactions and behavioralcontext in a largely solitary bee, Lasioglossum

(Dialictus) figueresi (Hymenoptera, Halictidae).Insectes Sociaux 44: 199–208.

Wenzel, J. W. 1990. A social wasp’s nest from theCretaceous period, Utah, USA, and its biogeo-graphical significance. Psyche 97: 21–29.


Social Behaviours in Solitary Bees: Interactions Among Individuals

Documents

Transcript of Social Behaviours in Solitary Bees: Interactions Among Individuals