Behav Ecl Scibil (1985) 17: 363-369 Behaviral Eclgy and Scibilgy it) Springer-Verlag 19R5 Parental energy expenditure: a prximate cause f helper recruitment in the pied kingfisher (Ceryle rudis) Heinz-Ulrich Reyer h and Klaas Westerterp 2 1 Max-Planck-Institut fur VerhaItensphysilgie, D-8131 Seewiesen, Federal Republic f Germany 2 Bimedisch Centrum, Pstbus 616, NL-62 MD Maastricht, The Netherlands Received Octber 15, 1984 Accepted February 25, 1985 Summary. Energy expenditure f adult Pied Kingfishers was measured with dubly-labeled water. Results were related t reprductive success f parents aided and unaided by helpers. Energetically stressed parents in a clny with pr fd supply accepted ptential helpers mre ften than unstressed birds in anther clny where fd was easily available. This treatment f helpers was reversed in bth clnies thrugh experimental manipulatin f clutch size and hence energetic stress. Results are discussed in relatin t the csts and benefits that helpers incur n the parents' fitness. Intrductin In recent years several authrs have attempted t explain cperative breeding in birds as an adaptatin t eclgical cnditins (e.g. Brwn 1974; Orians et al. 1977; Gastn 1978; Kenig and Pitelka 1981; Emlen 1982, 1984). The critical test fr such a hypthesis is t cmpare inclusive fitness values f cnspecifics pursuing different behaviral strategies in identical envirnments and f thse pursuing identical strategies in different envirnments. Such a cmparisn is pssible in the Pied Kingfisher (Ceryle rudis). It was shwn previusly that males chse the strategy which yields the highest inclusive fitness under the prevailing eclgical and demgraphic cnditins (Reyer 198, 1984). Fr male breeders this chice means t reject ptential helpers (which are always males and therefre als ptential rivals) when fd cnditins are gd but t accept them when cnditins are t pr fr parents t raise all their hatchlings unaided. If birds are t make this ultimately cr * T whm ffprint requests shuld be sent rect decisin they require prximate mechanisms which are (1) cnstantly available and (2) gd predictrs f fitness, i.e., they must be clsely related t current reprductive success and the prbability f survival. Energy balance is a likely candidate. Althugh a negative relatinship between energetic stress in ne year and survival int the next has nly been shwn fr a few species (Nur 1984), including the Pied Kingfisher (Reyer 1984), current reprductive success in many species is clsely linked t energy balance and bdy cnditin (Drent and Daan 198). These in turn depend n envirnmental cnditins, particularly fd. The cnditins in a given seasn determine whether a bird will breed at all, hw early and hw ften eggs will be laid, hw big the clutch will be, hw much fd the parents can bring and cnsequently what the grwth and survival rate f the yung will be. It is cnceivable that the parents' decisin t accept helpers r nt als depends n energy balance and bdy cnditin. Perhaps a bird culd cmpare the fd requirements f the yung (as cmmunicated in their begging) t its wn feeding capacity and the amunt f cmpetitin frm helpers. This hypthesis cmbines cnclusins frm the abve mentined energetic studies with an idea, riginally suggested by Orians et al. (1977) and recently mdelled by Brwn (1982, 1984) and Emlen (1982), namely that parents shuld accept r reject helpers accrding t need. We tested this hypthesis by measuring first the daily energy expenditure (DEE) f feeding adult Pied Kingfishers in tw eclgically different clnies, using dubly-labeled water. We then related the DEE f parents t their reprductive success, their behavir twards ptential helpers and the begging duratin f their yung under nrmal cnditins. Finally we manipulated clutch size, changing the begging
364 f yung and the energetic stress f parents. We nly cnsider the treatment f "secndary helpers", unrelated t the breeders, and nt that f related "primary helpers". The latter apparently d nt cmpete with the breeders fr females and were tlerated under all bserved circumstances (Reyer 198, 1984 and in preparatin). Methds The study was carried ut between 1981 and 1983 n tw breeding ppulatins at Lake Victria and Lake Naivasha, Kenya (Reyer 198). Adult birds wh fed nestlings were caught in the late afternn, weighed and then injected with.5-.8 ml f 18% (excess atm percent) xygen-18 enriched and 1% deuterium enriched water ehso). After 1 h f equilibratin ca. 6 J.1l bld was extracted frm the wing vein befre releasing the bird. Birds were recaptured and weighed, and bld samples taken again 24 h (r a multiple theref) later. CO 2 prductin was calculated frm the reductin f zh and 1BO between the first and secnd sample as described previusly (Westerterp and Bryant 1984) and then cnverted int DEE (kj) with an RQ value f.8 fr fish prtein n which the birds feed almst exclusively. Additinally, the number, type, and size f fish prvided fr nestlings by the injected bird was mnitred during the whle activity perid. The daily fd cntributin f the zh z lbo-injected bird was calculated frm these data as describe~ earlier (Reyer 1984) by using the lengthweight relatinships in Table 1. In the same way the cntributins f all ther birds feeding at the particular nest were recrded. Frm the ttal amunt f fd and the knwn number f nestlings the average energy intakenestling and day was calculated. The effect f that intake n grwth was measured by weighing nestlings befre and after the experiment when in the pst-absrptive state. Weights were usually taken at 53 h, befre the first feed f the day. The nestlings', hunger was mnitred by placing a micrphne at the entrance f the nest chamber and cnnecting it t a tape recrder in the bservatin hide. Ca. los recrdings were made every 5 min thrughut the day t determine whether the chicks were begging. Begging duratin was expressed as the percentage f recrdings with begging, excluding cases in which the nestlings were being fed r had been fed in the previus minute. All the abve measurements n adults and yung were taken when the nestlings were 5 t 17 days ld. In Pied Kingfishers this is the perid f linear grwth (Reyer, unpublished data). Details f the manipulatin experiments will be mentined with the results. Results Energy expenditure andfeeding capacity As the daily energy expenditure f adults increases, the amunt f fd delivered t nestlings rises in a linear fashin at bth lakes but with significantly different slpes (Fig. 1; P=.37, analysis f cvariance; Sakal and Rhlf 1969). Thus, a Lake Victria bird will achieve a lwer feeding cntributin than ne frm Lake N aivasha fr the same amunt f energy expended. The eclgical reasns fr this difference include (Reyer 198): (1) :g, c 5 5 "R1 3 c E --------------------: l:,.. ~ : ' ' 1 - - - - - - - - -- -; - - - - - - - ~.; ' y.~ :.,, ~,) I i I I I 13 15 19 22 25 28 energy expenditure [kjd ) Fig. 1. Amunt f fd delivered t nestlings (kjad*d) in relatin t daily energy expenditure f feeding adults (kjd). - - -=Lake Naivasha, y=4.693 x-718.3; 11=17, r=.756, P<O.OO1. (e---e=lake Victria, y=i.485 x 29.9; n = 15, r=.545, P=.36.... = Upper limit f energy expenditure (frm Fig. 2) and resulting feeding capacities respectively a lwer energy yield per fish at Lake Victria which must be cmpensated fr by catching mre fish (cp. allmetric equatins in Table 1), (2) greater turbidity f the water (leading t lwer hunting success) and lnger distances between fishing grunds and the breeding clny at Lake Victria resulting in mre flying and hvering time per catch. The ranges and averages f bdy mass befre injectin were nt significantly different between birds frm the tw lakes (bth medians 76 g; Lake Victria: interquartile range 72.-8., n = 15; Lake Naivasha: interquartile range 7.5-78.5, n= 17; Z=.566, N.S., Mann-Whitney V-test, twtailed). Birds with a DEE f up t 21 kj maintain r even increase their bdy mass (Fig. 2). Abve that they increasingly lse mass (P <.1, Mann Whitney V-test, ne-tailed), as much as 9.2%day. Thus 21 kj seems t represent sme physilgically determined threshld which cannt be exceeded fr prlnged perids withut a reductin in bdy mass. Pssible reasns fr this reductin will be addressed in the Discussin. Entering the threshld f 21 kj int Fig. 1, we can predict that n average a pied kingfisher at Lake Naivasha can deliver as much as 267.2 kjday withut lsing
I r j I Table 1. Relatinship between standard length sf (mm), dry mass M d (g) and energy cntent (kjg M d ) fr cichlid fishes and Engraulicypris argenteus (Cyprinidae) frm Lake Victria and fr Tifapia ssp. (Cichlidae) and Micrpterus safmnides (Centrarchidae) frm Lake Naivasha. r=pearsn crrelatin cefficient between sf and M d (sample size=n 1 ). x and SD=mean and standard deviatin f energy cntentg M d as analyzed by bmb calrimetry (sample size=n 2 ) 365 Prey type Regressin n 1 x SD n 2 Engraulicypris Cichlidae Tipapia ssp. + Micrpterus lg M d =.37* sl- 2.32 lg M d =.26* sl-1.461 lg M d =,23* sl-1.258.98.99.97 52 95 44 19.95 18.89 19.52.61.23.96 3 3 8.1 D ' CJ) <1J CJ) c.c u <1l <1l -5 E >. D.Q.5 (J' 16 2 Cl. '< :::'3 ~ g us '" -1 n 1459841 I I [ I I kjd 13 17 21 25 Fig. 2. Medians and interquartile ranges f bdy mass change (gd) in feeding adults in relatin t their energy expenditure (kjd). Data frm bth lakes cmbined bdy mass (95% cnfidence limits: 193.1-341.3). A Lake Victria bird can nly deliver an average maximum f 11.9 kjday (7.-133.7) which is 61.8% less. Sme values lie abve these threshlds in bth clnies. Hwever, days with high fd deliveries are regularly fllwed by days with exceptinally lw nes. Althugh a bird may exceed the predicted threshlds fr shrt perids it apparently des nt maintain a higher perfrmance, with cncmitant reductin in bdy mass, fr several days r even weeks. We therefre cnsider the abve values t be the" average feeding capacities". Grwth f nestlings The effect f the different feeding capacities f the tw lakes n grwth and survival f nestlings is shwn in Fig. 3 (data frm bth clnies cmbined). The daily bdy mass changeyung is pltted against the average amunt ffd the nestling receives. One parent at Lake Naivasha has a feeding capacity f 267.2 kjparent and the average clutch size at hatching is 4.8 (Reyer 198 and un- received fd [kjnestling x dy I Fig. 3. Bdy mass change (Ym) f nestlings (e--e) and begging duratin (Yb) (+---+) in relatin t the amunt f fd received. Ym=.124 x-l1.6; n=22, r=.99, P<O.OO1. Yb= -.447 x+79.6; n=15, r= -.83. P<O.OO1. Arrws shw the average amunt f fdday and the average daily change in bdy mass fr a nestling at Lake Naivasha (white arrws) and Lake Victria (black arrws) published data). With tw parents, each nestling will therefre receive an average f 111.3 kjday (= 2*267.24.8). Accrding t Fig. 3, this results in a bdy mass gain f 2.7 gday (95% cnfidence limits: 1.1-4.3). Hand-reared Pied Kingfishers f the same age, fed ad lib., shwed a mean grwth rate f 4.3 gday (SD = 1.; based n five birds individually averaged ver 7 days within the perid flinear grwth). Thus, even withut helpers, Lake Naivasha parents can guarantee grwth f all their yung at a rate nt t far belw that f a nestling under ptimal cnditins. Hwever, nestlings raised at Lake Victria by parents alne will receive nly 44.3 kj as each parent has a feeding capacity f 11.9 kj and the average clutch size at hatching is 4.6. This is unsufficient and will lead t an average bdy mass lss f 5.6 g (95% cnfidence limits: -6.8 t -4.4). High cmpetitin amng nestlings is t be expected and sme will survive at the expense f thers.
366 These predictins, based n the energetically limited feeding capacities f adults, are brne ut by the results. Unassisted pairs at Lake Victria lse 61 % f their yung, mainly due t starvatin, as ppsed t 19% at Lake Naivasha althugh bth start with a similar number f eggs (prbably limited by incubatin cnstraints). While ne and tw helpers at Lake Victria can reduce lsses t 22% and % respectively, pairs with helpers as Lake Naivasha d nt fledge significantly mre yung than thse withut (Reyer 198, 1984, unpublished data). 1..5 LAKE NAIVASHA 6 ~ yip 2-3 4-5 n 9 9 1. LAKE VICTORIA Begging fnestlings The begging f yung prbably tells adults whether they are receiving sufficient fd and thus whether helpers are needed. It has a clear influence n the adults' feeding patterns. Parents resting in the clny were regularly bserved appraching the nest entrance and either resuming rest when the begging respnse was sft r immediately flying t the lake when it was intense. Daily duratin f begging decreases significantly as fd supply increases (Fig. 3). Cnsequently, the average begging duratin at Lake Victria is lnger than at Lake Naivasha. It appears pssible that the differences in the demands f the nestlings and the energetic stress n parents (Fig. 1) between the tw clnies were the prximate mechanisms respnsible fr the different treatment f helpers at Lake Victria and Lake Naivasha. Manipulatin experiments This hypthesis was further tested by reversing the energetic stress and begging duratin in the tw clnies thrugh manipulatin f clutch size and by cmparing the treatment f helpers under nrmal and manipulated cnditins. A ptential helper was labeled "rejected" when his attempts t feed adults and nestlings f a pair were aggressively prevented by the male breeder f that particular pair. Helpers which were greeted and allwed t feed were labeled" accepted". A mre detailed descriptin f the respective behavir patterns is given by Duthwaite et al. (in press). Observatins were cnfined t male breeders because it is they wh cmpete with male helpers ver the scarce females. Females prved t tlerate helpers much mre readily (Reyer, unpublished data). In unmanipulated cnditins, the male breeder's" decisin" whether t accept a ptential helper r nt is usually made within the first 7 days 5 yip 2-3 ::1 n 7 7 Fig. 4. Prprtins f encunters in which mated males (11) attack (hatched bars) r greet (white bars) ptential secndary helpers. Tp: Lake Naivasha; bttm: Lake Victria; left: nrmal clutch size (2-3 yungparent); tp right: increased clutch size (4-5 yungparent); bttm right: reduced clutch size (;;; 1 1 yungparent). The tw bars f each graph add up t 1.. Means and standard deviatins are given after hatching (Reyer, unpublished data). Birds which were nt tlerated by day 1 were nt accepted later either. The fllwing data cme frm breeders with nestlings lder than 1 days. All ptential helpers had been rejected befre hatching (Reyer 1984, unpublished data). At Lake Naivasha 88% f the breeding males (n=25) cntinued t reject them under nrmal cnditins (2 3 yungparent) as ppsed t % (n = 31) at Lake Victria (;(2=44.9, P<.1). Then sme clutches at Lake Naivasha, with 13-22-day-ld nestlings, were experimentally increased t 8-1 yung (i.e. 4-5 nestlingsparent), putting the parents int a psitin similar t that f Lake Victria birds: they culd n lnger prvide enugh fd (59.4 kj yung; Figs. 1 and 3). Nw nly 2% f the pairs (n = 1) rejected ptential helpers. This differs significantly frm nrmal cnditins (P<.1, Fisher test, ne-tailed). The reverse experiment was equally cnclusive. When clutch size at Lake Victria was reduced t 1-2 (i.e..5-1 nestlingparent and 135.9 kjjyung) ptential helpers were rejected in 87.5% f the cases (n=8), als different frm nrmal cnditins (P<.1). These differences d nt result just frm the fact that birds which spend mre time in getting fd have less time t chase away ptential helpers. Such an ex
367 planatin can be ruled ut by the results frm Fig. 4 which shw the prprtin f encunters in which the mated male either attacks r greets the ptential helper. Attacks prevailed at Lake Naivasha and greeting at Lake Victria when clutch size was nrmal. Hwever, when it was increased, Lake Naivasha birds switched frm attacking helpers t greeting them (P=O.Ol, Wilcxn-test, ne-tailed), whereas in Lake Victria birds with experimentally reduced clutch size attacking prevailed ver greeting (P<O.Ol, Mann Whitney U-test, ne-tailed). The applicatin f the Wilcxn-test t Lake Naivasha data and f the U-test t Lake Victria results, reflects the slightly different experimental design in the tw clnies. At Lake Naivasha sme birds were bserved befre and after increasing their clutch sizes. Thus, the data frm manipulated clutches represent a subset f the data frm the nrmal clutches. At Lake Victria nrmal and manipulated clutches represent independent samples. Here, clutch size was already reduced 1-2 days after hatching, befre any helper had been accepted. This differential design resulted frm bservatins f undisturbed grups, shwing that nce a helper has been accepted, he will remain accepted, independent f hw the brd size develps. Experimental reductin f clutch size in tw pairs with accepted helpers cnfirmed these bservatins. In summary: (1) male breeders reject ptential helpers when the parents can raise all their yung alne, (2) they switch frm rejecting t accepting when the nestlings fd requirements exceed the parents' feeding capacities, but (3) nce a breeder has accepted a helper, he des nt later reject him if the required feeding effrt decreases. The influence f the begging f the yung n the treatment f helpers was tested separately. Ludspeakers were hidden clse t the nest entrances f tw pairs at Lake Victria bth f which had tw yung (12-15 days ld) and had rejected ptential helpers s far. Begging recrdings were played thrughut the day fr 4 min fllwed by 1 min f silence (8% begging duratin). At first, bth pairs' feeding frequencies were higher than thse f the day preceeding the experiment. Hwever, they sn returned t nrmal and after 2 days had still nt accepted helpers. After entering the nest the parents prbably received cmpensating behaviral stimuli frm their yung which were well fed during this experiment. The nestlings may even nt have been gaping n ccasin, as parents seemed t leave the nest carrying fish mre ften than usual. Discussin Energy is ften cnsidered ne f the mst imprtant factrs in shaping cperative breeding (Brwn 1982, 1984), territriality (Gill and Wlf 1975; Ewald and Carpenter 1978; Davies and Hustn 1981), grup fraging (Jarman 1974; Krebs 1974; Krebs and Cwie 1976), and ther scial behavirs (Pulliam et al. 1974; Carac 1979; Carac et al. 198). Hwever, quantitative energy budgets are usually nly measured in relatin t an individual's size, age, sex, thermal envirnment r activities (see Walsberg 1983 fr a review) and rarely in a scial cntext. This discrepancy between prpsed imprtance and actual analysis f energy arises mainly frm the difficulties f recrding natural scial behavir under labratry cnditins and measuring energy budgets precisely under field cnditins. The develpment f the dubly-labeled water methd (Lifsn and McLintck 1966) has vercme the latter difficulty even within cmplex scial situatins like cperative breeding. The critical DEE beynd which Pied Kingfishers lse bdy mass is 21 kj (Fig. 2). This average value ignres individual variatins in bdy mass and s d the calculated regressins between parental feeding capacity and energy expenditure (Fig. 1). Such variatins culd be imprtant (a) because light individuals need less energy than heavy nes (cp. allmetric equatins in Aschff and Phi 197) and (b) because weight lss may be an adaptive strategy (e.g. Nrberg 1981). Argument (a) culd be held respnsible fr the bserved difference in energy expenditure between the tw lakes, if Lake Naivasha birds were lighter than Lake Victria birds. This, hwever, did nt apply (see Results). Argument (b) seems t find supprt frm the fact that the critical DEE f 21 kj is 22% lwer than expected fr a 76 g bird n metablic grunds (Kirkwd 1983). We are nt able t tell whether a Pied Kingfisher parent is unable t expend mre than 21 kj withut lsing bdy mass r whether he "chses" t reduce his mass t save energy. This, hwever, des nt alter the ntin that bdy mass change is related t energy expenditure and that there is a critical DEE which demands mass lss. The threshld f 21 kjday is almst identical t the value predicted fr a bird its size frm the 4*BMR (basic metablic rate) regressin line which Drent et al. (19781979) calculated fr 31 bird species. The majrity f these were fish-eaters like the pied kingfisher. In mst bird species investigated s far, 4*BMR seems t represent the upper limit f energy expenditure which cannt be ex
368 ceeded fr prlnged perids withut a decline in bdy cnditin (Drent and Daan 198; Westerterp and Drent, in press). Such a decline can have bth shrt-term effects n present and lng-term effects n future reprductin. There is a significant negative crrelatin fr the Pied Kingfisher and a few ther species between amunt f feeding in ne year and survival int the next (Nur 1984; Reyer 1984). Thus, energy expenditure is indeed a currency fr fitness csts, a reliable prximate indicatr f ultimate reprductive success. Selectin will cnsequently favr threshlds f energy expenditure which maximize current reprductive success with minimal csts fr the residual reprductive value. Where fr eclgical reasns there is a danger f crssing these threshlds, behaviral traits which reduce energy expenditure will be develped. Recruitment f helpers under pr fd cnditins is ne such trait. Hwever, the benefits f reduced energetic stress and imprved reprductive success thrugh helpers must be weighed against the csts which Pied Kingfisher helpers can impse n breeders. These are mainly frm cmpetitin fr their mates (Reyer 1984, in preparatin; Reyer et al., in preparatin). Male surplus (and cnsequently cmpetitin fr females) is similar in bth clnies but benefits frm helpers differ (Reyer 198). Thus the high reprductive success and lw DEE f parents at Lake Naivasha shuld shift the utcme f the breeder-helper cnflict twards rejectin f helpers, whereas the pr reprductive success and high DEE f parents at Lake Victria shuld favr acceptance. This indeed happens under nrmal cnditins. When cnditins in the tw clnies were reversed by experimental manipulatin f clutch size, the parents' behavir twards ptential helpers was als reversed. Clutch size culd nt be manipulated withut a crrespnding change in time budget. Thus time may appear as likely a limiting factr as energy. The limits culd arise frm the need t save time fr nest-guarding, plumage-maintenance and ther imprtant activities (Walsberg 1983). Hwever, even the mst active feeders withut helpers spent plenty f time "lafing" away frm the clny (Reyer and Westerterp, in preparatin). Mrever, the breeding seasn f Pied Kingfishers in ther areas cincided with lw wind velcities and lw water turbidities (Whitfield and Blaber 1978). These cnditins favur hunting frm a perch rather than by hvering flights (Duthwaite 1976), which saves energy but increases the timecatch rati (Reyer, unpublished data). We therefre cnsider differences in energy expenditures between the tw clnies t be mre imprtant fr the differential treatment f helpers than differences in time expenditures. This des nt imply any statement abut the precise physilgical mechanism. We d nt pretend that a parent actually mnitrs his energy state and uses variables such as fat reserves r CO 2 -prductin as the immediate cues t vary his treatment f helpers. Our results nly shw that the decisin between acceptance and rejectin is smehw influenced by the energy budget. Energy expenditure als differs cnsiderably between ppulatins and individuals f ther species (Bryant and Westerterp 1983; Karasv and Andersn 1984; Westerterp and Drent, in press). Additinally, intraspecific variatin can be seen in the behavir f cperative breeders and ther scial animals (Emlen 1984; Ltt 1984). Therefre the causal relatinship between energy expenditure and scial behavir reprted in this paper may hld fr many mre species. This is nt a new idea. Several authrs have already emphasized that cperative breeding is shaped by eclgical cnstraints such as habitat saturatin, fd supply, lack f sexual partners r envirnmental harshness leading t energetically intensive parental investment which can be reduced helpers (Brwn 1974, 1982, 1984; Orians et al. 1977; Brwn et al. 1978; Gastn 1978; Stallcup and Wlfenden1978; Vehrencamp 1979; Kenig and Pitelka 1981; Emlen 1982, 1984; Riedmann 1982). Hwever, they did nt measure energy expenditure and usually cnsidered nly ultimate cnsequences. This is the first attempt t ur knwledge that als deals with the prximate mechanisms by incrprating quantitative measurements f DEE under nrmal and manipulated field cnditins. We feel that this apprach will prve useful in many areas f scial behavir. Acknwledgements. This study wuld nt have been pssible, withut the cntinuus supprt f w. Wickler and funding frm the Max-Planck-Gesellschaft and the Deutsche Fr schungsgemeinschaft (Re 5531-1), nr withut the help f M.L. Mdha (Wildlife Cnservatin and Management Dept., Nairbi) and E.K. Ruchiami (Office f the President, Nairbi) wh issued and extended the research permit (Op 131C189114), nr withut the hspitality f thse Kenyans wh kindly allwed us t live and wrk n their prperties, S. and M, Higgins (Naivasha) in particular, We are mst grateful t all f them as well as t the many peple wh assisted in cllecting field data: J, Dittami, V. Haas, J. Heucke, H.A. Isaack, W. Riss, D. Schmidl, E. Snnenschein, H. Waterblk and the late W. Weidig, W.G, Mk (University f Grningen) allwed us the use f this Labratry fr Istpephysics where R. Blaauw analyzed mst f the bld samples. J.L. Brwn, J. Lamprecht, M. Tabrsky, P. Vgel, P. Ward and W. Wickler and an annymus referee made useful cmmcnts n an earlier
369 versin f the manuscript which was transferred frm German English int British English by L. Gardiner, illustrated by D. Schmidl, and typed in its final versin by T. Natli. We are afraid that the cstbenefit rati fr all these unrelated helpers was much wrse than fr the authrs. References Aschff J, PhI H (197) Del' Ruheumsatz vn Vgeln als Funktin der Tageszeit und KrpergrBe. J Ornithl 111:38-47 Brwn JL (1974) Alternate rutes t sciality in jays with a thery fr the evlutin f altruism and cmmunal breeding. Am Zl14:63-8 Brwn JL (1982) Optimal grup size in territrial animals. J Ther BiI 95: 793-81 Brwn JL (1984) The evlutin f helping behavir - an ntgenetic and cmparative perspective. In: Gllin ES (ed) The evlutin f adaptive skills: cmparative and ntgenic appraches. Academic, New Yrk Brwn JL, Dw DD, Brwn ER, Brwn SD (1978) Effects f helpers n feeding f nestlings in the grey-crwned babbler (Pmatstmlls tempralis). Behav Ecl Scibil 4:43-59 Bryant DM, Westerterp KR (1983) Shrt-term variability in energy turnver by breeding huse martins (Delichl urbica: a study using duble-labeled water. J Anim Ecl 52:525-543 Carac T (1979) Time budgeting and grup size: a test f thery. Eclgy 6: 618-627 Carac T, Martindale S, Pulliam I-IR (198) Avian time budgets and distance t cver. Auk 97: 872-874 Davies NB, Hustn AI (1981) Owners and satellites: the ecnmics f territry defence in the pied wagtail,!vitacilla alba. J Anim Ecl 5: 157-18 Duthwaite RJ (1976) Fishing techniques and fds f the Pied Kingfisher at Lake Victria in Uganda. Ostrich 47: 153-16 Duthwaite RJ, Reyer I-IU, Dunn EK (in press) The Pied Kingfisher (Ceryle rlldis). In: Cramp S, Simmns KEL (eds) Handbk f the birds f Eurpe, the Middle East and Nrth Africa. Oxfrd University Press, Oxfrd Drent R, Daan S (198) The prudent parent: energetic adjustments in avian breeding. Ardea 68: 225-252 Drent R, Ebbinge, B, Weijand B (19781979) Balancing the energy budgets f artie breeding geese thrughut the annual cycle: a prgress reprt. Verh Ornithl Ges Bayern 23:239-264 Emlen ST (1982) The evlutin f helping. I. An eclgical cnstraint mdel. Am Nat 119: 29-39 Emlen ST (1984) Cperative breeding in birds and mammals. In: Krebs JR, Davies NB (eds) Behaviral eclgy: an evlutinary apprach, 2nd edn. Blackwell, Oxfrd, pp 35 339 Ewald PW, Carpenter FL (1978) Territrial respnses t energy manipulatins in the Anna hummingbird. Oeclgia (Bed) 31 :277-292 Gastn AJ (1978) The evlutin f grup territrial behavir and cperative breeding. Am Nat 112: 191-11 Gill FB, Wlf LL (1975) Ecnmics f feeding territriality in the gldenwinged sunbird. Eclgy 56: 333-345 Jarman P (1974) The scial rganizatin f antelpes in relatin t their eclgy. Behaviur 48: 215-267 Karasv WH, Andersn RA (1984) Interhabitat differences in energy acquisitin and expenditure in a lizard. Eclgy 65:235-247 Kenig WD, Pitelka FA (1981) Eclgical factrs and kin selectin in the evlutin f cperative breeding in birds.!n: Alexander RD, Tinkle DW (eds) Natural selectin and scial behavir: recent research and new thery. Chirn, New Yrk, pp 261-28 Krebs JR (1974) Clnial nesting and scial feeding as strategies fr expliting fd resurces in the great blue hern (Ardea herdias). Behaviur 51 :99-134 Krebs JR, Cwie RJ (1976) Fraging strategies in birds. Ardea 64:98-116 Lifsn N, McLintck R (1966) Thery and use f turnver rates f bdy water fr measuring energy and material balance. J Ther BiI 12:46-74 Lu DF (1984) Intraspecific variatin in scia! systems f wild vertebrates. Behaviur 88: 266-325 Nrberg RA (1981) Temprary weight decrease in breeding birds may result in mre fledged yung. Am Nat 118:838-85 Nur N (1984) The cnsequences f brd size fr breeding blue tits. I. Adult survival, weight change and the csts f reprductin. J Anim Ec! 53: 479-496 Orians GH, Orians CE, Orians KJ (1977) Helpers at the nest in sme Argentine blackbirds. In: Stnhuse B, Perrins C (eds) Evlutinary eclgy. MacMiIlian, Lndn, pp 137 151 Pulliam HR, Andersn KA, Nisztal A, Mre N (1974) Temperature-dependent scial behaviur in juncs. Ibis 116:36-364 Reyer HU (198) Flexible helper structure as an eclgical adaptatin in the Pied Kingfisher (Ceryle rlldis). Behav Ecl Scibil 6: 219-227 Reyer HU (1984) Investment and relatedness: a cstbenefit analysis f breeding and helping in the Pied Kingfisher (Ceryle rudis). Anim Behav 32: 1163-1178 Riedman M L (1982) The evlutin f allparental care and adptin in mammals and birds. Q Rev BiI 57:45-435 Skal RR, Rhlf F.T (1969) Bimetry. Freeman, San Francisc Stallcup.TA, Wlfenden GE (1978) Family status and cntributins t breeding by Flrida scrub jays. Anim Behav 26:1114-1156 Vehrencamp SL (1979) The rles f individual, kin, and grup selectin in the evlutin fsciality. In: Marler P, Vandenbergh.TG (eds) Handbk f behaviral neurbilgy, scial behavir and cmmunicatin, vl 3. Plenum, New Yrk, pp 351-394 Walsberg GE (1983) Avian eclgical energetics. In: Farner DS, King.TR (eds) Avian bilgy, vl 7. Academic, New Yrk Westerterp KR, Bryant DM (1984) Energetics f free existence in swallws and martins (Hirudinidae) during breeding: A cmparative study using dubly-labeled water. Oeclgia (Berl) 62: 376-381 Westerterp K, Drent R (in press) Energetic csts and energysaving mechanisms in parental care f free-living passerine birds as determined by the Dz 18 -methd. Prc IVIII Cngr Int Ornithl, Mscw 1982 Whitfield AK, Blaber.TM (1978) Feeding eclgy f piscivrus birds at St. Lucia. Part 1. Diving birds. Ostrich 49: 185-198