Durham E-Theses. A study of some factors inuencing breeding of the kittiwake gull Rissa tridactyla (L.) Dixon, Fiona

Transcription

1 Durham E-Theses A study f sme fatrs inuening breeding f the kittiwake gull Rissa tridatyla (L.) Dixn, Fina Hw t ite: Dixn, Fina (1979) A study f sme fatrs inuening breeding f the kittiwake gull Rissa tridatyla (L.), Durham theses, Durham University. Available at Durham E-Theses Online: Use pliy The full-text may be used and/r reprdued, and given t third parties in any frmat r medium, withut prir permissin r harge, fr persnal researh r study, eduatinal, r nt-fr-prt purpses prvided that: a full bibligraphi referene is made t the riginal sure a link is made t the metadata rerd in Durham E-Theses the full-text is nt hanged in any way The full-text must nt be sld in any frmat r medium withut the frmal permissin f the pyright hlders. Please nsult the full Durham E-Theses pliy fr further details. Aademi Supprt Oe, Durham University, University Oe, Old Elvet, Durham DH1 3HP e-theses.admin@dur.a.uk Tel:

2 The pyright f this thesis rests with the authr. N qutatin frm it shuld be published withut his prir written nsent and infrmatin derived frm it shuld be aknwledged. A study f sme fatrs influening breeding f the kittiwake gull Rissa tridatyla (L.) Fina Dixn A thesis submitted t the Faulty f Siene, University f Durham, fr the degree f Dtr f Philsphy 1979

3 CONTENTS page ABSTRACT i ACKNOWLEDGEMENTS i i 1. INTRODUCTION 1 2. BACKGROUND 6 The study area 6 3. NESTING DENSITY OCCUPATION OF THE COLONIES 24 Methds 24 Pani flights 24 The return 25 The effet f windspeed 36 Individual variatin in the date f return t the lny at the start f the seasn 36 Attendane 39 The vaatin f the lny at the end f the seasn PLUMAGE CHANGES THE DISTANCE OF REACTION HATCHING SUCCESS 60 Methds 60 The effet f nesting density n the date f hathing 61 The effet f nesting density n the prprtin f nests in whih eggs hath EGG PREDATION FLEDGING SUCCESS 82 Differenes between the lnies and between the years 82 The effet f nesting density n the prprtin f nests frm whih hiks are fledged 88

4 page The effet f hathing date n the number f hiks fledged frm eah nest 92 The effet f nesting density n the number f hiks fledged frm eah nest 93 The effet f the previus years' breeding suess n the number f hiks fledged DUKBAR AND ST ABB'S HEAD DISCUSSION 110 SUMMARY 125 REFERENCES!29 APPENDICES Results 137 Speifi names f animals mentined in the text Statistis used 143 Symbls and abbreviatins used in the text

5 i ABSTRACT The kittiwake is restrited t breeding within lnies. The effet f nesting density n the time f breeding and n breeding suess was investigated, at lnies in Nrtheast England. Evidene exists that the female kittiwake requires stimulatin frm the mate, and frm surrunding pairs, befre breeding. The psitive effet f nesting density, that is, sial stimulatin, is mediated thrugh i t s effet n laying date, resulting in larger luthes f birds breeding at high density. There is a seasnal deline in luth size. Nesting density has a negative effet n the number f yung prdued frm eggs that hath, whih may be due t the negative atin f aepted density dependent fatrs. The effet f sial stimulatin is nt restrited t any ne year, but is arried frward t the next and subsequent seasns. Birds whih breed at high density return t the lny earlier in the fllwing year, and at a mre advaned stage f the pre-nuptial mult. Behaviural studies during the pre-egg laying phase have indiated that the lny, unless very small, never funtins as a whle, but as a series f interating and interlinking grups. The effet f nesting density is nt ne f mean lny density: the psitin f the nest within a lny is f imprtane. Althugh i t is aepted that there are differenes between reruits t high and lw density areas, i t is prpsed that many f the subsequent differenes are mediated thrugh the effet f sial stimulatin n hrmne seretin.

6 sr. ( A -

7 I J i / / / I 5 I i 1?

8 i i AC KNOWLEDGEMENTS I shuld like t rerd my gratitude t Dr J.C. Culsn whse guidane, enuragement and stimulating ideas thrughut the study were invaluable. In additin, he made available muh unpublished data and early phtgraphs f the Marsden lnies. Withut his help the field wrk at the Dunbar and St Abb's Head lnies wuld have been extremely d i f f i u l t. Fr a l l these things I am muh indebted t him. I am grateful t Prfessr D. Barker fr the f a i l i t i e s made available t me in the Department f Zlgy, University f Durham. I shuld like t thank Vaux Breweries, Sunderland, and in partiular Mr May, the manager, and the staff f The Grtt at Marsden whse peratin was freely given thrughut the perid f study. I am greatly bliged t Mr Eri Hendersn wh prvided bth tehnial advie and assistane with phtgraphy. Pat Mnaghan, Rn Wller and Callum Thmas have helped me with muh useful disussin, and I am partiularly grateful t Callum fr the use f unpublished data fr mparative purpses, and fr assistane with final mputer analyses. Ray Heller kindly prvided the illustratins. My thanks als g t Mrs R.L. Reed wh typed the manusript, and t Mr David Huthinsn wh prepared the figures. Lastly, I thank Nihlas Sykes whse patiene and ntinued supprt thrughut the perid f study were unfailing. The study was finaned by a Siene Researh Cunil researh studentship.

9 1 INTRODUCTION The kittiwake, Rissa tridatyla, is a lnial seabird: i t is bligatrily lnial, being unable t breed in islatin. Clniality (breeding grups f aggregated individuals) is a widespread phenmenn in the animal kingdm, urring in bth invertebrates and vertebrates. Amng the lwer invertebrates the term implies an interdependent physial unin; individuals are prdued frm the same parent stk, and are thus f the same geneti nstitutin. The membeis f inset lnies are a l l lsely related; the distinguishing feature f these lnies is the divisin f labur: individuals are physially separate but dependent n the lny. They shw a high degree f altruism, whih is better explained by studying the degree f relatedness between individuals (e.g. the lse relatinship between sisters in Hymenptera due t the hapidiplid mde f sex determinatin (Hamiltn 1964)). Gruping, flking and lnial breeding in vertebrates have been the subjets f muh study, and several advantages in suh behaviur have been presented t aunt fr i t. Suh aggregatins, whih lak diret physial links, require behaviural adaptatins whih ensure the hesin f the grup. This may require a mplex f speial behaviur patterns, as in the truly lnial speies, r simpler presses in the less permanent aggregatins f animals. There is nsiderable evidene t shw that grups f individuals are mre suessful at deteting r deterring predatrs than islated pairs r individuals. Mst yprinid fish shl in suh a way as t nfuse predatrs (Parr 1927); grund squirrels and blak-tail prairie dgs have been shwn t detet predatrs earlier when in lnies (Carl 1971, King 1955). Densely nesting guillemts are less vulnerable t predatin (Birkhead 1977), and dereased predatin is imprtant in maintaining lniality in sand martins (Hgland and Sherman 197 6) SCIENCE SECTION Library

10 2 Anther advantage in flking is in the use f peripheral individuals as a barrier. As predatrs will tend t take a marginal individual there is an advantage in eah individual trying t gain a entral psitin within the grup. Mvement int the entre f the grup is a widespread phenmenn urring in fish shls, attle herds, bird flks and gull lnies. Sme bird speies flk nly when threatened by a predatr (Tinbergen 1951, Gss-Custard 1970). In general, fish frm mre mpat shls when fed, but beme less aligned when hungry: the advantage in predatr avidane is partially sarified fr the inreased pssibility f finding fd. Fraser Darling (1938) first prpsed the nept f sial stimulatin as an advantage in lnial breeding in birds. He suggested that 'The sial grup and its magnitude in birds whih are gregarius at the breeding seasn are themselves extereptive fatrs in the develpment and synhrnisatin f reprdutive nditin in the members f individual pairs, and thrughut the flk.'- In effet, there is an enhanement f reprdutin by individuals ther than the mate; birds nesting at similar densities breed at the same time; thse nesting at higher densities breed earlier. It has been shwn that fr ring dves under experimental nditins the stimulatin frm the surrunding lny is apable f aelerating reprdutive nditin ver and abve that indued by day length and interatin with the mate (Ltt, Shlz and Lehrman 1967). Lak and Ealen (1939), in their study f Amerian tri - lured redwings, fund that breeding was mre synhrnus within three lnies a few miles apart; their breeding perids were nt ntempraneus, indiating that the physial states f the birds were influened by fatrs ther than thse f the physial envirnment. The Darling effet has als been reprted in the blak-headed weaver (Cllias, Vitria and Shallenberger 1971), the gannet (Nelsn 1978), Viellt's blak weaver (Hall 1970), the

11 3 red-winged blakbird (Smith 1943), the equatrial swallw-tailed gull (Hailman 1964) and in the kittiwake (Culsn and White 1960). In all these speies synhrnisatin f breeding ativities is greater in lal areas. Fisher and Watersn (1941) reprted that the sial stimulus f numbers in a lny enables fulmars t g thrugh their full breeding yle. Hwever Lak (1943) nted that this effet may be explained by the differenes in the ages f the birds in the different sized lnies. Several studies have shwn that synhrny in breeding may result in mre suessful breeding due t a swamping effet f the predatrs (Parsns 1975, Pattersn 1965) and/r t benefits derived frm sial fraging (Emlen and Demng 1975, Hrn 1968). Synhrnised breeding f unknwn rigin urs in sial ungulates, and has been reprted in the wildebeest (Estes 1966) and Afrian buffal (Sinlair 1974). The mst favured reasn fr this is in the swamping f predatrs: this thery, althugh attrative, has nt yet been adequately tested. Being the member f a grup will redue an individual's risk f enuntering a predatr. The advantage f the grup inreases when the size f the grup is abve that number taken by a predatr n any ne enunter. Ward and Zahavi (1973) presented evidene t supprt their thery that breeding lnies and ther bird assemblages have been evlved primarily fr the effiient explitatin f unevenly distributed fd sures, ating as infrmatin entres. Clnial nesting is primarily adaptive t the variable nature f the fd supply in Brewer's blakbird (Hrn 1968). The advantage in sial fraging is an imprtant fatr influening within-lny breeding synhrny in sand martins (Emlen and Demng 1975). The advantage in gruping fr the effiient explitatin f fd resures is nt restrited t bird assemblages. O'Cnnell (1960) has shwn that, under experimental nditins, shls f Paifi sardines

12 4 are mre effiient in finding fd than islated individuals. Similarly, sme large predatry fish frm shls: they themselves have l i t t l e reasn t fear predatin. Wynne-Edwards (1962) has put frward the thery that animal ppulatins an regulate their wn numbers in relatin t the fd available. He has argued that ppulatins rarely inrease t suh a level that starvatin bemes imprtant, and that behaviural mehanisms have evlved whih prevent a speies utrunning i t s fd supply. Mrever, he suggests that the rigin f a l l sial behaviur lies in i t s funtin t prvide infrmatin abut ppulatin density. At present evidene f this type f system is laking. In summary, the fur prpsed advantages in lnial breeding in birds are sial stimulatin, ppulatin regulatin, antipredatin and infrmatin entres fr fd finding. The distributin f breeding birds is generally lassified as slitary, semi-lnial r lnial; lnial speies may als be mmunal r -perative breeders. The ause f any ne speies being lnial may be fr ne speifi reasn, r mre likely fr a mplex f mre than ne f the reasns listed abve. The kittiwake, whih nests n l i f f ledges, is relatively free frm bth avian and mammalian predatrs. I t is prbably derived frm a grund nesting gull, the advantage in l i f f nesting being that i t redues predatin (Cullen 1957). In ntrast t many ther avian speies, the maximum density f any kittiwake lny is determined by the tpgraphy f the rk. All lnies have areas f lw density and hene f late breeding. Areas f high nesting density, and hene early breeding, and thus a greater spread f breeding fr that lny, are gverned by the struture f the l i f f. Breeding synhrny is greater in lw density lnies, and thse with high density have a lnger breeding seasn (Culsn and White 1960).

13 5 Kittiwakes, being bligatrily lnial, require mre stimulatin than an be btained frm their mate. Pairs are unable t breed suessfully in islatin. The advantage in sial stimulatin resulting in mre synhrnised breeding must be a funtin rather than a ause f lnial breeding. An assessment f the effets f nesting density n the time f breeding and n breeding suess was the main bjetive f the present study, in an attempt t gain greater insight int the Fraser Darling effet f sial stimulatin.

14 6 BACKGROUND The breeding bilgy f the kittiwake has been studied fr nany years by Culsn at a warehuse lny at Nrth Shields, Tyne and Wear (55 C I'N : 1 25W). This lny f individually lur-ringed birds, whih has been studied sine its fundatin in 1949, has prvided muh infrmatin n the speies. Assumptins made in the-present study are upheld by data lleted frm this lny. The kittiwake is a lng-lived seabird with a lw annual zrtality: this enables many individuals t breed fr several years (Culsr. 1966). The age f first breeding is three r fur years fr tr.e female and fur r five years fr the male, and breeding suess inreases with age f the female (Culsn 1966). There is a strng tendeny fr individuals t retain their mate in suessive years: rj-.is is nre marked in lder birds, and in thse whih bred suessfully ir. the previus year (Culsn 1972). In additin, breeding birds breeding tgether in suessive years tend t hld the same nest site. Thus, urtship ntinues after pairing, and in eah year at the nset f the breeding seasn. The study area The main study area was Karsden Bay, Tyne and Wear (Nat. Grid ref NZ ) where a maqnesian limestne stak and mainland liffs prvide nesting ledges fr sme fur thusand breeding pairs f kittiwakes. Figure 1 shws the situatin f the lnies. Culsn and White (1956) reprted that kittiwakes were first seen in the area in 1930, and that breeding had begun by Muh f the histry f the lnies is knwn (Culsn and White 1958), and the verall grwth in breeding pairs is shwn in Figure 2. Crmrants and herring gulls nest n the tp f the main stak, and fulmars nest n the larger ledges n the stak and liffs.

15 CS \ CS / \ \

16 'igure 2. The grwth f the kittiwake ppulatin at Marsden Bay. The linear relatinship f the lwer graph between 1937 and 1957 indiates that during this time the ppulatin inreased at a nstant rate. Sine then the rate f inrease in the number f nest sites has prgressively dereased. (The 1937 t 1965 data were supplied by Dr J C Culsn)

17 YEAR r I I I I t YLAR

18 9 The area was divided int twenty lnies arding t the tpgraphy f the rk. Many f the lnies desribed by Culsn and White (1956) have inreased nsiderably in size, and in sme ases the initial entre f lnisatin is n lnger the mst dense area f the lny. These fatrs, r hanges in rk frmatin, aused sme f the lnies t be given different bundaries. Althugh sme f the lnies are still inreasing, the mre dense lnies are dereasing r have reahed a saturated level, smewhat belw the maximum attained (Table 1). In the densest area, where this was bserved, the derease in nesting density may have been due t the inrease in breeding herring gulls immediately abve the area, and thus in predatin. Similarly, a few sites previusly used by kittiwakes were taken ver by fulmars. Five areas were hsen beause f their age, density and aessibility at high tide, fr detailed study. These were the West Fae f the main stak (WF), part f the Suth Fae f the main stak (SFa), tw areas f the Nrth Clny (NCn and NC), and the Suth Clny One (SCI). The West Fae was hsen as there was nly ne site suitable fr munting a time-lapse amera, and this lked ut nt the West Fae. Beause f the size f the Suth Fae lny (733 nests in 1974), nly a prtin was studied intensively. The area hsen (SFa, 129 nests in 1974), was the area f greatest nesting density at Marsden and frmed the tp right hand rner f the Suth Fae. It was therefre bunded n nly tw sides by breeding kittiwakes. Kittiwakes at Marsden first lnised the main stak: as the lny grew, the mainland liffs were lnised. The Nrth Clny Original area (NC) was the initial area f liff lnisatin, and the Nrth Clny New area (NCn) the densest area f the Nrth Clny. The Suth Clny One area (SCI) was the mst aessible lw density area. The beah belw all lw density lnies was ut ff at high tide:

19 10 the beah leading t the area belw the Suth Clny One was the first t be expsed by the reeding tide. These five areas hsen fr intensive study are referred t as the main study areas thrughut. Figure 3 shws the distributin f nest densities in the five nair. study areas (all lnies have areas f lw density, the high density areas have a greater spread f nesting densities). It is ntable that the range in nesting density fr eah lny is greater than that reprted by Culsn and White (1960). In their study the maximum number f nests rerded within a five feet radius was twelve: in three f the lnies there are nests with twie this number f nests within a five feet radius. This is disussed further under Nesting Density (see page 13 ).

20 Table 1 Nest unts fr the 20 Marsden lnies fr 1974 t 1977 inlusive Year Clny SF WF NF KC (1) (2) NCn (3) (4) (5) (6) NC (7) (8) (9) GC GCC SCI CAVE BUT BUT ' ST N ST S SC FSB FSC MVB MVBs MVBst W MVBst S EF (SFa) Ttal In 1974, 1975 and 1976 lny nest unts were taken three times between May and June, and the mean number f nests taken as the number f nests in eah lny. In 1977 nests were unted ne nly in late June.

21 12 Figure 3 shws the distributin f nest densities in the five main study areas in The number f ther nests within a five feet radius f eah nest was rerded and the number in eah density grup expressed as a perentage f the ttal number f nests in the lny. The dtted line indiates the 10% level. It is ntable that the range in density is greater in the high density lnies, and that the range in nesting density fr eah lny is greater than that reprted by Culsn and White (1960). (In their study the maximum number f nests rerded within a five feet radius was twelve.)

22 10 SFa 10 N Cn 10 WF C=3 NC 10 0) CD 0) SC I 10-0 "1 2 3 '4 ' '8 '9 10 " ' ' 15' 16' 17 ' '20 21 '22*23"24 "25 "26 ^7 1 Nests in 5ft. radius

23 13 NESTING DENSITY The basis f this researh was t study the effet f nesting density n breeding suess and n the time f breeding. As suh, the Dethd used in measuring nesting density is f nsiderable imprtane. Varius methds are pssible, but eah has its wn limitatins. Althugh kittiwakes aggregate fr breeding, they are nt evenly distributed within an area. In general, nests in high density areas have nsistently high values fr the number f nests at suessive distanes frm them, and lw density areas have nsistently lw values. Figure 4 shws the psitin f nests in parts f tw f the main study areas (Nrth Clny New area, NCn, and Suth Clny One, SCI). Clumping f the nests is mre evident in the lw density lny and here the desribed relatinship des nt hld. Beause f lal lumping, suh methds as nearest neighbur and lsest individual tehniques, giving a mean value f nests per unit area, will tend t give exaggerated density values fr lw density lnies. Further, the bilgial signifiane f nesting density must be nsidered. A mean value f nests per unit area fr eah lny takes n aunt f the distributin f nests within that unit area, althugh suh values will serve t mpare high and lw density lnies. It must be nsidered, and uld be expeted, that nests in lse prximity with anther will have a greater effet than thse further away. Fr these reasns the methds used in measuring nesting density were as fllws. The main study areas were phtgraphed and eah nest numbered n an enlarged print. The nrmal kittiwake nest is 12 inhes (ne ft) in diameter, and beause f this all density measurements were alulated in feet. The diameters f twenty nests were measured n eah phtgraph, and the mean value f nest diameter taken as a standard f ne ft (0.305m). With the use f a perspex verlay n whih nentri irles were ethed

24 14 Figure 4. The psitin f nests in areas f tw f the lnies (Nrth Clny, NC, and Suth Clny One, SCI), shwing lal lumping f nests. Clumping is mre evident in the lw density lny (SCI) and is due in part t the struture f the rk

25 SCI NC

26 15 at the equivalent f 1, 2, 3, 4, 5 and 10 feet frm the entre, the number f nests within eah area was rerded fr eah nest (i.e. within the areas f 0-1, 1-2, 2-3, 3-4, 4-5, 5-10 feet f eah nest). Additinally fr the main study areas, and fr all ther lnies, a density measure was btained by the bundary strip methd (Suthwd 1966). On an enlarged phtgraph f eah lny a line was drawn jining the peripheral nests, and a line drawn at the equivalent f five feet utside the frmer bundary. The area f the lny was alulated frm this, and the mean number f nests within 78.5 square feet (the area f a irle with a radius f five feet) was alulated (76.5 sq ft = 7,3 sq m). 3ehaviural studies f the infetius nature f the greeting eremny, during the pre-egg laying perid, have indiated that pairs respnded t thers ver a distane f apprximately five feet (Chapter 6, page 52). AS it is likely that this behaviural respnse is an imprtant sure f stimulatin fr the kittiwake, the mean number f ther nests within a radius f five feet was used as a measure f effetive nesting density. Beause f the methd used in taking the standard f ne ft, frm the diameter f twenty nests, an element f subjetivity is intrdued. Althugh nsistent within this study, the atual bundary (diameter) f the nest may have been greater than that used by Culsn and White (1960). The area used fr kittiwake nest sites n the main stak has remained mre nsistent than the areas f mainland liff lnies, the latter having inreased in size nsiderably sine the study by Culsn and White. In an attempt t assess hw muh the methd emplyed aused the differenes in density reprted here and thse reprted by Culsn and White (1960), the nesting density f the Suth Fae lny was alulated fr 1958 frm phtgraphs. The Suth Fae lny as desribed by Culsn and White (1960)

27 16 inrprated the West Fae lny whih in 1958 was a small lw density peripheral area f the Suth Fae. (The Suth Fae and West Fae tgether will be desribed as the Suth Fae C Clny.) Initially a lny density measure fr the Suth Fae C Clny in 1958 was alulated, by again taking a standard f ne ft, and by using the bundary strip methd. The mean nesting density f the Suth Fae C Clny was alulated t be nests/78.5 sq ft (f reprted as the nesting density f the Suth Fae C Clny in 1958 by Culsn and White (1960))- Their methd invlved alulating the area f a retangle f liff, whih was drawn ten feet utside ne bunding peripheral nests. The area f liff used was thus greater than that used in the present alulatin fr Nnetheless, alulatins using the bundary strip methd shw that the nesting density f the Suth Fae Clny Inreased by 5.3% between 1958 and 1975, and the nesting density f the West Fae by 225% ver the same perid. Bth inrease in nesting density and the methd emplyed have given rise t the fat that the present density measures are higher than thse reprted by Culsn and White (1960). Lw density lnies at Marsden uld be expeted t have inreased in density mre than the high density Suth Fae lny, in that the verall number f nests at Marsder. inreased by 74% between 1958 and T assess the nsisteny f the ne ft standard, the number f nests with n thers within a radius f five feet, and the number f nests with ne ther within a radius f five feet, were alulated. A small disrepany in the measure will bviusly be f signifiane when determining the atual number f nests within a five feet radius. If, fr example, a standard f nt ne ft but 1.1 feet is mpared with a standard f ne ft, the result is in an inrease f 21% in the area f a 'five feet' radius irle. Culsn and White (1960) reprted 1.5% f

28 17 nests in the Suth Fae C lny with n ther nests within a radius f five feet, and 4.5% f nests with nly ne ther within a five feet radius; by using the same methd, I btained 1.0% f the nests with n thers within a radius f five feet, and 1.4% f nests with nly ne ther within a radius f five feet. The standard f 'ne ft 1 used in the present study is therefre greater than that used by Culsn and White. This disrepany in the measure must be nsidered when any literal mparisn between nesting density in the tw studies is made, althugh the evidene shws that bth the methd emplyed, and the fat that nesting density has inreased, have given rise t the differenes in the values f nesting density reprted by Culsn and White, and thse in the present study. Clumping urs in nly a small prprtin f nests: beause f this there are psitive rrelatins between suessive density values btained fr eah nest in the main study areas (Table 2 fr 197 5; Table 3 fr 1976). It is ntable that all rrelatins are psitive, and that values inrease dwn the table. As the area f nentri irles inreases as the distane frm the entre (nest) inreases, the number f nests inreases. Fr example, there are mre nests tw t three feet frm a nest than ne t tw feet frm it, and this is shwn in Table 4 whih gives the mean nesting densities at suessive distanes frm eah nest. Beause f this, there is ptentially greater variatin in the number f nests at inreasing distanes, and thus greater pssibility f inreased rrelatins between the higher density values f areas further frm a nest. In additin, the perentage f nests with n nests within eah area dereases as the area inreases (Table 5). The suessive inrease in rrelatin effiients at inreasing distanes frm the nest are due t this, as are the inreases in regressin effiients arss the Tables 2 and 3.

29 18 Tables 2 and 3 shw the rrelatin effiients and regressin effiients between density values fr 1975 and As the area f nentri irles inreases with distane frm the entre (nest), the number f nests within eah area inreases. There is, therefre, ptentially greater variatin in the number f nests at inreasing distanes, and the inreased pssibility f inreased rrelatins between higher density values f areas further frm the nest. In additin, the number f nests with n nests within eah area dereases as the area inreases. The suessive inrease in rrelatin effiients a.z inreasing distanes frm the nest, and the inreases in regressin effiients arss the tables, are due t this.

30 19 Table 2. A) shws the rrelatin effiients (all psitive) between density values f all nests in the main study areas in 1975, tgether with the mean number f nests in eah area, and the standard deviatin, based n 814 nests. B) shws the regressin effiients (all psitive) between density values in 1975, based n 814 nests. A) Density 1-2' 2-3' 3-4' 4-5' 5-10' x ± S.D. 0-1' * ' ' ' ' ' B) Density 1-2' 2-3' 3-4' 4-5' 5-10' 0-1' ' ' ' ' = the number f nests within ne ft f a nest 1-2' = the number f nests between ne and tw feet frm a nest et. x ± S.D. = the mean number f nests within eah area and its standard deviatin * = the mean nest density f the 0-1" area is a measure f the mean number f ther nests within that area, fr eah nest. The true value shuld thus be inreased by ne, t inlude the entral nest frm whih measurements were taken.

31 20 Table 3. A) shws the rrelatin effiients (all psitive) between density values f nests in the main study area in 1976, tgether with the mean number f nests in eah area and the standard deviatin, based n 614 nests B) shws the regressin effiients (all psitive) between density values in 1976, based n 614 nests A) Density 1-2' 2-3' 3-4' 4-5' 5-10' x ± S.D. 0-1' * ' ' ' ' ' B) Density 1-2* 2-3' 3-4' 4-5' 5-10' 0-1' " ' " " ' = the number f nests within ne ft f a nest 1-2' = the number f nests between ne and tw feet frm a nest et. x ± S.D. = the mean number f nests within eah area and its standard deviatin * = the mean nest density f the area is a measure f the mean number f ther nests within that area, fr eah nest. The true value shuld thus be inreased by ne, t inlude the entral nest frm whih measurements were taken.

32 21 5 H S U 01 4J 01 s in rl tn nj a) (0 4-1 in A 10 a > > r! UH 0) A 4J VI m (0 a> 3 H >* 4J H (0 0) > ih V IB 0) to CO E 0) 6 in CD id 0) H u H V) 0) rh u n u w 4J d! u <»H 10 n (0 0> 4-> in < N H tn 0) rh TO (0 S 4-> 10 r( > > T3 rt (0 g 3 m Q +J w 0) 0) 10 V u (0 V r< 10 J: u r 01 rl 4J 4J (0 0> VW 0 M v -i 3 Cl> 4-) 4J U) 0> 0) JJ 0) 0) x; 4J tn 4J tn 0) 0 0) 3 Z in I r- O VO m V0 VO r in rh CM rh i-h rh rh vd O in in > rh r r- ID * # rh <N O r VD i-l rh rh r-h r CM VD Q in rh in 00 O CM a CM rh rh CM VD 3 1 rh (M CM VD rh r T r rh. rh in O r-- rh Q CO rh in n CM U) rh CM rh O CM I r vd VD O CM VD CM vd CM r CM CM r I CM CM I CO 00 Q r- r W rh H *H O rh IX a CM VD CO CM r CM CO v CM CM CM 6 r r rh CO r r- 6 CM CM vo O r CM VD in r vo CM CO en in rh rh d rh rh VM vo r- O rh v rh O CO CO CM tn a> 1 X rh i-h r rh (0 4J tn -rj T CM r- CO TJ a in CM O O 01 rh I w d d d d d d 4J 1 x: vd n O u LO in r 1 O vo X it O O O O d 10 0) E 4J in tv MH 0 4J 0 O MH 01 C 4J 05 4J in UH 0 u 0) x> E 3 C 0) J= 4J u 4J 01 4J tn 01 s U-i 4-J > 01 U-l 4J (0 u 01 0> ^ 4J 0) S3 V> 4-: tn t' UH O rl 01 X) 0) x: 4J 0) rh XI 5-i C 0 rh 8 u z <0 5

33 22 Table 5. The perentages f nests in eah lny whih have n ther nests within the areas f 0-1, 1-2, 2-3, 3-4, 4-5 and 5-10 feet frm them. As the area f nentri irles inreases with distane frm the entre (nest) the number f nests within eah area inreases. Assiated with this, the perentages f nests with n ther nests within eah area derease as the area inreases. The lw density areas have mre nests with n ther nests within eah area Distane frm the nest Clny 0-1' 1-2' 2-3' 3-4' 4-5' N SFa WF NCn NC SCI Mean

34 23 In bth the analyses f fatrs affeting hathing date and f thse affeting the number f hiks fledged, multivariate stepwise regressin analyses have been used with density values as independent variables. Althugh they remain true independent variables, beause f the rrelatins between them, the results btained were in sme ases misleading, and are has been taken t avid errneus nlusins being drawn. Fr example, in the multivariate stepwise regressin analysis f fatrs affeting hathing date in 1976, bth ne t tw feet density and five t ten feet density were signifiant fatrs, but within ne ft density, tw t three feet density, three t fur, and fur t five feet densities were mitted. Althugh the analysis is statistially rret, i t makes little sense bilgially t have hsen the ne t tw feet density value in preferene t the within ne ft density value. When any ne density value is intrdued as an independent variable int the equatin f a multivariate stepwise regressin analysis, the rrelatin f ther density values with the residual variane is redued. If a send density value subsequently enters the equatin, i t may be that i t desribes a negative relatinship when the initial density desribed a psitive relatinship. The initial hsen value having a psitive relatinship will have remved muh f the effet f the fllwing density value, and the latter, when later intrdued, may desribe the spread f the relatinship and thus exhibit negative rrelatin and regressin effiients. Mrever, the hie f different density values in equivalent analyses in suessive years may nt indiate real differenes in the tw situatins; fr example, by hane variatin three feet density may f i t the data f hathing date better than fur feet density whih was hsen in the previus year. This effet is again due t the rrelatins between suessive density values, and higher rrelatins assiated with an inrease f nests in the larger areas.

35 24 OCCUPATION OF THE COLONIES Previus studies f kittiwake breeding bilgy have indiated that nesting density is an imprtant fatr in determining the time f breeding (Culsn and White 1960). As nesting density an nly have a stimulatry effet when the birds are present at the lny, the annual reupatin and vaatin f the lnies were investigated. Methds The number f sites upied and the number f pairs present at ali lnies were unted at least ne eah week between and G.hJ.T. during 1974 and 1975 thrughut the perid f upatin, and in 1976 during the perid f reupatin. The nests in the five main study areas were eah numbered n an enlarged phtgraph f eah area. These areas were phtgraphed weekly t give data n the upatin f the individual sites. The West Fae was als studied with the use f time-lapse phtgraphy. Phtgraphs were taken every ten minutes frm dawn t dusk twie weekly thrughut the 1974 breeding seasn, and muh f the seasn. Several days' data were lst in 1975 due t a failure f the autmati light meter en the amera. On eah visit t the lny the amera was set t film n the fllwing day. During the early part f the seasn sme f the films were f pr quality ue t heavy seamist, rain and asinally ther adverse weather nditins. Standard 35mm, blak and white, 400 ASA film was used, and the prjeted negatives prvided enugh detail fr diret analysis. Pani flights Pani flights were first desribed by Kirkman (1937) as upflights f the blak-headed gull: they als ur in mmn and arti terns, and are used t deter avian predatrs. In the kittiwake, during the early part

36 25 f the seasn, these flights appear spntaneusly withut an bvius stimulus, r in respnse t the appearane f a lw flying airraft r helipter, whih presumably simulate an avian predatr. The birds leave the nest site with a harateristi dive, and then fly rapidly ut t sea: the whle ativity takes plae in mplete silene, in ntrast t the nrmal nditins at a lny. The return Many seabird speies spend lnger at their nest site than required fr nest building and breeding (Culsn and VThite 1958, 3elpl'skii 1961, Birkhead 1977), and this is true f the kittiwake. The reupatin f the lnies is a gradual press whih takes plae ver several weeks. Cuisn and White (1956) have desribed this fr the lnies at Marsden, and althugh the general pattern f behaviur remained the same, the time and rder f return differed frm their findings. The first stage f the reupatin.is nt the presene f birds n their nest sites, but the presene f a raft f birds sme yards ffshre. These rafts frm early in the mrning and are present until abut nn, when the birds fly ut t sea tgether. The rafts were nt frmed in windy weather. The size f the rafts inreases daily until as many as three hundred birds may be present (Table 6). Frm the time that the rafts ntain abut thirty individuals, birds le ave the raft in a synhrnised grup and fly in twards the liffs, fly past, and then ut t rejin the raft. This is repeated until eventually sme f the birds land n the sites, where they remain fr a few minutes nly. When n the sites the birds are extremely alert and uneasy: they sn leave the lny in a pani flight and return t the raft. By the time that 60% f the sites are regularly upied, the rafts are n lnger frmed: the birds fly straight in frm the sea in the early mrning.

37 26 Table 6 The size f rafts seen ff the Suth Fae Clny in 1975, and the number and perentage f nest sites upied n the Suth Fae. The rafts frm early in the mrning and are present until nn when the birds fly ut t sea tgether. Rafts are nt frmed in windy weather. (It is ntable that n , when n raft was frmed, the mean daily wind speed was 16.7 knts.) The size f the rafts inreases daily: by the time that 60% f the sites are regularly upied, the rafts are n lnger frmed. Number f Perentage f Mean daily wind Date Raft size sites upied sites upied speed (knts)

38 27 Never was a single bird seen t fly in frm a raft and upy an empty lny. At least ten, and usually abut frty birds wuld fly in tgether, and sme upy sites. The departure frm the lny was less synhrnised, with single birds asinally being seen n large areas f liff, but fr shrt perids nly. Birds appear t require the stimulatin frm ther birds t verme the urge t stay at sea, and t satisfy the breeding drive t me ashre and upy the nesting ledges. N rrelatin was fund between the numbers f birds initially lnising an area and either the density f birds, r the nest density f the lny. As the birds invlved in the initial reupatin are the lder birds, and thse returning t their sites f the previus year, suh a relatinship wuld be unexpeted. A lear threshld level f the number f birds reupying a lny was nt evident, althugh i t an be said that single birds were never seen t upy empty lnies. The return and build-up f numbers f birds in the main lnies are shwn in Figure 5, fr 1975, by expressing the number f sites upied as a perentage f the subsequent nests in that lny. The return is gradual, starting in early January eah year, and the rder f return was muh the same in the three years studied. This rder is nt that f initial lnisatin, but lsely fllws the rder f mean nesting density (Table 7). The differenes between the lnies are smaller than thse reprted by Culsn and White (1956) wh fund that the rder f return was the same as that f lnisatin in bth 1953 and During this perid, the lwer density lnies were rapidly inreasing in size and therefre had a higher perentage f yung birds, returning later, whih uld have given rise t the greater variatin between lnies. The effet f lny mean density n the date f return t the five main study areas was investigated. Figure 6, the perentage f sites upied against density fr eleven dates during January and February and

39 28 Figure 5 The return and build-up in numbers f birds in the main lnies in The number f sites upied is expressed as a perentage f the subsequent nests in that lny and pltted against date. The reupatin f the lnies is a gradual press whih takes plae ver several weeks: there are differenes in the dates f return t the different lnies. A shws the return t the Suth Fae lny (SF) and Marsden Village Bay (MVB) B shws the return t the Suth Fae lny, the West Fae lny (WF), and the Far Suth Clny (FSC) C shws the return t the Suth Fae lny, the Nrth Clny (NC), and the Suth Clny Tw (SC2) D shws the return t the Suth Fae lny, the Grtt Clny (GC), and the Suth Clny One (SCI) E shws the generalised pattern f return t tw hypthetial lnies.

40 CN u y l/l Z 1/1 I CO LU U in U u. i 1/1 CO LU \ LU CO 39VlN3D>J3d I O

41 29 Table 7 The rder f return t the lnies in 1974, 1975 and 1976 fllws that f mean nesting density and nt that f initial lnisatin. The differenes between the lnies are less marked than thse fund by Culsn and White (1956) when the lw density lnies were rapidly inreasing. The magnitude f the differenes are shwn fr in Figure 5, and tabulated belw. Breeding first Order f Order f return Clny reprted lnisatin ' density SF WF NC GC SC FSC MVB ? First date n whih 10%, 20%, 30% and 40% f Clny the sites were upied in 1975 (1 January = 1) 10% 20% 30% 40% SF WF NC GC SC FSC MVB

42 30 Figure 6 The relatinship between the perentage f sites upied and lny mean density f the five main study areas fr eleven dates in January and February and ne date in April The high density lnies had a higher perentage f sites upied n eah day during the reupatin phase. As the seasn prgressed the differenes beame less marked. (Dates : 1 = = = = = = = = = = = = 5.1)

43 i 1 h g m T- 3DVlN3DU3d

44 31 ne date in April in 1976, shws that the perentage f sites upied n eah day inreases with nest density. As nne f these lnies is steadily inreasing, i t seems unlikely that these differenes are due t differenes in the age struture. I t shuld be remembered that the differenes fund here in the dates f return t the different lnies are less marked than thse reprted by Culsn and White (1956). It may be that mrtality in lw density peripheral areas is higher than in the high density areas (Culsn 1968) but if this were t explain the bserved effet, a linear relatinship between annual mrtality rate and nesting density wuld be neessary. Evidene fr r against this is laking. The Nrth Clny was divided int five areas arding t the tpgraphy f the liff, and the return in 1975 and 1976 was rerded (Table 8). The rder f return did nt fllw that f mean density f eah area. The peripheral areas (1 and 5), whih were bth inreasing and may be expeted t have had a high prprtin f yung birds nesting in them, were the latest areas t be reupied. The densest area, and the mst stable in numbers, was, in eah ase, the first area t be reupied (nt the initial area f liff lnisatin i.e. area 3). These data indiate that bth the number f yung birds in the lny and als the nesting density are imprtant in influening the date f return t eah area. As the perentage f sites upied inreases thrugh the seasn, the time spent by individuals at the lny daily inreases. By the time that a quarter f the sites are upied, the birds arrive at the lny shrtly befre sunrise and depart arund sunset: the birds leave in pani flights and fly straight ut t sea. The numbers f sites upied inreases during the day, reahing a peak in the late afternn. By the time that fifty perent f the sites are regularly upied, the birds remain at the lny until several hurs after sunset, and nly bservatins

45 32 Table 8. The rder f relnisatin f five areas f the Nrth Clny at the start f the and 1976 seasns did nt fllw that f mean nesting density. The peripheral areas (1 and 5), whih were bth inreasing and may be expeted t have had a high prprtin f yung birds nesting in them, were the latest areas t be reupied. The densest area, whih was stable in numbers, was in eah year the first t be reupied. The riginal area f liff lnisatin was area 3. The magnitude f the differenes in the dates f return t the areas f the lny are shwn. Nrth Nest Order f Relnisatin 0-5' Perentage hange in Clny Cunt Mean nests Areas density First date n whih 10%, 20%, 30% and 40% f the sites were upied (1975 and 1976) (1 January = 1) Area 10% 20% 30% 40% 10% 20% 30% 40%

46 33 in the dark shwed that the lnies were in fat vaated at night. Cunts f birds present at night were impssible, althugh their presene r absene was deteted, using a strng flashlight. The lnies were vaated daily up until abut tw weeks befre egg laying. As the perentage f sites upied inreases thrugh the seasn, the perentage f sites upied by pairs inreases (Figure 7): thirty perent f the upied sites are upied by pairs nly by the time that frty five perent f the sites are upied. As the pair bnds stabilise and the seasn prgresses, pairs spend less time tgether. After the eggs are laid there is a marked deline in the time that pairs spend tgether, and thus a redutin in the perentage f upied sites upied by pairs (Figure 11). The five main study areas were divided int twenty five units, eah having at least twenty nests, and the mean date f return in 1976 f eah unit was pltted against the mean density f eah unit (Figure 8). The signifiant negative relatinship (r^^ = -0175, b = ± 0.077, =+23.4, p < 0.001) again shws the imprtane f lny density in determining the date f return. Frm this relatinship i t an be predited that fr an inrease f ne pair nesting within a five feet radius f any bird, that bird will return tw and a half days earlier in the fllwing breeding seasn. Nesting density in the ensuing breeding seasn an nt have a diret effet n the date f return t the lny. If density des have an effet, and the evidene strngly suggests that i t des, i t is nt restrited t its effet n any ne seasn, but affets the fllwing year, in that a high nesting density in ne year gives rise t an advaned, earlier, breeding date f return t the lny at the start f the fllwing seasn.

47 34 Figure 7 The perentage f upied sites upied by pairs against the perentage f sites upied, during the reupatin phase (January t April), fr the five main study areas. As the perentage f sites upied inreases thrugh the reupatin phase, the perentage f sites upied by pairs inreases. Thirty per ent f the upied sites are upied by pairs nly by the time that frty five per ent f the sites are upied.

48 (0 m O t 1 m i r- r CM r - O < in CT) U a: r- O in O CO (SUlVd) CM 3DVlN3Dy3d

49 35 Figure 8 The five main study areas were divided int twenty five units, eah having at least twenty nests, and the mean date f return t eah area in 1976 pltted against the mean nesting density f eah unit (nests within a 5' radius). Birds breeding in high density areas return t the lny earlier at the start f the fllwing seasn. The signifiant relatinship (r^^ = -0.75, p < 0.001) is desribed by the regressin equatin: y = -0.42x

50 LO e CO n LU CN LU s OO CN s A1ISN3Q

51 36 The effet f wind speed Wind speed was the sle envirnmental fatr fund t have a marked effet n the number f birds present n nseutive days. On days when the wind speed was abve eleven knts (12.7 mph) the numbers f birds were dramatially redued. Cunts frm suh dates (during the perid f reupatin f the lnies) were paired with thse made n the next r previus visit, when similar numbers f birds uld be expeted. Tnese data were graphed against the mean daily wind speed (Figure 9). Similarly, the effet f wind speed n the numbers f pairs present is shwn in Figure lo. It is seen frm these figures that n dates when the numbers f sites upied and the numbers f pairs present are redued, the wind speed is greater than eleven knts. Inreases in wind speed firstly ause a drp in the number f pairs present, and sendly in the number f sites upied. As there is nsiderable mpetitin fr nest sites it is an bvius advantage fr ne bird f a pair t remain at the nest site when ther birds are attempting t find sites. As the seasn prgresses the effet f wind speed dereases, and by May has little effet n the presene f breeding birds. Inreases in wind speed at this time d ause redutins in the number f yung prspeting birds. Individual variatin in the date f return t the lny at the start f the seasn As kittiwakes shw marked nest site tenaity, and lder birds return t the lny earlier, a psitive relatinship between the dates f arrival f individuals in suessive years may be expeted. The rrelatin effiient fr the dates f arrival t all sites in the main

52 37 Figure 9 The effet f wind speed n the perentage f sites upied in the main study areas, during the perid f reupatin ( t : t ). Cunts frm days n whih the wind speed was abve eleven knts (12.7 mph) were paired with thse taken n the subsequent, r previus, visit, when similar numbers f birds uld be expeted. The perentage f sites upied n eah day was pltted against the mean daily wind speed. On days when the wind speed was abve eleven knts the perentage f sites upied was redued.

53 eg eg 0> eg eg UJ UJ CO >- in LU eg lh< > 00 (0 1 3DVlN3.Dy3d

54 38 Figure 10 The effet f wind speed n the perentage f upied sites upied by pairs in the main study areas, during the perid f reupatin ( t : t ). Cunts frm days n whih the wind speed was abve eleven knts (12.7 mph) were paired with thse taken n the subsequent r previus visit (when a similar perentage f upied sites upied by pairs uld be expeted) and these data graphed against mean daily wind speed. On days when the wind speed was abve eleven knts the perentage f pairs present was redued.

55 CO CM CO LU LU n CO CO so CM CO 3DVlN3Dy3d

56 39 study areas fr 1975 and 1976 was alulated (r Q._ = +0.44, b = ± 0.017, p < 0.001). Similarly, the rrelatin between the date f departure in 1975 and the date f arrival in 1976 was alulated (r g l? = -0.33, b = ± 0.024, p < 0.001). Birds whih are early in arriving bak at the lny at the start f the seasn are nsistently early; mrever, they vaated the lny later at the end f the previus seasn. The relatinship between the date f departure and the date f arrival in the fllwing year predits that any bird will return fur days earlier in the next breeding seasn fr eah additinal day that it remained at the lny. Table 9 shws the rrelatin effiients fr eah main study area where the situatin is seen t be similar. The relatinship between the dates f arrival in suessive years is thus nt nly a funtin f lny density, but pints t the imprtane f individual variatin. Attendane The attendane at all lnies thrughut 1974 and 1975 was rerded. The mst striking differenes between the lnies were the perentages f yung prspeting birds present. Prprtinately mre suh birds were present in the lw density lnies (Table 10), indiating that in these lnies ptential nest sites were available. The perentage f sites upied by pairs was highest during the pre-egg laying perid when pair frmatin and urtship take plae. After egg laying, the number f sites upied by pairs drps dramatially. After the hiks have fledged, and befre the vaatin f the lny by breeding birds, there is a resurgene f urtship behaviur, when the perentage f sites upied by pairs shws a distint peak. Figure 11, the annual upatin f the Suth Fae lny in 1975, shws these pints.