Behavior of the Swallow-Tailed Gull of the Galapagos

BEHAVIOR OF THE SWALLOW-TAILED GULL OF THE GALAPAGOS

BARBARA K. SNOW AND D. W. SNOW British Trust for Ornithology Beech Grove Tring, Hertfordshire England

TWO species of gulls breed in the Galapagos, plumage (except dependent young) are seen and both are peculiar to the archipelago. One at the breeding colonies. of them, Larus fuliginosus, is apparently quite Swallow-tailed Gulls are social breeders, closely related to Larus modestus of the Peru- although the distance between nests varies vian coast (Murphy 1936); but the other, the widely. Eggs are laid in all months of the Swallow-tailed or Fork-tailed Gull, Creagmcs year, the reproductive cycle of the individual furcatus, differs strikingly from all other gulls bird being usually 9 to 10 months (Snow and in a number of ways, and its separation in a Snow 1967). A single egg is laid, and the in- monotypic genus has been almost unanimously cubation and fledging periods are unusually accepted. long (33-34 days and about 100 days, respec- Creagrus is a rather large gull, with long tively). The nest site is typically on more-or- pointed wings, long forked tail, and short less broken cliffs, and Hailman (1965) has tarsus, being in these respects more ternlike shown that the Swallow-tailed Gull shares than other gulls (table 1). Unlike a tern, it with the Kittiwake, Ri.ssatridactyla, some fea- has large webbed feet well adapted for swim- tures that Cullen (1957) showed were adapta- ming. Its beak is long and strong. The body tions to cliff-nesting. plumage is white, with a pale gray mantle, The calls and displays of Creagrus, which edged with white along the scapulars. The form the subject of this paper, are also un- wings have a conspicuous gray, white, and usual. Some are so different from those of black pattern. In nuptial plumage the hood is other gulls that in the following account they dark gray and contrasts sharply with a con- are given their own descriptive names. The spicuous patch of white feathers at the base of terminology of Tinbergen (1959) has been the beak, which is black with a pale-gray tip. used only for a few calls and movements that The gape and eye-ring are vermilion, and the seem to be clearly homologous with those of legs a more pinkish red. other gulls. Creagrus is a pelagic, nocturnal feeder on small fish and squid. [A total of 49 food items METHODS were identified during the study, of which 37 From February 1963 to April 1964 one or both were fishes (clupeoids) and 12 were squids of us paid fortnightly visits to South Plaza ( Symplectoteuthis ouulaniensis) . In Febru- Island, off the east coast of Santa Cruz Island, ary-July, 34 of 37 food items were fishes, in the center of the Galapagos group. From whereas in November, 9 of 11 were squids. February to December 1963 we spent 48 hours Hence there may be important seasonal vari- on the island at each visit, and so were able ations in diet. Hailman (1964b) concluded to observe all stages of the breeding cycle, at that squid were the main food, but his obser- all seasons. One of us also spent seven days vations were made in November and Decem- observing some breeding pairs on Tower ber only.] When it is not breeding it is totally Island in May 1964. pelagic, migrating eastward to the coasts of All the Plaza nests were numbered, and 54 Ecuador and Peru. No birds in immature adults were marked with color bands. Chicks were marked with a single color band that in- dicated the period within which it was born. TABLE 1. Relative tail length and tarsus length of Creagrus compared with typical gulls and large terns. Swallow-tailed Gulls are tame enough to be watched quite easily from a distance of 26 Tail/wing Tarsus/wing yards without a blind, and most observations (per cent ) (per cent) were made in this way. Some nests were also Typical LUTU-Sspecies 37-43 15-18 studied from blinds erected a few feet away. Creagrus 46 12.5 The technique of observation was essentially Large terns (Thalasseus) 44-54 9-9.5 that used by Tinbergen and his associates in Note. Percentages are based on mean measurements given by Dwight (1930) and Murphy (1936). their earlier studies of gull behavior. The TheCondor, 70:252-264, 1968 l%21 BEHAVIOR OF THE SWALLOW-TAILED GULL 253 various display actions were identified and de- months of the year, but small colonies or scribed as accurately as possible, and repeated groups of birds breed synchronously. Within observations were made of the circumstances these groups the most suitable sites, affording in which they were performed. Particular at- protection from the sea and cover for the tention was paid, over periods of several weeks, chick, were taken up first, and males estab- to groups of breeding pairs containing some lishing themselves later tended to be less suc- banded birds, whose behavior was followed cessful. All the marked males returned to the from the time they first occupied their nesting site where they had nested before, or to one area. A photographic record was obtained of within a yard of it, but several of the marked most of the displays. females changed their sites. There was usually competition between females over a male OBSERVATIONS with an established nest site, and sometimes NEST SITE, NEST DISPERSION, one ousted another that had already been AND PAIR FORMATION visiting a male for several days and appeared On South Plaza Island the Swallow-tailed to have been accepted by him, i.e., allowed to Gulls nest either on or near the cliffs on the remain at the site. One marked female estab- windward side. Nests on top of the cliff, lished preliminary pair bonds with two dif- which are comparatively uncommon, depend ferent males before finally pairing with a third on some protection provided by rocks or near the site used in her previous breeding bushes. On the cliff, ledges as small as one season, In each case the breakdown of the square foot may be used. Sites where there is earlier pair bond was caused by the interven- a rock overhang or a small cavern are much tion of another female; but no dissolution of a favored. As mentioned earlier, the distance pair was observed if it had reached the stage between nests varies greatly. It may be as of mutual head-tossing and copulation. little as 18 inches, while at the other extreme the occasional nest may be 200 yards or more VOICE AND DISPLAYS from its nearest neighbor. It is usual, how- Two calls, the Rattle and Whistle and the De- ever, for nests to be in groups, with the num- parture Call, with their associated behavior, ber of nests in the group and their distance are described first. These are the only two apart dictated by the terrain. Social behavior that are not directly related to reproductive undoubtedly plays a large part in the forma- activities, and so stand rather apart from the tion of these groups (Hailman 1964a), but on others. The other calls and displays then fol- the Plaza Islands their locations on the cliffs low in what is hoped is a logical order, al- also depended on the weather (Snow and though, as in other gulls, the circumstances in Snow 1967). which they are performed are varied and it is The Swallow-tailed Gull pairs at the nest not possible to present a neatly cut-and-dried site and does not use “clubs,” as do many other scheme. gulls. There is little doubt that this practice The Rattle and Whistle. This is the only call is related to the shortage of good nest sites. described by Moynihan ( 1962). (He also de- The male first acquires and defends a site, and scribed some variations such as the “Buzz” here one or more females will visit him and go and “Peep,” which he heard from nonbreeding through a form of greeting ceremony. On birds at sea. ) The call consists of a rattling South Plaza Island good nest sites are not trwr, the last part of which is audible only superabundant, and many pairs failed because at close range, followed by a long, loud they had a poor site. For instance, a consider- heeeeeuuuuuu. The second part of the call, able proportion of the losses (the percentage lasting 2-3 seconds, is more similar to the call varying from month to month) was due to the of a shorebird than a gull. The whole seems site being washed by heavy spray (Snow and to be quite unlike any call uttered by any Snow 1967). So, presumably, a female is other gull. The Rattle and Whistle was never selecting not just a mate but a mate with a used during intraspecific encounters, but solely good nest site. Although few fights were wit- as an alarm call at the approach of man or nessed, many males that were known to be frigatebirds. It was heard once or twice at taking up nest sites were seen with large scars night, presumably in response to the owl across their heads consisting of displaced and Asio flammeus galapagoensis, which regularly broken feathers. It is possible that the fights visited the breeding colony and was known to were seldom seen because they took place be- have taken some chicks. One of the striking fore dawn when the gulls first return to their features of the call is that the bird always sites after the nights’ fishing. turns its head from side to side while calling. As mentioned above, eggs are laid in all It is a conspicuous and deliberate movement 254 BARBARA K. SNOW AND D. W. SNOW

The alarm Rattle and Whistle is contagious, and when one bird calls many birds nearby will give the call, some without seeing the cause of alarm, Creagrus does not readily take flight, and some birds on the nest may be approached closely enough to be caught by hand or with a small net. At sections of the cliff where we did much catching, the birds quickly became less tame, but if one approached an undisturbed section the whole colony would give the alarm Rattle and Whistle without any birds flying or moving off their nests. The only movement would be from older chicks, not being brooded, which would take cover under the nearest rock or crouch against a rock face. Throughout the day there are frigatebirds, Fregata mugnificens and F. minor, over the Plaza Islands, patrolling up and down the cliffs 60 to 100 feet above the island. At this height the patrolling frigatebirds are ignored by the gulls, but the moment one of them swoops low over the colony, which they can do at great speed, the gulls at that colony utter the Rattle and Whistle and any exposed chicks hide. The Rattle and Whistle is the loudest and most frequently used call of Creagrus. It is also the only call that is completely unlike a FIGURE 1. Top: the Rattle and Whistle (from a photo); middle, left: Downward Piping (from a field gulls’ call. The fact that it is used exclusively sketch); middle, right: The Upward Jerk (from a as an alarm call and never in intraspecific en- photo ) ; bottom: the Head-up Tail-up posture; two counters suggests that it was evolved to meet birds in the Drocess of forming a Dair, the female a special situation when the predecessors of (right) showing the posture in &ore-pr&ounced degree (from a photo). Creagmcs first colonized the Galapagos. It seems that the presence of frigatebirds in Galapagos may have been the critical factor. Frigatebirds do not breed in the Humboldt through an arc of about 130”. Exactly the Current region of South America, the most same call and movement are made whether likely place of origin of the ancestors of the bird is sitting on its nest or standing. Creagrus; nor in fact do they seem to breed When both birds of a pair are at the nest, the in close proximity to gulls anywhere in the male does the greater part of the calling. The world, except in the Galapagos and parts of call is also uttered in flight, and then the head the Pacific coast and offshore islands of is not turned but the head and neck are Mexico. stretched obliquely upward, and the bird flies It is probably significant that the three with deep slow wing beats. Both the head- other loud calls of Creagrus, the Downward turning and the flight-display make the call- Piping, Upward Mew, and Departure Call, are ing bird visually conspicuous, and this is pre- all characterized by starting with a “con- sumably important when, as often happens, the seas are rough on the windward side of the sonant” and consisting of short staccato notes islands where Creugru-s nests and the call is frequently repeated. They are easily distin- not easily audible. guished from the long whistle of the alarm The typical position of a bird uttering the call, thus making the latter even more distinc- Rattle and Whistle is shown in figure 1 (top). tive and therefore effective as an alarm. If one makes a closer approach, the bird be- The Departure Call. This call, heard only on gins to rotate on the spot, turning round with the wing, is a loud four- or five-syllabled kew- little steps and at the same time raising the you-you-you. While the bird is calling the posterior part of the body. The uptilting of head and neck are stretched up and the wing the tail appears to be a flight-intention move- beats are slower than in normal flight. We ment and is a component of other displays. could see no difference in the appearance of BEHAVIOR OF THE SWALLOW-TAILED GULL 255 birds uttering this call or the alarm call in It is used as a greeting between members of flight. With one exception, the Departure Call a pair, being performed by the bird that has was heard only in the evening from birds fly- remained at the nest and often also by the re- ing off for the nights’ fishing. turning bird. When the latter is flying back, The Departure Call is the third loudest of it may start the piping while still in the air. the Swallow-tailed Gulls calls, and it is also When collecting stones for nest-building, the quite distinct from its other calls. Hence one male often does the Downward Piping while would expect to find it fulfilling some impor- walking up to the nest with a stone in his tant social function; but during the breeding beak. This call is also used when a parent season it does not obviously do so. Only about feeds a chick or the male feeds the female. one quarter of the birds give the call when The bird first regurgitates the food into the they are leaving, and then it is often given beak and, while holding it, does the Down- when the bird is out of earshot of its mate; ward Piping. hence it does not appear to function as a When a colony resettles after the alarm at- signal of departure between members of a tending an intrusion, many birds do the Down- pair. Neither does it function as a synchroniz- ward Piping, and when uttered thus by a par- ing signal enabling birds to leave as a flock ent bird it is the signal for its chick to come for the nights’ fishing, as they normally leave out of hiding and return to the nest. Often in singly or in twos or threes over a period of encounters between members of a pair and nearly an hour. However, on one threeday between a parent and its chick, the full Down- visit to the island during May a flock of 20-30 ward Piping is not uttered but only the muted birds collected each evening for an hour or eerk eerk or the creaking trr trr. more before departure, and these birds left in Finally, when a male first mounts the female bigger groups, up to 15 at a time, and most he may give some Downward Piping calls, gave the Departure Call as they flew up. which sometimes continue during cloaca1 con- This suggests that the Departure Calls’ main tact. function is related to flocking behavior. Only Possibly Downward Piping is homologous once, as described above, was flocking seen at with the Mew Call and Arched Posture of the breeding colony, although during four other gulls (Tinbergen 1962). The call itself months of the 9- or lo-month cycle when the is very different, being staccato and not long- birds are not breeding they are known to form drawn-out, but the position of the body, par- flocks off the coast of Peru (Murphy 1936). ticularly of the neck, shows similarities to the Downward Piping. Downward Piping is typ- Arched Posture of Larus argentatus (Tin- ically performed with the body horizontal, the bergen 1953). The body is more horizontal in neck curved forward and downward, and the Creagmcs, but this may be related to its rather beak pointing vertically to the ground and temlike proportions, with long wings and tail nearly touching it (fig. 1, middle left). The and short legs; in all its postures the body is call is a loud, rapid kweek kweek kweek, the either horizontal or the rear end is uptilted. beak opening and closing with each kweek. A rare preliminary to Downward Piping Frequently after about 8 to 10 kweeks the was the Upward Mew. It is accompanied by beak nearly closes, and the call becomes muted a call kew kew, slower than Downward Piping to an eerk eerk eerk, and then to a creaking and with a distinct silence between each call. trr trr trr. As will be mentioned later, Down- From the posture adopted, it seems that it ward Piping frequently follows three display may well be homologous with the Long Call movements in which the head is pointed up- of other Laridae, but unlike these it was not ward (Upward Mew, Upward Jerk, and Head- seen to be used at all in male advertising. On forward Rattle), and it seems that the curving the five occasions when it was observed, it was down of the head is an integral and significant followed by Downward Piping. It was per- part of the display. formed either by the paired male alone or by After the Rattle and Whistle, Downward the pair in unison. On one occasion, during Piping is the loudest and most frequently used courtship feeding, a male regurgitated a lump call, but surprisingly, with one possible excep- of fish and then held it while performing the tion, it was never heard from unmated males. Upward Mew. However, once the pair is formed, Downward The rareness of the Upward Mew suggests Piping is performed frequently by the male or that it is being lost to the species, presumably the female, or by both in unison. It is com- because there is no necessity for a loud vocal monly performed after border disputes with accompaniment to male advertising. neighbors, when it may be done at the nest The Upward Jerk. In this movement (fig. 1, site or at the borders of the territory. middle right) the body of the bird is horizon- 256 BARBARA K. SNOW AND D. W. SNOW tal with the wings folded, the neck is stretched with the Upward Choking of Kittiwakes upward, and the closed beak is also pointed (I&a triductyla); and in this species Choking upward. From this position the beak is jerked is the advertisement display of males, so that slightly upward and backward five or six the contexts are similar (Cullen 1957). times. It is occasionally performed by a bird In origin Upward Jerking looks like a ritu- sitting on the nest site. At a range of a few alization of the jerking movement made by a feet a few low, creaking calls may be heard to gull or cormorant when swallowing a fish. accompany the movement. These calls are not This swallowing movement was observed fairly synchronized with the jerks but are of a slower frequently in Creagmcs, as during courtship tempo. As the majority of observations were feeding the male normally reswallows the food made beyond the range of audibility of the (fish or squid) which he has regurgitated, and call, we do not know how frequent it was, but again when a small chick is being fed the certainly sometimes the movement was silent. same fish may be regurgitated and reswal- The Upward Jerk is a common display, and lowed many times until the whole of the head the contexts of 70 of them were recorded. Of end is digested by the parent and the fish be- these, 50 were performed by males alone, 14 comes a suitable size for the chick to eat. by females alone, and 6 by pairs in unison. All The Head-up Tail-up Posture and Facing but 14 of the total were performed at the nest Away. As mentioned in the description of the site; occasionally unpaired males were seen to Rattle and Whistle, the greater the fear of a do it at two potential nest sites within the bird and its tendency to flee, the more its tail territory. Where the nest site is under an over- and body are uptilted. This flight-intention hanging rock or largely hidden because sur- position has apparently become stylized into rounded by rocks, the Upward Jerk is per- the Head-up Tail-up posture (fig. 1, bottom), formed from some vantage point on top of a which is frequently seen during courtship and rock near the site. Thirteen of the 14 remain- occasionally when birds are paired. ing records were from such vantage points. When a female lands in the territory of an The Upward Jerk is nearly always termi- unpaired male she orientates laterally and as- nated by bringing the head back through an sumes the Head-up Tail-up posture. The male arc of 160” so that the beak is pointing down- does likewise, then either just the female or ward. In one quarter of the detailed records both birds run up and down or round each the head was brought down to perform Down- other in this posture while still maintaining ward Piping, and in most of the remainder the lateral orientation. The type of running is Foot-watching followed; occasionally the bird characteristic of the posture and consists of simply nibbled at stones or sat on the nest and small, rapid steps. Alternatively, the male may shaped it, and males sometimes performed the run to the nest site in the Head-up Tail-up Regurgitation Bow. posture, and the female may then follow in The Upward Jerk is one of the main adver- the same manner and stand laterally orien- tisement displays of unmated males. In this tated beside him. In this position both will context, when a female lands somewhere frequently Face Away, the female usually nearby, the male usually runs to the nest site more often than the male. The Facing Away and gives the display. It is also frequent after appears to be the typical movement of other territorial disputes with neighbors, which may hooded gulls. terminate with the male going to the nest site It is noticeable that the uptilting of the and performing the Upward Jerk. When a females’ body and tail is often more acute than pair is formed the female may do so as well. the males,’ which may be only slightly above During nest-building, which is done mainly the horizontal. The females’ body usually be- by the male but occasionally by the female, comes more horizontal the longer she remains either sex after adding a new stone to the nest in the males’ territory, presumably reflecting usually does the Upward Jerk. a decrease in her escape tendency. The Upward Jerk is not obviously homolo- During the prolonged courtship of a par- gous with any of the described displays of ticularly nervous pair, the male was seen to other gulls, but it does show similarities of Face Away while his beak was open at the posture with the Throw-back component of beginning of the Regurgitation BOW, and the the Long Call, particularly as described for female to intersperse Facing Away with Head- Hemprichs’ Gull (Larus hem@&) by Fogden tossing. After pair formation either sex may (I=4), with the obvious difference that occasionally Face Away, usually at the return Creagrus has the beak closed and consequently nf a partner who has been briefly absent. makes a very reduced call. It also shows simi- The Landing Gape. The Landing Gape (fig. larities, particularly in the head movement, 2, upper left) is a silent movement performed BEHAVIOR OF THE SWALLOW-TAILED GULL 257

During courtship or just after pairing, one or both members of a pair will run about their territory and up to the nest site in the Breast- lowered position. It largely replaces running about in the Head-up Tail-up posture, which is characteristic of the earliest stages of courtship. It is especially common late in the evening, before the birds leave the nest site to go fishing. Occasionally one or both birds of a courting pair stand in the Breast-lowered position at the nest site and swing slightly from side to side. The position is fairly certainly homologous with the Choking of other gulls, although it is silent and without any rhythmic head movement. If the rhythmic head movement of \ - _/ other gulls is derived from shaking nest ma- terial out of the mouth (Tinbergen, personal , /!!lJt* communication), then its absence in Creagrus, FIGURE 2. Upper left: the Landing Gape (from a which builds only with stones, is not surprising. field sketch); middle: the Breast-lowered position The Regurgitation Bow. The Regurgitation (from a field sketch); lower left: the Regurgitation Bow (fig. 2, lower left) shows considerable Bow, initial position (from a photo). variation in the extent of the bowing and in the position of the body, but great consistency in that the gape is opened and the tongue is by a bird landing from flight or from a hop arched in just the same way as when food is forward, or at the end of a run forward. The regurgitated to the chick or during courtship beak is briefly opened and closed as the bird feeding. lands or comes to a halt, the position of the Typically in the early stages of courtship an head and body being similar to that of the unpaired male, who has assumed the Head-up Head-up Tail-up posture. It is performed al- Tail-up posture at the arrival of a female at most exclusively within a birds’ territory, or in the edge of his territory, then opens his beak the case of an unpaired female, in the territory and arches his tongue, bows down, and returns of a prospective mate, so that the nearest bird to his original position in approximately one is the mate or prospective mate, and as would second. In between the bows the beak may be expected there is always lateral orientation. be closed or may remain open. The bow may The Landing Gape is seen most frequently be repeated many times, especially when the in the early stages of courtship, particularly female is nervous and does not readily come when a female is visiting the territory of an any closer. The display is attractive to the unpaired male; she will terminate every hop female, who will usually in time go nearer and and movement with a gape, especially if she is may join in and perform the Regurgitation nervous. But throughout the nesting cycle a Bow just after the male. Occasionally the gap- bird will usually do the Landing Gape when ing is accompanied by the production of a it lands in its territory beside its mate. little mucus. Sometimes a nervous bird, male Showing the gape is also a component of or female, will Face Away between bows, or one of the main aggressive displays, but in the may even Face Away while the gape is open Landing Gape, when it is combined with the and the tongue arched. A more frequent vari- ritualized flight intention of the Head-up ation, which appeared to be correlated with Tail-up posture and lateral orientation, the a more advanced stage of courtship, was a conciliatory elements predominate. Regurgitation Bow from the Breast-lowered The Breast-lowered position. In the Breast- position; it was always seen at the nest site, lowered position (fig. 2, right) the breast is performed by the male only. The birds body held just clear of the ground while the remainder of the body tilts obliquely upwards and is held in the Breast-lowered position and the the head and neck form an S with the beak head is then bowed, while at the same time pointing downward. A similar posture is either the head or the head and breast are adopted by a bird that is just about to lower swung from side to side through an arc of one itself onto the nest in order to incubate or to or two inches. This type of Regurgitation Bow shape the nest. No sound was heard to accom- sometimes develops into courtship feeding. pany the movement. While the male is doing this, the females’ 258 BARBARA K. SNOW AND D. W. SNOW behavior may vary. She may Head-toss in the Breast-lowered position, at the same time swinging the breast from side to side, or she may swing from side to side in the Breast- lowered position but with the beak closed and no special head movement; or she may pick up and put down pebbles at the nest. When the Regurgitation Bow is performed in the Breast-lowered position, the combination much resembles the advertisement Chok- ing of the male Kittiwake ( Cullen 1957), except that it lacks the rhythmic action and‘ is silent. In Creagrus, as in the Kittiwake, the brightly colored gape is used in hostile displays, but with the arching of the tongue, which accompanies the Regurgitation Bow of Creagrus and the Choking of the Kittiwake, the amount of colored gape that is visible is reduced, and this may be of importance in minimizing the aggressive aspect of the display. Courtship Feeding. In the form of Court- ship Feeding characteristic of the early stages of pair formation, the male proceeds from the Regurgitation Bow to the actual regurgitation of food into his beak. The bow is followed by swelling of the neck and by throat movements, until a fish or squid is brought up into the beak. Often the male regurgitates the food and reswallows it several times, while the female is still some way off and hesitant in approaching. If the first regurgitation is away from the nest site, later ones are usually nearer to or at the site. At a slightly later stage in pair formation the male does the Downward Piping while holding the regurgitated food in his beak, in the same way as parents call the chick for food. In most of the observations of Courtship FIGURE 3. Top: Head-tossing; middle: mutual preening; bottom: Foot-watching ( all from photos ). Feeding the female got no more than a token taste of the food by pecking at it while it was held in the males’ beak. Often he turned away and reswallowed it before she received any. was recorded only from pairs or courting birds The few more substantial feedings observed and was never seen to occur in encounters were given to females within about three between males or in hostile encounters be- weeks before laying. Probably many more tween neighboring pairs. substantial Courtship Feedings took place The male, when Head-tossing, has his body when the birds first returned from fishing, be- parallel to the ground and usually raises his head only halfway to the vertical. He often fore it was light. Hailman (196413) states that does it over the females’ back, so that after chicks are fed from midnight onward. We each toss his beak rests briefly in the feathers noticed that frigatebirds were quick to take of her back. The male may initiate Head-toss- advantage of daylight feedings to chicks at ing, but more frequently the female starts and exposed nests, so they probably also take ad- the male joins in, and they then toss their vantage of Courtship Feeding in daylight. heads alternately. This mutual Head-tossing Head-tossing of adults. The Head-tossing of is almost always the preliminary to mounting Creagrus (fig. 3, top) is basically similar to and continues for several minutes with an in- the movement in other gulls. Each movement crease of tempo (the fastest recorded was is usually accompanied by a thin plaintive 6 tosses in 9 seconds) until the male mounts. kew, but sometimes it is silent. Head-tossing The female does two different forms of BEHAVIOR OF THE SWALLOW-TAILED GULL 259

Head-tossing. When she does it as an invita- appeared to culminate in cloaca1 contact. All tion to copulation she stands, usually just in took place within the pairs’ territory and most front of the male, with her body parallel to actually on the nest site (not on neutral the ground and raises her head to a vertical ground away from the nesting cliff, as sug- position with each toss. While her head is gested by Hailman [ 19651). All but four of raised she frequently pecks at the males’ the mountings occurred within an hour before beak, usually at the white tip, but sometimes or after sunset. she grasps the middle of his beak. If she is The normal sequence of mating, which was facing the same way as the male, she twists always preceded by mutual Head-tossing, was her head first over one shoulder then over the for the male to hop onto the females’ back and other in order to reach the males’ beak. She stand between her shoulders, sometimes pad- often continues to Head-toss while the male dling with his feet and waving his wings to is mounted. aid balancing. When attempting cloaca1 con- When a female Head-tosses to beg for food, tact, the male rhythmically beats his wings she keeps her head lower and seldom pecks and moves to a more posterior position on the at the males’ beak; sometimes she swings her females’ back; here he rests on his tarsi and body from side to side. If the male mounts moves his closed tail from side to side, then she dislodges him by pecking at his feet and fans his tail and depresses it so that it presses continues to Head-toss with her head at a still against the ventral surface of his mate. The lower position. She may even utter some kew female raises her tail over her back and rhyth- calls without any head movement. The food- mically opens and closes the cloaca as contact begging Head-tossing was always accompa- is achieved. Several (usually 4-7, maximum nied by the call, but the premating Head-toss- recorded, 8) cloaca1 contacts are attempted at ing was sometimes silent. each mounting. Between each contact the In the Herring Gull (Lurus urgent&us) the male moves forward to a standing position females’ Head-tossing for food is indistinguish- between his mates’ shoulders. Usually on first able from her Head-tossing as an invitation mounting the male does some Downward Pip- to copulation (Tinbergen 1953). The fact that ing and sometimes he calls throughout mount- they differ in Creagrus suggests that they may ing, but at other times he is entirely silent. be functionally more distinct. Creugrus lays a The female stands throughout, although some- particularly large egg for its size (average of times a mounting is unsuccessful because she 401 eggs, 76.9 g). It flies to distant feeding overbalances. She usually Head-tosses intermit- grounds. In addition it was found that before tently while the male is mounted. the egg is laid, in one instance 11 days before, When the male dismounts, both birds are one of the pair stays behind at night, appar- silent and usually preen themselves for a few ently to guard the nest. It was not discovered minutes. This may be followed by allopreen- how the sexes share this duty, but the female ing or by mutual preening. must sometimes stay behind. These consider- The Threat-gape and the Upright Threat ations suggest that the female may need sup- posture. The aggressive displays described in plementary feeding while her egg is develop- this and the following section are functionally ing, and hence that courtship feeding may distinct from the behavior dealt with in the have real nutritional importance. previous sections. The latter were concerned Mutual preening. When a pair has been with the relations between members of a pair formed, both the members may preen each and form a more-or-less orderly sequence other (fig. 3, middle). Sometimes both birds leading from the beginning of pair formation preen each other at the same time, but more to copulation. The aggressive displays con- often one bird preens the other. The preening cern the relations between rival or neighbor- is usually confined to the head or chin, but ing territory owners. It may be recalled that occasionally other parts of the dorsal surface nests may be as little as 18 inches apart, are preened. This behavior was only rarely though more often they are a few feet apart. seen after the egg was laid. Thus neighbors may be in close contact with At a nest where it was possible to erect a each other. blind a few feet away, the male was heard to In the Threat-gape, which is silent, the feath- utter a soft call, trm trrr, while preening his ers of the head and neck are erected, the mate; but most observations were made at too head is thrust forward and the open gape is great a range for this call to have been audible, pointed directly at the object of hostility. Un- and so it is not known how general it is. like the Landing Gape, in which the beak is Mounting and copulation. A total of 25 briefly opened and then closed, the gape may mountings were recorded, about half of which be displayed for as long as a minute or more. 260 BARBARA K. SNOW AND D. W. SNOW

raised and the antagonists make little jumps into the air, at the same time pecking forward at the others’ bill. This stage of aggression was reached quite frequently without a fight ensuing; but it is probably the normal preliminary to a fight. In another, less aggressive form of the Threat-gape the head is lowered (fig. 4, top). It is also done by males facing each other at the borders of their territories, and the posture may be held for long periods, sometimes with the beak open and sometimes closed, but at all times with the eyes directed forward. In appearance it looks similar to the extreme defensive posture of the Herring Gull (Tin- bergen 1953), but in Creagrus it is adopted by both contending birds, and when one changes to the more aggressive Upright Threat- gape the other immediately follows suit. The Head-forward Rattle. In this display (fig. 4, bottom) the tail is raised, the head is also raised and thrust forward, the beak is slightly opened, and the bird makes a deep rolling rattle or growl, grrrrrk grrrrrk. Each rattle continues for about two seconds, during which time the throat and sides of the neck are distended, but the whole display may continue for several minutes with a long series FIGURE 4. Top: the Threat-gape with head low- of rattles. The males’ call is perceptibly lower ered (from a field sketch); bottom: the Head-for- pitched than the females,’ but neither is audi- ward Rattle (facing bird); two males at the boundary ble much beyond 10 yards. The display is of their territories (from a photo). usually performed in a standing position, with the head stretched toward the object of hostility, but it may also be done while sitting on When a strange bird intrudes on the fringes the nest. of a well-established territory, the owner of The Head-forward Rattle appears to be an the territory or his mate will usually run or fly exclusively hostile display used predominantly at the intruder doing the Threat-gape, but by pairs with established territories and mates. occasionally they may stand and gape, or if During the initial establishment of territo- the intruder is below them on the cliff they ries, males normally use the Upright Threat- may lean down and Threat-gape from above. gape, but a week or so later, when territories A static form of the Threat-gape which em- are fixed and the birds are paired, the same bodies an Upright Threat posture is frequent individuals, often accompanied by their mates, between males during the initial disputes over will do the Head-forward Rattle at any territories, particularly on crowded ledges stranger that lands on the outskirts of their where nest sites are close. Typically two males, colony. each on its own territory, stand facing each As an illustration of the relationship be- other a few inches apart with the gapes open tween the two main hostile displays, it was ob- wide, the carpal joints raised, and the feathers served that when one of two males, which for of the head and neck erected. The body is a minute or more had been doing the Upright held parallel to the ground and not in the Threat-gape at each other, was joined by its oblique position of other gulls. Another strik- mate, the solitary male rotated away from his ing feature of the Upright Threat is the forward rotating of the eyes to produce binocular antagonist, who immediately changed from vision directed straight at the rival. the Upright Threat-gape to the Head-forward Often this Upright Threat posture will be Rattle. held for minutes on end; occasionally the beak Foot-watching. Foot-watching, as in other is closed while all other features of the posture gulls, is performed either as a display or ap- are maintained. If there is an increase of ag- parently autochthonously, as a response to gression, the wings and tail are opened and some irritation affecting the leg or foot. The BEHAVIOR OF THE SWALLOW-TAILED GULL 261 body is held in the normal resting posture, frequently by the male, consists of pressing and the head and neck are curved downward the breast into the ground with the tail cocked (fig. 3, bottom). almost vertically upward, and at the same As a display, Foot-watching terminates time scraping backward with the feet. This other more “emotionally intense” postures. movement has no effect on the many nest sites Most commonly (in 29 of the 67 recorded situated on rock, but where there is soil a instances) it followed the Upward Jerk, but it slight depression is made. followed a variety of other displays as well, Stones for the nest are collected either from some of them undoubtedly aggressive such near the nest or by flying outside the territory, as the Threat-gape and Head-forward Rattle, by both sexes but mainly by the male. Males and others that are not aggressive such as do not usually start building until they are Head-tossing, the Head-up Tail-up posture, paired, but two unpaired males with poor and displacement preening. In all cases ob- sites made small collections of stones at more served, Foot-watching was followed by stand- than one place in their territories. ing in the resting posture or by displacement Nest-shaping, as already described, contin- preening of the scapulars. ues as the stones are collected. A well-ad- A more exaggerated form of Foot-watching vanced nest is also shaped by the bird sitting was occasionally observed. The bird looks and slowly rotating on it, and at the same time backward through its legs, and may hold the stretching out to pick a stone from the perim- position for several seconds. This was some- eter of the nest and to place it beside its flank. times done by females, who occasionally Foot- An incubating bird will maintain the structure watch in unison with their mates, on occasions of the nest by taking stones from the perim- when they had not performed any previous eter and placing them in the cup. During in- movement, and on two occasions it was per- cubation more stones may be added, usually formed by males after prolonged hostile en- at the time of nest relief. counters with neighbors. There seems to be Parental care. Both sexes incubate, but as little doubt that Foot-watching is an appease- incubation behavior was not studied in detail, ment display and that the exaggerated form the share of the sexes and the lengths of the indicates a greater degree of fear. attentive periods were not ascertained. It is Autochthonous Foot-watching is frequently noteworthy that in spite of its single-egg seen in birds standing idle, apart from their clutch, Craagrus has two lateral incubation fellows, and in such circumstances it is often patches. The most probable explanation is followed by the shaking of one foot, or of that its ancestor once laid a clutch of three both, one after the other. The stimulus-per- eggs, like most other gulls, and had three in- haps irritation by mites or some other small cubation patches, one median and two lateral. invertebrate-was not discovered. When the clutch was reduced to two, the median patch was lost, but when the clutch was NESTING BEHAVIOR further reduced to one egg, no further reduc- The nest and nest-building. The nest of the tion of the incubation patches was possible. Swallow-tailed Gull is typically composed of Soon after hatching, while the chick is still a substantial pile of small stones (of the order damp, the egg shell disappears. The removal of 399), with a well-shaped cup somewhere was not witnessed, so it could have been done near the center. A big nest measures 15 inches by the ubiquitous crab, Grapes grapsus, across, but many were much smaller, and a which quickly removes any scraps of food few eggs (all failing) were laid on the bare dropped by the gulls. Up to at least seven rock. Where the nest is on a slight slope it is days old the chick stays in the nest, and one built up with extra stones on the downhill side parent either broods it or shades it from the in compensation. One pair that nested on a sun, according to its need. At this age the fairly steep slope collected many sticks as well chick does not take cover at the Alarm Rattle, as stones, and a few of the sticks were incorpo- but remains in the nest covered by the parent, rated into the nest. As mentioned earlier, the male acquires and who defends it. A week or so later the chick defends a territory, where by the Upward takes cover at an alarm, but one of its parents Jerk display he indicates to the female the stays with it day and night until it is 21 days potential nest site. Most territories contained old or more. only one site of optimum suitability, but a few Although we paid hundreds of visits to nests contained two possible sites, and in such cases with chicks, we saw only two cases of aerial some nest-building was sometimes done at attack from a parent-the swoop and strike both. typical of many other Laridae. Nor was there The preliminary nest-shaping, done most any distinctive call associated with the attacks, 262 BARBARA K. SNOW AND D. W. SNOW although once a short, high-pitched note was What were almost certainly the first short uttered by the attacking bird. flights of a number of young birds were ob- There were insufficient observations to show served at ages of 58-65 days. any difference in the share taken by the sexes in feeding the chick. When the chick is only DISCUSSION a few days old, the parent regurgitates the The foregoing account has aimed at straight- fish or squid onto the ground and picks small forward description of the behavior of Crea- pieces up which the chick takes from its beak. gms at its breeding colonies. This bird differs The chick may also pick pieces up from the from typical gulls as much in its behavior as ground. Later, a whole fish or a fish with the it does in its morphology, and we have not head digested is passed directly from the attempted a detailed comparison of its dis- parents’ to the chicks beak. plays with those of Larus species. It may, Voice and food-begging of th.e chick. The however, be pointed out that the evidence very young chick has a cheeping call that it from behavior strongly suggests that any uses when in need of brooding or in discom- morphological resemblance between Creagrus fort. It also uses this cheep when it is hungry. and Sabines’ Gull, Xemu sabini, is convergent. The early food-begging behavior consists of Their displays as well as their breeding biol- this cheep accompanied by pecking at the ogy are utterly dissimilar (Tinbergen, personal parents’ beak and sometimes at the red eye- communication; R. G. B. Brown, in litt.). ring and the white feathers at the base of the Lack (1967) has recently drawn attention beak. If the parent responds by opening the to the need to consider all aspects of the beak, which it often does, the chick then pulls breeding biology of seabirds as interrelated at its red tongue. By the age of three weeks parts of an adaptive whole; no feature can the food-begging call is a squeak accompa- properly be understood in isolation from the nied by occasional indefinite Head-tossing others. Tinbergen (1967) further illustrated movements, and the chick also often takes its the same theme with examples drawn from parents’ white beak-tip in its mandibles. When the Black-headed Gull, Lams ridibundus. To it is fed a fish or half a fish, a three-week-old what extent can the various features of the chick is able to toss and turn it in order to breeding biology of Creagrus be viewed as swallow it tail last. interrelated parts of a single adaptive system, When the chick is feathered at about 45 days and, more particularly, to what extent can the old, the food-begging is more ritualized. The form of the displays and the other behavior chick stands in the hunched, submissive pos- patterns dealt with here be causally related ture with its head lower than its back and to other aspects of the biology of the species? calls kew, kew, similar to the adults’ Head- A complete answer is not possible at this stage, tossing call except that it is more husky. It and discussion will be confined to a number of makes many calls with only tiny upward points that appear to us to be significant. movements of the beak, but periodically the The Galapagos Islands apparently do not head is raised higher in a pronounced Head- provide very favorable living conditions for tossing movement. Once or twice chicks of a typical gull. The shores are mainly barren this age were seen to take the parents’ beak in lava, with few beaches where detritus collects their mandibles, but it was unusual and evi- and a complete absence of anything in the dently no longer an essential part of food- nature of an estuary. This kind of coast prob- begging. Often these older chicks begged for ably affords a poor food supply for a typical as much as an hour during the day without gull, as is suggested by the fact that Lams receiving any food. fuliginosus, a coastal scavenger, is very sparsely Other activities of the chick. The chick be- distributed there. Creagrus, the only other gins to preen itself when as little as six days breeding gull, is abundant but feeds far out at old, and also begins to use the typical stretch- sea. Two other gull species occur occasionally, ing movements of gulls, such as the stretching in very small numbers (tiveque, Bowman, back of the wing and leg on the same side. and Billeb 1966). By this age, the chick does not defecate hap- If, as seems likely, the ancestor of Creagrus hazardly when in the nest, but backs to the was a plunge-diver, then in its search for good edge and ejects its feces clear of it. Chicks feeding areas it may well have had to range under 40 days of age were not observed to do fairly widely, with the consequent necessity wing exercises, although they may have done for long flights between feeding grounds and so at night. But between 40 and 45 days nesting colonies. Whether frigatebirds were (about two to three weeks before their first or were not present when the ancestral Crea- flight) occasional wing-exercising was seen. grus was establishing itself on the Galapagos BEHAVIOR OF THE SWALLOW-TAILED GULL 263

is of course impossible to say; but in any case discussed some of those connected with noc- it seems certain that it was the presence of turnal behavior and cliff-nesting. Courtship frigatebirds that forced it-certainly an off- behavior too has been molded by the special shore and perhaps already a pelagic feeder- ecological conditions under which the Swal- to become nocturnal, as Hailman ( 1964b) has low-tailed Gull exists. also suggested. Even the strong-flying boo- A striking behavioral attribute of Creagrus bies suffer badly from piracy from frigatebirds, is the absence of a loud male advertising call, and a day-flying pelagic gull would probably which is characteristic of many ground-nesting find life intolerable. This may be one of the (as opposed to cliff-nesting) gulls. The Kitti- reasons, as suggested earlier, why the breed- wake, also a cliff nester, lacks such a call too. ing distributions of gulls and frigatebirds are, The unpaired Kittiwake advertises by a Chok- with very few exceptions, mutually exclusive. ing display (Cullen 1957), which may be Terns, on the other hand, coexist extensively accompanied by a variety of calls; but these with frigatebirds, presumably because they are not the loudest of its calls and appear not are too small to be worth a frigatebirds’ atten- to be a significant part of the display (Cullen, tion. personal communication). The unpaired It is also likely that the adoption of cliff- Swallow-tailed Gull advertises by the Upward nesting habits resulted from predation and Jerk and the Regurgitation Bow, both of piracy by frigatebirds at the nesting grounds, which are silent or accompanied by a call and perhaps also from predation by Asia audible for only a few yards. In both the flammew, for there are no native mammalian Swallow-tailed Gull and the Kittiwake the predators and no other apparent reason why male acquires and defends a nest site be- Creagrus should not nest on the flat tops of fore pairing, and as there is a shortage of good islands. In fact, it appears that the chief sites, a male who is in possession of one has requirement is that the nest should be near a thereby proved his fitness as a mate. All he boulder or crevice, where the chick can hide, needs to do is demonstrate the site, which is and this requirement is best met on the broken what both the Swallow-tailed Gull and the cliffs. We found that nests on the cliff tops Kittiwake do. On the other hand the males were invariably unsuccessful unless there were of ground-nesting gulls, in which pair forma- hiding places nearby, and nests on the cliff tion takes place at a “club,” have to advertise itself were more successful if they were pro- themselves only, and therefore use displays, tected by an overhang than if they were ex- such as the Oblique posture and the Long posed. Call, which provide the maximum of visual Colonial breeding and the consequent syn- and auditory stimulus. chronization of nesting activities within groups In Creugrus the early stage of courtship, (Nelson 1967) may also be adaptations against which is mainly silent, is followed after pair- predation, since the concerted alarm of the ing by a vocal period when Downward Pip- adult gulls at the approach of a frigatebird ing is frequent. This advertises the paired must be a more effective warning the more state of the birds concerned, and may be ad- birds there are on watch. But another, facili- vantageous in reducing interference and abor- tating factor is probably involved. Lack (1967) tive attempts at pairing by unpaired females. has pointed out that there is a general con- The Swallow-tailed Gulls’ long wings and nection in seabirds between colonial breeding tail are presumably an adaptation to pelagic and offshore or pelagic feeding. If the feeding habits, and its short tarsus is probably corre- and breeding areas are far apart, the extra lated with the fact that it perches mainly on distance to be flown if such birds nest in colo- cliffs, and never walks. (The Kittiwake also nies instead of singly is negligible, especially has a relatively short tarsus for a gull.) The as they are bound to be strong fliers. Thus combination of long wings and tail with a pelagic feeding habits probably facilitate co- short tarsus results in a more temlike stance lonial nesting, if there are advantages to be than is usual for a gull, and facilitates the gained from it for other reasons. raising of the posterior half of the body. Up- Pelagic feeding, cliff-nesting in colonies, right body postures are correspondingly less and nocturnal habits, all of which, it is argued, easy to maintain. This probably explains why resulted primarily from a combination of poor in many of the displays the posterior half of inshore feeding conditions with the presence the body is elevated, whether the birds are of frigatebirds, must themselves have entailed aggressive (e.g., Head-forward Rattle) or con- a complex of subsidiary adaptations. A ciliatory (e.g., Head-up Tail-up, Regurgitation thorough treatment of all these adaptations is Bow), and why there is no equivalent of the not yet possible; Hailman (1964b, 1965) has Oblique posture of other gulls. 264 BARBARA K. SNOW AND D. W. SNOW

SUMMARY ritualized preliminary to courtship feeding; A breeding colony of Swallow-tailed Gulls, Courtship Feeding, which is most frequent Creagrus furcatus, was studied on South Plaza in the three weeks before egg-laying; and Island in the Galapagos Archipelago from Head-tossing, which is the usual preliminary February 1963 to April 1964, and supplemen- to copulation. tary observations were made on Tower Island Purely aggressive displays, connected with in May 1964. the acquisition and defense of territories, in- Pair formation takes place at the nest site, clude a silent Threat-gape and the Head-for- which is acquired by the male. This is in con- ward Rattle. Foot-watching occurs both as a trast to pair formation at “clubs,” which occurs display and autochthonously, as in other gulls. in typical gulls, and probably results from the Nest-building, parental care, and the devel- shortage of suitable sites. opment of the chick are briefly described. All marked males returned to the sites that It is suggested that pelagic feeding, colonial they had used previously, but some females cliff-nesting, and nocturnal habits resulted changed their site. from a combination of poor inshore feeding Calls and displays are described, and are conditions with the presence of frigatebirds. classified as follows: The Rattle and Whistle, These in turn have led to a complex of subsid- an unusual and unique call, used only in re- iary adaptations, and courtship and associated sponse to alarm from potential predators; the behavior have been profoundly affected in Departure Call, used only when birds are various ways. leaving the colony in the evening; Downward ACKNOWLEDGMENTS Piping, uttered by paired birds on their terri- We are grateful to the Charles Darwin Foun- tory, in both agonistic and courtship situations; dation for the Galapagos and to UNESCO for the Upward Jerk, one of the main advertise- making possible this work, which was carried ment displays used by unpaired males, but out while one of us was Director of the also used by paired birds; the Head-up Tail- Charles Darwin Research Station in the GalL- up posture, Facing Away, and the Landing pagos, and to N. Tinbergen, Esther Cullen, Gape, characteristic of the early stages of and R. G. B. Brown for advice and information courtship; the Breast-lowered posture, which that have helped us in preparing this account, tends to replace the Head-up Tail-up posture which is Publication no. 86 from the Charles later in courtship; the Regurgitation Bow, a Darwin Foundation for the Galapagos.

LITERATURE CITED CULLEN, E. 1957. Adaptations in the Kittiwake to cliff-nesting. Ibis 99:275-302. DWIGHT, J. 1925. The gulls (Laridae) of the world; their plumages, moults, variations, relationships and distribution. Bull. Amer. Mus. Nat. Hist. 52:63408. FOGDEN, M. P. L. 1964. The reproductive behaviour and taxonomy of Hemprichs’ Gull Law hmprichi. Ibis 196:299-320. HAILMAN, J. P. 1964a. Breeding synchrony in the equatorial Swallow-tailed Gull. Amer. Nat. 98:79-83. HAILMAN, J. P. 196413. The Galapagos Swallow-tailed Gull is nocturnal. Wilson Bull. 76:347-354. HAILMAN, J. P. 1965. Cliff-nesting adaptations of the Galapagos Swallow-tailed Gull. Wilson Bull. 77: 346-362. LACK, D. 1967. Interrelationships in breeding adaptations as shown by marine birds. Proc. XIV Internat. Ornithol. Congr., 3-42. LBVBQUE, R., R. I. BOWMAN, and S. L. BILLEB. 1966. Migrants in the Galapagos area. Condor 68: 81-101. MOYNIHAN, M. 1962. Hostile and sexual behavior patterns of South American and Pacific Laridae. Be- haviour, Suppl. 8. MURPHY, R. C. 1936. Oceanic birds of South America. Amer. Mus. Nat. Hist., New York. NELSON, J. B. 1967. Colonial and cliff nesting in the Gannet compared with other Sulidae and the Kitti- wake. Ardea 55: 60-90. SNOW, D. W., and B. K. SNOW. 1967. The breeding cycle of the Swallow-tailed Gull Creagrus furcat~. Ibis 199: 14-24. TINBERGEN, N. 1953. The Herring Gulls’ world. Collins, London. TINBERGEN, N. 1959. Comparative studies of the behaviour of gulls ( Laridae) : a progress report. Be- haviour 15: l-70. TINBERGEN, N. 1962. An introduction to the behaviour and displays of British gulls. Pp. 191-200, in D. A. Bannermau, The birds of the British Isles, vol. XI. Oliver and Boyd, Edinburgh and London. TINBERGEN, N. 1967. Adaptive features of the Black-headed Gull Lams ridibundus L. Proc. XIV Inter- nat. Omithol. Congr., 43-59.

Accepted for publication 8 June 1967.