The Use of Food Samples from Sea Birds in the Study of Seasonal Variation. 10 the Surface Fauna of Tropical Oceanic Areas MYRTLE J. ASH MOLE AND N. PHILIP ASHMOLEI

ABSTRACT: Many parts of the tropical oceans appear to be relatively seasonless, but, because of the difficulty of samplin g mobile and patchily distributed animals and the cost of oceanographic investigations, few data are available on the extent of seasonal changes. By regularly collecting regurgitations from sea birds, and identifying and measurin g the food items, seasonal data could be obtained on the availability, size classes, and perhaps reproductive cycles of the fish and squid char­ acteristic of the surface layer of tropical seas. Flying fish (Exocoetidae), juvenile tunas (Scombridae) , and squid of the family Ommastrephidae are especially easily obtainable. Experience gained dur ing a recent study of the comparative feeding ecology of sea birds on Christmas Island (Pacific Ocean) makes it possible to assess the char­ acteristics of bird species which affect their suitability for such study. Potentially useful species include terns (especially Sterna fuscata, Anous stolidas, A. tenuirostris, and Gygis alba) and boobies (especially Sula SIt/a) . Samples could be obtained from several bird species in the same period, and a program could include sampling of inshore waters with the neuston net and making basic oceanographic observations. Investigations of this kind could be carried out economically on any of a large num­ ber of tropical oceanic islands.

IN ASSESSING the results of a study of the feed­ important commercial fisheries are good seasonal ing ecology of some sea birds of Christmas data available. Even in these areas, however, Island , in the central equatorial Pacific, we there are generall y few data on changes in the attempted to make use of published information abundance of the forage animals which form concerning seasonal variations in the fauna of the food of most sea birds and of many fish the surface layer of the sea in this region. It used for human food. This is because nets used soon became evident, however, that this type for routine plankton sampling do not catch of information is very sparse, and that those many nekton animals, and even modern nets workers who have made seasonal comparisons like Isaacs-Kidd trawls catch few fast-swimming have often had to depend on data collected in animals near the surface (King and Iversen, different areas, in different years, or by different 1962; Pearcy, 1965). methods (see, for instance, King and Demond, It would be of considerable inter est to ob­ 1953 ; King and Hida, 1957 ; King and Iver­ tain more information on the extent to which sen, 1962). seasonal fluctuations in physical and biological The absence of information on seasonal vari­ characteristics actually occur in those parts of ation in many tropical oceanic areas results the tropical oceans which appear to be more or mainly from the expense of running oceano­ less constant throughout the year. For instance, graphic ships, and the conflicting demands on orni thologists working on tropical islands in their time, which generally prevent adequate several different areas have found that among sampling in a single area over a period as long the sea birds breeding on a single island (or as a year. Only in a few areas where there are group) some species have evolved regimes under which individuals breed at intervals of less than 1 Peabody Museum of N atur al Hi story, and D epart­ ment of Biology, Yale Un iversity, N ew H aven, Con­ a year, suggesting that seasonal variati on in the necticut 06520. Manuscript received January 23, 1967. environment is not of great importance to them. 1 2 PACIFIC SCIENCE, Vol. XXII, January 1968

But on the same islands other species have sampling depends on considerable knowledge annual breeding regimes, indicating that for of the ecology of the available species. them seasonal influences are sufficiently strong In the course of our study of the food of sea to outweigh the selective advantages of breed­ birds on Christmas Island during the period ing at shorter intervals (Ascension Island: pa­ March 1963 to June 1964, we examined 800 pers in Ibis 103b, 1962-63; Christmas Island: samples (mainly regurgitations) from eight spe­ Gallagher, 1960 ; Ashmole, 1965 ; Ashmole and cies of sea birds. We were mainly concerned Ashmole, 1967 ; Ashmole, in press; Galapagos with comparing the food of different species, Islands : Leveque, 1964; Snow, 1965) . Some ad­ and we did not have the opportunity to collect vances are now being made in understanding very large numbers of samples from anyone the differences in feeding ecology which lead species, so that our data are not suitable for to dependence of different birds on different detailed seasonal analysis. However, they do groups of prey species, but lack of knowledge of demonstrate the potentialities of the method. the importance of seasonal effects in populations A complete account of the study is available in of the various prey species prevents further Ashmole and Ashmole (1967), and here we progress. The reproductive cycles of oceanic fish shall discuss only those aspects which are perti­ and cephalopods in relatively seasonless areas nent to the planning of a more general inves­ are even less well understood, but are of both tigation, and in particular to the choice of bird theoretical and practical importance. species from which samples should be obtained. The purpose of the present paper is to show The samples which we collected on Christmas that a very economical investigation of seasonal Island consisted of entire or partially digested (and year-to-year) variation in the surface fauna fish and squid, while two bird species also pro­ of a tropical oceanic area could be carried out vided various invertebrates other than squid, by making use of the sampling of the oceanic mainly water striders (Halobates) and crusta­ environment which is done by sea birds, and ceans. Nearly all the squid belonged to a single at the same time by collecting certain other data. genus (Symplectoteuthis) of the family Om­ The method which we propose depends on mastrephidae, but the fish were distributed the fact that tropical sea birds of many species among 33 families. The state of the samples are easy to catch and, when caught, will often varied considerably, but there were fairly con­ regurgitate the whole or part of their latest sistent differences among the bird species (Table meal. Regurgitations provide, on average, far 1). These differences resulted partly from the more items in good condition than do stomachs fact that in some species it was easiest to obtain of shot or netted birds and their collection does samples from young birds which had previously not appreciably affect the bird population. Re­ been fed by adults, while in others it was easy gurgitations are stored in formalin and the to catch adults. The best samples were those ob­ individual food items identified and measured tained as the birds arrived to feed their chicks, at leisure. Since the feeding methods of each while those obtained from roosting adults were species are reasonably constant, regular sampling generally more digested and thus the items were on an adequate scale would permit detection of more often unidentifiable . It would be impor­ changes in the relative availability of different tant, therefore, to plan future work in such a prey species. Furthermore, each of the bird spe­ way as to obtain the food items as soon as pos­ cies takes prey covering a considerable range in sible after they were caught by the birds. size, so that the data should reflect changes in The usefulness of a particular species would the size-frequency distributions of the various depend largely on the ease with which large prey species. Deductions could doubtless also numbers of identifiable fish could be obtained be made about the timing of reproduction in from it, so we have calculated (Table 1) the some of the fish and squid species. The method number of fish which we were able to identify could be used on almost any tropical island on (to family level) per 100 samples collected. In which appropriate species of sea birds are avail­ the following short accounts of the most promis­ able for a large part of the year, but the decision ing bird species we have also commented on the as to which bird species should be used for abundance of each species, the ways in which it Sea Bird Food and Ocean Surface Fauna-ASHMoLE AND ASHMOLE 3 may be caught, the methods and zones which it Before breeding starts, a few regurgitations can uses in feeding, and the size of the animals be obtained by catching birds at night when which it obtains. they are roosting on the ground, but the samples are generally small and largely digested. On EVALUATION OF THE BIRD SPECIES most tropical islands the species has one fairly short breeding season, but on Christmas Island Of the eight bird species whose food was and a few other central Pacific islands there are sampled on Christmas Island (Table 1), four two breeding seasons each year. would not be very suitable for marine biological Sterna fuscata catches its prey either while investigations. It is time consuming to obtain flying or by plunging to the surface, but it prob­ samples from the Christmas Island Shearwater, ably hardly ever submerges completely. It has Puffinus nativitatis Streets, the Phoenix Petrel, been recorded as feeding at night (Gould, in Pterodroma alba (Gmelin), and the Red-tailed press) , but this probably occurs only when the Tropic-bird, Phaethon rubricauda Boddaert, and moon is nearly full. It is an oceanic species the number of identifiable fish which can be capable of feeding hundreds of miles from land, expected in 100 samples (Table 1) is rather even when breeding. However, the actual dis­ low for all these species. The Blue-grey Noddy, tance traveled regularly probably varies con­ Procelsterna cerulea (F. D. Bennett), provides siderably from colony to colony. large numbers of items per sample, but the Our 242 samples were mostly from adults populations on many islands are rather small, feeding chicks. They contained, on the average, and the species is not widely distributed. Fur­ 5.6 items each, of which 60% were fish, 40% thermore, it feeds close inshore on very small squid. On the whole, the samples were in good items which could probably be sampled more condition, and 79% of the fish were identified efficiently with a neuston net (David, 1965) to the family level. This means that about 266 towed behind a small boat. The other four spe­ identifiable fish could be expected per 100 sam­ cies studied-the Sooty Tern, Sterna fuscata ples. The identified fish belonged to 21 families, Linnaeus, the Brown Noddy, Anous stolidus the ones which occurred most regularly being (Linnaeus), the Black or Lesser Noddy, Anous Exocoetidae and Scombridae, Gempylidae, Ser­ tenuirostris (Temminck), and the White or ranidae, and Emmelichthyidae. The identified Fairy Tern, Gygis alba (Sparrman)-aU have squid were nearly all Ommastrephidae . Of the certain advantages for regular study. Salient fish, 93% were between 2 and 10 cm in length numerical details are given in Table 1. (measured to the base of the tail), while 98% of the squid were between 2 and 8 em in Sterna fuscata (Sooty Tern) mantle length. W e obtained sufficient data from This is the most abundant of all tropical sea this species to suggest that important seasonal birds, breeding in enormous colonies on islands changes exist in the availability of certain fish in the tropics around the world (see Ashmole, families, but a more intensive sampling program 1963 for details of its breeding distribution and would be necessary to demonstrate such changes breeding seasons) . The eggs are laid on the convincingly. ground in the open, and sometimes there are In summary, the advantages of this species as many as five nests per square meter. The are that it is available in great numbers, and birds may desert whole areas if they are dis­ samples can easily be obtained ; the samples are turbed too much at the start of nesting, but often in good condition, and contain a high later they sit tightly and may be caught with a proportion of identifiable items ; Scombridae hand net. However, incubating birds rarely re­ are especially well represented in the samples; gurgitate when caught, and the only way in and the feeding range of the species is such which large numbers of samples can easily be that it provides samples from a large oceanic collected is to catch adults (with a long-handled area. However, one cannot determine precisely hand net, or mist nets) as they arrive to feed the area from which each sample comes, and their chicks, especially in the late afternoon, or it is not easy to obtain many samples at times to catch chicks shortly after they have been fed. when the birds are not breeding. ~

TABLE 1

S UMMARY OF D ATA ON THE FOOD OF EIGHT SP ECIES OF SEA BIRDS ON CHRISTMAS ISLAND

Phaethon Puffinu s Pterodroma Sterna A nous A nous Gy gis Procelsterna ITEM rubricauda nativitatis alba fuscata stolidus tenulrostris alba* cerulea Mean weight of bird (gm) 665 324 269 173 173 90.9 101 45.4 Mean number of items per sample 3.8 10.5 4.4 5.6 5.1 18.3 1.3 41.8 - 4.3 % composition by number Fish 33 63 26 60 71 95 59 49 Squid 64 36 48 40 29 4 41 9 'i:i Other invertebrates 3 0.2 26 0.1 0 1 0.3 42 >(J ...... >-rJ .. % of fish identified to ...... family 55 20 1 79 67 27 87 38 (J % of squid identified to Ul (J family** 100 62 24 88 100 71 100 74 ...... tTl Estimated number of fish Z identifiable to family (J per 100 samples 69 132 1 266 243 470 67 778 ..tTl - 221 < ~ • Wh en two figures are given for Gyg;s alba the upper refers to food samples carried in th e bill and the lower to regurgitations . • • Some additional squid were tentatively identified. ~ ~ ......

'-< l'O ~ l: ..,l'O '< ..... \0 co0\ Sea Bird Food and Ocean Surface Fauna-c-Asnxrot. s AND ASH MOLE 5

Anous stolidus (Brown N oddy) Anous tenuirostris (Black N oddy) Anous stolidus breeds on many of the tropical Thi s species (i n which we include Anous islands where Sterna [uscat« is found, but gen­ minutus], although it breeds only on islands erally is less numerous. The birds are fairly where bushes, trees, or cliffs are available to easy to catch with a hand net at night, especially provide nest sites, often occurs in large colonies. when there is no moon ; mist nets could doubt­ Young birds often regurgitate when handled, less also be used, and unfledged young can be but more consistent sampling can be carried caught by hand. No general account of the out by catching roosting birds with a hand net breedin g seasons of this species is available, but early in the night, since on Christmas Island and in many colonies breeding is probably less many other islands some birds are present at tightly synchronized than that of S. [ascata, all times of year. while in some areas (i ncluding Christmas Is­ A. tenuirostris feeds by the same methods as land ) individuals roost on the breeding islands A . stolidus and S. [uscata; there is no conclusive even when they are not breeding. evidence that it feeds at night. This species A. stolidus uses feeding methods similar to seems usually to feed even closer to its shore those of S. [uscata, but sometimes also feeds base than does A . stolidus: the Christmas Island while swimming on the surface ; there is no birds fish very largely within 5 miles of the evidence that it feeds at night. In contrast to shore, but in some other areas (for instance S. [nscat«, A . stolidus apparently does not nor­ Ascension Island, Atlantic Ocean) they go mally feed more than about 50 miles from th e farther out to sea. colony, although the size of the zone utilized Our 110 samples, which were mostly regurgi­ may be different in different areas. tations from roosting adults, contained on the N early all our 38 samples from A. stolid/IS average 18.3 items each. Of these items, 95% were obtained from roosting adults early in the were fish, nearly all the remainder being squid. night, soon after they had return ed from feed­ Th e regurgitations were often in the form of ing. They contained, on the average, 5.1 items tightly-packed masses, many of the items being each, 71% of these items being fish, the re­ in a rather advanced state of digestion. We mainder squid . Over half of the fish were identified only 27% of the fish to the family identified to the family level, so th at about 243 level, partly because a large numb er of the fish identifiable fish could be expected per 100 were fry that we could not identify. Neverthe­ samples. The identified fish belonged to nine less, some 470 fish were identified per 100 samples. The identified fish belonged to 17 families, of which Exocoetidae, Scombridae, families, among which the most regularly rep­ Holocentridae, and Gempylidae occurred most resented were Exocoetidae, Gempylidae, Scorn­ regularly. All the identified squid were Om­ bridae, Blenniidae, Holocentr idae, and Emmel­ mastreph idae. Of the fish, 94% were between 2 ichthyidae. The few identified squid were all and 12 em in length, while all but 1 of the 50 Ommastreph idae. Of the fish, 98% were less measurable squid had mantle lengths between 2 than 6 em long, and 77% between 1 and 4 em. and 8 em. Of the few squid, 83% were between 2 and 6 em The advantages of A . stolidus as a sampler in mantle length. of the surface fauna of tropical seas are that the A. tenuirostris was the only species which species is widely distributed, is available on showed any dramatic differences in the diet at some islands even outside the breeding season, different seasons. In May 1963 and June 1964 we and will often provide regurgitations when found in the diet a far higher proportion of fish caught early in the night; its relatively limited less than 2 em long than were present in our feeding range means that animals obtained from other sampling periods. it could normally be assumed to come from This species has the advantages that it is within 50 miles of the island. However, it is abund ant, easily caught while roosting, and on less abundant than S. [uscat« and rather more some islands is available all the year round ; the difficult to catch. samples provide large numbers of fish, and these 6 PACIFIC SCIENCE, Vol. XXII, January 1968

are from inshore waters; the small size of its that further identification would be much easier prey would make study of its food an excellent than in most of the other birds. The fish which complement to that of one of the larger species, we identified represented 22 families, 5 of for instance A . stolidus. Its main disadvantage which were not found in the food of any of is the difficulty of identifying many of the fish the other bird species. Blenniidae occurred most fry. regularly, followed by Exocoetidae, Myctophi­ dae, Scombridae, Gempylidae, and Gonostorn­ Gygis alba (White or Fairy Tern ) atidae. Squid of the families Ommastrephidae Gygis alba, another widespread tropical spe­ and Enoploteuthidae were identified. Of the cies, generally is less abundant than the preced­ fish, 96% were less than 8 cm in length ; ing three species, but many tropical islands have the samples carried in the bill contained a popu lations of a thousand or more birds. The lower proportion of small fish than did the nests are dispersed, either in forks or hollows in regurgitations. Of the squid, 93% were between branches of trees, on the tops of coral blocks, 2 and 6 em in mantle length. or on cliff ledges where these are available. This species ate far more squid during one G . alba differs from the other species in that of our sampling periods than at other times, the adults rarely regurgi tate, and they carry and during this same period the squid taken food for the young in their bills. Hence the included Enoploteuthidae ( Abralia sp.) which technique for obtaining food samples is different. were not found in samples obtained in the earlier The birds usually come quite close to a human part of the study. intrud er, even when carrying food, and may The main advantages of this species are the often be caught with a long-handled net ; the excellent condition of the items carried in the food items dropped by the birds as they are bill and the fact that a wide variety of prey caught can then be retrieved from the ground. are eaten. However, it is difficult to obtain Our samples were mainly obtained from adults many samples, since usually the populations are in this way, although we also collected regurgita­ not very large, and some of the birds are too tations from juveniles and from adults. wary to be caught when carrying fish, so that On Christmas Island G. alba breeds at all a quantitati ve study would probably have to times of year. In many other areas, however, depend primarily on other species. breeding is more seasonal and, since the individ­ uals leave the colonies when they have finished Other Species breeding, it would be impossible to obtain The species just discussed are those with samples at all times of year. which we have most experience, but other spe­ This species catches most of its prey in flight, cies also might be used for sampling, and we and there is some evidence that it feeds exten­ have tried to indicate some of the criteria which sively in the half light at dawn. It sometimes should be used in deciding whether a given feeds close inshore, and most of the prey carried species would be appropriate. The only large back to the young in the bill are probably caught species included in our study was Pbaetbon fairly close to the colony; however, the species rubricaud« (weights are included in Table 1), sometimes ranges hun dreds of miles from land. which takes large prey, but from which it is Among our 152 samples, those carried in generally not easy to obtain many samples. the bill contained, on the average, 1.3 items The obvious additional candidates, available on each, while regurgi tations contained 4.3 items. many tropical islands, are the various species of Fish made up 59% of the items, the remainder frigate birds,' Fregata spp., and boobies, Sula being squid. The samples were generally in spp. We would reject the frigate birds on the excellent condition, and 87% of the fish were grounds that they obtain some of their food by identified to the family level. Collecting only piracy, which complicates interpretation, and also food carried in the bill, one could expect to the food generally is in poor condition. Of the obtain appr oximately 67 fish identifiable to three widespread tropical boobies, the Brown the family level per 100 samples.H owever, Booby, Snla leucogaster (Boddaert) , is mainly all of these would be in excellent condition , so an inshore feeder, is rather difficult to catch, and Sea Bird Food and Ocean Surface Fauna-AsHMoLE AND ASHMOLE 7 is not often available in large numbers. However, stance, to investigate the effects of seasons in either the Red-footed Booby, Sula sula (Lin­ the ocean around Christmas Island, it would naeus), or the Masked (or Blue-faced, or probably be best to obtain a series of regurgita ­ White) Booby, Sula dactylatra Lesson, might tions, perhaps twice each month , from the three be useful. species Anous tenuirostris, A. stolidus, and S. S. sula nests in colonies in bushes or trees, sula. These samples would include animals and regurgitates readily, at least when with ranging in size from less than 1 em to at least young. The birds roost on many islands through­ 20 em. The regurgitations from A. ten uirostris out the year; on Christmas Island, where the probably would represent sampling largely species nests on the main island, there is some within 5 miles of the island, those from A. stol­ breeding at all times of year. S. sula catches idus sampling largely within 50 miles, and those some of its prey below the surface by diving from S. sula might have been obtained as much from a height, but has also been recorded catch­ as 100 miles away. On the other hand, for a ing flying fish in the air ; it is thought some­ long-term study concerned primarily with the times to fish at night (Murphy, 1936). S. sula detection of differences in the fauna of an probably fishes mainly within a hund red miles oceanic area at the same season in a series of suc­ of its home island (Murphy, 1936; Royce and cessiveyears, it would probably be most econom­ Otsu, 1955, who almost certainly mistook ical to collect samples from Sterna fuscata, S. sula for S. dactylatra). In a small series of choosing those months when this species has samples from Oahu, Hawaii (Ashmole and young. In any study, collection of samples from Ashmole, in press) we found an average of nine Gygis alba would be useful in providing animals items per sample, of which 64% were fish, in excellent condition and wide variety, to form 36% squid. Nearly 75% of the fish were the basis of a reference collection of the fish and identified, so that about 400 identifiable fish squid available at the surface in the area. could be expected per 100 samples from this If one is to use sea birds as samplers of their species. The fish found most frequently were marine environment, it is important to have Exocoetidae and Gempylidae, while all the some understanding of the part which they identified squid were Ommastrephidae. Half play in the ecology of the surface layer of the of the fish were between 8 and 12 em long, sea. All the terns (including "noddies") con­ while there were appreciable numbers in the sidered here obtain their food within a few 12-16 and 16-20 em groups ; however, only centimeters of the surface of the sea; boobies, a few squid with mantle length of more than 8 Sula spp., tropic birds, Pbaetbon spp., and some em were present in our samples. The main shearwaters (Puffinus spp.) , can penetrate a advantages of this species are that it is available little deeper. Thus the terns should be sampling in some places throughout the year, and takes the fauna characteristic of the surface film, or fish (and perhaps also squid) larger than any "neuston." In fact, only one of the species for taken by the terns previously discussed. which we have data-the Blue-grey Noddy, We have had little experience with S. dactyl­ Procelsterna cerulea- has a diet with a composi­ atra, but relevant data have been presented by tion conspicuously similar to hauls made with Dorward (1962). It would have advantages the "neuston net" in the Indian Ocean (David, similar to S. sula, but probably ranges even 1965): both contain large proportions of water farther from its breeding colonies. striders (Halobates), pontellid copepods, and very small fish larvae. Among the other terns, only A. tenuirostris sometimes took many small DISCUSSION fish larvae; and small crustaceans and H alobates It is clear that the use of analysis of the food were absent from the diet of all the larger of each of these birdspecies has certain merits, species. Although there is little information but also certain disadvantages. We suggest that available as to how many larger animals are valuable data could be obtained by making use normally present within a few centimeters of of a group of bird species selected for the the surface during the day, observations suggest particular problem under investigation. For in- that, except for Procelsterna cerulea (and per- 8 PACIFIC SCIENCE, Vol. XXII, January 1968

haps Gygis alba feeding at dawn on animals originating fish than did any of the birds, which visit the surface during the night), all and presumably this reflects the fact that tunas the terns studied are dependent on schools of can quickly and easily change from feeding tuna and other predatory fish to drive fish and right at the surface to feeding at considerable squid to the surface: many of the prey animals depths or around reefs, while the birds must are caught by the birds while they are actually wait for their prey to come to the surface. jumping out of the water in efforts to escape the After they have completed larval life, typical predators below. The larger birds like Phaeth on reef-inhabiting fish such as Acanthuridae evi­ rubricaud« and S. sula, which can penetrate the dently do not often come to the surface even water to a depth of at least 1 meter, may be able when pursued, and so are rarely eaten by bird s ; to catch prey which are not disturbed by other however, they were eaten by the tunas in con­ predators, and so may be partially independent siderable numbers . (The pelagic larvae were of the presence of predatory fish. also eaten by tunas; they may also sometimes Tuna schools are by no means randomly dis­ be taken by birds, and it must be remembered tributed over the tropical oceans, but occur that we were able to identify only a small largely in those areas where the animals on proportion of the partly digested larval fish which they feed are most concentrated. In an found in samples from A . tenuirostris at certain earlier paper (Ashmole and Ashmole, 1967) times of year.) In contrast, Exocoetidae were we have discussed the evidence that in the open unimportant in the diet of the tunas, but are of ocean convergence at "fronts" produces local great importance in the diets of all the sea concentrations of plankton and nekton, even in birds studied ; probably the ability of these fish areas where the overall density of organi sms is to escape submarine predators by flying renders very low, and that these concentrations attract them especially vulnerable to aerial predation by tuna schools. The existence of large popula­ birds. tions of sea birds-like Sterna fllScata-which Attempts to sample forage animals near the can range great distances from their islands surface of the central Pacific dur ing the daytime, even when breeding, probably depends on the by means of oblique hauls with Isaacs-Kidd presence of rich feeding areas of this kind. and other trawls, yielded so little material that Clearly a knowledge of the areas where the wide­ they were abandoned in favor of night hauls, ranging species are feeding will help in inter ­ which produced catches containing far more preting the data obtained by the analysis of organisms of generally larger size (King and their food . The inshore-feeding sea birds, on Iversen, 1962) . Although the greater number the other hand , make use of the concentrations of organisms in night hauls partly reflects the of prey animals, and of surface-feeding tunas, movement to the . surface layers at night of which are often present in the lee of islands. animals which are mesopelagic during the day, Thinking that there might be a close corre­ it is of interest that very few Exocoetidae, spondence between the food of surface-caught Scombridae, or Gempylidae, or squid of the tunas and the food of the birds, we studied family Ommastrephidae, were identified in records of stomach contents of Yellowfin Tuna, either the day or night samples studied by N eoth lt111Z11S macropterus (Temminck and King and Iversen. Since members of all these Schlegel) , kindly made available by the Hono­ groups are commonly caught by birds feeding lulu Biological Laboratory of the U.S. Fish and at the surface, it may be that they are such fast Wildlife Service. The data used were for Yel­ swimmers that they can generally dodge the lowfin caught at the surface within 10 miles of nets (ef. Pearcy, 1965:266) . Alternatively, they Christmas Island and neighboring islands, but in may be extremely patchily distributed, or so different years from our sampling of the food strictly confined to the surface that they are of the birds. We found that there were strik­ only rarely encountered by nets used in oblique ing differences between the diets of the tunas hauls. In any case, it is clear that members of and of the birds, even if the comparison was these groups are much more efficiently sampled restricted to the birds which fed close inshore. by the birds. The tunas took a far higher proportion of reef- In view of the commercial importance of tuna Sea Bird Food and Ocean Surface Fauna-ASHMoLE AND ASHMOLE 9 fisheries, and consequent special interest in the Malcolm R. Clarke kindly undertook the identi­ biology of tunas, the effectiveness of tropical fication of the cephalopods. Garth I. Murphy sea birds in catching juvenile tunas is worth and Richard S. Shomura generously allowed us emphasizing. King and Iversen (1962:301) to make use of their unpublis hed paper on "The said, " It was our hope that by means of mid­ abundance of tunas in the Central Equatorial water trawls we would capture juvenile tunas Pacific in relation to the environment." We of lengths above 12 mm. which were able to thank Donald W . Strasburg and Philip Helfrich elude the plankton nets. In 274 hauls made with for their helpful comments on the manuscript. the four midwater trawls described in this re­ port we captured only six juvenile tunas, which ranged from 18 to 60 mm. in length ." In con­ REFERENCES trast, in the 800 food samples which we ob­ ASHMOL E, MYRTL E J., and N . PHILIP ASH­ tained from birds on Christmas Island we found MOLE. (In press) Notes on the breeding 247 young scombrids of which 166 were in season and food of the Red-footed Booby 243 samples from Sterna fu scata. Of the latter (SIt/a Stlla) on Oahu, Hawaii. Ardea. group, 77 were examined in detail, and 69 ASHMOLE, N. P. 1963. The biology of the proved to be Yellowfin Tuna. The scombrids Wideawake or Sooty Tern Sterna fuscata from S. fu scata ranged from about 2 cm to 11 cm on Ascension Island. Ibis 103b(3) :297-364. in length (measured to the base of the tail) , --- 1965. Adaptive variation in the breed- . but over 70% were between 4 and 8 cm. Those ing regime of a tropical sea bird. Proc. Nat. obtained in samples from A. tenuirostris in­ Acad. Sci. 53:311-318. cludedabout a third which were less than 2 cm in length. --- (In press) Breeding and molt in the Since the birds eat mainly animals more than White Tern (Gygis alba) on Christmas 0.5 cm in length, a program using birds to detect Island, Pacific Ocean. Condor. seasonal or longer-term changes in the environ­ - -- and MYRTLE J. ASHMOLE . 1967. Com­ ment could profitably be combined with other parative feeding ecology of sea birds of a methods (such as use of the neuston net) of tropical oceanic island. Peabody Mus. N at. sampling the smaller organisms inhabiting the H ist., Yale Univ. , Bull. 24. sea surface. If a version of this net could be DAVID, P. M. 1965. The Neuston net : a device arranged to fish clear of the bow wave of a for sampling the surface fauna of the ocean. small boat, it would permit quantitative sam­ J. Mar. BioI. Assoc. UK. 45 :313- 320. pling of the waters immediately around an DORWARD, D. F. 1962. Comparative biology of oceanic island, to complement the sampling of the White Booby and the Brown Booby larger animals, and a larger area, by the birds. SIt/a spp. at Ascension. Ibis 103b ( 2) :174­ The use of a small boat would of course also 220. make it possible to obtain basic physical and GALLAGH ER, M. D. 1960.Bird notes from chemical data requisite to an und erstanding of Christmas Island, Pacific Ocean. Ibis 102 : changes detected in the biota. 489-502. GOULD, PATRICK J. (In press) Nocturnal feed­ ACKNOWLEDGMENTS ing of Sterna fuscata and Plt/finus pacificus. Condor. The work on Christmas Island and in H awaii was supported by a Yale University-Bishop KING, JOSEPH E., and JOAN D EMOND. 1953. Museum fellowship awarded to the second Zooplankton abundance in the Central Pacific. author. Grateful acknowledgment is made to US. Fish and W ildlife Serv., Fish. Bull. 54 H. Laurence Achilles, S. Dillon Ripley and (No. 82) :111-144. Roland W . Force, and the others who helped to --- and THOMAS S. HIDA. 1957. Zooplank­ make this arrangement possible. For assistance ton abundance in the Central Pacific, Part II. in identifications of fish we are especially in­ US. Fish and Wildlife Serv., Fish. Bull. 57 debted to Gareth J. Nelson and James W . Atz; (No. 118) :365-395. 10 PA CIFIC SCIENCE, Vol. XXII, January 1968

--- and ROBERT T . B. IVERSEN. 1962. Mid­ PEARCY, WILLIAM G. 1965. Species composi­ water trawling for forage organisms in the tion and distribution of pelagic cephalopods central Pacific 1951-1956. U.S. Fish and from the Pacific Ocean off Oregon . Pacific Wildlife Serv., Fish. Bull. 62 (No. 210): Sci. 19:261- 266. 217-321. ROYCE, WILLIAM F., and TAMIO OTSU. 1955 . LEVEQUE, RAYMOND. 1964. Notes sur la repro­ Observations of skipjack schools in Hawaiian duction des oiseaux aux Iles Galapagos. waters, 1953. U.S. Fish and Wildlife Serv., Alauda 32:5-44, 81-96. Spec. Sci. Rept. , Fisheries No. 147. 31 pp. MURPHY, ROBERT CUSHMAN. 1936 . Oceanic SNOW, D . W . 1965 . The breeding of Audubon's Birds of South America. American Museum Shearwater (PufJi nus lherminieri) in the of N atural History, New York. 2 vols. Galapagos. Auk 82:591-597.