The Food, Feeding, and Development of Young Tufted and Horned Puffins in Alaska

Home , Horned puffin, Sand lance, Tufted puffin

Condor 85~421442 ic: The Cooper Ornithological Society 1983

THE FOOD, FEEDING, AND DEVELOPMENT OF YOUNG TUFTED AND HORNED PUFFINS IN ALASKA

D. H. S. WEHLE

ABSTRACT. -Aspects of nestlingTufted Puffin (Fraterculu cirrhata) and Horned Puffin (8’. comicdata) feeding ecology, growth rate, and fledging successwere studied on Buldir Island, Alaska, in 1975 and on Ugaiushak Island, Alaska, in 1976 and 1977. Growth rates were measured for single wild chicks given supplemental food, singlechicks raised in captivity and fed ad libitum, and for artifically twinned chicks with and without supplemental food and for twinned chicks raised in captivity. Data from these colonies and years (colony years) were compared with those from other colony years for these speciesand for Atlantic Puffins (F. arctica). For most aspectsof nestling feeding ecology that have been measured, available data are too variable (seasonally,annually, or geographically) to give a reliable correlation with feeding conditions. Growth rates equal to 1.5% of adult body weight/day or lesswere associatedwith poor feeding conditions while growth rates of 2.5% or more represented optimum feeding conditions. Nestling growth rates currently provide the best means of assessingfeeding conditions. The im- ulications of puffins’ abilitv to raise two chicks, rather than their normal clutch of one, are also discussed._

Local seabird populations in Alaska may well lagic seabirdsthat measure about 390 and 350 be affected by the development of offshore oil mm in length, respectively. They nest in un- and gas resources.Recognizing this, the U.S. derground burrows or rock crevices on steep Bureau of Land Management, through inter- sea slopes,along cliff-edges, or in talus slopes. agency agreement with the National Oceanic Both sexes have two brood patches, but fe- and Atmospheric Administration (NOAA), males typically lay only one egg, which is in- sponsoredthe Outer Continental Shelf Envi- cubated for six to seven weeks. Puffins capture ronmental AssessmentProgram (OCSEAP) to prey (primarily small fish and invertebrates) obtain baseline data on the distribution and by pursuit diving, and feed whole prey to their abundance,breeding biology, and feeding ecol- semi-precocial young for six to seven weeks. ogy of certain Alaskan seabird species.These Information on their habits is given by Daw- data were to be used in making management son (1913), Willett (1915), Bent (1919) De- decisionsand as a standardagainst which post- ment’ev and Gladkov (195 l), Kozlova (1957), development data could be compared. If the Gabrielson and Lincoln (1959), Swartz (1966) consequencesof petrochemicalexploitation are and Sealy (1973a). to be predicted and safeguards established My purposeshere are: (1) to present the re- against potential problems, the ecology of the sults of studies I conducted for OCSEAP on speciesmost likely to be affectedmust be known the feeding ecology (food, feeding, and devel- (M&night and Knoder 1979). opment) of nestlingTufted and Horned puffins Between 40 and 50 million seabirds of at on UgaiushakIsland, Alaska, in 1976 and 1977 least 35 speciesbreed in Alaska (Sowls et al. and from a study I conducted on Buldir Island, 1978). Alcids (Family Alcidae) comprise at Alaska, in 1975; (2) to compare my resultswith least 65% of these, of which approximately those for these speciesin other colonies and 20% are puffins (Sowls et al. 1978): Tufted years (colony years) and with those for Atlantic Puffins (Frutercula cirrhata), Horned Puffins Puffins (F. arctica); (3) to describe and sum- (F. corniculuta),and Rhinoceros Auklets (Cer- marize the current data for the major com- orhinca monocerata)-actually misnamed puf- ponents of nestling Tufted and Horned puffin fins (Storer 1945). Of the 176 speciesof birds feeding ecology in Alaska; and (4) to identify using marine habitats in Alaska, British Co- those components of nestling feeding ecology lumbia, and Washington, Tufted and Horned that might best be used in future studies to puffins rank among those most vulnerable to detect possible effects of offshore oil and gas oil pollution (King and Sanger 1979). In ad- development on puffin feeding conditions (e.g., dition, they may be affected by changesin the availability and abundance of suitable prey). distribution and abundanceof their prey caused Fledging successprovides the best measure by offshore oil and gas development. of the reproductive health of a population but Tufted and Horned puffins are diurnal, pe- it is sometimes influenced by factors unrelated 14271 428 D. H. S. WEHLE

60‘ $90 60.

BERING SEA 2 *. 60 BULDIR So *- ISLAND

PACIFIC OCEAN

FIGURE 1. Location of Ugaiushak Island and Buldir Island, Alaska. to feeding conditions. Becauseyoung birds ap- supports lush Calamagrostis grassland and pear to be sensitive to regional and annual mixed-meadow communities; these also occur variations in feeding conditions (Ricklefs and on West Island with an extensive heath com- White 1975) their growth rates provide a bet- munity. Sea cliffs rising to 100 m comprise ter criterion by which feedingecology data from most of Ugaiushak’s western and southern pre- and post-offshoredevelopment studiescan coastlines. Tufted Puffins nest almost exclu- be compared. Since puffin growth rates varied sively in earthen burrows along the tops of among colony years, I had to develop a stan- these cliffs where the island’s soil is deepest. dard for making comparisons. I attempted to Gradually sloping ledges,beach boulders, and provide such a standard by correlating ob- talus slopesmake up the eastern and northern served growth rate with apparent feeding con- coasts; here, Homed Puffins principally use ditions and by determining the maximum rock crevices for nest sites. growth rate chicks could obtain. Feeding con- Buldir Island (52”21’N, 175”56’E), the west- ditions were assessedand maximum growth ernmost member of the Rat Island group of rates were estimated by measuring the growth the Aleutian Islands, is the most isolated of of single chicks and artificially twinned chicks the Aleutians, being 113 km west of the nearest raised in situ and given supplemental food, island (Fig. 1). Over 1.8 million seabirdsof 20 and of chicks raised in captivity and fed ad species breed on Buldir, including approxi- libitum. The twinning experiments also pro- mately 20,000 each of Tufted and Horned puf- vided information on the birds’ ability to raise fins (Sowls et al. 1978). My study of puffins on a second young. the island wasancillary to an ongoingU.S. Fish and Wildlife Service research project on the STUDY AREAS endangered Aleutian Canada Goose (Branta Ugaiushak Island (56”47’N, 156”41’W), in the canadensisleucopareia). western Gulf of Alaska, is one of many small With an area of about 2,000 ha, Buldir’s islands harboring seabird colonies along the topography is dominated by three volcanic Alaska peninsula (Fig. 1). It was chosen as an peaks, which range in elevation from 539 to OCSEAP study site because of its proximity 65 5 m; the remainder of the island being lesser to areas being considered for offshore oil and peaks and rolling hills. Buldir is heavily veg- gas development and because of its rich di- etated, except at the highest elevations, with versity and abundance of breeding seabirds. moss-willow tundra and Elymus-umbel com- Fifteen speciesof seabirds,totaling over 56,000 munities whoseplants reach over 2 m in height birds, breed on Ugaiushak, including approx- by late summer. The 20-km coastline consists imately 14,000 Tufted Puffins and 18,000 chiefly of large boulder or gravel beaches or Horned Puffins (Sowls et al. 1978). Ugaiushak talus slopes,backed by steepsea slopes or cliffs. (ca. 175 ha) consistsoftwo land masses-here- Most Tufted Puffins nest in earthen burrows after called East Island and West Island-con- on sea slopes,but some nest in the rock crev- nected by a narrow isthmus. East Island is rel- ices of talus slopes.Most Horned Puffins nest atively flat with a maximum elevation of 30 in rock crevicesof talusslopes and amongbeach m; West Island has three peaks located on the boulders, with some nesting in earthen bur- southwesternend of the island (the tallest being rows on sea slopes or on the banks of inland 170 m), a central ridge, and a gently sloping creeks. My study plots for Tufted Puffins were plain acrossits northwestern end. East Island primarily on the sea slopes at the northwest PUFFIN FEEDING ECOLOGY 429 end of the island and for Horned Puffins on Daily feeding activity of Tufted and Horned the Main Talus, located on the north-central puffins on Buldir was studied during all-day side of Buldir. Observations of daily feeding watchesfrom a blind on Main Talus on 12 and activity were made on the Main Talus where 22 August. On these dates, the average chick several thousand pairs of both puffin species age was approximately 2.5 weeks for Tufted and tens of thousandsof pairs of Aethia auklets Puffins and approximately 4 weeksfor Horned nest. Puffins (Wehle 1976) roughly corresponding with the agesof Atlantic Puffin chicks during METHODS their period of peak growth (Harris 1978). Investigations were conducted on both species Thus, I expected that feeding frequency would at Buldir from 17 May to 5 September 1975, probably increase between these two dates to and at Ugaiushak from 24 May to 2 September meet the increasing nutritional requirements 1976 and from 23 April to 29 August 1977. of the chicks. Watches were made at 15-min In 1976, investigationsof Horned Puffins were intervals from approximately 0.5 h before sun- secondary to those of Tufted Puffins. rise to 1 h after sunset. (Preliminary obser- Eight components of nestling feeding ecol- vations indicated that puffins did not feed their ogy were studied for both speciesbut not in young at night.) Each watch consistedof count- each colony year: food delivery, feeding fre- ing for 1 min the numbers of puffins of each quency, prey speciescomposition of food loads specieswith and without food in the bill that in the bill, prey size, food load size and weight, flew past a given point on a colony. In the growth rates, and fledging success.I selected analysis, data from two consecutive watches, these components because, with the possible beginning at sunrise, were pooled into 0.5-h exceptionof food delivery, they seemedto have intervals and the few birds observed before the potential to vary in response to feeding sunrise and after sunset were included in the conditions. first 0.5-h interval after sunriseand before sunset, respectively. FOOD DELIVERY Canvas blinds erected within the breeding col- PREY SPECIES COMPOSITION OF BILL onies of both speciesbefore egg-laying began LOADS, PREY SIZE, AND BILL LOAD SIZE AND WEIGHT allowed me to view, at close range (1 - 10 m), nest site entrancesand to monitor the coming Adult puffins carry their prey crosswisein their and going of adults. Tufted Puffin burrows, bills back to their nests. These loads of food which measuredfrom 0.5 to > 3.0 m long, were can be collected and their contents weighed, marked with numbered tongue depressorsand measured, and identified. colored flags; at each, I made a small access On Ugaiushak, I collected Tufted Puffin bill hole (ca. 10 cm diameter) into the enlarged loads from burrows provided with accessholes nest chamber, usually located at the rear of the and monitored from a blind. When a food- burrow. Each hole waskept pluggedby a wedge- carrying adult entered one of these burrows, I shaped piece of sod that could readily be re- immediately left my blind, opened the access moved. Horned Puffin nest siteswere similarly hole, and retrieved the food sample. To de- marked; accessto nests was provided by nat- termine if prey speciescomposition of loads, ural openingsamong the rocks. When a food- prey size, or load size and weight varied during carrying adult entered a marked nest site, I the nestlingperiod, I collected 4 1 samplesfrom immediately left the blind and looked into the 10 monitored burrows from 3-14 and from nest. 19-30 August in 1976 and 31 samples from 44 monitored burrows from 1O-l 5 and 2 l-27 FEEDING FREQUENCY August in 1977. Collections were made during Estimates of feeding frequency for individual all hours of the day and sometimes from the Tufted Puffin chicks on Ugaiushak were based same burrow in the same day. In 1977, I also on observations of birds delivering food to determined, using 10 x 50 binoculars,the prey marked burrows and on the results of moni- speciesand number of prey for 33 bill loads toring burrows using toothpick “gates” (see carried by Tufted Puffins. Boersma and Wheelwright 1979). I watched On Buldir, I collected 15 loads each from from a blind during sessionsof 2-8 h con- Tufted and Horned puffins between 14 and 21 ducted at least three days per week throughout August. Loads were obtained primarily by the nestling period in both years. Toothpick catching food-carrying adults in mist-nets gates were placed at the entrances to 10 bur- placed over nest entrances;a few food-carrying rows containing chicks, once in early and again adults were also shot and their bill loads col- in late August each year. Toothpicks were left lected. in place until displaced. General information on the nestling food of 430 D. H. S. WEHLE

Horned Puffins on Ugaiushak came from ob- I provided would moderate the chick’s behav- servations of dropped fish found within their ior so that adults would deliver food only if colonies and from observations of bill loads the chick indicated that it was still hungry. In beingcarried by adults.Accessible Horned Puf- this situation, the chick would receive as much fin nest sites on Ugaiushak were usually far food as it could use and grow at a nearly max- apart, thereby allowing only a few nest sitesto imum rate. Without a feedback mechanism, be monitored simultaneously from a blind; I the amounts of food delivered by adults would therefore decided not to collect Horned Puffin be inherent to their feeding behavior as influ- food loads. enced by feeding conditions. A chick reared All loads collected were weighed to the near- under thesecircumstances would grow at nearest 0.5 g; individual prey were counted and ly the maximum rate if feeding conditions were measured to the nearest 1 mm (fish from the optimal, but more slowly if feeding condi- tip of the snout to the tip of the caudal fin, and tions were less favorable. For fed twins, I ex- squid from the anterior end of the mantle to pected the resultsto be similar to those for fed the tip of the longesttentacle). Most of the fish singles,except that the growth rates of twins were identified by me; other prey were pre- would be even more sensitive to food short- served in 5% formalin and later identified by ages. the staff of the Aquatic Collections, University For unfed twins, if a feedback mechanism of Alaska Museum, Fairbanks. was operative, then the adultswould have twice the normal pressureto supplyfood; thesechicks GROWTH RATES would grow more slowly than unfed singles During the three colony years on Ugaiushak unlessfeeding conditions were excellent. In the and Buldir combined, I measured the growth absenceof a feedbackmechanism, adults would rates of 67 Tufted Puffin and 26 Horned Puffin deliver only the amount of food normal for a chicks reared under natural and experimental singlechick and the twinned chickswould grow conditions. Chicks raised under natural con- more slowly than unfed singles,regardless of ditions hereafter are called “unfed singles.” feeding conditions. Experimental chicks included single chicks I measured growth rates by weighing chicks given supplemental food (“fed singles”), arti- to0.5g(under100g)orto1g(100+g)every ficially twinned chicks with and without food second or third day on Ugaiushak and daily (“fed” and “unfed” twins, respectively), and on Buldir from hatching until either fledging single and twinned chicks raised in captivity or my departure from the island. and fed ad libitum (“captive singles”and “cap- Food supplements consisted of fresh fillets tive twins,” respectively). Growth rates of of black rockfish (Seabastesmelanops) and chicksraised under the various conditionswere small whole fish, primarily Pacific sand lance intended to provide information on feeding (Ammodyteshexapterus) and capelin (Mallo- conditions,the maximum rate that chickscould tusvillosus). Fed singlesreceived 50 g of food attain, and the ability of adults to feed more supplement once daily beginning at three days than one young. Interpretation of these data of age;fed twins together received 50 g of food was basedon assumptionsthat: (1) the amount supplement once daily when three to seven of supplementalfood given to chicksalone was days old and 100 g daily thereafter. Supple- insufficient to enable chicks to reach maxi- mental food was placed at the entrance to the mum growth rates; (2) the amount of supple- nest chamber for Tufted Puffins and near the mental food plus the amount of food delivered nest for Horned Puffins. by adults would be equal to or greater than Twins were artificially created by adding a that needed for chicks to attain maximum chick from one nest to another containing a growth rates; and (3) captive chicks fed ad li- single chick; chicks were twinned when less bitum would grow at maximum rates. For than two days old. Captive chicks were re- chicksreceiving supplementalfood, I also con- moved from their nestswhen two to five days sidered whether or not a feedback mechanism old and raised outside in wood and burlap sim- existed between chicks and adults such that ulated burrows. All captives were fed ad libi- certain chick behavior stimulated adults to de- turn the same prey used for food supplements liver food. plus a commercial brand multi-vitamin and Fed singles would presumably grow faster multi-mineral supplement. than unfed singlesif feeding conditions were suboptimal (i.e., if adults were unable to pro- FLEDGING SUCCESS vide chicks with sufficient food for maximum Due to ship scheduling,I left each island each growth) and at the same rate as unfed chicks year before all puffins had fledged. Therefore, if feeding conditions were optimal. With a I estimated fledging successfrom (1) the pro- feedback mechanism, any supplemental food portion of nestlings initially monitored that PUFFIN FEEDING ECOLOGY 431

had fledged or were of fledging weight at the HORNED PUFFIN time of my departure, and (2) a subjectiveeval- uation of the growth rates of chicks not of fledging weight at the time of my departure. I consideredfledging weight of Tufted Puffins to be 496 g, the minimum weight of any chick known to have fledged on Ugaiushak in 1976 or 1977, and of Horned Puffins to be 400 g, TUFTED PUFFIN the lowest fledgingweight reported by Amaral (1977). STATISTICAL ANALYSIS J4 A chi-square test was used to test for homo- geneity in the relative numbers of (1) birds observed carrying food during all-day watches on 12 and 22 August 1975, and (2) individual FIGURE 2. Percent of total numbers of food-carrying sand lance, capelin, and gadids in bill loads Tufted Puffins (n = 80) and Homed Puffins (n = 307) ob- between early and late August and between servedduring half-hour intervals from sunrise(SR) to sun- years. set (SS). Prey size and bill load size and weight were compared by the Mann-Whitney test (W) for samples within the same colony year and by addition to the Kvalue, I also expressedgrowth the Kruskal-Wallis (H) and a posteriori Dunn rates in terms of Hussell’s (1972) instanta- tests (Dunn 1964) for samples among colony neous growth rate, KR/4 x 100 (percent of years. adult body weight/day), which occurs at the Growth rates for chicks in my studies were inflection point of the fitted logistic curve. analyzed by the Mann-Whitney test and Krus- RESULTS AND DISCUSSION kal-Wallis test on the median weight-gain/day (g), during the interval from 5 to 30 days of FOOD DELIVERY age, for chicks raised under each of the various I was often able to remove the sod plugs of conditions. For the analysis of twins, I aver- Tufted Puffin burrows or to position myself aged the weight gains of the two chicks in each within rock crevicesbeside Homed Puffin nests set, so that the appropriate error was the nest- so as to observe puffin activity within the nests to-nest variability. Growth rates of unfed sin- while remaining undetected. Tufted and gle chicks were also analyzed by Ricklefs’s Homed puffin chicks received their first loads (1967) graphical method so that they could be of food when lessthan one day old. In each of compared more easily with previously pub- two Horned Puffin first-food deliveries that I lished growth rates of Tufted and Horned puf- observed, the load consistedof two sandlance, fins and other alcidsanalyzed by this technique each approximately 60 mm long. Adults (Sealy 1973b, Manuwal and Boersma 1977, dropped food on the nest floor near the chick Vermeer and Cullen 1979). for the first week and usually near the nest site I computed growth rates presented in Table entrancethereafter. The amount of time adults 8 from data provided by other investigators spent underground when delivering food de- (cited in other tables), except for those on the creased from several minutes during the Barren Islands, which were calculated by Ma- brooding period (O-3 days for Tufted Puffins nuwal and Boersma (1977). and 5-7 days for Horned Puffins [Wehle 19801) By Ricklefs’s (1967) method, growth data to less than 10 s in later weeks. Occasionally, for all colony years best fit the logistic equa- I saw adult Tufted and Horned puffins accom- tion. However, aspointed out by Sealy (1973b), pany their food delivery by specific behavior, the actual growth curves of alcids in the later particularly when chicks were less than one stagesof development do not conform well to week old. For both species,this behavior in- the logistic equation and therefore, the good- cluded the bird lowering its head and moving nessof fit dependsto some extent on subjective it slowly from side to sidewhile uttering a faint, consideration of which data to include. In low-pitched “err? sound. Whenever an adult Ricklefs’s terminology, the equation for the entered the nest, chicks (both species)called logisticcurve is: d Wldt = KW( 1 - w/a), where loudly and frequently and assumed a food- W is the weight of the growing bird, a is the beggingposture, crouching low to the ground final weight (asymptote), t is time, and K is the and sometimespartially spreadingtheir wings. growth constant (representing overall growth Once a chick had eaten, it usually ceasedcall- rate) derived by the fitted logistic equation. In ing.

PUFFIN FEEDING ECOLOGY 433

TABLE 2. Number and percent occurrenceof prey in bill loads delivered by Tufted Puffins on Ugaiushak Island, 1976 and 1977.

1976 1977 Early August Late August TOtd Early August Late August TOtal Prey species NO. % NO. % NO. % NO. % NO. % NO. %

Pacific sand lance 319 88.8 51 87.9 370 88.8 249 92.6 38 46.9 287 82.0 Capelin 10 2.8 1 1.7 11 2.6 13 4.8 29 35.8 42 12.0 Gadids 20 5.6 : 10.4 26 6.2 3 1.1 14 17.3 17 4.9 Other 10 2.8 0.0 10 2.4 4 1.5 0 0.0 4 1.1 Total 359 100.0 58 100.0 417 100.0 269 100.0 81 100.0 350 100.0

Food delivery by Tufted and Horned puffins The lack of parental feeding for one or two was similar to that of Atlantic Puffins (Myr- days has been noted elsewherefor Tufted Puf- berget 1962, Corkhill 1973) except that in over fins (Baird and Moe 1978, Vermeer et al. 1979) a dozen instancesof food delivery to nestlings but has never been reported for Horned Puf- (both species combined) less than one week fins. During the nestlingperiod, Tufted Puffins old, I never saw chicks take food directly from forage in inshore, offshore, and oceanic waters, their parent’s bill. depending on colony location, while Horned Puffins feed primarily in inshore waters (Wehle FEEDING FREQUENCY 1982). Seabirdsthat feed farther offshore tend Tufted Puffin chicks on Ugaiushak received to have longer intervals between feeding visits from none to at least four loads of food daily; to their young than do speciesthat forage closer occasionally, some chicks were not fed for two to shore (Lack 1968). consecutive days. My data on the relative proportions of Tuft- Parental feeding of Tufted and Horned puf- ed and Horned puffins carrying food on 12 and fin chicks was loosely synchronized within a 22 August were in variance with the tendency population. Daily feeding activity began at for puffins to increasethe number of daily food sunriseand tended to occur in peaks every l- deliveries during the period of peak nestling 2 h throughout the day, with most food deliv- growth (Corkhill 1973, Ashcroft 1977, Ma- eries made in the morning. On Buldir, over nuwal and Boersma 1977, Harris and Hislop 60% of all deliveries occurred between sunrise 1978). (approximately 07:OO)and 13:00, with lessthan 3% occurring between 19:00 and sunset (ap- PREY SPECIES COMPOSITION OF proximately 22:00) (Fig. 2). FOOD LOADS On Buldir, the proportion of food-carrying In food loads of Tufted Puffins on Ugaiushak adults to non-food-carrying adults observed in 1976 and 1977, sand lance comprised over during all-day watches on 12 and 22 August 80% of all prey items (Table 1). In 1976, most was similar for Tufted Puffins (12 Aug. = 3.2%, other prey consisted of nearly equal numbers n = 1,338; 22 Aug. = 2.3%, n = 1,608; x2 = of capelin and gadids(walleye pollock and saf- 2.18, P > 0.05) but decreasedsignificantly for fron cod); in 1977, capelin accounted for two- Horned Puffins (12 Aug. = 5.6%, y1= 3,154; thirds of all other prey. The relative propor- 22 Aug. = 2.2%, n = 5,821; x2 = 65.39, P < tions of sandlance, capelin, and gadidsin early 0.00 1). versus late August loads were similar in 1976 Weather appeared to affect feeding activity (x2 = 2.91, df = 2, P > O.OS),but differed sig- only slightly. Even under the worst conditions nificantly in 1977 (x2 = 98.28, df = 2, P < (i.e., heavy rain and fog, waves over 2 m, or 0.000 1) when more capelin were taken later in winds in excessof 80 km/h), some adults de- the month (Table 2). The relative proportions livered food. of thesethree prey groupsfor the whole month Other studies have found that Tufted and were significantly different between the two Horned puffins generally receive from two to years (x2 = 26.23, df = 2, P < O.OOOl),again, four feedingsdaily, with a range of none to six largely becauseof the variation in numbers of for Tufted Puffins and two to six for Horned capelin taken. On Buldir, sand lance and squid Puffins (Manuwal and Boersma 1977, Ver- each comprised approximately 35% of all prey meer et al. 1979). The mean number of daily in Tufted Puffin bill loads, with the remaining feedings for Atlantic Puffins during various prey consistingprimarily of sculpin and Atka colony years ranges from 2.3 to 15.0, with a mackerel (Table 1). maximum of 24 bill loads delivered to a single In food loads of Horned Puffins on Ugaiu- burrow in one day (Harris and Hislop 1978). shak, sand lance appeared to be the most nu- 434 D. H. S. WEHLE

TABLE 3. Number and percent occurrenceof prey in bill loads delivered by Homed Puffins during various Alaskan colony years. See Table 1 for scientific names of prey species.

Buldir Shuma ‘n Barren Islands Island Man a s 1975 1976 1976 1977 Prev soecies NO. % NO. % NO. % NO. 96

Pacific sand lance 22 42.3 107 66.5 3 15.8 51.7 Capelin 34 21.1 14 73.1 44.8 Atka mackerel 22 42.3 Pacific cod 17 10.6 1 1.7 Pacific sandfish 1 0.6 1 5.2 1 1.7 Whitespotted greenling (Hexagrammos stellerz] 1 5.2 Yellow Irish lord 1 1.9 Unidentified fish 2 1.2 Squid (Gonatidae) 7 13.5 Source this study Moe and Day Manuwal and Manuwal and (1979) Boersma (1977) Boersma (1977)

merous prey in both years, with the remainder ferences in the availability or abundance of of prey being capelin and gadids. On Buldir, prey species,in the selection of specific prey sand lance and Atka mackerel each accounted by the puffins, or both. The causeof variation for about 42% of all prey; most of the remain- can best be determined by sampling the waters ing prey were squid (Table 3). within the puffins’ foraging range and concur- In other OCSEAP studies,the predominant rently analyzing the prey speciescomposition prey speciesin Tufted and Horned puffin bill of food loads. In Britain, the composition of loads in all colony years were either sand lance North Sea trawl catchesbroadly resemblesthe or capelin (Tables 1 and 3); on Chowiet Island composition of food loads of nearby Atlantic in 1976, sandlance were the predominant prey Puffins (Harris and Hislop 1978). This study in 12 Tufted Puffin and 20 Horned Puffin loads also found that puffins tend to select more nu- (Leschner and Burrell 1977; G. Burrell, pers. tritious, but apparently less abundant, prey comm.). Of 13 colony years for Tufted and species.Coupled with the need for a similar Horned puffins in Alaska (Chowiet plus those study in Alaska is the need for greater knowl- given in Tables 1 and 3) 11 of them showed edge of the life histories of puffin prey species sand lance and/or capelin accounting for at in Alaskan waters. With the exception of the least 85% of all prey, while for both specieson commercially exploited gadids,life history pa- Buldir, they accounted for lessthan 45% of all rameters of puffin prey are poorly understood. prey. Although sand lance and capelin were the primary prey speciesin both Tufted and PREY SIZE Horned puffin bill loads, the composition of Individual prey in Tufted and Horned puffin subordinate prey differed between puffin food loads were generally 50-100 mm long species.The most numerous lesser prey were (Table 4). Tufted Puffin loads contained sig- (in decreasingorder of abundance) for Tufted nificantly smaller sand lance on Ugaiushak in Puffins: cephalopods, gadids, sculpin, and 1976 than in 1977, or on Buldir in 1975; dur- greenling; and for Horned Puffins: greenling ing both years on Ugaiushak, sand lance were and gadids, squid, and sandfish(Tables 1 and significantly larger in late versus early August. 3). Capelin in Tufted Puffin loads were also sig- Seasonal variation in prey speciescompo- nificantly smaller on Ugaiushak in 1976 than sition of bill loads has been noted for Tufted they were in 1977. Puffins at Sitkalidak Strait (Baird and Moe Tufted and Horned puffins on Buldir deliv- 1978) and for Horned Puffins on the Shumagin ered nearly the same size sand lance, but Tuft- Islands (Moe and Day 1979); in both colony ed Puffins delivered significantly larger squid years the occurrence of capelin in food loads than did Horned Puffins (Table 4). decreasedover the nestling period. Basedon the relationshipbetween sandlance Seasonal, annual, and/or geographic varia- body length and age class (Blackburn 1979), tion in the prey species composition of bill most sand lance in puffin bill loads were prob- loads occurs for Tufted Puffins in British Co- ably age classes0 and 1 fish, with few fish in lumbia (Vermeer 1979, Vermeer et al. 1979) age class 2 present. Most capelin taken were and for Atlantic Puffins in Britain (Harris and probably age class 1 fish, with a few age class Hislop 1978). Variation may result from dif- 0 and 2 fish present (Jangaard 1974). All squid PUFFIN FEEDING ECOLOGY 435 in bill loads were probably subadults (Berry 19 12, Roper and Young 1975). Seasonalvari- ation in prey size may reflect either the summer growth of the fish or predation on different populations of fish as they move into the birds’ foraging areas (Blackburn 1979) rather than the birds’ selectionof progressivelylarger fish. The sizes of sand lance and capelin I found in Tufted and Horned puffin food loads are similar to those reported for other colony years (Hatch et al. 1979, Moe and Day 1979, Ver- meer 1979), with the exception of Triangle Is- land, B.C., in 1978; there, Tufted Puffins delivered mostly age class 2 sand lance (Vermeer 1979). Seasonal and annual variations in the length of prey have been noted elsewhere for Tufted Puffins (Cody 1973, Baird and Moe 1978, Hatch et al. 1979, Vermeer 1979) and for Atlantic Puffins (Myberget 1962, Corkhill 1973, Harris and Hislop 1978).

SIZE AND WEIGHT OF FOOD LOADS Tufted Puffinsdelivered significantlymore prey per load on Ugaiushak in 1976 than in 1977, or on Buldir (Table 5). The size of loads was similar between early and late August in 1976, but decreased significantly during August in 1977. Tufted Puffin loads were significantly heavier on Ugaiushak in 1977 than in 1976, and were similar between early and late August within each year (Table 5). Tufted and Horned puffins on Buldir delivered loads of similar size and weight (Table 5). Means of both size and weight of food loads in this study fall within the ranges of those observedelsewhere in Alaska (Table 6). In other colony years where the loads of Tufted and Horned puffins were weighed, the weights did not differ significantly between species (Ma- nuwal and Boersma 1977, G. Burrell, pers. comm.). Although seasonaland annual variation in load size and weight have been reported for Tufted Puffins at other colonies (Cody 1973, Manuwal and Boersma 1977, Baird and Moe 1978,Vermeer et al. 1979), the statistical significance of the observed differ- encesis unknown. Among Atlantic Puffins, the size of loads varies (Myrberget 1962, Corkhill 1973, Harris and Hislop 1978), whereas the weight of loads is generally similar between years and throughout the seasonat the same colony (Harris and Hislop 1978).

GROWTH RATES Growth rates of Tufted and Horned puffin nestlings raised under various conditions on Buldir and Ugaiushak are given in Table 7. Unfed Tufted Puffin single chicks on Ugaiu- shak grew at the same rate in 1976 as in 1977, and significantlyfaster on Ugaiushakboth years 436 D. H. S. WEHLE

TABLE 5. Size and weight of bill loads delivered by Tufted (TP) and Homed (HP) puffins on Buldir Island, 1975, and Ugaiushak Island, 1976 and 1977.

PUffi” Size (no. prey) Weight (g) Location Time species n Median Range Significance n Median Range Significance Buldir 1975 TP 15 3.0 l-29 > * ;; ;;:; ** HP 15 2.0 4.9-24.5 l-11 W = 244, NS W = 184, NS 5.5-l 5.5> Ugaiushak 1976 early Aug. TP 30 8.5 l-25 W = 663, NS ;y 1;:; ::;I;;:; W = 608, NS late Aug. TP 11 8.0 l-13 I > combined TP 41 8.0 l-25 * 41 9.0 2.5-28.0 ** Ugaiushak 1977 late Aug. TP 21 5.0 l-9 P < 0.05 16 15.5 5.0-22.5 W = 202, NS early Aug. TP 43 6.0 l-13 > w= 154 15 13.5 7.0-34.0 > combined TP 64 6.0 l-13 * 31 14.0 5.0-34.0 **

NS = not significantat P 5 0.05. * Differenceamongcolony years: H = 13.2,df = 2, P < 0.005; 1975vs. 1976,expaimental-w&P < 0.05; 1975vs. 1977,NS; 1976vs. 1977,expeimental- wise P < 0.05. ** Differenceamong colony years: H = 18.8, df= 2, P < 0.001; 1975 vs. 1976, NS, 1975 vs. 1977, NS; 1976 vs. 1977, experimental-wiseP < 0.05. than on Buldir (H = 6.11, df = 2, P -C 0.05; adults of unfed nestlingswere delivering close experimental-wiseP < 0.0 1). The growth rates to the maximum amount of food the chicks of unfed and fed Tufted Puffin singlesand twins could use. During my daily feeding of fed nest- in 1977 did not differ significantly among the lings, I never saw uneaten food in the nest. four groups (H = 5.48, df = 3, P > 0.05). Becausethe fed chicksgrew as rapidly as unfed Unfed Horned Puffin singlesalso grew sig- chicks, the adults apparently compensatedfor nificantly faster on Ugaiushak in 1977 than my food supplements by delivering less food those on Buldir (W = 3.00, P < 0.05). Growth themselves. This would be expected if feed- rates did not differ significantlyamong Horned back exists between chicks and adults, as has Puffin unfed and fed singlesand unfed twins been observed in Pied Flycatchers (Ficedulu on Ugaiushak in 1977 (H = 4.14, df = 2, P > hypoleuca;von Haartman 1949). 0.05). Among various colony years in Alaska, the The similar growth rates among the four growth rate (KY)of Tufted Puffins varied by groups of Tufted Puffins in 1977 suggestthat 107% and Horned Puffins by 92% (Table 8).

TABLE 6. Size and weight of bill loads delivered by Tufted and Homed puffins during various Alaskan colony years (n = number of bill loads).

Weight SEX ““EE Location n 1 SD Range n R SD Range SOUrCe Tufted Puffin 1975 Buldir Island 13 11.9 6.2 5.0-24.5 15 4.3 4.1 l-29 this study 1976 Ugaiushak Island 41 9.7 6.1 2.5-28.0 41 9.4 5.5 l-22 this study Barren Islands 24 14.9 - 2.0-36.5 79 3.8 - l-8 Manuwal and Boersma (1977) Chowiet Island 12 7.5 3.6 2.3-16.8 12 10.1 4.2 l-15 G. Burrell 1977 Ugaiushak Island 31 14.4 5.2 5.0-34.0 64 5.6 3.4 l-13 (pers. comm.) this study Barren Islands 28 20.4 - 9.0-35.0 58 3.4 - l-8 Manuwal and Boersma (1977) Sitkalidak Strait 10 19.3 6.2 13.5-35.0 - - R. A. Moe (pers. comm.) Homed Puffin 1975 Buldir Island 15 11.0 3.3 5.5-15.7 15 3.5 3.1 l-l 1 this study 1976 Barren Islands 9 10.7 - 3.0-19.0 26 1.5 - l-3 Manuwal and Boersma (1977) Chowiet Island 19 7.9 4.4 2.5-19.0 19 12.6 14.5 2-65 G. Burrell (pers. comm.) Shumagin Islands 18 13.8 3.9 7.4-25.4 18 6.0 4.7 l-16 Moe and Day (1979) 1977 Barren Islands 13 17.0 - 3.0-35.0 20 3.2 - l-7 Manuwal and Boersma (1977) PUFFIN FEEDING ECOLOGY 437

TABLE 7. Growth rates of Tufted and Horned puffin chicks reared under various conditions on Buldir Island, 1975, and Ugaiushak Island, 1976 and 1977.

Tufted Puffin Homed Putlin Median Median growth growth Location Year Condition n rate (g/day) n rate (g/day)

Buldir 1975 unfed singles 2 5.2 2 4.9 Ugaiushak 1976 unfed singles 27 15.7 - - unfed twins la 7.5 - 1977 unfed singles 10 15.9 10 10.9 fed singles 5 16.2 4 10.9 captive singles 2 8.9 unfed twins : 17.214.2 3 7.8 fed twins 7 14.6 - captive twins - - 1 9.0 s For twns, n is the numberof twinned pairs.

However, most of this variation was attributed of the unfed chicks’ comparatively high growth to the much slower growth rates of Tufted Puf- rates (Tables 7 and 8), I suspectthat all unfed fins on Buldir and Chowiet islands and for single chicks fledged after my departure. Horned Puffins on Buldir than in all other col- No experimental chicksfledged in either year ony years. Excluding Buldir and Chowiet, K on Ugaiushak before I left. Of 28 sets of non- values for Tufted and Horned puffins varied captive chicks twinned, one chick in each of by only 39% and 27%, respectively. 14 sets either disappeared or died (being re- jected by the adults or forcibly kept from food FLEDGING SUCCESS by its nest-mate); all other chicks were alive Few Tufted or Horned puffin chicks fledged when I left the island. Except for the one set before I left Ugaiushak in 1976 or 1977 (Table of unfed Tufted Puffin twins in 1976, growth 9). All observednestling mortality during these rates of non-captive experimental birds com- years occurred among chicks less than two pared favorably with those of unfed single weeksold. The most commonly reported caus- chicks (Table 7); consequently, I assume that es of death in chicks older than that are pre- all of these chicks eventually fledged. In 1977, dation by terrestrial predators (Amaral 1977; I found the banded remains of one of the unfed Mickelson et al. 1977, 1978) and insufficient Tufted Puffin twins of 1976 in its burrow; the food (Vermeer et al. 1979). In the absenceof fate of the other chick is unknown. terrestrial predatorson Ugaiushak and because On Buldir in 1975, I found only three young

TABLE 8. Specificgrowth parametersof Tufted and Homed puffins during various Alaskan colony years.”

Growth rate Asymp- Adult tote weight wg“d~y;oo Year Location n (9) @) R=dW K a Tufted Puffin 1975 Buldir Island 2 360 755 0.48 0.074 0.9 1976 Ugaiushak Island 27 600 792 0.76 0.125 2.4 Barren Islands 12 600 784 0.77 0.111 2.1 Chowiet Island 12 330 780 0.42 0.091 1.0 Shumagin Islands 15 520 830 0.63 0.145 2.3 Wooded Islands 2 550 792b 0.69 0.120 2.1 1977 Ugaiushak Island 10 555 792 0.70 0.153 2.7 Barren Islands 23 595 784 0.76 0.110 2.1 Sitkalidak Strait 10 590 792b 0.75 0.126 2.4 Cathedral Island 24 580 792b 0.73 0.127 2.3 Homed Puffin 1975 Buldir Island 2 300 491 0.61 0.075 1.2 1976 Barren Islands 3 440 600 0.73 0.122 2.2 Chowiet Island 12 280 519 0.54 0.113 1.5 Shumagin Islands 9 405 549 0.74 0.144 2.6 1977 Ugaiushak Island 10 380 526 0.72 0.139 2.5 Barren Islands 7 445 600 0.74 0.114 2.1 =Growth parameterswere calculated from data suppliedby OCSEAP investigatorsat eachcolony (see introduction and othertables), except for the Barren Islandswhich are given in Manuwal and Boersma(1977). CathedralIsland data wereprovided by R. A. Moe (pen. comm.). b Mean adult we&t of Tufted Puffinson UgaiushakIsland. 438 D. H. S. WEHLE

TABI .E 9. Estimated fledging success,fledging weight, and weight of nestlingsnot fledged by last check for unfed single Tufted and Homed puffins on Ugaiushak Island, 1976 and 1977.

Tufted Puffin Homed Puffin 1976 1977 1977 n 2 SD n K SD n R SD Total number of nestlings monitored 50 21 11 Nestlings fledged 19 (38%) 567 37.2 6 (29%) 556 37.3 0 Nestlings of fledging weightbnot fledged 21 (42%) 563 36.6 7 (33%) 572 44.8 1 (9%) 403 Nestlings below fledging weight not fledged 3 (6%) 469 35.5 7 (33%) 385 64.5 9 (82%) 328 59.2 Known nestling mortality 7 (14%) 1 (5%) 1 (9%) Estimated fledging success 86% 95% 9 1% a 1976 = 2 September;1977 = 27-28 August. bTufted Puffin= 496 g; Homed Puffin= 4M) g. each of Tufted and Horned puffins. Of these, ed and Horned puffin chicks on Buldir was one chick of each species had been deserted associatedwith poor feeding conditions in that by its parents. None of the four remaining year is based on three points. First, despite chicks, ranging in age from 42 to 46 days, had extensive searching,I found few chicks; many fledged when I left the island. All four chicks adults had apparently deserted their eggs or grew much more slowly than chicks in other chicks early in the nestling period. A similar colony years (Table 8) and when last weighed desertion by Tufted Puffins during incubation were at least one-third lighter than chicks of on Triangle Island may have been caused by equivalent ages on Ugaiushak in 1976 and food shortageor weather interference with for- 1977. Becauseof their very slow growth rates, aging (Vermeer 1978). Second, contrary to an I doubt that any of these chicks fledged. expected increase in numbers of adult puffins Estimates of fledging successon Ugaiushak observed carrying food to their young on 12 for Tufted Puffins in 1976 and 1977 and for versus22 August, I saw no changein the num- Horned Puffins in 1977 were the highest re- ber of Tufted Puffins and proportionally fewer ported for any colonies during those years (Ta- Horned Puffins. A decrease in feeding fre- ble 10). Breeding failures comparable to those quency was associated with apparent food suspectedfor Tufted and Horned puffins on shortageand the starvation of young Atlantic Buldir and thought to be related to poor feed- Puffins (Harris and Hislop 1978). Third, cap- ing conditions have been reported for Tufted elin were absentand sand lance comprised less Puffins on Triangle Island, British Columbia than 45% of all prey in puffin food loads on (Vermeer et al. 197 9) and for Atlantic Puffins Buldir; in other colony years, where chicksalso in Britain, Norway, and Newfoundland (Bod- grew faster, sand lance and/or capelin made dington 1960, Mills 198 1, Working Group on up at least 85% of all prey. The presence of Atlantic Seabirds 1982). specific prey species in Atlantic Puffin food Aspects of food delivery are primarily in- loads resulted in significant differences in herent characteristicsof puffin feeding behav- fledgingweight (Harris and Hislop 1978). The ior, presumably not subject to variation in re- slow growth rates of Tufted and Horned puffin sponse to environmental change. Other chicks on Chowiet Island were also associated componentsof nestlingecology may, however, with poor feeding conditions, as indicated by be affected, and thus may serve to indicate 10% of the 48 chicks monitored having died changesin feeding conditions. Of these, fledg- of starvation (Leschner and Burrell 1977). ing successis overall the best measure of a The growth rates of Tufted and Homed puf- population’s reproductive health, but it is often fins on Buldir and Chowiet islands(0.9 to 1.5% affected by factors other than feeding condi- of adult weight/day) provide data on rates of tions (e.g., weather [Manuwal and Boersma nestlinggrowth during years when feeding con- 19771, predation [Amaral 1977, Lehnhausen ditions were apparently poor. Data on growth 19801).Growth rate, on the other hand, is sen- rates of chicks raised under apparently favor- sitive to variations in feeding conditions able feeding conditions are provided by the (Ricklefs and White 1975) and is largely de- experimental chicks on Ugaiushak in 1977. termined by the combined effects of feeding Growth rates of Tufted Puffin unfed and fed frequency, prey size, and the prey speciescom- singlesand twins and Horned Puffin unfed and position, size, and weight of food loads. fed singlesand unfed twins did not differ in- My interpretation that slow growth in Tuft- traspecifically. These results indicate that PUFFIN FEEDING ECOLOGY 439

TABLE 10. Estimatedfledging success ofTufted and Homed puffinsduring various Alaskancolony years(n = number of chicks monitored).

Tufted Puffin Homed Puffin Fledgin Year Location n S”CCeSS( k ) n sE..~@o, SOlKlX

1975 Buldir Island 2 0 this study 1976 Ugaiushak Island 50 8; _2 - this study Barren Islands Manuwal and Boersma ( 1977) Chowiet Island 16 6956 3211 z: Leschnerand Burrell(1977) Shumagin Islands 32 ;;’ 10 90 Moe and Day (1979) 1977 Ugaiushak Island 21 11 91 this study Barren Islands 39 13 69 Manuwal and Boersma (1977) Sitkalidak Strait 41 ;;: - - Baird and Moe (1978) Wooded Islands 93 0 -- Lehnhausen(1980) ’

a Number of burrows monitored. weight gains of 2.5% adult body weight/day That uniparity in puffins may be a result of were near maximum and that feeding condi- such constraints is suggestedby the results of tions were nearly optimal. Although sample egg replacement experiments I conducted on sizeswere too small for statisticalcomparison, Ugaiushak in 1977 (Wehle 1980). In 10 each my interpretation is supported by the fact that of Tufted and Horned puffin nestsfrom which captive chicksand unfed singlesgrew at similar I removed the original egg soon after it was rates. The data also suggestthat a feedback laid, 7 and 3 nests, respectively, contained a mechanism in food delivery may exist between replacement egg 10-2 1 days later. Only if a chicks and adults for both puffin species. female was able to lay a second egg soon after Data on the major components of the feed- the first eggwas lost (within several days rather ing ecology of nestlingTufted and Horned puf- than 10-2 1 days) could constraintson the lay- fins for various colony years in Alaska are sum- ing female be discounted. marized in Table 11. They are insufficient to For the data to support Nelson’s (1964) con- establisha reliable correlation between feeding tention that pairs rearing two chicks had a sub- conditions and feeding frequency, prey size, or stantial reproductive advantageover pairs with prey speciescomposition, size, and weight of single chicks, several questions must be con- food loads. Nevertheless, changes in feeding sidered: (1) Is sufficient hardship imposed on conditions should at least be suspectedif data the female by laying a second egg or on the in future studiesdiffer significantly from those pair by rearing two chicks to reduce their po- presented in Table 11. Currently, however, tential contribution of young?(2) Is post-fledg- growth rate alone is the most reliable criterion ing survival of two chicks less than that of by which feeding conditions can be assessed. singles?(3) Does the attempt to rear two chicks The high growth rates and estimated fledging lower fledging success?No information is successof unfed Tufted and Horned puffin available on puffins to answer the first two twins on Ugaiushak in 1977 neither refute nor questions, and current data are insufficient to supportLack ’s (1954) hypothesisthat the clutch answer the third (Nettleship 1972; Corkhill size of a bird speciescorresponds with the larg- 1973; M. P. Harris, pers. comm.; this study). est number of young for which the parents can, on the average, provide enough food. In addition, my findings do not strengthen or con- CONCLUSIONS tradict Nelson’s (1964) conclusion that, al- Pacific sand lance and capelin are the most though some uniparous seabirdscan rear two important prey speciesfor nestling Tufted and young sufficiently well to give those birds a Horned puffins in Alaska. More information substantialreproductive advantage over pairs is needed on the life histories of these fish in having single chicks, they do not do so natu- Alaskan waters and on the relationship be- rally. Although both Tufted and Horned puffin tween certain aspectsof their life histories(e.g., pairs were apparently able to fledge two young age, size, seasonalmovements) and predation during that colony year, the data are too few by puffins and other seabirds. to determine whether or not they could do so Data for some aspectsof the feeding ecology “on the average.” In exception to Lack’s (1954) of nestling Tufted and Horned puffins cur- view, the clutch size of many speciesof birds rently are too variable to make a reliable cor- is determined by constraints on the laying fe- relation with feeding conditions. The data sug- male rather than the inability of pairs to feed gest that feeding frequency and prey species more young (see Klomp 1970, von Haartman composition of food loads may be better in- 1971). dicators of feeding conditions than prey size 440 D. H. S. WEHLE

TABLE 11. Summarized data for major components of the feeding ecology of nestling Tufted and Homed puffins during various Alaskan colony years.

Compo”e”tof feedingecology Tufted Puffin Homed Puffin

Food delivery -begins when chicks are one day old -as for Tufted Puffin -food dropped on nest site floor, often -as for Tufted Puffin near entrance -specific feeding behavior of adults, -as for Tufted Puffin particularly when chicks are young -feedback mechanism may exist between chick -as for Tufted Puffin behavior and stimulation of adults to deliver food Feeding frequency -0-6 food loads/day, chicks sometimes - 2-6 loads/day not fed for two consecutivedays -loosely synchronizedwithin population -as for Tufted Puffin -throughout daylight hours, most -as for Tufted Puffin deliveries in morning Variation’ -seasonal, tending to increasewith -unknown chick age (?) Prey species Primary prey -sand lance and capelin -as for Tufted Puffin Subordinate prey -cephalopods, gadids, sculpin, greenling -greenling, gadids, cephalopods,sandfish Variation -seasonal, annual, geographic -as for Tufted Puffin Prey size/Age -usually 50-100 mm -as for Tufted Puffin -sand lance mostly age classes0 and 1 -sand lance and squid fewer age class2; capelin mostly age as for Tufted Puffin class 1, fewer age classes0 and 2; squid all subadults Variation -seasonal (no trend), annual, geographic -unknown Bill load size (no. prey) K f SD, n’ (n”)b -5.3 f 2.3, 6 (269) -5.2 f 4.0, 5 (98) Range of n’ (n”) -3.4-10.1 (l-29) -1.5-12.6 (l-65) Variation -seasonal (no trend), annual, geographic(?) -annual (?), geographic(?) Bill load weight (g) K f SD, n’ (n”) -13.9 + 4.3, 7 (159) -11.9 f 3.2, 5 (74) Range of n’ (n”) -7.5-20.4 (2.3-36.5) -7.9-17.0 (2.5-35.0) Variation -annual, geographic(?) -geographic (?) Growth rates (%/day) K f SD, n’ (n”) -2.2 * 0.5, 10 (137) -2.1 f 0.5, 6 (43) Range of IZ’ -0.9-2.7 -1.2-2.7 Variation -geographic -as for Tufted Puffin Fledging successE(%) .X * SD, n’ (n”) -82 * 11, 8 (217) -65 + 21, 6 (79) Range -0-95 -0-91 Variation -annual (?), geographic(?) -as for Tufted Puffin * Seasonalvariation is that observedduring nestlingperiod at samecolony in sameyear; annual variation is that observedat samecolony among years; geographicvariation is that observedamong colonies. Question marks indicate non-statistically documented variation. bWeighted mean and standarddeviation for all colonyyears combined, n ’ is numberof colonyyears; n” is total numberof samplesfor all colonyyears combined. CPercent of chickshatched

or load size and weight. Nevertheless, growth shore development and changes in feeding rate (as the combined product of all compo- conditions as indicated by growth rates. nents) is currently the best indicator of feeding In yearsof favorable feedingconditions, pairs conditions for puffins. Growth rates of 1.5% of both Tufted and Horned puffins can suc- adult body weight/day or lessreflect poor feed- cessfully rear two chicks; that they do not do ing conditions and, in general, are correlated so naturally is probably the result of con- with low fledgingsuccess; growth rates of 2.5% straints on the laying female. or more reflect the best feeding conditions and, in the absence of other mortality factors, are ACKNOWLEDGMENTS associatedwith high fledging success.Growth Field work on Buldir was sponsoredby the Aleutian Is- rates have varied considerably among colony lands National Wildlife Refuge, U.S. Fish and Wildlife Service, Adak, Alaska, and on Ugaiushak by OCSEAP, years in studies thus far. Hence, data from a U.S. Fish and Wildlife Service, Office of Biological Ser- number of colony years will be needed in order vices-coastal Ecosystems,Anchorage, Alaska. to establisha credible relationship between off- R. H. Day, B. E. Lawhead, K. Powers, E. P. Hoberg, G. PUFFIN FEEDING ECOLOGY 441

V. Byrd, J. L. Trapp, D. L. Woolington, and G. E. Keiser JANGAARD,P. M. 1974. The capelin, Mallotus villosus. assistedin the field. K. 0. Coyle, G. E. Keiser, G. Mueller, biology, distribution, exploitation, utilization, and and K. Vogt (all at The Aquatic Collections, University composition. Fish. Res. Board Can. Bull. 186. of Alaska Museum, Fairbanks) identified prey specimens. KING, J. G., AND G. A. SANGER. 1979. Oil vulnerability M. J. Amaral, P. A. Baird, 6. Burrell, R. Fi. Day, S. A. index for marine-oriented birds, p. 227-239. In J. C. Hatch, W. A. Lehnhausen.L. L. Leschner.D. A. Manuwal. Bartonek and D. N. Nettleship [eds.], Conservation R. A. Moe, and K. Vermker kindly sharkd with me thei; of marine birds of northern North America. U.S. Fish unpublished data on the biology of Tufted and Horned Wildl. Serv. Wildl. Res. Rep. 11. puffins. Amaral, Baird, Day, Hatch, Leschner, Manuwal, KLOMP, H. 1970. The determination of clutch size in and Vermeer, as well as F. C. Coleman, M. P. Harris, D. birds. Ardea 58:1-124. N. Nettleship, R. E. Ricklefs, and S. G. Sealy, reviewed KOZLOVA,E. V. 1957. Charadriiformes, SuborderAlcae. various drafts of the manuscript. I especiallythank F. C. In Fauna of USSR: Birds. Vol. 3, No. 2. Israel Prog. Coleman for editing and typing the final draft of the manu- Sci. Transl.. Jerusalem(196 1). Clearinghousefor Fed- script. era1Scientific and Technical &format&, Springfield, VA. LITERATURE CITED LACK, D. 1954. The natural regulation of animal numbers. Clarendon, Oxford. AMARAL,M. J. 1977. A comparative biology of the Tuft- LACK, D. 1968. Ecological adaptations for breeding in ed and Horned puffin in the Barren Islands, Alaska. birds. Methuen and Co., London. M.Sc. thesis, Univ. of Washington, Seattle. LEHNHAUSEN,W. A. 1980. Nesting habitat relationships ASHCROFT,R. E. 1977. Breedingbioloav and survival of of four speciesof alcids at Fish-Island, Alaska. M.%. Puffins. Ph.D. diss., Univ. of Oxf%d, Oxford, En- thesis. Univ. of Alaska. Fairbanks. gland. LESCHNER;L. L., AND G. &JRRELL. 1977. Populations BAIRD,P. A., AND R. A. MOE. 1978. The breedingbiology and ecology of marine birds on the Semidi Islands, and feedingecology of marine birds in SitkalidakStrait D. 13-109. In Environmental assessmentof the Alas- area, Kodiak Island, 1977, p. 3 13-524. In Environ- kan Continental Shelf. Annual reports of principal mental assessmentof the Alaskan Continental Shelf. investigators. Vol. 4. NOAA, Environmental Re- Annual reports of principal investigators. Vol. 3. searchLab., Boulder, CO. NOAA, Environmental ResearchLab:, Boulder, CO. MANUWAL,D. A., AND D. BOERSMA.1977. Dynamics of BENT.A. C. 1919. Life historiesof North American div- marine bird populationson the Barren Islands, Alas- i& birds. U.S. Natl. Mus. Bull. 107. ka, p. 294-420. In Environmental assessmentof the BERRY;S. S. 1912. A review of the cephalopodsof west- Alaskan Continental Shelf. Annual reports of prin- em North America. Bull. U.S. Bur. Fish. 30. cipal investigators, Vol. 4. NOAA, Environmental BLACKBURN,J. E. 1979. Pelagic and demersal fish as- ResearchLab., Boulder, CO. sessmentin the Lower Cook Inlet Estuary System, p. M&NIGHT, D. E., AND C. E. KNODER. 1979. Resource 289-446. In Environmental assessmentof the Alas- development along coastsand on the ocean floor: po- kan Continental Shelf. Annual reports of principal tential conflictswith marine bird conservation,p. 183- investigators. Vol. 4. NOAA, Environmental Re- 194. In J. C. Bartonek and D. N. Nettleship [eds.], searchLab., Boulder, CO. Conservation of marine birds in northern North BODDINGTON,D. 1960. Unusual mortality of youngPuf- America. U.S. Fish Wildl. Serv. Wildl. Res. Rep. 11. fins on St. Kilda, 1959. Scott. Birds 1:218-220. MICKELSON,P. G., W. A. LEHNHAUSEN,AND S. E. QUINLAN. BOERSMA,D., AND N. T. WHEELWRIGHT.1979. Egg ne- 1978. Community structureof seabirdson Wooded glect in the Procellariiformes: reproductive adapta- Islands, Alaska, p. 680-772. In Environmental as- tions in the Fork-tailed Storm-Petrel. Condor 8 1:157- sessmentof the Alaskan Continental Shelf. Annual 165. reports of principal investigators.Vol. 4. NOAA, En- CODY,M. L. 1973. Coexistence,coevolution, and con- vironmental ResearchLab., Boulder, CO. vergent evolution in seabird communities. Ecology MICKELSON,P. G., W. A. LEHNHAUSEN,S. E. QUINLAN, 54:3l-44. AND J. M. SHERWOOD.1977. Seabirdsofthe Wooded CORKHILL,P. 1973. Food and feedingecology of Puffins. Islands, Alaska, p. 421-499. In Environmental as- Bird Study 20: 207-220. sessmentof the Alaskan Continental Shelf. Annual DAWSON,W. L. 19 13. The Tufted Puffin. Bird-Lore 15: reports of principal investigators.Vol. 4. NOAA, En- 268-271. vironmental ResearchLab., Boulder, CO. DEMENT’EV,G. P., AND N. A. GLADKOV[EDS.]. 1951. The MILLS, S. 1981. Graveyard of the puffin. New Sci. 9 1: birds of the Soviet Union. Vol. 2. Israel Proa. Sci. 10-13. Transl., Jerusalem(1969). Clearinghousefor Federal MOE, R. A., AND R. H. DAY. 1979. Populations and Scientificand Technical Information, Springfield,VA. ecology of the seabirds of the Koniuji Group, Shu- DUNN,0. J. 1964. Multiple comoarisonsusine. rank sums. magin Islands,Alaska, p. 395-49 1. In Environmental Technometrics 6:24i-252. _ assessmentof the Alaskan Continental Shelf. Annual GABRIELSON,I. N., AND F. C. LINCOLN. 1959. The birds reports of principal investigators.Vol. 6. NOAA, En- of Alaska. Stackpole Co. and Wildlife Management vironmental ResearchLab., Boulder, CO. Inst., Harrisburg, PA. MYBERGET,S. 1962. Contribution to the breeding biol- HARRIS,M. P. 1978. Supplementary feeding of young ogy of the Puffin (Fratercula arctica L.): eggs, incu- Puffins, Fratercula arctica. J. Anim. Ecol. 47: 15-23. bation and young. Medd. StatensViltunders. (2) No. HARRIS,M. P., AND J. R. G. HISLOP. 1978. The food of 1 l:l-51. youngPuffins, Fratercula arctica. J. Zool. (Lond.) 185: NELSON,J. B. 1964. Factors influencing clutch-size and 213-236. chick growth in the North Atlantic Gannet Sula bas- HATCH,S. A., T. W. PEARSON,AND P. J. GOULD. 1979. sana. Ibis 106~63-77. Reproductive ecology of seabirds at Middleton Is- NETTLESHIP,D. N. 1972. Breeding successof the Com- land, Alaska, p. 233-308. In Environmental assess- mon Puffin (Fratercula arctica L.) on different habi- ment of the Alaskan Continental Shelf. Annual re- tats at Great Island, Newfoundland. Ecol. Monogr. ports of principal investigators. Vol. 2. NOAA, 421239-268. Environmental ResearchLab., Boulder, CO. RICKLEFS,R. E. 1967. A graphicalmethod of fitting equa- HUSSELL,D. J. T. 1972. Factors affecting clutch size in tions to growth curves. Ecology 48:978-988. Arctic passerines.Ecol. Monogr. 42:317-364. RICKLEFS,R. E., AND S. C. WHITE. 1975. A method for 442 D. H. S. WEHLE

constructingnestling growth curves from brief visits VERMEER,K., L. CLJLLEN,AND M. PORTER. 1979. A pro- to seabird colonies. Bird-Banding 46: 135-140. visional explanation of the reproductive failure of ROPER,C. F. E., AND R. E. YOUNG. 1975. Vertical dis- Tufted P&ins (Lunda cirrhata) on Triangle Island, tribution of pelagiccephalopods. Smithson. Contrib. British Columbia. Ibis 121:348-354. Zool. 209. - - - - VONHAARTMAN, L. 1949. Der Trauerfliegenschnapper. SEALY,S. G. 1973a. Breeding biology of the Homed I. Acta Zool. Fenn. 56:1-104. Puffin on St. Lawrence Island, Bering Sea. with zoo- VONHAARTMAN, L. 1971. Population dynamics, p. 39 l- geographicalnotes on the North Pa&c puffins. Pac. 459. In D. S. Famer and J. R. King reds.], Avian Sci. 27:99-l 19. biology. Vol. 1. Academic Press, New York: SEALY,S. G. 1973b. Adaptive significanceof post-hatch- WEHLE.D. H. S. 1976. Summer foods and feedinr!ecol- ing developmental patterns and growth rates in the ogy of Tufted and Homed puffins on Buldir Gland, Alcidae. Omis Stand. 4: 113-l 2 1. Alaska- 1975. M.Sc. thesis, Univ. of Alaska, Fair- SOWLS,A. L., S. A. HATCH, AND C. J. LENSINK. 1978. banks. Catalog of Alaskan seabird colonies. U.S. Fish and WEHLE,D. H. S. 1980. The breeding biology of the puf- Wildlife Service, Washington, DC. fins: Tufted Puffin (Lunda cirrhata), Homed Puffin STORER.R. W. 1945. Structuralmodifications in the hind (Fratercula corniculata),Common Puffin (F. arctica), limb in the Alcidae. Ibis 87:433-456. and Rhinoceros Auklet (Cerorhincu monoceruta). SWARTZ.L. G. 1966. Sea-cliff birds. D. 61 l-678. In N. Ph.D. diss., Univ. of Alaska, Fairbanks. J. Wilimovsky and J. N. Wolfe iehs.], Environment WEHLE,D. H. S. 1982. Food ofadult and subadultTufted of the Cape Thompson region, Alaska. U.S. Atomic and Horned puffins. Murrelet 63:51-58. Energy Comm., Oak Ridge, TN. WILLETT,G. 1915. Summer birds of Forrester Island, VERMEER,K. 1978. Extensive reproductive failure of Alaska. Auk 32:295-305. RhinocerosAuklets and Tufted Puffins. Ibis 120:112. WORKINGGROUP ON ATLANTIC SEABIRDS. 1982. Capelin VERMEER,K. 1979. Nesting requirements, food, and and the marine environment of Atlantic Canada. Pac. breeding distribution of Rhinoceros Auklets, Ceror- Seabird Group Bull. 9:41-45. hinca monocerata,and Tufted Puffins, Lunda cirrhata. Ardea 67:101-l 10. Division of Life Sciences,University ofAlaska, Fairbanks, VERMEER,K., AND L. CULLEN. 1979. Growth of Rhi- Alaska 99701. Present address:R.D. 3 Box 346, Tru- nocerosAuklets and Tufted Puffins, Triangle Island, mansburg,New York 14886. Received 3 July 1982. Final British Columbia. Ardea 67122-27. acceptance4 April 1983.

(hndoras:442 0 The Cooper Omithologml Society 1983

RECENT PUBLICATIONS

Handbook of the Birds of Europe, the Middle East, and previous papers. Distribution and zoogeography, com- North Africa/The Birds of the Western Palearctic. Vol- parative habits, and anatomy (chiefly skeletomuscular)are ume III. Waders to Gulls.-Edited by Stanley Cramp treated, concentratingon the four best-known species.A with K. E. L. Simmons, Duncan J. Brooks, N. J. Collar, brief appendix tabulates the state of sixty morphological Euan Dunn, Robert Gillmor, P. A. D. Hollom, Robert and life history charactersfor thesesame speciesand gives Hudson, E. M. Nicholson, M. A. Ogilvie, P. J. S. Olney, a phenogram based on these data. The report is useful as C. S. Roselaar,K. H. Voous, D. I. M. Wallace, Jan Wattel, a compilation but it shedslittle new light on the systematic and M. G. Wilson. 1983. Oxford University Press,Oxford, position, ecology, or behavior of whistling-ducks. Their London, and New York. 913 p. E 49.50. This volume whistling is not mentioned. Maps, drawings, references. follows the same basic plan as its predecessors(noticed in Condor 80:253 and 83:141). Owing to the large number of charadriiform speciesin the region and the great deal Anatomy of the Domestic Birds.-R. Nickel, A. Schum- that is known about them, it has been necessaryto relegate mer. and E. Seiferle. 1977. Verlaa Paul Parev. Berlin and the terns, skimmers, and alcids to the next volume, to- Hamburg. 202 p. Source:Springer-Verlag N& York, 175 getherwith the remainder of the non-passerines.The treat- Fifth Ave., New York, NY 10010. This is a textbook on ment of certain topics has been revised to overcome prob- the anatomy of the fowl and other domestic birds, which lems that became manifest in the previous volumes and, has been skillfully translated from the German original more importantly, to adjustto the habits of the birds under edition (1973) by two British poultry anatomists, W. G. consideration.The book is illustratedwith 105 color plates Siller and P. A. L. Wight. Conventional in plan and treat- as well as countlessdrawings, maps, sound spectrograms, ment, it marchesthrough the organ systemsfrom skeleton and diagrams of the annual cycle. Staggeringlycompre- to skin. The approach is heavily descriptive and lightly, hensive, this superbreference work is essentialfor anyone if at all, physiologicalor functional. The level of detail is working on charadriiforms, whether palearctic speciesor mostly gross,with little attention paid to microanatomy. not. References,corrections to volumes I and II, and in- Many drawings(a few in color) are provided, the majority dex. of them original; although well executedand reproduced, they show no new views. References,index. As a sound, convenient source of basic descriptive information, this Whistling-ducks:Zoogeography, Ecology, Anatomy.-Eric book compares well with those by King and McLelland C. Bolen and Michael K. Rylander. 1983. Special Publi- (noticed in Condor 78:148), T. Koch (Anatomy of the cations No. 20, The Museum, Texas Tech University. 67 Chickenand DomesticBirds. 1973. Iowa StateUniv. Press, D. Paoer cover. $12.00. Source: Texas Tech Press, Sales Ames), and Schwarze and Schrijder (noticed in Condor bffi&, Texas Tech University Library, Lubbock, Texas. 82:9), yet each of these has certain advantages.For stan- 79409. The eight speciesof Dendrocygnaare distributed dardized terminology and more details, particularly on around the world, largelybetween the 30th parallels,north microanatomyand physiology,one must consultthe works and south. This report collectspublished and unpublished edited by Baumel (noticed in Condor 82:397) or King and information on them, principally from the authors’ own McLelland (noticed in Condor 82:327, 84:21).