DIET AND POSTNATAL ENERGETICS IN CONVERGENT TAXA OF PLANKTON-FEEDING SEABIRDS

DANIEL D. ROB¾• Departmentof Biology,University of Pennsylvania, Philadelphia,Pennsylvania 19104 USA

ABSTRACr.--Toevaluate the influence of diet compositionon seabird postnataldevelop- ment, I examinedthe relationshipsbetween feeding rate,energy intake, energypartitioning, and chick growth in three speciesof high latitude, plankton-feedingseabirds: Least Auklet (Aethiapusilia), South Georgia Diving Petrel (Pelecanoidesgeorgicus), and Common Diving Petrel (P. urinatrix).Average dietary lipid contentvaried from 22 to 46%of dry massamong the three species.Least Auklet chickswere fed meals rich in wax esters;on average63% of total energy intake was in the form of wax. Interspecificcomparisons of chick dietary com- positionand energybudgets indicated that calanoldcopepods, a wax ester-richfood, wasa balancedsource of protein and lipids. However, the lipid contentof Antarctickrill (Euphausia superba)was insufficientto meet maintenanceenergy requirementsof CommonDiving Pe- trels. Lipid-rich diets were associatedwith shorterbrooding periods, higher ratesof nestling fat deposition,and larger lipid reservesat fledging. Accumulationrates of lipid-free dry matter were similar in the three speciesdespite differences in energy intake related to diet. The energy costof growth was a relatively minor componentof nestling energy budgets; most assimilatedenergy was allocatedtoward maintenanceand fat deposition.Selection apparentlyfavors high latitude seabirdsthat meet the high energyrequirements of their chicksby providing a lipid-rich diet. Differencesin diet compositionexplained much of the variation in postnataldevelopment among the three species,but there was no evidencethat energy limited growth per se. Received11 December1989, accepted 30 July1990.

PELAGICseabirds share a suite of life history posit larger fat reserves,and fledge before they traits that includes single chick broods, slow attain adult mass. Ricklefs and White (1981) growth, and long development periods (Lack showedthat growth in the oceanicSooty Tern 1968). These reproductive traits have been at- (Sternafuscata) is slower than growth in the tributed to limitations in providing food to the nearshore-feeding Common Tern (S. hirundo), chick. Lack (1968) inferred that the reproduc- and that daily energy requirements of Sooty tive pattern in seabirdsserves to reduceenergy Tern chickspeak at a lower level than those of requirements at the nest to a level that a breed- the Common Tern. Also, the lipid reservesof ing pair can consistentlymeet. Researchersas- Sooty Tern chickswere approximately 4 times sumedthat the availabilityof energyto a breed- greater than those of Common Terns, which is ing pair is constrainedby both the scarcityand consistentwith longer fasting periods for the the unpredictability of potential prey at sea young of the offshorespecies. These results sug- (Ashmole 1963, 1971; Harris 1977; Nelson 1977; gest that the slower growth and larger fat de- Case 1978). Energy limitation is viewed as a pots of SootyTern chicksare related to low and pervasiveforce selectingfor observedpatterns variable rates of energy provisioning. in seabird reproduction (Drent and Daan 1980). Ricklefs and White (1981) estimated that dou- Much of the variation in post-hatchingde- bling the growth rate (accumulationof lipid- velopmental patterns in the seabirdfamily A1- free dry matter) of Sooty Tern chicks would cidaeis apparentlya consequenceof differences result in only a 20% increase in total energy in feeding ecology (Sealy 1973). Speciesthat requirementsof the chick and only a 5% in- forage farther offshore and feed their young creasein adult energy requirements. They con- less frequently have slower growth rates, de- cluded that energy could not limit growth rate in this species.Similarly, in a study of the en- ergeticsof slow growth in Leach'sStorm-Petrel t Present address:Cooperative Wildlife Research (Oceanodromaleucorhoa), Ricklefs et al. (1980a) Laboratoryand Departmentof Zoology,Southern estimated that the accumulation of lipid-free Illinois University, Carbondale,Illinois 62901 USA. dry matter (LFDM) amounted to <5% of the 131 The Auk 108: 131-146.January 1991 132 DANIEL D. ROI•Y [Auk, Vol. 108 total daily energy requirement. The energy I gathered data on chick meal sizes, feeding equivalent of daily lipid deposition was 3-8 frequency,and diet compositionfor three spe- times that of LFDM, indicating that variation ciesof high latitude,plankton-feeding seabirds. in growth rate would be expectedto have little The three speciesare from two convergenttaxa'. affecton total energy requirementsof the nest- auklets (Charadriiformes:Alcidae) and diving ling. The regularityin feedingof Leach'sStorm- petrels(Procellariiformes: Pelecanoididae). All Petrel chicks implies that the large lipid re- three are plankton-feedersthat captureprey by serves of nestlings are not an adaptation to pursuit-diving (sensuAshmole 1971), yet taxo- survive fasts (Ricklefs et al. 1985), but instead nomic compositionof the diet differs consid- they may representan "energysink" for chicks erably (B•dard 1969, Payne and Prince 1979). I fed lipid-rich but protein-poordiets (Ricklefs describethe energy requirementsof chicksand 1979). These findings contradict the hypothesis how thoserequirements are partitioned among that slow growth in offshore-feedingseabirds maintenance, growth, and lipid deposition is an adaptationfor reducingenergy require- functions throughout the nestling period. I ments at the nest and imply that other factors comparedthe compositionof chick diets with are responsiblefor long developmentperiods. the patterns of energy allocation among the Ricklefs (1979, 1983) has made a strong casefor three speciesto test the hypothesisthat differ- constraintsat the tissue level as the primary encesin postnataldevelopment relate to diet. factorlimiting growth in semiprecocialseabird chicks. METHODS AND MATERIALS Studiesof the relationship of diet to seabird developmenthave emphasizedthe quantityof I studiedLeast Auklets on St. GeorgeIsland (56ø35'N, food fed to chicks.Little attention hasbeen paid 169ø35'W) in the Pribilof Islands, Alaska, from June to variation in the quality of chick diets,yet the to August, 1981 and 1982. Nests were located and biochemicalcomposition of seabirdprey varies marked in a large (ca. 129,000breeding individuals; considerably(Clarke and Prince 1980, Ricketts Hickey and Craighead1977) colony on the slopesof and Prince 1981,Prince 1985).Analyses of high- Ulakaia Hill. The study area was describedin detail latitude zooplanktonindicate large variation in by Roby and Brink (1986a,b). Work on diving petrels was conducted on Bird Island (54ø00'S, 38ø02'W), lo- lipid content (Lee et al. 1972, Bensonand Lee cated at the western end of South Georgia, during 1975). Lipids constituteup to 77% of dry mass Januaryand February 1982.Locations of study colo- in somezooplankters (Lee 1974),which implies nies are describedin Roby and Ricklefs (1983). Nest that certain zooplanktondiets may be energy- burrowsand nestinghabitats of the two diving petrel rich but protein-deficient. Considering the speciesare describedby Payneand Prince (1979). variability in composition of high-latitude Nestswere checkeddaily during the hatching pe- zooplankton,prey selectionby adult plank- riod to determine the date of hatching, and known- ton-feedingseabirds may have important con- age chickswere periodicallyweighed and wing chord sequencesfor chick energy intake and the measuredduring the nestling period. A series of partitioning of assimilatedenergy among com- known-age chicks, in approximately equally spaced age classesfrom hatching to fledging, were removed peting functionsof maintenance,growth, and from their nestsaround midday. The actual date of fat deposition. hatching was known for only a few Common Diving If seabird growth is largely energy limited, Petrel chicks,so age was estimatedearly in the nest- and interspecificdietary differencesresult in ling period from the relationship of wing length to significantvariation in diet quality, then much age(Payne and Prince1979). Chicks were transported of the variation in seabird postnatal develop- to field laboratories where metabolic trials were con- ment would be a function of diet composition. ducted. Oxygen consumption was measured volu- Specifically,seabird chicks with relativelyhigh metricallyusing a temperature-controlled,closed sys- lipid diets would grow faster and fledge at a tem (Ricklefsand Roby 1983).Chicks were weighed younger age. Conversely, if chick growth is at the beginning and end of each respirometrytrial on an Ohaus triple beam balance(+0.1 g). Oxygen constrainedby factorsthat operateat the tissue consumptionrates were convertedto cm3.g-• .h • by level, differences in diet composition should dividing by the massof the subjectat the beginning bear little on development. High-lipid diets of the trial. would be reflected in higher rates of lipid de- After respirometrytrials, subjectswere humanely position by chicks,but not higher growth rates sacrificedby thoraciccompression, placed in plastic and shorter nestling periods. bags,and frozen at -20øC for later determination of January1991] SeabirdDiet and Chick Energetics 133 body composition.After partial thawing, eachcarcass sumed diving petrel chicks maintained SMR for 4 h was plucked, skinned, and dissectedinto 15 body each day when adults were normally in attendance parts. Feathers, skin, and body parts were placed in during the night. LeastAuklet chickswere brooded separatepreweighed aluminum pansand immediate- continuouslyuntil 4 dayspost-hatch and then inter- ly weighed on a SartoriousModel 2463 electronic mittently during the day until 18 days post-hatch balance (+0.001 g) to determine wet mass.Then each (Roby and Brink 1986a).I used data on age-specific componentwas oven-dried to constantmass at 50- broodingfrequency of LeastAuklet chicksto estimate 60øC.The dried skin and body parts were wrapped the proportion of the day at standardmetabolic rate separatelyin cotton gauze, and extractedfor 48 h in and I assumedthat all chicks,regardless of age, were each of three baths of 5:1 (v/v) petroleum ether and at SMR during the night, when both parents were chloroform.This solventsystem removes storage lip- normally in attendance.I made no attempt to incor- ids (triacylglycerols)and structural lipids (phospho- poratepotential energy costsof activity into estimates lipids), but very little nonlipid material (Hagan et al. of daily respired energy. Such costsare likely to be 1967,Dobush et al. 1985). In the caseof fatty carcasses, insignificant until nestlings begin to exercisetheir it was sometimesnecessary to increasethe number wings just before fledging. The energy equivalent of of successivebaths to as many as six before achieving oxygen was assumedto be 20.1 kJ/1 O2 consumed constantmass. Following extraction,carcasses were (Ricklefs 1974). oven-dried at 50-60øCfor 24 h to evaporate solvents, I calculatedthe energyequivalent of growth during and reweighed to determine massof lipid-free dry each phase of the nestling period from the energy matter (LFDM). Lipid masswas determined by sub- equivalent of accumulatedLFDM plus the estimated traction. cost of LFDM biosynthesis. I calculated the energy I estimated rates of accumulation of LFDM (includ- equivalentof lipid depositionfrom the energy equiv- ing skin and plumage)and lipid during the nestling alent of accumulatedlipids plus the estimatedcost of period for each speciesfrom changes in chick body adipose tissuebiosynthesis. I assumedthe following compositionas a function of age. Growth was defined combustibleenergy equivalentsof accumulatedtis- as accumulation of LFDM and did not include accu- sue:LFDM (includesprotein, carbohydrate,and skel- mulationof lipid. Dataon age-specifictotal body mass, etal tissue)= 20.5 kJ/g; lipid = 37.7 kJ/g (Ricklefs LFDM mass,and lipid masswere each fitted to linear, 1974). Biosynthesiswas estimated as 0.33 times the quadratic,cubic, logistic, and Gompertzmodels using energy equivalent of accumulated tissue (Ricklefs proceduresin the SAS package(SAS Institute, Inc., 1974). I calculatedthe energy expended for mainte- 1982). The model that minimized the residual mean nanceby subtractingthe total energy costof biosyn- squarewas used to estimatethe rate of accumulation thesis(LFDM + lipid) from the energy equivalent of of LFDM or lipid during each phaseof the nestling oxygen consumption. Total energy requirement was period. The predicted values from each model were the sum of growth, lipid deposition,and maintenance comparedwith the actual measurementsto checkfor requirements. biases in the distribution of residuals. The frequency of chick feeding in Least Auklets To estimate energy expenditure rates of chicks, I was measuredby continuousobservation of a small multiplied the average mass-specificrate of oxygen colonyof markedadults (Roby and Brink 1986a),while consumptionat eachage times the averagebody mass feeding frequencyof diving petrelswas determined at that age (Robyand Brink 1986a;Roby and Ricklefs by an overnight weighing technique (Roby 1989). I unpubl. data). Theseoxygen consumptionrates were estimateddaily food intake of chicksfrom the product adjusted for age-specificdifferences in the incidence of averagefrequency of meal delivery x averagechick of brooding. I assumedthat, when brooded by their meal mass.Chick meals from all three specieswere parents, chickswere within the thermal neutral zone collectedby mist-netting adults as they returned to and consumed oxygen at standard metabolic rate the colony to feed chicks.In the laboratory,the lipid (SMR). Standard metabolic rate for each chick was content of chick mealsas a percentageof wet mass assumed to be the minimum oxygen consumption was determined by the extractionand quantification rate recordedduring metabolictrials. I also assumed of lipid classes(wax esters,triacylglycerols, free fatty that unattended chickswere exposedto averageam- acids,phospholipids) from a subsampleof chick meals bient temperatures recorded in nest sites (7.8øC for (n = 14 LeastAuklet meals,n = 12 eachSouth Georgia LeastAuklets, 8.80C for SouthGeorgia Diving Petrels, and Common diving petrel meals) by methods de- 7.0øC for Common Diving Petrels). Oxygen con- scribedby Roby et al. (1986). A secondsubsample of sumption rates at these ambient temperatures were chick meals (n = 12 Least Auklet, n = 6 each South estimatedfrom the averageof rates measuredat 5øC Georgiaand Common diving petrel) was oven-dried and 10øC.I assumedthat diving petrel chickswere at to constant mass at 50-60øC to determine water con- SMR continuouslyuntil 6.5 days(South Georgia Div- tent. Average LFDM contentwas determinedby sub- ing Petrel) or 11 days(Common Diving Petrel) post- tractingaverage lipid contentand averagewater con- hatch,the averagelength of their respectivebrooding tent from average wet mass. Protein content was periods (Roby and Ricklefs 1983).Subsequently, I as- determined for the chick meals that were dried to 134 D,m•ELD. ROB¾ [Auk,Vol. 108 determine water content. Following extractionof lip- larger meal size compensatesfor lower fre- ids from the dried meals, total nitrogen content was quencyof mealdelivery. The averagedaily food determined by the Kjeldahl method at the Forage intake was lowest in Least Auklets (Table 1). Analysis Laboratory, New Bolton Center, University Daily food intake of chicksas a percentageof of Pennsylvania.I estimatedrates of lipid, LFDM, and adult mass was 42% in Least Auklets, 37% in protein intake (g/day) for each speciesfrom their averageproportions in chick meals (% of wet mass) SouthGeorgia Diving Petrels,and 33%in Com- times the average rate of food intake (g wet mass/ mon Diving Petrels (Table 1). day). The taxonomic composition of chick meals Carbohydrate, chitin, and ash contents of chick differed among the three species,and the bio- meals were not measured and were estimated from chemical composition of chick diets reflected published values. Carbohydrateis a minor constitu- those differences (Table 2). The Least Auklet ent in planktonic crustacea.For E. superba,published diet wasdominated by calanoidcopepods (95% valuesfor carbohydrateare only 5% (Raymontet al. of mealvolume on average).South Georgia Div- 1971) and 2.7% of LFDM (Clarke 1980). Ash was the ing Petrel chick mealswere dominatedby the major nonprotein constituentof LFDM: ash averaged 21.6% in three speciesof boreal euphausiids(Falk- euphausiidsThysanoessa sp. (63%)and Euphausia Petersen1981), and 19%and 18.3%in E. superba(Ray- superba(26%). Common Diving Petrel chick mont et al. 1971and Clarke 1980,respectively). Lipid- mealswere dominatedby E. superba(78%) with free dry matter in E. superbaconsisted of 5% chitin copepodssecond in importance.Because of the (Raymont et al. 1971) and 10.9%chitin (Clarke 1980). preponderance of calanoid copepods, Least Clarke (1980) did not account for 11.4% of LFDM in Auklet mealshad significantly higher lipid con- E. superba,but it probably consistedprimarily of free tent (t = 3.36, P < 0.05), and most of the lipid fatty acidsand other metabolicbreakdown products. was in the form of wax esters (Table 2). Chick For purposesof estimatingthe energycontent of chick meals of Common Diving Petrels consisted diets, I used the averageof published values to esti- mostlyof E. superbaand had significantlylower mate the compositionof nonprotein LFDM: 52% ash, lipid content (t = 5.43, P < 0.05) and little wax. 19%chitin, and 10%carbohydrate. ! assumedthat the unidentified portion of LFDM had the sameenergy Associatedwith the preponderanceof euphau- content as protein. siids in the diet, South Georgia and Common Chitin is generally consideredan indigestiblecom- diving petrel meals had significantlyhigher ponent of the diet, but there is growing evidence that protein levels than LeastAuklet meals (t = 10.20, among birds in general, and seabirds in particular, at P < 0.05). South Georgia Diving Petrel meals least a portion of dietary chitin is metabolized (Jeu- had relatively high average lipid content in ad- niaux and Cornelius 1978, Stemmler et al. 1984). Al- dition to high protein content, apparentlybe- though digestive efficiencies for chitin have been causethe main dietary item (Thysanoessa)had measuredat 40-60% (A. R. Placepers. comm.), for the larger lipid storesthan E. superba.The prepon- purposeof estimatingenergy contentof seabirddiets I assumedthat chitin was digested completely. To derance of Thysanoessasp. in the chick meals estimatethe energy contentof the diet, ! assumedthe of South Georgia Diving Petrels and their ab- following grossenergy equivalents:lipid = 39.5 kJ! sencein Common Diving Petrel mealsmay be g, protein = 18.0kJ/g, carbohydrate= 17.3kJ/g (Rick- a consequenceof differences in either selectiv- lefs 1974), chitin = 17.2 kJ/g (Clarke 1980). ity or availability. Common Diving Petrelsnest earlier than South Georgia Diving Petrels on RESULTS Bird Island, and although it is unlikely that E. superbais less available when South Georgia Nestlingdiets.--Parents of the three study spe- Diving Petrels raise their young, it is possible ciesmust supply all the prefledgingnutritional thatThys'anoessa isscarce during the chick-rear- requirements of their nestlings. Least Auklets ing period of Common Diving Petrels. are only active at the nestingcolony during the Differencesin lipid and protein content of day while diving petrelsare strictly nocturnal. mealsresulted in differencesin the energyden- Short nights at high latitudes have important sity of chickdiets. The averageenergy densities consequencesfor chick feeding frequency in of LeastAuklet (5.48 kJ/g wet mass)and South nocturnal seabirds. Least Auklet chicks were Georgia Diving Petrel chick (5.77 kJ/g) diets fed an average of 5.6 meals each day, while were similar. In Common Diving Petrels, en- diving petrel chickswere fed at mosttwo meals ergy density was significantly lower (4.16 kJ/g each night, one from each parent (Table 1). In wet mass, t = 4.294, P < 0.05). The average diving petrels and other nocturnal seabirds, energy density of Common Diving Petrel meals January1991] SeabirdDiet and Chick Energetics 135

T^BLœ1. Average food intake for Least Auklet, South Georgia Diving Petrel, and Common Diving Petrel chicks. Means are +SD; sample sizes are in parentheses.

South Georgia LeastAuklet Diving Petrel Common Diving Petrel Adult mass(g) 84.5 + 5.23 (44) 115.3 + 6.99 (24) 145.0 + 6.85 (32) Peak chick mass(g) 91.5 + 11.67(11) 142.0 + 15.71(52) 134.0 + 13.75(124) % of adult mass 108.3 123.1 92.4 Feedingfrequency (meals/day)_ 5.60+ 1.08 (10) 1.83+ 0.38 (78) 1.90+ 0.35 (78) Chick meal mass(g) 6.30 + 2.27 (28) 23.3 + 3.92 (24) 25.5 + 4.54 (32) % of adult mass 7.5 20.2 17.6 Food Intake (g/day) 35.3 + 14.42 42.6 + 11.39 48.5 + 12.41

was within the range of values reported for differencesin the contribution of lipid depo- fresh E. superba(3.8-5.5 kJ/g; Clarke and Prince sition to overall growth obscurethe pattern of 1980).Of the total estimatedenergy contentof LFDM accumulation. chick diets, lipid contributed69% in LeastAuk- Of the five models used to fit data on LFDM let meals, 60% in South Georgia Diving Petrel as a function of age, logistic equationsresulted meals,and 43%in CommonDiving Petrel meals. in the lowest residual mean square for each Chickgrowth.--Logistic models had the low- species.The growth rate constantfor the logis- est residual mean squarewhen fitted to data on tic equation fitted to LFDM data was highest in total body massas a function of age for eachof Least Auklets (smallestspecies) and lowest in the studyspecies (Table 3). The logisticgrowth Common Diving Petrels(largest species; Fig. 1, rate constant(K) was significantly greater and Table 3), but the 95% confidence limits about the inflection point (I) and asymptote(A) sig- the value of K overlappedfor the three species. nificantly lessfor LeastAuklets comparedwith Also, a trend of decreasinggrowth rate with the two diving petrel species(no overlapin 95% increasingasymptotic mass is expectedbased confidence limits). For the two diving petrel on the allometric scalingof growth rate to body species,K and I were not significantlydifferent size describedby Ricklefs (1979). Consequently, (P > 0.05), but the asymptoteof the fitted lo- growth rate of LFDM was similar among the gistic equation was significantly greater (no three study speciesonce K values were adjusted overlap in 95% confidence limits) for South for differencesin adult body mass.The differ- Georgia Diving Petrels.Although these results ence in the length of the nestling period be- suggestthat Least Auklets grow more rapidly tween LeastAuklets and the two diving petrels than either of the diving petrels, interspecific was primarily a function of differencesin as-

TAI•LE2. Average taxonomiccomposition (% volume) and biochemicalcomposition of chick diets in three speciesof high latitude plankton-feeding seabirds.Means are +SD; samplesizes are in parentheses.

South Georgia Least Auklet Diving Petrel Common Diving Petrel Taxonomic composition Calanoid copepods 94.8 ñ 3.2 (12) 3.5 ñ 11.5 (12) 21.0 ñ 20.6 (12) Euphausiasuperba -- 26.3 ñ 37.5 (12) 78.0 ñ 20.5 (12) Thysanoessasp. 2.4 ñ 1.96(12) 63.3 ñ 32.5 (12) 0.1 ñ 0.58 (12) Other Taxa 2.8 ñ 2.53(12) 5.5 ñ 7.0 (12) 1.4 ñ 1.73 (12) Biochemicalcomposition Water a 78.6 ñ 3.28(12) 75.8 ñ 1.00 (6) 79.4 ñ 1.41 (6) Total lipid b 45.5 ñ 8.33(14) 36.0 ñ 6.02 (12) 21.7 ñ 6.85 (12) Wax esters b 39.0 ñ 1.24(6) 14.5 ñ 6.76 (6) 5.4 ñ 3.94 (6) Protein b 34. ! ñ 3.12(12) 44.2 ñ 1.01 (6) 52.2 ñ 1.07 (6) Chitin and carbohydrate' 6.2 5.8 7.3 Ash and unrecovered c 15.6 14.1 18.8

• Percent fresh mass. bPercent of dry mass. ßEstimated from published values in the literature (see text). 136 DANIELD. ROBY [Auk, Vol. 108

TABLE3. Parametersof fitted equationsfor lipid-free dry matter (LFDM) massand lipid massas a function of age for nestlingsof LeastAuklet, SouthGeorgia Diving Petrel, and CommonDiving Petrel. Means are +SE: K is the rate constant(days-'), I is the inflectionpoint (days),and A is the asymptote(g).

Logistic parameter Residual K I A r2 mean square Least Auklet Total Mass 0.239 _+ 0.007 8.6 + 0.15 95.8 + 1.07 0.990 32.69 Lipid 0.215 + 0.039 14.5 _+1.3 20.9 + 1.8 0.964 5.49 LFDM 0.186 + 0.012 11.4 + 0.48 22.1 + 0.63 0.996 0.84 South Georgia Diving Petrel Total Mass 0.145 _+ 0.0036 14.9 + 0.30 147.5 + 1.93 0.985 73.38 Lipid 0.163 + 0.031 19.3 + 1.70 30.4 + 2.16 0.969 11.69 LFDM 0.153 + 0.011 16.3 _+ 0.76 31.1 + 0.94 0.994 2.61 Common Diving Petrel Total Mass 0.146 + 0.0081 15.6 + 0.37 138.5 + 1.71 0.985 190.48 LFDM 0.128 + 0.017 20.0 + 1.23 35.6 _+ 1.31 0.989 8.03

RegressionParameter Age F df P ta Lipid 0.39 + 0.051 57.6 1,23 <0.0001 0.715 ymptotic LFDM mass rather than differences Fitted logistic equations had the lowest re- in the accumulationrate of LFDM (Fig. 1). The sidual mean square for data on lipid accumu- maximum accumulation rates of LFDM at the lation in LeastAuklets and South Georgia Div- inflection point of the fitted curveswere similar ing Petrels(Table 3). The residualmean square for the three species(0.97 g/day for LeastAuk- for lipid content as a function of age in Com- lets, 1.10 g/day for South Georgia Diving Pe- mon Diving Petrels was minimized by a linear trels, and 1.07 g/day for Common Diving Pe- model (Table3, Fig. 2). The 95%confidence lim- trels). Predicted LFDM mass at fledging was its aboutthe growth rate constants(K) of lipid 86.6% of adult LFDM for Least Auklets, 91.6% accumulation in Least Auklets and South Geor- for adult LFDM for South Georgia Diving Pe- gia Diving Petrelsoverlapped considerably, and trels, and 84.4% of adult LFDM for Common maximum rates of lipid accumulation at the in- Diving Petrels. flection point were similar (1.0 and 1.1 g/day,

45 ÷

• 35

3o

r'• 25

m 20 -• •5

• lO • 5

I- o JI•g• •i•i• [:g sLe•••:ømmøn '11vPvie•rgelPetre' 0 10 20 30 40 50 Adults

Age (days post-hatching) Fig. 1. Age-specificlipid-free dry matter massfor nestling LeastAuklets (a), South GeorgiaDiving Petrels (b), and Common Diving Petrels (c). Curves represent logistic equationsfitted to the data for each species. Values for adults are averages(n = 6) for each species. January1991] SeabirdDiet and Chick Energetics 137

40' 35'[] Least Auklet ß SouthPetrelGeorgia Diving J ß * ß

•• 25'30' •' Common D• o • b

0 I I

Age (days post-hatching) Fiõ. 2. A•e-speci•c tipid aass {or •estli• LeastAu•lets (a), SouthOeor•ia Divi• •etrets (c). Curvesrepresent 1o•istic eguado•s fitted to the data {tom •estt[• LeastAu•tets a•d South Georgia Divi• •etrels, a•d the teastsguares re•ressio• {or data {tom •estti• Commo• Divi• •etrels. respectively). The predicted lipid content of gy were similar in the three study speciesdur- South Georgia Diving Petrelsat fledging (30 g) ing the first 5 days after hatch, when chicksof was 50% higher than in Least Auklets (20 g). all specieswere brooded continuously (Fig. 3). However, the ratio of lipid massto LFDM mass After day 5 the curvesdiverged, primarily as a in fledglingswas similar in the two species(Least consequence of interspecific differences in Auklet: 0.91,South Georgia Diving Petrel:1.12). length of the brooding period. At day 8 post- Common Diving Petrel chicksaccumulated lip- hatch, South GeorgiaDiving Petrel chickswere id at a significantlylower rate than the other broodedfor approximately17% of the day, Least two species(slope = 0.39 g/day, SE = 0.051) Auklet chicksfor approximately79%, and Com- and, despitelarger fledgling massand a nesting mon Diving Petrel chicks for the entire 24-h periodalmost twice aslong asthat of LeastAuk- period. Estimateddaily energy expenditure for lets,fledged with similar lipid stores(21 g). The diving petrel chicksincreased sharply when ratio of lipid massto LFDM massin Common continuousbrooding ended. Unlike the diving Diving Petrelsat fledging was only 0.50. In all petrels,the incidenceof brooding in LeastAuk- three species,average fledgling lipid reserves lets declinedgradually and estimateddaily en- were considerablyhigher than thoseof adults; ergy expenditure did not increasesharply. lipid contentof adultsduring the chick-rearing In all three species,daily respired energy in- period averaged4.9 g for LeastAuklets, 11.5 g creasedduring the first half of the nestlingpe- for South Georgia Diving Petrels,and 8.7 g for riod and remained approximately constant Common Diving Petrels, only 25-41% that of thereafter(Fig. 3). LeastAuklet chicksreached their respectivefledglings. an asymptoticrespiration rate at approximately In the three species,residual mean square 120kJ/day, South Georgia Diving Petrelchicks values indicated that measurements of accu- at 155kJ/day, and CommonDiving Petrel chicks mulatedlipid varied more aboutthe fitted curves at 145 kJ/day. than measurements of accumulated LFDM (Ta- Energy budgetsof nestlings.--The energy ble 3) (i.e. accumulation rates of LFDM varied equivalentof growth (lipid-free dry matterplus within narrow limits). This supportsthe con- biosynthesis) for Least Auklets increased to a clusionthat lipid depositionwas dependenton maximum of 27 kJ/day at day 11 post-hatch the level of surplusdietary energy once main- (18% of total energy budget) and then declined tenance and LFDM accumulation requirements graduallyuntil fledging.Growth declinedfrom had been met. a maximumof 33%of total energybudget short- Respiredenergy.--Estimates of respired ener- ly after hatch to approximately 5% at fledging 138 DANIELD. ROBY [Auk,Vol. 108

•40'

120'

100'

60'

40' SouthGeorgia Diving Petrel 20' CommonDiwngPetrel

o 0 s 10 lS 20 2s 30 35 40 45 so ss

Age (days post-hatching) Fig. 3. Age-specificdaily respiredenergy for nestlingLeast Auklets, South Georgia Diving Petrels,and CommonDiving Petrels,calculated from temperature-specificoxygen consumption rates and parentalbrood- ing schedules.

(Fig. 4a). In South Georgia Diving Petrels, en- marily due to lower rate of lipid deposition. ergy equivalent of growth peaked at 31 kJ/day Maximum energy requirement for fat deposi- on day 14 (17%of total budget). As a proportion tion coincidedwith high energy requirements of total daily energy budget, growth declined for maintenance and growth in Least Auklets from 35% shortly after hatching to about 1%at and South GeorgiaDiving Petrels.The peak in fledging (Fig. 4b). In Common Diving Petrels, total daily energy requirementwas less pro- the energy equivalent of growth peaked at 31 nounced in Common Diving Petrels,which de- kJ/day on day 21 (19%of total budget).Growth positedlipid at a constantrate throughout the declined from 29% of total energy budget at nestling period (Fig. 4: a-c). hatching to 1% at fledging (Fig. 4c). Estimatedtotal energy intake of chicks was The energy equivalent of lipid accumulation 193 kJ/day for LeastAuklets, 246 kJ/day for was twice the energy equivalent of growth dur- South GeorgiaDiving Petrels,and 202 kJ/day ing the latter half of the nestling period in Least for CommonDiving Petrels(Table 4). Estimates Auklets and South GeorgiaDiving Petrels(Fig. of maximumenergy requirements for the three 4: a, b). In contrast, the energy equivalent of species were 91%, 83%, and 88% of estimated lipid accumulationin Common Diving Petrel totalenergy intake, respectively. Consequently, nestlingswas less than that for growth during interspecificdifferences in estimateddaily en- much of the nestling period (Fig. 4c). ergy intake were matchedby differencesin es- Estimatesof total daily energy requirements timated daily energy requirement.If parentsof peaked at 176 kJ/day in Least Auklets, at 204 the three study speciescannot increasethe rate kJ/day in South Georgia Diving Petrels,and at of energy delivery to the nest, then chick en- 178 kJ/day in Common Diving Petrels.Despite ergy budgetsare apparentlyconstrained by av- the similarity in peak nestling massof South erage energy intake. Georgia Diving Petrels (142 g) and Common Diving Petrels (134 g), estimated maximum en- DISCUSSION ergy requirementwas 15%higher in the former. The peak nestling massfor LeastAuklets (91.5 Diet composition.--Waxesters are a readily as- g) was 32% less than for Common Diving Pe- similable component of chick diets (Place and trels, yet maximum energy requirementswere Roby 1986, Roby et al. 1986). Wax esters com- virtually identical. The low relative energy re- prised 86% of lipids in the diet of LeastAuklet quirement of Common Diving Petrels was pri- chicks,equivalent to 115 kJ/day. This is more January 1991] SeabirdDiet and ChickEnergetics 139

a 200 '

Lipid 160 ' LFDM 140' Maintenance

100'

80'

40'--

0 1 3 5 8 11 14 17 20 23 26

b 220 -

200 '

120'

100'

80-

60-

40'

20'

0 1 3 5 8 11 14 17 20.5 24,5 29 33 37 40 43

c 180'

160'

140'

• 100 • 8o

C 60 ILl 40

20

0 , 1 3 5.5 9.5 11.5 14.5 17 21 24 27 30 33 35.5 40 45.5 49.5 53.5

Age (days post-hatching) Fig. 4a-c. Age-specificdaily energy budgetsof LeastAuklet (a), South GeorgiaDiving Petrel (b), and CommonDiving Petrel (c) chickscalculated from ratesof accumulationof lipid and lipid-free dry matter (LFDM) and daily respiredenergy. Values for energyequivalents of lipid and LFDM accumulationinclude the estimatedenergy costof biosynthesis. 140 DANIELD. ROBY [Auk, Vol. 108

TABLE4. Calculationsfor averagedaily energy intake of nestlingLeast Auklets, South Georgia Diving Petrels, and Common Diving Petrels. Means are +SD.

South Georgia LeastAuklet Diving Petrel Common Diving Petrel Lipid a Content (% of wet mass) 9.6 + 1.75 8.7 + 1.46 4.5 + 1.43 Intake (g/day) 3.39 + 1.52 3.71 + 1.17 2.18 + 0.89 Energy (kJ/day) 133.9 + 60.04 146.5 + 46.22 86.1 + 35.16 Protein b Content (% of wet mass) 7.2 -+ 0.66 10.7 + 0.24 10.8 + 0.22 Intake (g/day) 2.54 + 1.06 4.56 + 1.22 5.24 + 1.34 Energy (kJ/day) 45.9 + 19.08 82.3 + 21.96 94.1 + 24.12 Chitin and carbohydratec Content (% of wet mass) 1.3 1.4 1.5 Intake (g/day) 0.47 + 0.188 0.59 + 0.160 0.74 + 0.186 Energy (kJ/day) 8.1 + 3.23 10.1 + 2.75 12.7 + 3.20 Total energy Intake (kJ/day) 193 + 63.1 246 + 51.25 202 + 42.,"6 Energy equivalentof lipid = 39.5 kJ/g. Energyequivalent of protein= 18.0kJ/g. Estimatedfrom publishedvalues; energy content of chitin and carbohydrate= 17.2 kJ/g.

than sufficient to meet the peak maintenance haug 1980, Evans 1981, Roby et al. 1981). Like requirements of nestlings (112 kJ/day). Ap- Least Auklets, Dovekie chicks deposit large fat proximately 60% of the estimated energy con- reserves(21 g) before fledging (Taylor and Ko- tent of chick meals was in the form of wax es- narzewski 1989). Parakeet Auklets (Cyclorrhyn- ters, and 65% of estimated assimilated energy chuspsittacula) and CrestedAuklets (Aethiacris- was from wax esters. Adult Least Auklets spe- tatella)are large planktivorousauklets (250-300 cialize on calanoidcopepods during the breed- g), and their diet consistsof approximately20% ing season(B•dard 1969) and the high energy and 10% of the larger calanoid copepods,re- requirementsof chicksare met by the high lipid spectively(B•dard 1969). In the SouthernOcean, content of calanoidcopepods. The ability to me- Euphausiasuperba replaces calanoid copepods as tabolize wax estersis an essentialadaptation for the dominant grazing zooplankter,and no small utilization of this food source. planktivorousseabirds are known to specialize Approximately40% of dietarylipids for South on calanoid copepods. Georgia Diving Petrels and 25% for Common The lipid content of LeastAuklet meals(9.6% Diving Petrelsconsisted of wax esters.This is of wet mass)is high comparedwith other plank- equivalent to a daily intake of 59 kJ (29% of ton diets. However, the lipid content of cal- total energy intake) for South Georgia Diving anoid copepodsis exceededby other nonplank- Petrel chicks and 21 kJ (12%) for Common Div- tonicseabird prey (e.g.herring, Clupeaharengus, ing Petrel chicks.Dietary wax esterssupplied is 13.6%lipid, and mackerel, Scornberscornbrus, more than the energy equivalent of deposited is 18.6% lipid; Montevecchi et al. 1984). The lipid plus the energy costsof adipose tissue lipid content (4%) and energy density (3.89 kJ/ biosynthesisin both diving petrel species.Al- g) of krill (Clarke 1980) is similar to that of though less prevalent in diving petrel chick capelin (Mallotus villosus;3.3% lipid and 3.99 meals, wax esters nevertheless provide an im- kJ/g energy density), an important food source portant,highly assimilablesource of energyfor for piscivorousseabirds breeding in the North both taxa of seabirds. Atlantic region (Montevecchi and Piatt 1984, The small size of calanoidcopepods (4-8 mm) Montevecchi et al. 1984). Thus the lipid con- may limit their suitability as high-energy prey tents and energy densities of high latitude for all but the smallestplanktivorous seabirds. planktonic crustaceafall within the range of Dovekies (Alle alle), the smallestNorth Atlantic values reported for important fish prey of sea- alcid, specializeon calanoidcopepods (Norder- birds. January1991] SeabirdDiet and Chick Energetics 141

Protein content was only 7.2% of wet massin lower energy requirements for either mainte- auklet meals, comparedwith 10.7%in diving nance or growth. Instead, lipid depositionwas petrel meals. The protein index (100 x g pro- enhancedby the high lipid contentof the diet. tein/kJ) for auklet chick meals was 1.3, com- The diet of Common Diving Petrels was dom- pared with 1.9 in South Georgia Diving Petrel inated by E. superbaand had a lower lipid con- meals,and 2.6 in Common Diving Petrel meals. tent and a lower energy density than the diets Protein content for a variety of nonplankton of the other studyspecies. The energysupplied seabirdprey (fish and squid) rangesfrom 14- by dietary lipids in LeastAuklets (134 kJ/day) 20%of wet mass,higher than either the auklet and in South Georgia Diving Petrels(147 kJ/ or diving petrel meals,and protein indicesrange day) was sufficientto meet the maximum com- from 1.4-3.7 (Clarke and Prince 1980, Monte- bined energy requirements for both mainte- vecchi et al. 1984). The low protein content and nance and biosynthesis (128 kJ/day in Least high lipid contentof high latitudecalanold co- Auklets and 146 kJ/day in South Georgia Div- pepodsraises the possibilityof a dietary protein ing Petrels).Presumably, once growth require- deficiencyfor growing Least Auklets. Never- ments for protein had been met, any additional theless,the growth rate of LeastAuklet chicks assimilated protein was metabolized to meet at the inflection point was 1.0 g LFDM/day, maintenance costs,and the energy saved was much less than the estimated 2.55 g of crude stored as fat. In Common Diving Petrels, the protein that chicks ingested on a daily basis. energy supplied by dietary lipids (86 kJ/day) (Williams et al. [1977] estimatedthat LFDM was was only 57% of the respired energy (mainte- 94%protein in penguins.)Whereas estimates of nance and biosynthesis).The difference(65 kJ/ crude protein from measurementsof total ni- day) must have been supplied by catabolizing trogen may overestimateactual protein content other diet components,primarily protein. The (Clarke 1980),the assimilationefficiency for di- important conclusionis that, even if Common etary protein is probably closeto 95% (Robbins Diving Petrel chicks assimilatedall the avail- 1983).This indicatesa dietary surplusof protein able energy in the form of lipid, carbohydrate, and suggeststhat more than half of assimilated and chitin, approximately half of the dietary protein was metabolizedby LeastAuklet chicks protein must have been metabolized to meet as an energy source.Also, LeastAuklet chicks energy requirementsfor maintenanceand bio- grew at a similar rate comparedwith the diving synthesis. petrel chicks despite higher protein levels in The nestlingsof all three speciesaccumulated the diets of the latter (Table 3). I believe that substantiallipid reservesby fledging age.Both dietary protein does not limit growth in the LeastAuklets and SouthGeorgia Diving Petrels three study species. had sufficientenergy reservesat fledging to sur- Diet andchick lipid reserves.--I have shown,by vive in the nest ca. 1 week without food. carcassanalysis, that Least Auklets and South Common Diving Petrels could have survived Georgia Diving Petrels deposited lipid at sim- ca. 5-6 days without food. Lack (1968) and ilar ratesand fledgedwith similar lipid reserves O'Connor (1978) suggestedthat nestling lipid relative to LFDM mass. However, Common reservesare adjustedto variation in feeding fre- Diving Petrel nestlings accumulatedlipid at a quencyand that the large fat depotsof seabird lower rate than either of the other study species chicks insure survival during periods when and, despite a longer nestling period, fledged adults fail to deliver food. The energy reserves with lower lipid reserves. of the three speciesI studied were far in excess Lipid deposition rates are presumably deter- of those required for chicks normally fed on a mined by the daily surplus of assimilateden- daily basis.The large fat depotsof LeastAuklets ergy, once maintanenceand growth require- relative to Common Diving Petrels were par- ments have been met. Consequently, ticularly unexpected,given the high frequency interspecificdifferences in lipid deposition rate of feeding in Least Auklets. Similarly, Leach's were a function of interspecific differences in Storm-Petrelchicks accumulate large fat depots energy intake, maintenance requirements, (up to 30 g) sufficientto meet maintenancere- growth requirements,or somecombination of quirementsfor >2 weeks(Ricklefs et al. 1980a), thesefactors. I haveshown that the higher lipid but intervals between feedings rarely exceed3 deposition rates of Least Auklets and South days (Ricklefset al. 1985).Instead, the charac- Georgia Diving Petrels were not a function of teristic large fledging fat reservesof many sea- 142 DANIELD. ROB¾ [Auk, Vol. 108 birds may be an adaptation to enhance post- Petrels (Fig. 2). By the end of the brooding pe- fledging survival. riod (7 days), South Georgia Diving Petrels had Both auklets and diving petrels receive no accumulated sufficient fat reserves to survive postfledgingparental care, and large energyre- for more than a day without food. Common servesmay be important for postfledging sur- Diving Petrels did not accumulatecomparable vival. No data are available on the relationship energy reserves until approximately 11 days of fledgling fat reservesto postfledgingsurviv- post-hatch, when brooding ceasedin this spe- al, but somestudies indicate that fledgling mass cies. The length of the brooding period appar- is positively correlated with postfledging sur- ently matchesthe time required for the chick vival (Perrins 1966, Perrins et al. 1973, Jarvis to accumulatesufficient energy reservesto sur- 1974). The large prefledging fat storesof other vive the period betweenmeals without a parent seabirdsthat receive no postfledging parental in attendance.The longer brooding period of feeding suggest that energy reserves play an Common Diving Petrelswas necessitatedby the important role in postfledgingsurvival (Burger lower energy intake rates of chicks, which in 1980, Ricklefs et al. 1980a, Montevecchi et al. turn was a function of the lower lipid content 1984). In seabirds with extended periods of of the diet. postfledgingparental care, lipid reservesat Dietary energy and growth.--Ricklefs (1979, fledging are substantiallysmaller (Burger 1980, 1984) has shown that most of the variation in Shea 1985). avian growth rate can be explained by differ- Diet and broodingperiod.--A major problem in encesin adult body mass.Much of the residual the rearing of pelagic seabirdchicks is the tran- variation in growth rate is a function of mode sition to thermal independence,particularly in of development,with young that are altricial at speciesthat breed in cold climates.The chick- hatch growing at a higher rate than precocial brooding period imposesmaximum demands young. The three study speciesshare a semi- on food acquisitionby adults, which must ob- precocialmode of development.They are down- tain sufficient food for (1) the round-trip be- covered at hatch, are capable of maintaining tween feeding area and nest site, (2) mainte- body temperatureat an early age, and remain nanceand chick-broodingduring the brooding in the nest until fully grown. Consequently, shift, and (3) the growing chick (Ricklefs 1983). once growth rate is adjustedfor differencesin This implies that adults should seek to termi- adult body mass, differences in growth rate nate brooding as early as possible. The transi- among the study specieswould be expectedto tion to thermal independence is particularly be largely a function of dietary differences. abrupt for the young of seabirdsthat visit the While the higher lipid contentof LeastAuklet nest site only at night, such as diving petrels. and South Georgia Diving Petrel chick meals During the day, a parent can either remain in wasresponsible for higherlipid depositionrates the nest and brood its chick, thus limiting to comparedwith CommonDiving Petrels,growth one the number of meals delivered, or it can rate (accumulationof LFDM) was not signifi- feed at sea and leave the nestling unattended cantly higher. This indicates that differencesin for the entire day. South GeorgiaDiving Petrel energy intake were not reflected in differences chickswere broodedcontinuously until 6-7 days in the nonlipid componentsof growth. old and CommonDiving Petrel chicksuntil 11- Nevertheless, the low energy content of 12 days old, which matched closely the respec- Common Diving Petrel chick meals may pro- tive periods required for developmentof ho- long the nestling period by failing to supply a meothermy (Ricklefs and Roby 1983). sufficient daily surplus of energy to meet min- Chicks must store sufficient energy reserves imal lipid depositionrates. A large proportion to supportmaintenance costs between feedings of dietary protein mustbe metabolizedby Com- before they can be left unattended in the nest. mon Diving Petrel chicks to meet maintenance These energy reserves can be expressedas es- energy requirements,because of the low lipid timated survival time (i.e. the period that stored content of the diet. Common Diving Petrels lipids could provide maintenancerequirements fledge at a lower percentageof adult masscom- in the absenceof brooding). During the first pared with South Georgia Diving Petrels and week post-hatch,South GeorgiaDiving Petrels little additional growth occursduring the last accumulatedlipid reservesat a rate ca.0.55 g/day 10 days of the nestling period (Figs. 1 and 4c). comparedwith 0.39 g/day in Common Diving On the other hand, lipid depositionrate is ap- January1991] SeabirdDiet and Chick Energetics 143 proximately constant throughout the nestling straintson postnataldevelopment. This results period (Fig. 2), and a major advantage to the in (1) earlier cessationof brooding due to higher longer nestling period of Common Diving Pe- rates of lipid deposition in neonates and the trels may be the additional accumulation of fat consequentfreeing of adults to forage full-time, reservesprior to fledging. (2) shorter nestling periodsdue to sufficientdi- Postnataldevelopment in aukletsand diving etary lipid to meet maintenance and biosyn- petrels may not be energy-limited if adults can thesis requirements,and (3) larger chick lipid provide additional food for their chicks.Twin- reservesat fledging. ning experiments (Nelson 1964, Harris 1970) The postnataldevelopment of AntarcticPrions and cross-fostering experiments (Shea and (Pachyptiladesolata) and Blue Petrels (Halobaena Ricklefs 1985) indicate that some seabirds re- caerulea)supports the hypothesis that supple- spond to increasednestling demand with in- mentation of plankton diets with stomach oils creasedfood delivery rates. However, there are results in shorter nestling periods. Antarctic indicationsthat the three studyspecies feed their Prions feed their young mostly E. superba,but chicks at closeto the maximum rate. Twinning chick meals include on average 20% stomach experimentswith alcid specieshave invariably oils (Prince 1980b). The chicks are brooded for resulted in failure to successfullyraise both only 0-5 dayspost-hatch (Tickell 1962)and meet young to normal fledging mass (reviewed in the energy demands of homeothermy by me- Gaston 1985). Also, the considerable intercolo- tabolizing stomachoils supplied by the parents ny variation in growth ratesand fledgling mass (Simonsand Whittow 1984).Despite larger adult in somealcid species(reviewed in Gaston 1985) mass (160 g), Antarctic Prions have a shorter suggeststhat food supply is frequently insuf- average nestling period (50 days;Tickell 1962) ficient to support maximal growth. For noctur- than Common Diving Petrels. Nestlings de- nal seabirds,like diving petrels, it is probably posit lipid at higher rates (1.5 g lipid/day; D. impossible for adults to deliver more than one D. Roby unpubl. data) than do diving petrel chick meal each night. Compared with other chicks,and fledging massaverages 13% higher procellariiforms,diving petrel chick meals are than adult mass (Tickell 1962), which indicates larger (Roby 1989) and wing loadings higher that chicks fledge with considerable fat re- (Warham 1977a).This suggeststhat meal volume serves. Blue Petrels are larger than Antarctic is constrained. Finally, the adults of the three Prions (averageadult mass:193 g), yet the nest- study speciesexhibit relatively high free-rang- ling period is shorter and chicks fledge at a ing metabolicrates during the chick-rearingpe- larger massrelative to adults. Blue Petrel chick riod (Roby and Ricklefs 1986). This indicates meals consist of approximately equal propor- that parents forage at close to a maximum rate. tions of E. superbaand fish, but the proportion Consequently,planktivorous seabirdsthat for- of stomachoils in chick mealsis higher than in age farther offshoreand cannotfeed their young Antarctic Prions (39% of wet masson average; each night must either provide meals with a Prince 1980b). higher energy density or reduce energy re- However, many procellariiform seabirdsthat quirements at the nest. feed their chicks stomach oils do not exhibit Auklets and diving petrels feed their young high growth ratesand short nestling periods. essentially fresh food, and parents do not en- For example, Leach'sStorm-Petrels are smaller hance the lipid content of chick meals by se- than the speciesin this study (45 g), but growth lectively assimilatingor excreting nonlipid di- rates are considerably lower (maximum accu- etary components.However, all procellariiform mulation rate of LFDM = 0.25 g/day) and nest- seabirdsother than diving petrels form stomach ling periodsare longer (65-70 days;Ricklefs et oils in the proventriculusduring foragingbouts al. 1980b). Nestlings deposit up to 35 g of fat, at sea, and they feed their young a diet with a a portion of which must be metabolizedbefore higher energy content than fresh food (War- going to sea (Ricklefs et al. 1980a). Postnatal ham 1977b). The formation of stomach oils al- development in Leach'sStorm-Petrel is clearly lows parent seabirdsto consumefoods, such as not energy-limited. Instead, tissue-level con- the abundant E. superba,with a low lipid-to- straints apparently limit growth to levels far protein ratio relative to the requirementsof their below what energy intake will permit. Addi- young (Obst 1986). Chick meals that contain tional experimentalwork on slow-growing sea- stomach oils potentially avoid energy con- birds is needed in order to elucidate the relative 144 DANIELD. ROBY [Auk,Vol. 108 contribution of diet and tissue constraints for growth. Further studies of feeding rates and limiting growth. diet composition in relation to chick energy My study supports the hypothesis that pat- budgets will enhance our understanding of the terns of postnataldevelopment in seabirdsare ecologicalfactors that shape observedpatterns related to diet. Differences in diet composition, in seabird reproduction. specificallylipid content,among the studyspe- cies were associated with differences in chick ACKNOWLEDGMENTS energy budgets and development. Energy budgetsindicated that maintenanceenergy re- Logisticsupport for this researchwas provided by quirements of chicks were high relative to en- the British Antarctic Survey and the Pribilof Islands ergy provided by the parents. Parentswere ap- Program, National Marine Fisheries Service. I am parently providing food to their nestlings at or grateful to K. L. Brink, R. H. Day, J. A. L. Hector, B. near the maximum rate. E. Lawhead, P. A. Prince, and R. E. Ricklefs for assis- tancein the field; and to the Royal Navy for retrieving Calanoid copepods,which formed the bulk my samplessafely from war-torn South Georgia. I of the Least Auklet diet, contained considerable thank T. E. Bowman (Smithsonian Institution) and P. quantities of wax esters,which provided most Ward (British Antarctic Survey) for identification of of the energy requirementsof chicks.The diet zooplanktonspecimens. A. R. Place provided valu- of Common Diving Petrel chickswas relatively able information on laboratoryanalyses. S. Abraham low in dietary lipid and considerabledietary and W. Blaufausassisted with laboratoryanalyses and protein was metabolized in order to meet main- K. Eberle assistedwith data analysis.R. J. Gates,J. C. tenanceenergy requirements.The study species Haney, A. R. Place,R. E. Ricklefs,R. E. Shea,and T. lack postfiedging parental care; large lipid re- C. Tachaoffered helpful comments,which improved serves at fiedging presumably enhance post- earlier drafts. This researchwas supportedin part by the National Science Foundation (USA): DPP80-21251 fiedging survival. and DPP82-17608to R. E. Ricklefs and by a grant from While differences in diet composition were the Shell Foundation. sufficient to explain much of the interspecific variation in chick development in these sea- birds, there was no significant interspecific dif- LITERATURE CITED ference in accumulation rate of LFDM. This supports the conclusion that growth (accumu- ASHMOLE,N. P. 1963. The regulation of numbersof lation of LFDM) is not constrained by energy tropical seabirds.Ibis 103:458-473. intake. High lipid diets were associatedwith 1971. Seabirdecology and the marine en- higher rates of lipid deposition by chicks, but vironment. Pp. 223-286 in Avian biology, vol. I not higher growth rates. Instead, constraints (D. S. Farner and J. R. King, Eds.). New York, operating at the level of tissuesare apparently Academic Press. responsiblefor most of the variation in growth BfiDARD,J. 1969. Feeding of the Crested, Least, and rate among seabirds. Parakeet auklets on St. Lawrence Island, Alaska. The formation of stomach oils enables breed- Can. J. Zool. 47: 1025-1050. BENSON,A. A., & R. F. LEE. 1975. The role of wax in ing procellariiform seabirds(except diving pe- oceanic food chains. Sci. Am. 232: 76-86. trels) to enhance the lipid content of the diet BURGER,J. 1980. The transitionto independenceand and supply the high energy requirements of postfiedgingparental care in seabirds.Pp. 367- their young. However, many seabirdsthat feed 447 in Behavior of marine animals. Vol. 4. Marine their young stomach oils have lower growth birds (J. Burger, B. Olla, and H. E. Winn, Eds.). rates and longer nestling periods, indicating New York, Plenum Press. that postnatal development is frequently not CASE,T.J. 1978. On the evolution and significance energy-limited. Twinning and cross-fostering of postnatal growth rates in the terrestrial ver- experiments also indicate that some seabirdsare tebrates. Q. Rev. Biol. 55: 243-282. capableof increasingthe rate of energydelivery CI•RKE, A. 1980. The biochemicalcomposition of krill, Euphausiasuperba Dana, from SouthGeorgia. to the nest, suggestingthat not all seabirdsare J. Exp. Mar. Biol. Ecol. 43: 221-236. foraging at the maximum rate when feeding -, & P. A. PRINCE.1980. Chemical composition young. Apparently, tissue-level constraintsare and calorific value of food fed to mollymauk the primary factorlimiting growth rate in many chicksDiomedea melanophris and D. chrysostomaat seabird species,while others may experience Bird Island, SouthGeorgia. Ibis 122:488-494. energy intake constraintsthat militate maximal DOBUSH,G. R., C. D. ANKNEY, & D. G. KREMENTZ. 1985. January1991] SeabirdDiet and Chick Energetics 145

The effect of apparatus,extraction time, and sol- bird predators.Comp. Blochem. Physiol. 78A: 15- vent type on lipid extractionsof Snow Geese.Can. 20. J. Zool. 63: 1917-1920. --, R. E. RICKLEFS,I. R. KIRKHAM, & D. GABALDON. DRENT,R. H., & $. DA•N. 1980. The prudent parent: 1984. Growth energeticsof Northern Gannets energetic adjustmentsin avian breeding. Ardea (Sula bassanus).Auk 101: 334-341. 68: 225-252. NELSON,J. B. 1964. Factorsinfluencing clutch-size EVANS,P. G. H. 1981. Ecologyand behaviourof the and chick growth in the North Atlantic Gannet Little Auk Alle alle in West Greenland. Ibis 123: Sula bassana. Ibis 106: 33-77. 1-18. 1977. Somerelationships between food and FALK-lVETERSEN,S.1981. Ecologicalinvestigations on breedingin the marine Pelecaniformes.Pp. 77- the zooplankton community of Balsfjorden, 87 in Evolutionaryecology (B. Stonehouse and C. northern Norway: seasonalchanges in body Perrins, Eds.). London, Macmillan. weightand the mainbiochemical composition of NORDERHAUG,M. 1980. Breedingbiology of the Lit- Thysanoessainermis (Kroyer), T. raschii(M. Sars) tle Auk (Plautusalle) in Svalbard. Oslo, Norsk and Meganyctiphanesnorvegica (M. Sars)in rela- Polarinst. Skrifter 173. tion to environmental parameters.J. Exp. Mar. OBST,B.S. 1986. The energeticsof Wilson's Storm- Biol. Ecol. 49: 103-120. Petrel (Oceanitesoceanicus) breeding at Palmer GASTON,g. J. 1985. Developmentof the young in Station, Antarctica. Ph.D. dissertation, Los An- the Atlantic Alcidae. Pp. 319-354 in The Atlantic geles, Univ. California. Alcidae (D. N. Nettleship and T. R. Birkhead, O'CONNOR,R.J. 1978. Growth strategiesof nestling Eds.). New York, Academic Press. passerines.Living Bird 16: 209-238. HAGAN, $. N., E. W. MURPHY, & L. M. SHELLEY. 1967. ß 1984. The growth and developmentof birds. Extractionof lipids from raw beef lean by using New York, Wiley and Sons. various solvent systems.J. Assoc.Official Anal. PAYNE, M. R., & P. A. PRINCE. 1979. Identification Chem. 50: 250-255. and breeding biology of the diving petrels Pe- HARRIS,M.P. 1970. Breedingecology of the Swal- lecanoidesgeorgicus and P. urinatrixexsul at South low-tailed Gull, Creagrusfurcatus. Auk 87: 215- Georgia. New Zealand J. Zool. 6: 299-318. 243. PERRINS,C.M. 1966. Survivalof young Manx Shear- 1977. Comparativeecology of seabirdsin waters Puffinuspuffinus in relation to their pre- the GalapagosArchipelago. Pp. 65-76 in Evolu- sumed date of hatching. Ibis 108: 132-135. tionary ecology (B. Stonehouse and C. Perrins, --, M.P. HARRIS, & C. K. BRITTON. 1973. Survival Eds.). London, Macmillan. of Manx ShearwatersPuffinus puffinus. Ibis 115: HICKEY, J. J., & F. L. CRAIGHEAD. 1977. A census of 535-548. seabirdson the Pribilof Islands. Pp. 96-195 in PLACE,A. R., & D. D. ROBY. 1986. Assimilation and Environmental assessment of the Alaskan con- depositionof dietary fatty alcoholsin Leach's tinental shelf, annual reportsof principal inves- Storm-Petrel,Oceanodroma leucorhoa. J.Exp. Zool. tigators, vol. 2. Boulder, Colorado, Natl. Oceanic 240: 149-161. & AtmosphericAdministration / Environ. Res.Lab. PRINCE,P.A. 1980a. The foodand feedingecology JARVIS,M. J.F. 1974. The ecologicalsignificance of of Grey-headedAlbatross Diomedea chrysostoma chick size in the South African Gannet Sula ca- and Black-browedAlbatross D. melanophris.Ibis pensis(Lichtenstein). J. Anim. Ecol. 43: 1-17. 122: 476-488. JEUNIAUX,C., & C. CORNELIUS. 1978. Distribution and ß 1980b.The food and feeding ecology of Blue activity of chitinolytic enzymesin the digestive Petrel (Halobaenacaerulea) and Dove Prion (Pa- tract of birds and mammals. Proc. First Int. Conf. chyptiladesolata). J. Zool. London 190: 59-76. on Chitin/chitosan.(M.I.T. Sea Grant Program 1985. Populationand energeticaspects of MITSG 78-7, index 78-307-Dmb.) therelationship between Blackbrowed and Grey- LACK,D. 1968. Ecologicaladaptations for breeding headed albatrosses and the southern ocean ma- in birds. London, Methuen. rine environment. Pp. 473-477 in Antarctic nu- LEE,R. F. 1974. Lipid compositionof the copepod trient cyclesand food webs (W. R. Siegfried,P. Calanushyperboreas from the Arctic Ocean: changes R. Condy, and R. M. Laws,Eds.). Berlin, Spring- with depth and season.Mar. Biol. 26: 313-318. er-Verlag. , J. HIROTA, J. C. NEVENZEL,R. $AUERHEBER,A. ., & C. RICKETTS.1981. Relationshipsbetween A. BENSON,& A. LEWIS.1972. Lipids in the ma- food supply and growth in albatrosses:an inter- rine environment. California Mar. Res. Comm., specieschick fosteringexperiment. Ornis Scan- CalCOFI Rep., 16: 95-102. dinavica 12: 207-210. MONTEVECCHI,W. A., & J. PIATT. 1984. Composition RAYMONT,J. E.G., R. T. SRINIVASGAM,& J. K. B. and energycontents of matureinshore spawning RAYMONT. 1971. Biochemical studies on marine capelin (Mallotusvillosus): implications for sea- zooplankton.IX. The biochemicalcomposition of 146 DANIELD. ROBY [Auk, Vol. 108

Euphausiasuperba. J. Mar. Biol. Assoc.U.K. 51: 581- ments,chick meals, and breedingdistribution of 588. Dovekies (Alle alle) in northwest Greenland. Arc- RICKETTS,C., & P. A. PRINCE. 1981. Comparison of tic 34: 241-248. growth of albatrosses.Ornis Scandinavica12: 120- ß A. R. PLACE, & R. E. RICKLEFS. 1986. Assim- 124. ilation and depositionof wax estersin planktivo- RICKLEFS,R. E. 1974. Energetics of reproduction in rous seabirds.J. Exp. Zool. 238: 29-41. birds. Pp. 152-292 in Avian energetics (R. A. •, & R. E. RICKLEFS.1983. Some aspectsof the Paynter, Ed.). Publ. Nuttall Ornithol. Club No. breeding biology of the diving petrels Peleca- 15. noidesgeorgicus and P. urinatrixexsul at Bird Island, 1979. Adaptationßconstraintß and compro- South Georgia. Br. Antarctic Surv. Bull. No. 58: mise in avian postnataldevelopment. Biol. Rev. 29-38. 54: 269-290. --., & . 1986. Energy expenditurein adult 1983. Some considerationson the repro- LeastAuklets and diving petrelsduring the chick- ductiveenergetics of pelagicseabirds. Stud. Avi- rearing period. Physiol. Zool. 59: 661-678. an Biol 8: 84-94. SASINsTITUTE,INc. 1982. SAS user'sguide: statistics. ßC. H. DAY, C. E. HUNTINGTON, & J. B. WILLIAMS. Cary, North Carolina, SAS Inst. 1985. Variability in feeding rate and meal size SEALY,S. G. 1973. Adaptive significanceof post- of Leach's Storm-Petrel at Kent Island, New hatchingdevelopment patterns and growth rates Brunswick. J. Anita. Ecol. 54: 883-898. in the Alcidae. Ornis Scandinavica 4:113-121. ß A. R. PLACE,& D. J. ANDERSON. 1987. An SHEA,R.E. 1985. Growth and developmentof trop- experimental investigation of the influence of ical island-breeding terns. Ph.D. dissertation, diet quality on growth in Leach'sStorm-Petrel. Philadelphia, Univ. Pennsylvania. Am. Nat. 130: 300-305. SHEA,R. E., & R. E. RICKLEFS.1985. An experimental ß& D. D. ROBY. 1983. Developmentof homeo- testof the idea that food supplylimits growth in thetroyin the diving petrelsPelecanoides urinatrix a tropical pelagicseabird. Am. Nat. 126: 116-122. exsuland P. georgicus,and the Antarctic prion SIMONS,T. R., & G. C. WHITTOW. 1984. Energeticsof Pachyptiladesolata. Comp. Blochem.Physiol. 75A: breeding Dark-rumped Petrels. Pp. 159-181 in 307-311. Seabird energetics(G. C. Whittow and H. Rahn, --, & S.C. WHITE. 1981. Growth and energetics Eds.). New Yorkß Plenum Press. of chicksof the Sooty Tern (Sternafuscata) and STEMMLER,J., R. P. HERWIG,J. T. STALEY,& K. NAGY. Common Tern (S. hirundo).Auk 98: 361-378. 1984. Chitin degradation in Adelie Penguins. --, , & J. CULLEN. 1980a. Energetics of Antarctic. J. (U.S.) 19: 161-162. postnatal growth in Leach's Storm-Petrel. Auk TAYLORßJ. R. E., & M. KONARZEWSKI. 1989. On the 97: 566-575. importanceof fat reservesfor the Little Auk (Alle --, & --.. 1980b. Postnatal devel- alle) chicks.Oecologia 81: 551-558. opmentof Leach'sStorm-Petrel. Auk 97:768-781. TICKELL,W. L. N. 1962. The Dove Prion Pachyptila ROBBINS,C.T. 1983. Wildlife feeding and nutrition. desolataGmelin. Falkland IslandsDependencies New York, Academic Press. Surv. Sci. Rep. 33: 1-33. ROBY,D.D. 1989. Chick feedingin the diving petrels WARHAM,J. 1977a. Wing loadingsßwing shapesßand Pelecanoidesgeorgicus and P. urinatrixexsul. Ant- flight capabilitiesof Procellariiformes.New Zea- arctic Sci. 1: 337-342. land J. Zool. 4: 73-83. ß & K. L. BRINK. 1986a. Breeding biology of 1977b. The incidenceß functionsß and eco- Least Auklets on the Pribilof Islands, Alaska. logical significanceof petrel stomachoils. Proc. Condor 88: 336-346. New Zealand Ecol. Soc. 24: 84-93. ß & 1986b. Decline of Least Auklets WILLIAMS, A. J., W. R. SIEGFRIEDßA. E. BURGER,& A. on the Pribilof Islands, Alaska. J. Field Ornithol. BERRUTI.1977. Body compositionand energy 57: 57-59. metabolismof moltingeudyptid penguins. Comp. ---, & D. N. NETTLESHIP. 1981. Measure- Biochem.Physiol. 56A: 27-30.