GENERAL NOTES

Observationson Growth of Blue-footed Boobiesand Development of Temperature Regulationin PeruvianGuano .--Growth rates of seabirdshave been related to the availabilityof foodsuch that pelagic usually grow more slowly than coastal species (Lack1968, Harris 1977,Nelson 1977). Differences in growthrates between populations of the samespecies have been correlated with distance to foodsupplies or productivityof the surroundingocean (Nelson 1978), but morecomparative data are neededto fully describethese relationships. In thispaper, we reportupon the growthof the Blue-looted (Sula nebouxii), and the developmentof temperatureregulation in the (Sula variegata), Guanay ( bougainvillii), and Brown Pelican (Pelecanusoccidentalis) on two guano islands in the Peruviancoastal current, formerly one of the richest fisheries in the world. Studysites and methods.--Duffy visited Islandbetween 24 and 27 February1979, and MacabiIsland between 18 and 21 February1979. Chickswere weighedwith Pesolaspring balances and the wing chord,tarsus, and exposedculmen weremeasured at the beginningand end of a 3-dayinterval on eachisland. Growth increments were used to construct corn.. P osite growth . .curves accordin g to the method of Ricklefsand White (1975). In add•uon,proventricular temperatures of chicksisolated fromthe nestat prevailingambient temperatures (23-28øC) were measured with a therm- istor thermometer(YSI) at 10-min intervalsover a period of 30 min. A coefficientof homeothermywas calculated by the expressionH = (Tso- TA)/(To- TA)where To and Ta0are the bodytemperatures at 0 and 30 min, and T,• is the ambienttemperature. Results.--Ofthe 4 species,there were sufficient data only for the Blue-footedBooby on Lobosde Tierra to constructa compositegrowth curve. The relationshipbetween the 3-dayincrement of winglength and the initialwing length is shownin Fig. 1. Of the 20 observationsof winglength increment, 4 deviatedgreatly from the trendand weredis- regardedas either measurement errors or resultingfrom poornutrition. From Fig. 1, we constructedthe followingaging criterion (age in days,wing lengthin mm): (0, 20), (3, 23), (6, 28), (9, 34), (12, 40), (15, 48), (18, 58), (21, 70), (24, 83), (27, 97), (30, 114), (33, 132), (36, 152), (39, 173), (42, 196), (45, 219), (48, 243), (51, 267), (54, 288), (57, 306), (60, 319), (63, 329), (66, 336), (69, 342), (72, 346). The smallestchick encountered had a winglength of 20 mm and weighed40 g. Becauseeggs of the Blue-footedBooby on HoodIsland, Galapagos, weigh 65 g on average(Nelson 1978), and assuming

Blue-lootedBooby 30 - Lobosde Tierra 24-27II 1979 ß _ß

20- ß

0 100 200 300 Wing length(ram) FIGURE1. Incrementsin thewing length of Blue-footedBooby chicks over a 3-dayperiod plottedas a functionof initialwing length. The solidcurve, drawn by eye,was used to constructthe compositegrowth curve.

332 Vol.52, No. 4 GeneralNotes [333 that Peruvianbirds lay eggsof similarsize, this chick must have just hatched.We therefore estimatedthe wing lengthat day 0 as 20 mm. The largestchick measured had a wing lengthof 360 mm andwas judged to be about80 daysold. On Hood Island,the fledgling period averaged102 daysin 1964 (95-107 days)(Nelson 1978). It is difficult to placeconfidence limits on our estimatesof age; similar increment analysesof known-ageseabird chicks have producedestimates that are generally-+3 days of true age (e.g., Ricklefsand White 1975). Furthermore,the compositegrowth curve is intended to representthe growth processaveraged over the sampleas a whole during a brief period. The agingcriterion can changefrom period to period as the growth per- formanceof the populationis affected by environmentalconditions. Increasein body massand the lengthsof the tarsusand culmen of the Blue-footed Boobyare presentedin Fig. 2, where they are comparedto Nelson'sdata from Hood Island. We fitted Gompertzgrowth equationsto the individualdata pointsfor Peruvian birds and to weeklyaverages of Nelson'sdata for the GalapagosIslands. The Gompertz equation has the form M(t)= A exp(-exp(-K(t - i))) where M(t) is the mass(g) at age t (days),A is the asymptoteof the growth curve (g), i is the age at the inflectionpoint (days),and K is the growth rate constant(days -•) (see Ricklefs1967, 1968).The curveswere fitted by a nonlinearleast-squares regression (SAS procedureNLIN) that employsa modified Gauss-Newtonmethod. The estimatedparam- eterswere A = 1618 ñ 70 SE) g, K -- .66 (-+.008 SE) days-•, and i -- 16.6 (ñ 1.4 SE) days for the Peruvianbirds, and A -- 1741 g, K = .062 days-•, and i = 27.9 daysfor the Ga- lapagosbirds. Because averages were usedto fit curvesto the latter sample,we couldnot obtain meaningful estimatesof the standard errors of estimatedparameters. If we as- sumedsimilar standarderrors in both samples,the two populationswould differ signif- icantlyonly in the age at inflection. The major differencesin growth in the samplesof Peruvianand Galapagoschicks were that the phaseof rapid growth in massof the latter was delayedby about 10 days comparedto the former, and that growth of the tarsuswas delayedby up to 3 weeksin the Galapagoschicks (Fig. 2). Thesedifferences may reflect either the qualityof the season or that of the locality.The data from Hood Islandwere collectedduring a period of food shortageand high chick mortality. The growth rate of the Blue-footedBooby (K• -- .062, .066) is intermediateamong speciesin the genusSula. Three studiesof the smallerRed-footed Booby (S. sula)yielded estimatesof K• of .039, .052, and .056; the larger MaskedBooby (S. dactylatra)has an intermediategrowth rate (K• -- .060, .094) (Ricklefs1973). Each of the speciesexhibits marked variation in growth rate among studies,which may reflecteither localityor sea- sonaldifferences. The Brown Boobyrarely and the Blue-footedBooby commonly rear twochicks. Both feedin inshorewaters. The moreslowly growing Masked and Red-footed boobiesare pelagicfeeders and rear only one chick:distance to feedingarea, brood size, and growth rate are stronglyassociated in the (Lack 1968, Nelson 1977, 1978). The 25 growth incrementsobtained for the PeruvianBooby were so variablethat it wasimpossible to constructa reliablegrowth curve. The largestincrements in winglength were similarto thoseof the Blue-footedBooby, but growth performancewas very poor in many nestlings,especially on MacabiIsland, where 6 of 12 chicksdied or disappeared during the 3-dayperiod and 4 of the remaininglost 150 to 575 g. Many chickson Macabi Island were desertedfollowing infestation of the colonyby the tick Ornithodorusamblus (Duffy 1980). We obtained7 growthincrements of Brown Pelicanson MacabiIsland, mostlyfrom very young birds (•3 kg; weightsof 1 & and 1 •? were 7.0 and 5.1 kg, Murphy 1936). The smallestchick, which had just hatched,weighed 66 g and had a wing lengthof 20 min. We did not attemptto constructa compositegrowth curve. In the Blue-footedBooby, the coefficientof homeothermy,determined at 23-28øC, was .46 (-+.16 SD, n -- 3) for hatchlings,and increasedto 1.0 by 10-15 days of age. Accordingto Nelson(1978), Blue-footed Booby chicks on Hood Islandare broodedcon- tinuouslyduring the first 2 weeksafter hatching.The coefficientof homeothermyis 334] GeneralNotes' j. FieldAutumnOrnithol. 1981 Blue-footedBooby /ß ß,.,•"- ...... 1500 / ,"

ß/ // •1000 øo/.//

500 /,.,' ßLobos deTierra 1979 /./ Galapagos1964: •.// ..... average ße/d./ •- maximum

120 60

• I00 aaa a a a•/.•'-a ß 50 E t * .'" * ß • 80 40 • ß / • 60 30 O a Lobos de Tierra 1979 • 4o - • ,'" ß culmen 20 • •0 , , a tarsus • ./ Galapagos1964 • 20•., - .... tarsus 10

0 20 40 60 80 Age (days) FIGURE2. Increasein bodymass (top) and tarsus and culmen (bottom) as a functionof estimatedage in theBlue-footed Booby on Lobos de Tierra. Nelson's (1978) data for the GalapagosIslands are indicatedfor comparison.

relatedto bodymass for all the speciesin this studyin Fig. 3. Developmentof ho- meothermyappears to followa similarcourse in the 2 speciesof ,which attain H = 1.0 at a weightof 200-300 g. The few datafor the GuanayGormorant suggest that developmentof homeothermyisdelayed with respect to theboobies. We were unable to obtainmeasurements on Brown Pelicanchicks less than 200 g body mass. Boobiesappear to achievehomeothermy at a lowermass than does the GuanayGor- toorant(Fig. 3). This differenceis supportedby the followingdata in Table 1 of Dunn (1975). When testedat 28øG,the MaskedBooby achieved a valueof H = .75 at 300 g, Vol.52, No. 4 GeneralNotes [335

1.0 o 0 /•

ß ß o o 0.8 ß o o /x H

o 0.6 ß ß n ß Blue-footedBooby o PeruvianBooby 0.4 ßo ß "GuanayCormorant o Brown Pelican

0.Z I I I I I I [ [ I I t • t • t t 50 100 200 500 I000 Mass (g) FIGURE3. Relationshipof the coefficientof homeothermy(H) to bodymass in 4 species of in Peruß comparedto 150-200 g for the Blue-footedand Peruvianboobies. When testedat 20øC, the Double-crestedCormorant (Phalacrocorax auritus) achieved H = .75 at 500 g. In general, grow more rapidly(values of Ke, .078-.133)than do boobies (.039-.094, Ricklefs1973) which is consistent with the generalinverse relationship between precocityof thermoregulationand growth rate (Ricklefs1979a,b). There are not yet suf- ficientdata for boobiesto allow us to determinewhether or not this relationshippertains to specieswithin the genusSula. We suggestthat studiesinvolving visits as brief as 3 days,as in the presentstudy, or separatevisits at similarintervals, could yield much valuableinformation on variationin rate of growthand developmentamong seasons, years, populations, and species.These data might be used to test ideasabout factorsdetermining growth rates in seabirdsand to providean index to conditionsfor reproduction,particularly the availabilityof food.

LITERATURE CITED DuFFy,D.C. 1980. Comparativereproductive behavior and populationregulation of seabirdsof the Peruviancoastal current. Ph.D. thesis,Princeton University, Princeton, N.J. DUNN,E. 1975. The timing of endothermyin the developmentof altricialbirds. Condor 77:288-293. HARRIS,M.P. 1977. Comparativeecology of seabirdsin the GalapagosArchipelago. Pp. 65-76, in B. Stonehouseand C. M. Perrins(eds.), Evolutionary ecology. Macmillan, London. LACK,D. 1968. Ecologicaladaptations for breedingin birds. Methuen,London. MURPHY, R. C. 1936. Oceanic birds of South America. Am. Mus. Nat. Hist., New York. NELSON,J. B. 1977. Somerelationships between food and breedingin the marine Pele- caniformes.Pp. 77-87, in B. Stonehouseand C. M. Perrins(eds.), Evolutionary ecol- ogy. Macmillan,London. ß 1978. The Sulidae: and boobies.Oxford UniversityPress, Oxford, En- gland. RICKLEFS,R. E. 1967. A graphicalmethod of fitting equationsto growthcurves. Ecology 48:978-983. 336] GeneralNotes J.Field Ornithol. Autumn 1981

1968. Patternsof growth in birdsßIbis 110:419-451ß ß 1973. Patternsof growth in birdsßII. Growth rate and mode of developmentß Ibis 115:177-201. ß 1979a. Patternsof growth in birdsßV. A comparativestudy of growth and de- velopmentin the Starling,Common Tern, and JapaneseQuailß Auk 96:10-30. ß 1979b. Adaptation,constraint, and compromisein avianpostnatal developmentß Biol. Rev. 54:269-290ß --, and S.C. Whiteß 1975. A methodfor constructingnestling growth curvesfrom brief visitsto seabirdcoloniesß -Banding 46:135-140ß D^wr• C. DuFFy,Percy Fitzpatrick Institute of AfricanOrnithology, University of CapeTown, Rondebosch7700, SouthAfrica, ^•) ROBr•R•rE. RICKLEFS,Department of Biology,University of Pennsylvania,Philadelphia, Pennsylvania 19104ß Received 20 Mar. 1981;accepted 23 July 1981.

Rat Predation on Bonin Petrel Eggson Midway Atoll.--Rats (Rattusspp.) are native to few oceanicislands and where introducedhave generallycaused extensive damage to the vegetationand avifauna.Introduction of rats to Sand and EasternIslands, Midway Atoll (28ø13'N,177ø23'W), in the LeewardIslands of the HawaiianArchipelago occurred in 1943 with the buildup of military activities(Alsatt 1945). Insular tamenessdue to the historicallack of predatorsrenders many speciesvulnerable to predationby rats. Rats have been very successfulon islandsand are extremely difficult to exterminateonce established.A vigorousrodent control program during and prior to 1970 (Howell 1978a,b;T. R. Howelland G. A. Bartholomew,pers. comm.) by the U.S. Navyon Midway Atoll reducedthe impactof ratson birds'eggs. During our stayon MidwayAtoll during the springsof 1980 and 1981 we noted high roof rat (Rattusrattus) populations on both Easternand Sand Islandsand intensepredation on the eggsof Bonin Petrels(Pterodroma hypoleuca).Predation by the Polynesianrat (R. exulans)on eggs,young, and/or adult al- batrosses,petrels, shearwaters, tropicbirds, , and ternson nearbyKure Atoll hasalready been documented (Robbins 1966, Kepler 1967,Fleet 1972,Woodward 1972). Materialsand methods.--BoninPetrels excavate burrows .6 to 3 m long; at the end of each burrow is an enlarged chamber containinga well-lined nest. A single white egg weighingca. 39 g is laid. To quantify egg predation we dug verticalshafts 15-20 cm in diameter to the nest chamber which was .3 to 1.3 m vertically below the surface.The shaftswere coveredwith plywood.The two main studyareas chosen on Sand Islandwere a colonyof 654 nestburrows south of the Chapel(hereafter designated the chapelcolony), and a colony of 225 nest burrows north and east of the Cannon School(hereafter des- ignated the schoolcolony). Egg neglect(adult not present)and orientationin the nest chamberwith respectto the long axisof the tunnel were recorded.The shaftcovers were lifted and the orientationwas noted without disturbingthe incubatingbird. Results.--In 1980, 26 eggswere monitoredperiodically in the chapelcolony during the courseof incubation.Of these, 25 were eaten by rats and one hatched.Of 21 eggs monitored periodicallyin the schoolcolony, 19 hatchedand 2 were broken by incubating petrels.In addition, 94 other partially eaten eggswere found scatteredon the ground in the chapelcolony. Some eggs were consumedby rats insidethe burrows(partially eaten eggswere found in the nest chamber), thus evidencethat predation had occurred was not alwaysavailable on the surfaceof the ground. Only one egg in the chapelcolony of 654 burrows is known to have hatchedin 1980. In 1981, 130 partially eaten petrel eggs werefound in the chapelcolony. These two coloniesare about 100 m apart and separated by a narrow paved road. Ratswere frequentlyseen in both coloniesat night. Rat traps were set in the chapel colonyalmost every night from 31 Januaryto 29 February1980. We caught16 ratsinside the chapelcolony (area = 2548 m2) during 269 trap nights(number of trapsset x number of nights).Traps were set insidethe schoolcolony (area = 1766 m2) from 1-11 March 1980.We caught7 ratsduring 130 trap nights.Thus the numbersof ratscaught per trap night are similar in the two colonies. Egg neglectoccurs regularly in someof the smallerprocellariiform birds (Boersma