RELATIVE PASSAGE RATES OF AND AQUEOUS DIGESTA IN THE FORMATION OF STOMACH OILS

DANIEL D. ROBY,• KAREN L. BRINK,2 AND ALLEN R. PLACE3 •CooperativeWildlife Research Laboratory and Department of Zoology, SouthernIllinois University, Carbondale, Illinois 62901 USA, 2P.O. Box 571, Carbondale,Illinois 62903 USA, and 3Centerof MarineBiotechnology, University of Maryland,Baltimore, Maryland 21202 USA

ABSTRACT.--Weused tritium-labeled glycerol triether as a nonabsorbablelipid-phase mark- er and carbon-14labeled polyethylene glycol as a nonabsorbableaqueous-phase marker to examine gastrointestinaltransit of a homogenized fish meal fed to 4-week-old chicks of AntarcticGiant- (Macronectes giganteus) and GentooPenguins (Pygoscelis papua). Both aqueous-phaseand lipid-phase markers passedthrough the gastrointestinaltract without being metabolized.Label recoveries from the two specieswere statisticallyindistinguishable. Mean retention time was significantlylonger for lipid-phasecomponents than for aqueous- phasecomponents in both species.In the , mean retention time for lipid-phaseand for aqueous-phasewas significantlylonger than in the penguin. Interspecificdifferences in retention were largely the result of differing ratesof gastricemptying. Both markersemptied rapidly from the proventriculusand gizzard of the penguins,while in giant-petrelsthe lipid- phase was retained for extended periods in the stomach.Differential transit of lipid and aqueousphases coupled with the lower rate of gastricemptying in giant-petrelchicks provides a physiologicalbasis for accumulationof dietary in the proventriculus. The large, distensibleproventriculus and the ventral positionof the pyloric valve relative to the gizzard and proventriculusare morphologicaltraits which enhance the formation and retention of stomachoils. Received31 May 1988,accepted 19 December1988.

OFall avian internal organs,the range of mor- (Matthews 1949, Duke et al. 1989). In other phologicalvariation in the stomachis the great- the much smaller proventriculus is cranial to est. This reflectsboth widely differing dietary the gizzard. habits and the importance of the stomach in With the exception of diving petrels (Pele- processingsolid food items (Ziswiler and Far- canoididae), all members of the Procellariifor- ner 1972). The avian stomach consists of two messtore lipids ("stomachoils") in the enlarged chamberswhich are externallydistinguishable proventriculus. The origin of stomachoils was in most species(McLelland 1979). The glan- previously thought to be secretory(Matthews dular proventriculus is continuous with the 1949, Lewis 1966), but evidence is now com- esophagusand secretesgastric juice. The giz- pelling for a dietary origin (Cheah and Hansen zard or ventriculusis caudalto the proventric- 1970, Warham et al. 1976, Clarke and Prince 1976, ulusand functions as the siteof gastricproteoly- Imber 1976). The function of stomach oils has sisand, in manyspecies, of mechanicaldigestion been viewed primarily as an energy and water (Duke 1986b). reserve and secondarily as a defense against In procellariiforms (petrels, , shear- predators (Warham 1977, Jacob1982). Previous waters,and ),the mostpelagic avian work (Roby et al. 1986a,Place and Roby 1986, order,the division betweenthe proventriculus Duke et al. 1989) indicated that petrels retain and gizzard is readily apparent. Forbes(1882) lipids in the proventriculus longer than sea- describeda distensibleproventriculus that con- birds without stomachoils. This suggeststhat, sistedof a largesac with a fundus,a smallgiz- for seabirdsthat store stomachoils, gastroin- zard twisted so that the pylorus facesback in- testinal passagerates for lipid digestaare con- steadof forward,and an ascendingloop of the siderablylower than thosefor aqueousdigesta. duodenumbefore the descendingand ascend- However, there are no publishedreports of rel- ing limbs.The proventriculusmay be so large ative passagerates of lipid and aqueousdigesta when distended with food that it extends cau- in birds;hence, there is no basisfor comparison dally in the bodycavity well beyondthe gizzard with those speciesthat store stomachoils.

303 The Auk 106: 303-313. April 1989 304 RoB¾,BRINK, AND PLACE [Auk,Vol. 106

Nonabsorbablemarkers are widely used to ratesand gut morphologyin penguinsand pe- monitor movementof a specificphase along the trels, we hoped to elucidate mechanismsre- gastrointestinaltract. They are also used to es- sponsible for stomach oil formation in the lat- timate nutrient absorptionfrom or secretioninto ter. a marker specifiedphase. It is important that the marker have properties similar to compo- MATERIALS AND METHODS nents of the phaseunder study (Wiggins and Radiolabelsand fiuors.--We used [•4C]-polyethylene Dawson 1961; Carlson and Bayley 1972a, b). glycol(molecular weight 4000,15 mCi/g) from Amer- Polyethylene glycol (MW 4000) is a widely ac- sham (Arlington Heights, Illinois) without further cepted aqueous-phasemarker (Wingate et al. purification.Fluors were ACS II (Amersham,Arling- 1972)and glyceroltriether, a neutrallipid-phase ton Heights, Illinois) and BiosafeII (ResearchProd- marker, is used extensively in studies with uctsInternational, Mount Prospect,Illinois). Dupli- mammals(Morgan and Hofmann 1970;Carlson cate sampleswere counted on a BeckmanLS 3801 and Bayley 1972a,b; Meyer et al. 1986),but has scintillationcounter. A correction("quench") curve not been used with birds. The triethers will wasderived to determinecounting efficiency for dif- mark only the neutral lipid phase;polar lipids, ferentextracts and tissuetypes using Compton edge suchas phospholipids and fatty acids,partition ("H number") calibration (BeckmannInstruments). into the aqueousor miceliar phase.These com- Countingefficiency for •4Cin thesamples varied from 88.0-75.4%and that for 3H from 22.0-6.0%.Counting pounds are nonabsorbable, nontoxic, nonde- times (2-10 min) were chosen to ensure at least 95% gradableby digestiveor bacterialenzymes, and counting accuracy.The coefficient of variation for do not influence the normal absorptionof di- replicatesamples averaged 3.0% (SD = 1.3) for tritium etary aqueousnutrients or fat. and 1.9% (SD = 0.9) for carbon-14. All radioactivities In mammals, passagerate of lipids through are expressedin disintegrationsper min (DPM). the pyloric valve and into the small intestine is Tracer synthesis.--BachemBioscience Inc. (Phila- nearly half that for aqueouscomponents (Jian delphia, Pennsylvania)synthesized the glycerol et al. 1982,Meyer et al. 1986).If a meal is ho- triether[1-(9 cis-octadecenyl)2,3 didodecylglycerol mogenized prior to ingestion, however, both triether]as described by Morganand Hofmann (1970). lipids and aqueous components empty the The tritiatedglycerol triether (3H-GTE) was prepared by reductionwith platinum as a catalyst(New En- stomachtogether (Cortot et al. 1979). gland Nuclear, Boston, Massachusetts).Purified 3H- Using lipid-phase and aqueous-phasemark- GTE(> 98%radiopurity) was obtained by chromatog- ers, we examined gastrointestinal transit in raphy on a silicic acid column eluted with hexane/ chicks of the Antarctic Giant-Petrel (Macro- diethyl ether 85:15 (v/v). Solvent was removed with nectesgiganteus) and Gentoo Penguin (Pygoscelis nitrogen evaporationand the purified 3H-GTE dis- papua).These speciesrepresent the two major solvedin absoluteethanol to a specificactivity of 1 orders of seabirds in the Southern Ocean. The mCi/ml. young of both speciesare fed at the nest by Studyarea and subjects.--Feeding (including excreta their parents until full grown. Giant-petrels collection)trials were conductedduring Januaryand February,1986, at Ardley Island (62ø13'S,58ø55'W), a usually locate their nestsnear penguin breed- ca. 18-ha island off King George Island, South Shet- ing coloniesand feed their young a diet of (in land Islands,where ca. 9,000 pairs of GentooPen- decreasingorder of importance)penguins, guinsand 15 pairs of AntarcticGiant-Petrels nest (Roby (Euphausiasuperba), seals, squid, and small sea- et al. 1986b).Six nestlingsof eachspecies were re- birds (Hunter 1983, 1985). The chicks of both moved from nests and held in enclosures for 12-20 giant-petrelsand Gentoo Penguinsare usually h beforeingesting the labeledmeal. Subjects were ca. fed 1-2 meals daily. The diet of Gentoo Pen- 4 weeks old, judging from wing-length and body- guins consistsprimarily of krill and, to a lesser mass measurements (Hunter 1984, Volkman et al. extent,fish (Croxalland Prince1980). Nestling 1980).All weredeveloped sufficiently to thermoregu- giant-petrelsstore stomach oils at an early age late at ambienttemperatures (ca. 0-5øC)indepen- dentlyof their parents.Body mass of subjectsused in and are adept at regurgitatingoil on potential feedingtrials ranged from 0.98-2.2kg and averaged predators.Penguins do not storestomach oils. 1.5 kg. Averagemass of subjectsdid not differ sig- We measured relative transit times of the two nificantly between the two species(t = 0.344, P > phasesin penguins to ascertainthe extent of 0.05, Table 1). differential passagerates in an avian species Priorto feedingthe radiolabeled meal, we pumped which lacksstomach oils. By comparingpassage the stomachsof giant-petrelchicks to remove any April 1989] StomachOils and GI Transit 305

TABLE1. Morphological measurementsof the gas- Scornbetjaponicus peruvianus) was used to prepare the trointestinal tracts of Gentoo Penguin (n = 6) and labeled meal. Composition of the canned fish was Antarctic Giant-Petrel chicks (n = 6). analyzed later in the laboratoryin air-dried samples at 60øC(constant mass) and by determining total lip- Gentoo Antarctic ids in separatesamples (Bligh and Dyer 1959). The Penguin Giant-Petrel compositionof the fish was 24.9%water (SD = 1.78, Mass n = 3) and 3.7% lipid (SD = 0.04, n = 3). Canned fish, Total chick mass(kg) 1.56 + 0.39a 1.49 + 0.31a vegetableoil, and freshwater were thoroughly mixed Proventriculus (g) 16.3 + 3.90 22.3 + 1.94 in a 16:3:3ratio (w:v:v) which yielded a 40% water Gizzard (g) 11.5 + 2.12 8.0 + 0.69 and 13%lipid paste. Small intestine (g) 48.8 + 15.5 18.7 + 3.05 We warmed the fish paste to ca. 40øCand loaded Duodenum 20.3 + 6.97 6.36 + 1.26 ca. 30 ml into a 60-cm• disposable syringe. With a Ileum (g) 28.5 _+8.64 12.3 +_2.48 Colon and ceca(g) 3.99 + 1.33 1.47 _+0.20 1-cm* tuberculin syringe, we added 0.5 ml of the ra- diolabel-carrier mixture to the fish paste and filled Length (cm) the remainder of the 60-cm3 syringe with fish paste. Proventriculus 4.7 + 0.32 20.2 + 0.82 A single labeled meal was fed to subjectsby forcing Gizzard 5.6 + 0.97 4.9 + 0.58 the paste through a 20-cm length of polyethylene Small intestine 178 + 16.0 a 192 + 7.70 • Duodenum 59.8 + 5.77 38.0 + 3.44 tubing inserted into the esophagous.All chickstook Ileum 118 + 12.1 154 + 9.32 the feeding without any regurgitation. After inges- Colon 7.8 + 0.82 • 7.4 + 0.49 a tion, each chick was placed in an outdoor, covered Cecum 1.9 + 0.13 0.9 + 0.10 pen. At selectedtimes post-ingestionchicks were Area (cm2) quickly and humanely killed by stunning. Two giant- petrelsand two penguinswere killed at eachof three Proventriculus 40.4 + 4.34 292 + 44.9 Gizzard 44.3 + 8.56 28.8 + 5.34 trial periods:4, 12, and 24 h post-ingestion.Subjects Small intestine 297 + 48.2 a 254 + 18.1 • in the 24-h trials had their excreta collection bags Duodenum 120 + 21.9 72.6 + 7.54 replaced at either 10 h (penguins) or 12 h (petrels) Ileum 178 + 29.3 • 182 + 19.1 a post-ingestion. Colon 11.7 + 1.74 9.6 + 1.59 Immediately after death, we plugged the esophagus

a Not significantly different at the 0.05 level. of each subjectwith cotton to prevent regurgitation of the radiolabel. Excreta collection bags were re- residual stomach oils from the proventriculus. All movedand placedin doubleplastic bags. Abdominal subjectswere weighed to the nearest0.01 kg using a and thoraciccavities were opened,photographed, and Pesolaspring scale.We plucked the down feathers the entire gastrointestinaltract removed, measured, aroundthe cloacaand glueda plasticcylinder around and photographed.Each tract was divided into eight the cloacaof eachsubject with superglue.A plastic parts: proventriculus, gizzard, anterior duodenum, bag was taped to the cylinder to collect excretaand posterior duodenum, anterior ileum, roedial ileum, additionaltape was usedto help securethe cylinder posteriorileum, and colon.The transitionfrom duo- and collection bag to the chick. Changing to a new denurnto ileum wasnot externallyapparent, and was collectionbag was accomplishedby removal of the separatedarbitrarily at the splenic attachment. The tapefrom the plasticcylinder and retapinga new bag duodenumwas then divided at the midpoint and the onto the cylinder. ileum was separatedinto three equal segments.For Meal compositionand feeding studies.--We used sta- the purposesof this study, the ileum was considered bilized triglyceride emulsion (Sigma Lipase Sub- to be that portionof the intestinebetween the splenic strate, No. 800-1) as a carrier to assure consistent ratios attachmentand the paired ceca.Each part of the gas- of the two markersin mealsfed to subjectsin Ant- trointestinal tract was opened lengthwise and the arctica. Prior to departure for Antarctica, we added contentsscraped into separateplastic bags. (In addi- 10 •Ci of [3H-GTE),2.5 •Ci of [•4C-PEG],and 10 mg tion to digesta,scraping removes some intestinal ep- of PEG-4000 to each ml of the emulsion, mixed for ithelium). All sampleswere kept chilled in a snow- 10 rain on a vortexagitator, and countedaliquots in bank for a maximumof 48 h before being transported a liquid scintillationcounter. Upon return to the U.S., to freezer facilities on King George Island. Samples additionalaliquots were counted. The specific activity were kept frozen during shippingto the United States, (11.4 + 0.22 •Ci/ml [3H-GTE] and 2.91 + 0.028 •Ci/ where they were storedat -70øC until analyzed. ml [14C-PEG])of the two markersand their isotopic Measurementsof gastrointestinaltracts.--Each part of ratio (3.92 + 0.069, n = 10) were identical before and the gastrointestinaltract was thawed, weighed to the after the experimentswere conducted. nearest0.1 g, and measuredto the nearest min. We Effortsto induceadult GentooPenguins to regur- consideredthe proventriculusto be only that portion gitate chick mealswere unsuccessful.Consequently, of the gastricregion coveredwith secretorycells. The locallyavailable canned fish ("cabalia,"presumably gizzard was consideredthe part of the gastricregion 306 RoBY,BRINK, AND PLACE [Auk, Vol. 106 between the proventriculus and the pylorus. Prior to ageswere performed on arcsintransformed data. Dif- weighing,we removedthe mucuslining of the giz- ferenceswere consideredsignificant when P < 0.05, zard. For both proventriculusand gizzard, maximal exceptfor comparisonsinvolving ratios of disinte- lengthand width (opened)were used to estimatearea. grationsper minute(DPM) of two markers;in this Lengthof eachintestinal part wasmeasured with the case,P < 0.001 was chosen. In all cases,ratios were segmentfully extendedbut not stretched.Width of calculatedfrom a minimum of 1,000DPM per sample eachintestinal part was measuredon the openedseg- for either isotope.Fitted curveswere calculatedby ment at the midpoint. Area was estimatedfrom the the modified Gauss-Newton method (Johnson et al. product of length and width. 1981).Other statisticalprocedures used are identifed Marker recoveryand distribution.--We weighed each in the text. sampleof digestato the nearest0.1 g and homoge- nized with a Brinkmann Polytron homogenizerwith PTA 10TSgenerator until a uniform emulsionwas RESULTS obtained.Samples were diluted with deionizedwater to specifiedvolumes and aliquots removed immedi- Marker recoveryand phasespecificity.--Recov- ately for scintillationcounting. We measuredamounts ery of the two markers,as a percentageof in- of the two radiolabelsin digestafrom eachgastroin- gestedmarker, was statisticallyindistinguish- testinalsegment of eachsubject. Accumulated excreta able betweenspecies. Recovery of the aqueous in each collectionbag was diluted with deionized marker, •C-PEG, was 91.2% (SD = 18.9%, n = water and homogenizedas describedfor the digesta. 6) in the petrel chicksand 86.6%(SD = 18.9%, Data were expressedas cummulative percentagesper hour of the total amount of marker recovered. The n = 6) in the penguin chicks.Recovery of the mean retention time (t) was calculatedfrom the equa- lipid marker, 3H-GTE, was 78.5% (SD = 17.3%, tion: n = 6) in the petrels and 70.6% (SD = 11.1%,n = 6) in the penguins.A paired comparisonof marker recoveries within each individual in- t = • xttt/•xt, (1) dicatedthat recoveryof aqueousmarker (88.9%, SD = 18.2%, n = 12) was significantly greater where x, is the amount of marker excreted at time t, than recovery of lipid marker (75.3%, SD = (Warner 1981).Food retention time wasalso estimated 14.0%,n = 12, t• = 3.023, P = 0.017). Water used usingthe logisticfunction (Patton and Krause1972): to extractdigesta lowered recoveryof the lipid marker becauseit was water insolubleand prone Total % marker recovered to adsorptionby surfaces(glassware, digesta, •t= 1+ exp[-R(t- •)/25] ' (2) intestinal lining, etc.). To test the phasespecificity of each marker, where •1,is the percentageof markerexcreted from several samples of homogenized digesta and initial administration to time t, R is the maximum rate excretawere subjectedto the Bligh and Dyer of passageof the marker,and r is the time required for one-half of the marker to be excreted (mean re- (1959)extraction method. Ninety-six percent (SD tention time). In the originalequation of Pattonand = 5.3%, n = 8) of •4C-PEGradioactivity was re- Krause(1972), the percentageof recoveredmarker coveredin the aqueousphase while 95% (SD = (Total % marker recovered)was set to 100.We did not 2.5%, n = 8) of the 3H-GTE radioactivity was imposesuch a restrictionto fit curves.Estimates for recoveredin the lower chloroform layer. Thin R, r, and Total % marker recovered were obtained by layerchromatography of bothexcreted markers a nonlinearleast squares iterative procedure (modi- indicated that neither had been substantially fied Gauss-Newton method; Johnsonet al. 1981). Gas- metabolized(i.e. radiopurity was at .least95% tric emptyingwas modeledfrom the exponential of that prior to ingestion). function (Stubbs 1977, Smith et al. 1984): Gastrointestinaltransit of lipidand aqueous com- vt = V•(e-t'b) (3) ponents.--Ratioof lipid marker to aqueous marker (3H-DPM: •4C-DPM) was used as an in- whereV, is the gastricvolume at time t, V, the initial dicatorof differentialtransit of lipid and aqueous volume fed, and b the exponentialrate of emptying. componentsin the gastrointestinaltract. Re- We usedthe quantity(in •tCi) of the two markersat covered amounts of lipid-phase and aqueous- eachtime interval sampledto estimatethe meal vol- ume present in the stomachat time t. phasemarkers and the ratio of specificactivity Statistics.--Resultsare expressedas the samplemean of the markers are shown in Table 2. At 4 h _+the standarddeviation; n representsthe number post-ingestion,> 80% of eithermarker was still of measurements.Comparisons involving percent- found in petrel chick stomachs.In contrast, April 1989] StomachOils and GI Transit 307

2' SouthernGiant Petrel '1"] GentooPenguin 12 Hou rs Post-Ingestion

[] AqueousMarker

Southern Giant Petrel Gentoo Penguin 12 Hours Post-Ingestion 12 Hours Post-Ingestion 4O

Fig. 1. Distributionof lipid-phaseand aqueous-phasemarkers along the gastrointestinaltract 12 h after feedingAntarctic Giant-Petrel and GentooPenguin chicks. The top figuresrepresent the fractionof ingested markerrecovered from eachsegment of the gastrointestinaltract. The lower figuresrepresent the ratio of the lipid-phasemarker to the aqueous-phasemarker in eachsegment. The aqueous-phasemarker ([I-•C] PEG) was recoveredat 96.3%(SD = 8.3%;n = 4) and the lipid-phasemarker ([•I-I]-GTE)at 76.4%(SD = 10.6%,n = 4). The arrow indicatesthe ratio at the time of ingestion. much of the meal fed to the penguin chickshad markers was 2.4. Penguin chicks had excreted already emptied into the intestine. Distribution most of both lipid-phase and aqueous-phase of the two markers along the gastrointestinal markers and the ratio of the two markers was tract at 12 h post-ingestionis shown in Fig. 1. not significantly different from the ratio in the In petrel chicks >45% of the lipid marker was meal (3.6 + 0.4 vs. 3.9 + 0.07). recovered from stomach oils removed from the Gastricemptying.--We plotted the fraction of proventriculus (Fig. 1). In penguin chicks <5% recoveredmarker found in the stomach(pro- of the lipid markerwas recoveredfrom the pro- ventriculus and gizzard combined) over time ventriculus (Fig. 1). In both petrel and pen- (Fig. 2). In a pairwise comparisonwithin each guin chicks,ratios of lipid- to aqueous-phase , significantly more lipid-phase than markersin excretawere significantlylower than aqueous-phasemarker was found in the stom- the ratio in the meal (P < 0.001), even if cor- ach,which indicatesslower gastric emptying of rectedfor nonquantitativerecovery of the two lipids. In petrel chicks,aqueous-phase marker markers. This reflects the shorter transit time emptied at 11.3%.h -• while lipid-phase marker for aqueouscomponents compared with lipids emptied at 5.5%.h -•. The half-time (T•) for in both species.The decreasein the ratio of the aqueouscomponent emptying from the stom- two markers was gradual from the proventric- ach was 6.1 h while the T• for lipid emptying ulus to the colon in both species(Fig. 1). After was 12.5 h. In penguin chicks, the aqueous 24 h, only 46.4% of the lipid marker was re- marker emptied at 56.1%-h-• while the lipid covered in petrel excreta and the ratio of the markeremptied at 25.4%-h •. This is equivalent 308 ROBY,BmNK, AND PLACE [Auk, Vol. 106

TABLE2. Gastrointestinaltransit of lipid-phase and aqueous-phasemarkers in 4-week-old Antarctic 1.2 Giant-Petreland GentooPenguin chicks. The per- centageof marker recoveredin stomachvs. excreta at 4, 12 and 24 h post-ingestionare presented(stan- dard deviation in parentheses). E 0.8 Antarctic 8 0.8 Giant-Petrel Gentoo Penguin (n = 6) (n = 6) • 0.4 4h Proventriculus& gizzard g o.: Lipid 80.2% (12.0%) 32.2% (20.2%) Aqueous 80.7% (22.2%) 10.5% (6.8%) u. 0.0 10 20 3O Ratio" 3.99 (0.52) 9.52 (4.65) Excreta Hours Post-Ingestion Lipid 0.4% (0.1%) 8.4% (8.2%) Aqueous 0.5% (0.1%) 13.9% (16.1%) Ratioa 2.11 (1.05) 12.2 (0.78) Gentoo Penguin 12 h Proventriculus & gizzard Lipid 56.5% (3.3%) 2.7% (0.9%) • 1.0' Aqueous 16.7% (2.2%) 11.4% (10.3%) Ratio" 25.7 (13.3) 14.1 (6.96) Excreta Lipid 24.1% (3.9%) 51.9% (5.4%) oø 0.8- Aqueous 65.8% (1.3%) 88.0% (1.9%) Ratio" 1.43 (0.22) 2.30 (0.20) • 0.4- 24 h .o_ 0.2 Proventriculus& gizzard Lipid 25.3% (0.7%) 6.0% (5.9%) u. 0.0 , Aqueous 6.5% (5.8%) 5.3% (4.1%) 0 10 20 3O Ratioa 64.2 (48.0) 5.0 (1.2) Hours Post-Ingestion Excreta Lipid 46.4% (1.5%) 87.4% (9.7%) Fig. 2. Gastric emptying of the aqueous-phase Aqueous 78.3% (4.6%) 88.9% (7.6%) marker [1-•4C]PEG and the lipid-phase marker [3H]- Ratioa 2.4 (0.27) 3.60 (0.38) GTE by Antarctic Giant Petrel and Gentoo Penguin chicks. The lines represent the fitted exponential ' Representsthe ratioof lipid-phasemarker specific activity to aqueous- phasemarker specificactivity. function (Eq. 3) best describing the rate of gastric emptying.

to a T,• of 1.2 h for aqueouscomponents and a recovery of marker), estimated mean retention T• of 2.7 h for lipid components. time would be 19.5 h. Similar estimates (ca. 20 Cumulative excretion.--Cumulative excretion h) were obtained by linear interpolation be- data for the two markerswere comparedusing tween 12 h and 24 h data points and by using a pairwise comparison within each bird. Equation ! to estimate mean retention time. Aqueous-phasemarker was excreted at a sig- Mean retenton time (r) for aqueous-phaseand nificantly greaterrate than lipid-phasemarker lipid-phase markers in the gastrointestinaltract (Fig. 3). In petrel chicks (Fig. 3), mean reten- of penguin chicks(Fig. 3) was 7.6 h and 8.9 h, tion time for aqueouscomponents was 10.7 h respectively.Maximum ratesof excretion(R) of with a maximum excretion rate of 18.5%.h-L the two markers were statistically indistin- The fitted logisticcurve for lipid excretionwas guishable (12.9%.h-• and 12.2%.h-•, respec- suspectbecause only 60% of lipid marker was tively). excretedby 24 h and there was no accuratees- Results for cumulative excretion are consis- timate of the asymptote.Assuming nearly com- tent with findings for gastric emptying. First, plete excretionof marker by 196 h (i.e. 100% aqueouscomponents were excretedat a higher April1989] StomachOils and GI Transit 309

Southern Giant Petrel circledthe gizzard (Fig. 4A). In penguin chicks the proventriculuswas much smaller; similar 1.0 in size and massto the gizzard (Fig. 4B). Petrel o proventriculuswalls were thin (ca.0.5 mm); in penguinsthey were thick (ca. 5 mm) and mus- o o cular.Average mass of the petrel proventriculus was 35%greater than that of penguinsand av- eragesurface area of the petrel proventriculus was more than 7 times that of penguins. n- 0.4' In contrastto the proventriculus,gizzards of the two specieswere similar in size (Table 1). Grossanatomy of the gizzard wassimilar to that of other avian carnivores,such as raptors (Duke u. 1985) and herons (Rhoadesand Duke 1975), and 0.0 lackedthe opposingpairs of thin and thick mus- 0 10 20 30 cles characteristic of fowl and most other avian species(Duke 1986a).Most of the penguin giz- Hours Post-Ingestion zards were filled with stones whereas most of the petrel gizzards were filled with penguin Gentoo Penguin feathers.In petrelsthe pyloric valve wasventral 1.0' to both the gizzard and proventriculus, while • o in penguinsit wasdorsal (Fig. 4). Neither species had a well-developed pyloric sphincter, but in •a 0.8' penguins there was a glotus-like projection over the pylorus. The length of the small intestine was similar o o in the two speciesbut, due to a thicker wall, ß 0.6' //© the averagemass of penguin intestineswas more 0.4' than twice that of petrels (Table 1). Small in- testine volume was ca. 50% greater in penguins ._o 0.2 due to a larger duodenum (Table 1). The pen- guin duodenum was coiled tightly around a u. 0.0 I•-o. , , large pancreas(ca. 13 g), while the petrel duo- 0 10 20 30 denum was looped around a smaller pancreas Hours Post-Ingestion (ca. 7 g, Fig. 4). The colon was slightly larger in the penguin chicks,while the cecaof both Fig. 3. Excretion of the aqueous-phasemarker specieswere small and apparently nonfunc- [1-•4C]PEG and the lipid-phase marker [3H]-GTEby tional. six Antarctic Giant Petrel and six Gentoo Penguin chicks. The lines drawn represent the fitted logistic function (Eq. 2) best describingthe rate of gastroin- DISCUSSION testinal emptying. The avian proventriculusand gizzard play a major role in chemical and mechanical diges- rate than lipid componentsin both speciesand, tion of solid food. Consequently,gastric emp- second, excretion of both components, espe- tying is a major componentof gastrointestinal cially lipids,was slower in petrelsthan in pen- passagetime. In three domesticspecies (goose, guins. turkey, and chicken),48.5% ___24.5% (n = 6) of Gastrointestinalgross anatomy.--The proven- the total mean residence time of a meal (7.1 + triculus was the most striking difference be- 2.25h) involved gastricemptying (Warner 1981). tween the gastrointestinal tracts of the two The data from seabirds are limited, but in Jack- species.In petrel chicks,the proventriculuswas assPenguins (Spheniscus demersus) fed fish, 23% very large, filled mostof the abdominalcavity, of the mean residence time of 11 h involved extendedposterior to the cloaca,and partly en- gastricemptying (Wilson et al. 1985, Laugksch 310 RoB¾,BmNK, AND PLACE [Auk, Vol. 106

A

B

I I ------0

PF

Fig. 4. Ventral views of the gastrointestinaltract of 4-week-old Antarctic Giant Petrels (A) and Gentoo Penguins(B). The sternum,abdominal wall and liver have been removed.The marker is 1 cm. O = esophagus; Pr = proventriculus;V = gizzard; Py = pylorus;DI = duodenal loop; Pa = pancreas;I1 = ileum; Ca = ceca; CI = colon.

and Duffy 1986).Fish meals are known to empty transit time for lipids was a function of gastric the stomachmore rapidly than mealsof squid emptying.In general,it appearsthat food tran- or krill (Laugkschand Duffy 1986). sit time in seabirds(Jackass Penguins and Cape We have shown in birds that aqueouscom- Gannets[Morus capensis]) is slower than in other por/entsof a mealare emptiedfrom the stomach birds (Laugkschand Duffy 1986).A specialized at a higher rate than lipids. This differential gastrointestinalanatomy is associatedwith the emptying occursin both speciesdespite strik- petrel'slow gastricemptying rate. The petrel ing differencesin digestiveanatomy. Petrels ex- proventriculusis relatively large, and entire hibit a low overall passagerate; gastric emp- mealsreside in the proventriculusfor extended tyingcomprised 57.8% of the 10.6h meantransit periods,as reflectedin the low ratesof gastric time for aqueouscomponents and 62.5%of the emptying.The functonof the proventriculusin 20 h mean transit time for lipids. In penguins procellariiforms(i.e. thoroughchemical diges- only 16% of the 7.5 h mean transit time for tion of ingesta)is unique amongbirds. In other aqueouscomponents and 30.3%of the 8.9h mean species.,including penguins, food passes rap- April1989] StomachOils and GI Transit 311

idly through the proventriculusand chemical, et al. 1986a).For six commondiving petrel (P. as well as mechanical,digestion occursprimar- urinatrix)and four SouthGeorgia ily in the gizzard (Duke 1986b). chicks,the average gastric emptying rate for The pattern of gastricmotility in procellar- lipids was 30%.h-• (SE= 5.8%.h-•, n = 10) and iiforms is also unique among birds. The pro- the mean retention time for labeled lipid in the ventriculusis relatively inactive during a diges- stomach was 2.3 h (SE = 0.1 h, n = 10). This tive contractioncycle in Leach'sStorm-Petrel compareswith a gastricemptying rate of 5.5%' (Oceanodromaleucorhoa) chicks (Duke et al. 1989). h -• and a mean retention time of 12.5 h in the Proventricular contractionsare observedonly giant-petrelchicks we report.Diving petrelsare alongthe ventral surface(Duke et al. 1989).This an orderof magnitudesmaller in sizethan giant- is in contrastto the vigorous,coordinated mus- petrels,so higher passage rates for dietarylipids cle activity observedbetween the gizzard and are expectedin the former. However, in Ant- proventriculusin fowl (Dziuk and Duke 1972, arctic Prions (Pachyptiladesolata), a small pro- Duke 1982). The inactivity of the procellar- cellariiformsimilar in sizeto diving petrelsand iiform proventriculusallows gastriclipids and known to storestomach oils, the averagegastric aqueouscomponents to form and remain in sep- emptying rate for lipids was 4.6%.h-• (SE = arate phases.The denser aqueousdigesta ac- 1.1%.h -•, n = 2) and the mean retention time cumulates in the ventral portions of the pro- was 15.0h (SE = 3.5 h, n = 2; Robyet al. 1986a). ventriculusand in the gizzard. The pylorus is This provides strong evidence that both low ventral to the gizzard (and proventriculus)in passagerates and suitable gastricanatomy are the petrels and, consequently,aqueous digesta necessaryfor stomachoil formation. enters the duodenum first while lipid is re- Warham (1977) speculatedthat the formation tainedin the stomach.Low gastricmotility, slow of stomachoil is an adaptationfor exploitation gastricemptying, and the positionof the py- of a pelagicfood supplywhich is patchilydis- lorus relative to the proventriculus result in tributed,requiring long journeys between feed- stomachfunction analogous to a separatoryfun- ing sitesand necessitatingextended fasting pe- nel. riods for both adults and chicks. Others In penguins,like most birds, the proventric- (Ashmole1971, Laugksch and Duffy 1986,Obst ulus is neither large nor distensibleand is cra- 1986) emphasizedthe energeticadvantages to nial to the gizzard.Food items in the distensible both adultsand chicksof delivering chick meals crop are subjectedto acidic proteolysissequen- which consistmostly of stomachoils, particu- tially as they enter the small proventriculus. larly with regard to the increasein potential The relatively high gastric emptying rate re- foraging range of adults. Giant-petrel chicks, movesliquid digestafrom the stomachbefore however,are fed at leastas frequently as Gentoo separationinto a biphasicsystem can occur.The Penguin chicks, presumably becauseof the pylorus is situateddorsal to the gizzard, so if proximityof penguincolonies where mostchick phaseseparation occurred, lipids would tend to food is obtained. In the case of giant-petrels, passthrough the pylorus first. The large pan- there seems to be little advantage either for creasand muscularduodenum of penguinsare adults to concentratethe lipid content of chick consistentwith a rapid gastric evacuation of mealsor for chicksto storestomach oils for long food. fasts between meals. The crucialroles of both anatomyand phys- Diving-petrelsare apparentlyunique among iology in the formation of stomachoils are ex- procellariiformsin having lost the ability to emplified by diving petrels,the only procellar- form stomachoils (Roby 1986). Diving petrels, iiforms which do not store stomach oils. In adult like penguins,are pursuit-diversthat exploit a SouthGeorgia diving petrels(Pelecanoides geor- more concentratedand presumablymore pre- gicus),in contrastto giant-petrelchicks, the giz- dictablefood supply. The energeticallyineffi- zard and pylorusare situatedmore dorsalto the cient mode of flight in diving petrels suggests large,distensible proventriculus and would tend that they rarely, if ever, travel far to find food to empty lessdense digesta first. Moreover,pas- (Robyand Ricklefs1986). A potentialdisadvan- sagerates of digestathrough the gastrointesti- tage of stomachoil formationin pursuit-divers nal tracts of diving petrels are so high as to is the requisitelow passagerate of digestaand, preclude the formation of stomachoils (Roby consequently,low assimilationrate. Energy ex- 312 ROB¾,BRINKß AND Pr2,CE [Auk,Vol. 106 penditurerates of adultdiving petrelsare known , & . 1972b. Digestion of fat by young to be high relative to other procellariiformsea- pigs: a study of the amountsof fatty acid in the birds and similar to penguins(Roby and Rick- digestivetract using a fat-solubleindicator of ab- lefs 1986).Consequently, the ability to rapidly sorption. British J. Nutr. 28: 339-346. CHEAH, C. C., & I. A. HANSENß 1970. Wax esters in digest food and assimilate ingested energy stomachoil of petrels.Int. J. Blochem.1: 198-202. would be advantageous. CLARKE,A., & P. A. PRINCE.1976. The origin of stom- For seabirdsthat frequentlyexperience fast- ach oil in marine birds:analyses of the stomach ing periods,an energeticadvantage can be re- oil from six speciesof subantarcticProcellarii- alized by metabolizingstomach oils in prefer- form birds. J. Exp. Mar. Biol. Ecol. 23: 15-30. ence to fat from adiposetissue. Metabolizing CORTOT,A., S. F. PHILLIPS,& J. R. MALAGELADA.1979. stomachoils precludesthe energy costof syn- Gastricemptying of lipids after ingestion of an thesizingfat depotsfrom assimilatedfatty acids homogenizedmeal. Gastroenterology 76: 939-944. and of later mobilizing thoseenergy reserves, CROXALL,J.P., & P. A. PRINCE. 1980. The food of costs which amount to ca. 25-30% of the assim- GentooPenguins Pygoscelis papua and Macaroni PenguinsEudyptes chrysolophus at South Georgia. ilated energy (Ricklefs1974, Spady et al. 1976). Ibis 122: 245-253. This may be particularly significant for adults DUKE,G. E. 1982. Gastrointestinalmotility and its during the nonbreedingperiod if food supplies regulation. Poult. Sci. 61: 1245-1256. are patchilydistributed in time and space,ne- ß 1985. Raptorphysiology. Pp. 370-376 in Zoo cessitatingperiodic fasts and energy conser- and wild animal medicine, 2nd ed. (M. E. Fowler, vation. Ed.). Philadelphia, SaundersPubl. Co. 1986a. Alimentary canal:anatomyß regula- ACKNOWLEDGMENTS tion of feeding, motility. Pp. 269-288 in Avian physiology,4th ed. (P. D. Sturkie,Ed.). New York, We thank the Instituto Antartico Chileno, the Fuer- Springer-Verlag. za Aerea de Chile, and the Armada de Chile for their 1986b. Alimentary canal: secretion and logisticsupport. The following individuals were par- digestionßspecial digestive functionsßand ab- ticularly helpful during field studies:Luis Arias, sorption.Pp. 289-302 in Avian physiologyß4th tricio Eberhard, Alberto Foppiano, Siegfried King, ed. (P. D. Sturkie, Ed.). New York, Springer-Ver- Antonio Mazzei, Michel SallaberryßRuben Scheih- lag. ingßJos• Valencia, Jos• Yanez, and Gao Yaoting. We ß A. R. PLACE,& B. JONESß1989. Cineradio- alsoacknowledge the artisticassistance of Nina Stoy- graphicstudies of gastricemptying and gastroin- an and the generous gift of the tritiated glycerol testinal motility in Leach'sStorm-Petrel chicks triether from JamesMeyerß Veteran's Administration (Oceanodromaleuchorhoa). Auk 106: 80-85. Hospital, SepulvadaßCalifornia. David C. Duffy, Rob- DZIUK,H. E., & G. E. DUKE. 1972. Cineradiographic ert G. White, Alan WooIfßand an anonymousreview- studiesof gastricmotility in turkey. Am. J.Phys- er offeredvaluable suggestions for improvingearlier iol. 222: 159-166. drafts.This study was supportedby the National Sci- FORBESßW. A. 1882. Report on the anatomy of the ence Foundation (USA) PCM81-10819 and DCB87- petrels collected during the voyage of H.M.S. 09340ßa Lucille Markey Fellowshipßand a Depart- Challenger.Challenger Rep. 4: 1. ment of Navy Contract (N00014-86-K-0696)to A. R. HUNTER,S. 1983. The food and feeding ecologyof Place. This is contribution No. 115 from the Center the giantpetrels Macronectes halli and M. giganteus of Marine BiotechnologyßUniversity of Maryland. at South Georgia.J. Zool., London 200: 522-538. 1984. Breedingbiology and population dy- namicsof giant petrelsMacronectes at SouthGeor- LITERATURE CITED gia. J. Zool., London 203: 441-460. ß 1985. The role of giant petrelsin the South- ASHMOLE,N. P. 1971ßSeabird ecology and the ma- ern Ocean ecosystem.Pp. 534-542 in Antarctic rine environmentßPp. 223-287 in Avian biologyß nutrient cyclesand food webs (W. R. Siegfried, vol. 1 (D. S. Farner, J. R. King, and K. C. Parkes, P. R. Condy, and R. M. Laws, Eds.). Berlin, Eds.). New York, Academic Press. Springer-Verlag. BLIGH,E.G., & W. J. DYER. 1959. A rapid methodof IMBER,M. J. 1973. The food of Grey-faced Petrels total lipid extractionand purification.Can. J.Blo- (Pterodromamacroptera gouldii (Hutton)) with spe- chem. Physiol. 37: 911-917. cialreference to diurnalvertical migration of their CARLSON,W. E., H. S. BAYLEY.1972a. Preparation prey. J. Anim. Ecol. 42: 645-662. and use of glyceryl triether as an indicator of fat 1976. The origin of petrel stomachoils: a absorption. British J. Nutr. 28: 295-305. review. Condor 78: 366-369. April 1989] StomachOils and GI Transit 313

JACOB,J. 1982. Stomachoils. Pp. 325-340 in Avian latitude, plankton-feeding seabirds.Ph.D. dis- biology,vol. 6. (D. S. Farner,J. R. Kingßand K. sertation,Philadelphiaß Univ. Pennsylvania. C. Parkes, Eds.). New York, Academic Press. ß& R. E. RICKLEFS.1986. Energyexpenditure JIAN, R., N. VIGNERON,Y. NAIEAN, & J. J. BERNIER. in adult LeastAuklets and diving petrelsduring 1982. Gastricemptying and intragastricdistri- the chick-rearingperiod. Physiol. Zool. 59: 661- bution of lipids in man. A new scintigraphic 678. methodof study.Dig. Dis. Sci. 27: 705-711. ßA. R. PLACE, & R. E. RICKLEFS. 1986a. Assim- JOHNSONßM. L., J. J. CORREIA,D. A. YPHANTIS,& H. ilation and depositionof wax estersin planktiv- R. HALVORSON.1981. Analysis of data from the orousseabirds. J. Exp. Zool. 238: 29-41. analyticalultracentrifuge by non-linear squares ß M. SALLABERRY,& K. L. BRINK. 1986b. Notes techniques.Biophys. J. 36: 575-588. on petrels ()breeding on Ard- LAUGKSCH,R. C., & D.C. DUFFY. 1986. Food transit ley Island, South Shetland Islands. Ser. Cient. ratesin CapeGannets and Jackass Penguins. Con- INACH 34: 67-72. dor 88: 119-120. SMITH,J. L., C. L. JIANG,& J. N. HUNT. 1984. Intrinsic LEWIS,R.W. 1966. Studiesof glycerylethers of the emptying pattern of the human stomach.Am. J. stomach oil of Leach's Petrel Oceanodroma leuco- Physiol. 246: R959-R962. rhoa(Viellot). Comp.Blochem. Physiol. 19: 363- SPADY,D. W., P. R. PAYNE,D. PICOU,& J. C. WATERLOW. 377. 1976. Energybalance during recoveryfrom mal- MATTHEwS,L.H. 1949. The origin of stomachoil in nutrition. Am. J. Clin. Nutr. 29: 1073-1078. the petrels,with comparativeobservations on the STUBBS,D. F. 1977. Modelsof gastricemptying. Gut avian proventriculus.Ibis 91: 373-392. 18: 202-207. MEYER,J. H., E. A. MAYER,D. JEHN,Y. Gu, A. S. FINK, VOLKMAN,N.J., P. PRESLER,& W. TRIVELPIECE.1980. & M. FRIED.1986. Gastricprocessing and emp- Dietsof pygoscelidpenguins at King GeorgeIs- tying of fat. Gastroenterology90: 1176-1187. landß Antarctica. Condor 82.' 373-378. McLELLAND,J. 1979. Digestive system.Pp. 69-181 WARHAM,J. 1977. The incidence, functions and eco- in Formand functionin Birds,vol. 1. (A. S. King logicalsignificance of petrel stomachoils. Proc. and J. McLelland, Eds.). New Yorkß Academic New Zealand Ecol. Soc. 24: 84-93. Press. ß R. WATTS, & R. J. DAINTY. 1976. The com- MORGAN,R. G. H., & A. F. HOFMANN.1970. Synthesis position, energy content and function of the and metabolismof glyercol-3Htriether, a non- stomachoils of petrels(orderß Procellariiformes). absorbableoil-phase marker for lipid absorption J. Exp. Mar. Biol. Ecol. 23: 1-13. studies.J. Lipid Res. 11: 223-230. WARNER,A. C. I. 1981. Rate of passageof digesta OBST,B. S. 1986. The energeticsof Wilson's Storm- through the gut of mammalsand birds. Nutr. Petrel (Oceanitesoceanicus) breeding at Palmer Abstr. Rev. B 51: 789-820. Stationß Antarctica. Ph.D. dissertation. Los An- WIGGINSßH. S., & A.M. DAWSON. 1961. An evalu- geles,Univ. California. ation of unabsorbablemarkers in the studyof fat PATRON, R. A., & G. F. KR•USE. 1972. A maximum- absorption.Gut 2: 373-376. likelihood estimator of food retention time in WILSONßR. P., G. D. LA COCK, M.-P. WILSON, & F. ruminants. British J. Nutr. 28: 19-22. MOLLAGEE.1985. Differential digestionof fish PLACE, A. R., & D. D. ROBY. 1986. Assimilation and and squid by JackassPenguins. Ornis Scandi- depositionof dietary fatty alcoholsin Leach's navica 16: 77-79. Storm-Petrel,Oceanodroma leucorhoa. J.Exp. Zool. WINGATE,D. L., R. J. SANDBERGß& S. F. PHILLIPS. 1972. 240: 149-161. A comparisonof stableand •4C-labelledpolyeth- RHOADES, D. D., & G. E. DUKE. 1975. Gastric function ylene glycol as volume indicatorsin the human in a captive American Bittern. Auk 92: 786-792. jejunum. Gut 13: 812-815. RICKLEFS,R. E. 1974. Energeticsof reproductionin ZISWILER,V., & D. S. FARNER.1972. Digestion and birds. Pp. 152-297 in Avian energetics(R. A. the digestivesystem. Pp. 343-430 in Avian biol- PaynterJr., Ed.). Cambridge, Massachusetts, Nut- ogyßvol. 2 (D. S.Farner and J. R. King,Eds.). New tall Ornithol. Club, Publ. No. 15. York, Academic Press. ROBY,D. D. 1986. Diet and reproductionin high