J. Field Ornithol.

, 68(1):102-112

SEX DETERMINATION OF ADULT PENGUINS USING MORPHOMETRIC

HUMBOT CHARACTERS

CAgLOS B. ZAVALAGA 1 AND ROSANA PAREDES

Proyecto Conservaci6n Punta SanJuan de WildlifeConservation Society Parque Josg Acosta de 187
Lima 17 Perd

Abstract.-TheHumboldt Penguin (Spheniscus humboldti) one of the mostendangered is and least-studied species penguins.Studies its breedingecology needed to develop of of are
managementplans for its conservation,and sexdetermination is essential measurecertain to reproductive parameters. Humboldt Penguins, like other species penguins, of showed sexual sizedimorphism, with malesbeingheavierand largerthan females. One discriminant function wasderivedfrom externalmeasurements 297wild adultHumboldtPenguins known of of sex (146 femalesand 151 males) at Punta SanJuan, Peru. The sex of 95% of thesebirds could be correctlydetermined using the discriminantfunction D -- -38.98 + 3.16(WH) + 3.69(BL) where WH is width of the head and BL is bill length. In addition, the sexof 91% of the birdswerecorrectly classified means the bill lengthusingthe univariate by of function D,= -6.31 + BL. In both equations,if D > 0 the bird wasclassified male, if D < 0 it as was a female, and if D = 0, the sex could not be identified. The discriminant function D 1 wasnot accurate sexa groupof captive to HumboldtPenguins (71% successfully classified) because there weredifferences sizeof the head between in wild and captive birds.However,

weobtained 83%of cases successfully classified using function 2asbilllengthwas the D similar
betweenwild and captivebirds. The discriminantfunction may not be applicableto other wild and captivepopulations HumboldtPenguins. of
DETERMINACION DEL SEXO EN ADULTOS DE SPHENISCUS HUMBOLDTI UTILIZANDO CARACTERISTIC MORFOMTRICAS

Sinopsis.--Dentro las 17 especies pingtinos,el de Humboldt (Spheniscus de de humboldti) es uno de los ms amenazados menosestudiados. y Para 11evar cabo planesde manejo a dirigidosa su conservaci6n, necesarios son estudios sobresu ecologiareproductiva, siendo la determinaci6n sexoesencial del para medir algunos susparimetrosreproductivos. de Los pingtinos de Humboldtson sexualmente dim6rficos susmedidas en corporales, siendolos machosmilspesados grandesque las hembras.Una funci6n discriminante obtenidaa y fue partirde lasmedidas 297 adultos sexo de de conocido (146hembras 151machos) Punta y en SanJuan, Per6. E1sexode estas aves pudo ser identificadocorrectamente un 95% de los en
casosutilizando la funci6n discriminante D 1 = -38.98 + 3.16 (WH) + 3.69 (BL), donde

WH esel anchode la cabeza BL la longituddel pico. De igualmaneraseidentific6el sexo y del 91% de los pingtinos por medio de la ecuaci6nunivariada = -6.31 + BL. Para D2 ambas ecuaciones, D > 0 el avefue clasificada si comomacho,si D < 0 comohembray si
D = 0, el ave no era clasificada. La funci6n discriminante D no fue tan exitosa en la

determinaci6ndel sexode pingtinos cautivos (s61o 71% de los casos correctamente el fue clasificado) debido a que estas avesexhibieroncabezas milsgrandes que aquellas medidas en el campo.Sinembargo, sexodel 83% de lospinguinos el pudoseridentificado utilizando la ecuacion D,debido a que la longituddel pico fue similarentre los ping/iinos silvestres y los cautivos. funci6n discriminante La desarrollada necesariamente no pudiera aplicarse a otraspoblaciones pingtinos de Humboldt ya seasilvestres cautivos. de o

The Humboldt Penguin (Spheniscus humboldti)is one of the most threatened and least-studied speciesof penguins (Boersma 1991, Hays
Currentaddress: Apartado Postal180606, Lima 18, Perd.
102

Vol. No. 68, I

Sexing Humboldt Penguins

[ 103

1984). Becauseof a sharp decline in the wild population (Hays 1984), studiesof their ecologyand behaviorare needed for developingeffective conservationplans. To achieve most of these studies,the ability to determine the sex of individuals is essentialfor investigatingsex-specific behavior (Davis 1988) or measuring certain reproductive parameters (CCAMLR 1994). Determining sex of captiveHumboldt Penguinsin zoos is also important in order to maximize successful reproduction in captive birds (Cheney 1990, McGill and Perkins 1993, Steveson1993). Humboldt Penguins,like most seabirdspecies, lack plumage characters by which sexesmay be recognized. However, small differences in morphometric parametersreveal sufficient dimorphism to distinguishsexes

(Murphy 1936:452, Scholten1987). Characters usedto determinesexin penguinsincludevent measurements (Boersma and Davies1987), sexual
behavior (Davis 1988, Edgington 1989, Scholten 1992), morphometric differences in mated pairs (Edgington 1989) and temporal attendance during incubation (Kerry et al. 1993). Some of thesemethodshave been used successfully the wild and in captivity,but only on reproductive in individualsduring early parts of the breeding season. Cloacalexamination has been used in wild Adfilie Penguins (Pygoscelis adeliae,, Lishman 1985,

Sladen 1978), ChinstrapPenguins,(Pygoscelis antarctica; Lishman1985) and captive Humboldt Penguins (Yamazakiet al. 1994). However, this
technique requires trained researchers and specialized equipment to complete with minimal stress the animal. Other methods,suchaschroto mosome analysis,hormone analysis,and laparoscopyare effective but time-consumingand expensive (Edgington 1989). Sex determination by discriminant analysis external measurements of has been used successfully the wild in severalpenguin species in suchas MagellanicPenguin (Spheniscus magellanicus, Scolaroet al. 1983), Yelloweyed Penguin (Megadyptes antipodes, Darby and Seddon 1990), all pygoscelid penguins (Amat et al. 1993, Kerry et al. 1992, Williams 1990), Little Blue Penguin (Eudyptulaminor novaehollandiae, Gales1988), and others. Discriminant analysis uses morphometric differences between known male and female birds to calculate a function that predicts the sex of unknown individuals. This technique is reliable, practical, fast, inexpensive,non-invasive, and can be used outside of the breeding season. The objectives this studywere (1) to determine discriminantfuncof tion(s) using external measurements with which to determine the sex of wild adult Humboldt Penguinsat Punta SanJuan, Peru and (2) to test the applicabilityof thesefunctionson captiveHumboldt Penguins.
MATERIALS AND METHODS

Measurements wild penguins.--BetweenSeptember 1992 and June of 1993, the bodies of 223 dead adult Humboldt Penguinswere recovered from the port of San Juan de Marcona, 3 km north of Punta SanJuan (1522'S,7512'W),a 54-ha guano bird reserveon the southerncoastof Peru. This reserve holds one of the largest populations of Humboldt Penguinsin the country (Hays 1984). The birdswere accidentally caught

104]

c. B. Zavalaga R. Paredes and

j. Field Ornithol.
Winter 1997

and drowned in the nets of small boatsin the courseof normal fishing operationsbasedat Punta SanJuan along 50-60 km of the coast(within 5 km offshore).Thus, the penguinscaughtmay originatefrom PuntaSan Juan as well as from neighboring coloniessuch as Sombrerillo, San Nicolils, and San Fernando.

Immediatelyafter collection,carcasses wereweighedto the nearest100 g with a 10-kgPesolaspring balance. Head and bill were measuredwith vernier calipersto the nearest0.1 mm and the flipper wasmeasuredwith a ruler to the nearest 0.1 cm. The measurements included: total length of the head (LH) from supraoccipital the tip of the bill; width of the to head (WH) in a crevice just posteriorto the bulge behind the eyes; bill length (BL) from the edge of implantationof feathersto the tip of the

culmen;bill depth (BD) measured dorso-ventrally the nostrils at and flipper length (FL), maximumflattenedchord,from the humero-radialjoint
to the tip of the flipper (Fig. 1). The sex of each individual was determined by gonad examination after dissection.
Discriminant functions were derived from the measurements taken

from the 223 carcasses (analysis sample). These data were analyzedby stepwise discriminantfunction analysis with the SPSS/PC + 4.0.1 statistical package (Norusis 1990). The Wilks' lambda ratio wasused as a criterion for variable selection.All other statistical analysis were carried out using the SYSTAT5.0 statistical package(Wilkinson 1991). To testthe accuracy our results, measured breedingwild penof we 74 guins of known sex (validation sample). These animalswere banded and measuredbetween 1992 and 1994 at Punta SanJuan as part of a longterm studyof the breeding ecologyof this species. The sexof thesebirds wasidentified from their position during copulation. Body masswasexcluded from the discriminantanalysis becausethere was a significantamount of seasonal variability (CV = 13.6%). Also, the majority of dead penguins were wet and their lungs were full of water (due to drowning) at the time of weighing, which would bias this measurement(Table 1). Thesepenguins were 4.8% heavierthan a sampleof 288 living penguins of known sex (the sex of these individualswasdetermined from the discriminantfunction reported below) weighedat the sametime at Punta SanJuan (males: -- 4711.06 g, SE -- 31.55, range -- 3450-6000, n = 165; females: = 4047.39 g, SE -- 35.15, range = 2950-5400, n = 123;pooled: = 4427.62g, SE = 30.42, range = 29506000, n -- 288). These differences were significant (z-test,z = 98, P <
0.05).

All measurements were normallydistributed(Kolmogorov-Smirnov one sample test, P > 0.05 in all tests)with some overlap between sexes. In
order to determine a discriminant function, we established the same cri-

FIGURE Variables 1. measured sexadult HumboldtPenguins. = bill length;BD = bill to BL depth;LH = length of the head;WH = width of the head;FL = flipper length.

Vol. 68, No. I

Sexing HumboldtPenguins

[105

106]

C. B. Zavalagaand R. Paredes

Field Ornithol.
Winter 1997

Vol. No. 68, 1

Sexing Humboldt Penguins

[ 107

teria used by Kerry et al. (1992) for discriminant analysis: characters (1)
were not linear combinations of each other; (2) correlation coefficients
between characteristics used for the final discriminant function were less

than 0.60; and (3) the variance-covariance matriceswere not significantly

different (Box's M-statistic = 6.02; F = 1.98; P = 0.1140).

Measurements captive of penguins.--to test the efficiencyof the discriminant functions for determining the sex of captiveHumboldt Penguins,
we evaluated measurements of 35 birds of known sex. These birds were

measuredduring May 1995 by the bird keepersat the Metro Washington Park Zoo, Portland, Oregon, USA. The sex of these individualswas determined by genetic karyotyping.
RESULTS

Determination sexof wildpenguins.--Themeasurements 223 dead of of adult Humboldt penguins (112 males and 111 females) are presentedin Table 1. Maleswere significantly larger than females (z-test,P 0.05) for all morphometric characters (Table 1). Givensomeoverlapping maleand female distributions of each univariate character, the interception point between both curveswasdetermined from the analysis sample,obtaining the following valuesfor sex separation:LH = 13.09, WH = 4.94, BL =
6.31, BD = 2.42, FL = 15.29 (all measurements in cm). If the measure-

ment of some of thesevariableswashigher than its respective cutoff value the penguinwasclassified a male, if lower asa female, and if equal the as sex could not be determined. The percentageof cases correctlyclassified

for the validationsamplewere 91%, 89%, 88%, 78% and 70% for BL,
BD, LH, WH and FL, respectively. Following the criteria of Kerry et al. (1992), only WH and BL were significantly selected the stepwise by discriminant analysis (Wilks'Lambda = 0.321, X2 = 250.03, df = 2, P 0.001). On the basisof Wilks' Lambda values and the number of casescorrectly classified,the inclusion of the other variables wasnot justified. By usingWH and BL simultaneously, we obtained the following unstandardizeddiscriminantfunction:
D = -38.60 + 3.36(WH) + 3.48(BL), (1)

If D 0 the penguin wasclassified male and if D 0 the penguin as

wasclassified female. By usingthis function,we correctlysexed94% as of the wild penguinsin the analysis sample:103 of 111 females (93%)
and 106 of 112 males (95%) were correctly identified. When we applied the discriminantfunction to the validation sample (39 males,35 females),we correctlyclassified 97% of the birds,with two males and no femalesbeing misclassified.

The classification accuracy the analysis of sample(n -- 223, 94%) and the validationsample (n = 74, 97%) were similar,suggesting minimal sampling bias.The samples were combinedand a new functionbasedon the larger samplesizewasderived.This secondunstandardized function alsoaccurately discriminated betweenthe sexes (Wilks' Lambda = 0.3132,
X2 = 341.31, df -- 2, P 0.001):

108]
20

C. B. Zavalaga and R. Paredes

J.Field Ornithol.
Winter 1997

18

16

14

-4

-3

-2

-1

0
Discriminant

1
scores

FIGURE 2.

Distribution of discri finant scores males(filled bars) and females(open bars) of wild Humboldt penguinsfrc m Punta SanJuan, Peru.

D -- -3t

.98 + 3.16(WH) + 3.69(BL)

(2)

Again, if D waspositive he bird wasa male, if negative,a female. This

function correctlyclassified 95% of the penguins.The slightly higher accuracy and the larger sample size used to derive function 2 suggestit would be the mostusefulfunction. The averageof the discriminantscores
(group centroid) for femaleswas - 1.50 and for males 1.45. No birds with scoreshigher than 0.69 (males) or with scoresbelow -1.17 (females) were misclassified (Fig. 2). Thus, for greatestaccuracy, recommend we that sex of individualswhosediscriminantscore falls within this range
should not be identified.

Determination sexof captive of population.--Exceptfor bill length, captive penguins were significantlylarger than wild birds (t-test, P < 0.05) for all measurements. Bill depth showedthe highestdifference, captive birds having 18% thicker bills than wild penguins.On the other hand, except for the length of the head, there were significantdifferencesbetween sexesin all measurements(t-test, P < 0.05), males being larger
than females (Table 2).

We obtained poor sex discriminationof captivebirds using equation

Vol. No. 68, 1

Sexing Humboldt Penguins

[ 109

TABLE Morphometric data of captiveHumboldt Penguins(Spheniscus 2. humboldti) held in the Metro Washington Park Zoo. Rangeis givenin parentheses. Bodymass in grams, is
other measurements are in cm.
t-test

Males (n = 19) Females (n = 16) Pooled (n = 35)

Variable +_ SE +_ SE -+ SE t P

Bodymass

4802.6 _+100.5
(4100-5850)

4328.0 + 92.6
(3500-5000)

4585.7 _ 79.2
14.15 +_0.094
5.36 +- 0.042

3.42 <0.001
1.08
3.7

Length of the head 14.24 _+0.126 (13.33-15.03)

Width of the head 5.49 _ 0.045

14.04 -+ 0.145 (13.17-15.08)

5.22 -+ 0.055

0.28
<0.001

(5.19-5.99)

(4.83-5.64)

Bill length Bill depth

6.44 +_0.075
(5.68-6.92)

6.03 +- 0.057
(5.66-6.40)

6.25 -+ 0.059 2.85 - 0.037

4.16 <0.001 4.04 <0.001

2.97 _+0.045
(2.96-3.25)

2.71 +- 0.04
(2.51-3.04)

(2), derivedfrom wild penguins. This function correctlyclassified 71% of penguins. Maleswere successfully classified 84% of cases in while females had 56% successfulallocation. However, the sex of 83% of the birds

could be correctlyidentified using only bill length as the variablefor sex separation.We choseonly this variable becausethe means (t-test, t = 0.9, P = 0.36) and variances(F-test,F4,ee2 1.29, P = 0.14) were not signif= icantly different between captiveand wild birds.
DISCUSSION

All penguinsexhibit somesexualsizedimorphism,with malesgenerally being heavier and having larger flippers and bills than females (Farbain
and Shine 1993, Williams 1995). We have shown that for Humboldt Pen-

guins,malesare alsolarger than females. Discriminant analysis has been used successfully determine the sex to of penguinsand other seabirds suchasgulls (Evanset al. 1995, Fitzpatrick et al. 1988), cormorants (Glahn and McCoy 1995), and petrels (Lorentsen and Rov 1994, van Franeker and ter Braak 1993). The accuracyof our discriminant function for sexingwild adult Humboldt Penguinsis similar to thosereported for other penguin species (Amat et al. 1993, CCAMLR 1994:Appendix 2.1, Gales 1988, Scolaro et al. 1983, Williams 1990). Agnew (1992) has shown that if the discriminant function is greater than 80% successful the sample less and size than600 birds,the apparent mean
(derived from discriminant functions) of a character for males and fe-

malesis unlikely to be significantly different from the true mean (derived

where the sex of birds is known). In this case, the sexes can be distin-

guishedand sex determination by discriminantanalysis give acceptwill

able indices.

Although head width and bill length yielded the mostaccuratediscrimination of sex, a procedure requiring only one measurementcould be useful in some cases.For example, when sex determination of a large

110]

c. B. Zavalaga R. Paredes and

J.Field 1997 Ornithol. Winter

number of birds is required in a short period of time. In addition, birds cannot be sexedreliably using head measurements during the molt due to the deposition and depletionof fat asocurrsin other penguinspecies

(Boersma 1976, Cooper 1978). In thesecases recommendto use the we

equation D = -6.31 + BL as criterion of sex discriminationbecausethe bill length wasone of the simplest measurement take on living birds to and the numberof cases correctly classified (91% of successful allocation) was higher than other univariate characters. The discriminantfunction reported in this paper to sex wild adult Humboldt Penguinsis an easy,fast, inexpensive,and minimally stressful technique.However,studies other penguinspecies of (BostandJouventin

1990,Duffy 1987,Gales1988,Gandiniet al. 1992,Kerryet al. 1992,Murie et al. 1991), gulls (Evanset al. 1995), and petrels (van Franeker and ter Braak 1993) havenoted geographic variationin morphometry.Thus,geographicvariation mayinvalidatethe function'sapplicabilityto populations other than the one from which it was derived. Sex identificationby external measurements other wild populationsof Humboldt Penguins of must be interpreted with caution until a sample of known sex animals can be developed to act as a control. It was not appropriate to apply the two-variablediscriminant function derivedfrom wild penguinsat Punta SanJuan to sexcaptivebirdsbecause there were differences sizeof the head betweencaptiveand wild birds. in The extent by which captivebirdsdiffer from wild onesmaybe significant. Also,Humboldt Penguinsheld in the Metro Washington Park Zoo exhibited overgrown bills, not so much in length asin depth. We observed the samepattern in 30 captiveHumboldt Penguinsheld in the Parque de Las LeyendasZoo in Lima, Peru. The causes this extra depositionof of material are unknown. It may be nutrition-related or due to the absence of bill abrasionfrom digging,fighting or swimming, which is normal in wild penguins. However, obtained83% successful we allocation captive of birds using the univariatefunction D = -6.31 + BL obtainedfrom wild birds, as bill length wassimilar betweenwild and captivebirds. Thus, if the meansand variances the bill length are similarbetweenwild penof guinsat Punta SanJuan and other captivepopulations Humboldt Penof guins,the use of this function may be accomplished.
ACKNOWLEDGMENTS

We are deeplygrateful to William Alderete for his help in the collectionand dissection of deadpenguins. thank CynthiaCheneyand the bird keepers the Washington We at ParkZoo, Portland, Oregon, for providingmeasurements captiveHumboldt penguins.The sex of of thesebirdswasdeterminedby genetickariotypingin the Daniel F. and Ada L. Rice Conservation Biologyand Research Center,Brookfield,Illinois carrild out the karyotypeof these birds. Mike Cullen, Mike Schwartz,Ian Stirling, Patricia Majluf, and Juan Carios Riveros improved earlier draftsof this manuscript.To all of these,we are sincerely grateful.Also,we would like to thank Miguel Camposand Professor Jaime Garcia (Departmentof Physics and Mathematics, Universidad Peruana Cayetano Heredia) for advice on multivariate statistical analysis. This work wassupportedby a grant from Wildlife Conservation Societyto the authors. Fishingdata and dead penguinswere recoveredfrom the port of Punta SanJuan

Vol. No. 68, 1

Sexing Humboldt Penguins

[ 111

through the ArtisanalFisheries Monitoring Program,Punta SanJuan project, by a grant to Dr. P. Majluf from Wildlife ConservationSociety.
LITERATURE CITED

AGNEW, D.J. 1992. Can we use discriminant function analysisto sex penguins prior to calculating index of a morphometric an characteristic? 259-272, in Selected Pp. papers,
1992 (SC-CAMLR-SSP/9). CCAMLR, Hobart, Australia.

AMAX, A., J. VTNUEkA, J. ANDM. FERRER. 1993. Sexing Chinstrap Penguins(Pygoscelis antarctica)by morphological measurements. ColonialWaterbirds16:213-215. BOERSMA,D. 1976. An ecological P. and behavioral studyof the Galapagos Penguin. Living
Bird 15:43-93.

1991. Statusof wild and captivepenguinpopulations. Trendsin Ecol. and Evol.6:

381-382.

ANDE. M. DAXqES. 1987. Sexing monomorphicbirds by vent measurements. Auk

104:779-783.

CCAMLR. 1994. CCAMLR Ecosystem Monitoring Program:standardmethodsfor monitoring parametersof predatory species. Committee for the Conservationof Antarctic Marine Living Resources, Hobart, Australia. BOST,CH-A, ANDP.JOtWrNTTN.1990. Evolutionaryecologyof GentGoPenguins(Pygoscelis papua). Pp. 85-112, in L. S. DavisandJ. T. Darby,eds.Penguinbiology. Academic Press Inc., San Diego, California. CHENEY, A. 1990. Spheniscus C. penguins:an overviewof the world captivepopulation. Spheniscus Penguin Newsietter3:12-17. COO'ER, 1978. Moult of the black-footed J. penguin.International Zoo Yearbook18:22-27. DARY,J.T., ANDP.J. SEDDON. 1990. Breedingbiologyof Yellow-eyed Penguins(Megadyptes antipodes). 45-62, in L. S. Davisand J. T. Darby, eds. Penguin biology.Academic Pp. Press Inc., San Diego, California. DAXqS, S. 1988. Coordination of incubationroutinesand mate choicein Adelie Penguins L. (Pygoscelis adeliae).Auk 105:428-432. DUFFY, D.C. 1987. Ecological implications intercolonysize-variation Jackass of in Penguins.
Ostrich 58:54-57.

EDGINGTON, G. 1989. Behavioural D. and morphological sexingof the Humboldt Penguin (Spheniscus mboldti).SpheniscidPenguin Newsletter1:14-20. hu EvANs, R., P.M. CAVANAGH, W. FRENCH, B. G. BLODGET. D. T. AND 1995. Identifyingthe sex of Massachusetts Herring Gullsby linear measurements. Field Ornithol. 66:128-132. J. FARAIN, AND R. SHINE. 1993. Patternsof sexual size dimorphism in seabirdsof the J., Southern Hemisphere. Oikos 68:139-145. FITZPATRICK, C., C. G. GUEP, ANDT. L. KING. 1988. Sex determinationin gray gulls, L. RA, Larus modestus, usingexternal measurements discriminant and analysis. Estud.Oceanol.
7:71-74.

GALES, 1988. Sexingadult Blue Penguins externalmeasurements. R. by Notornis35:71-75. GANDINI, A., E. FeERE, P. ANDT. M. HOLIK. 1992. Implicancias lasdiferencias el tamafio de en corporal entre coloniaspara el uso de medidasmorfomtricascomo mtodo de sexado
en Spheniscus magellanicus. Hornero 13:211-213. El GLAHN, F., ANDR. B. McCoY. 1995. Measurements winteringDoubled-crested J. of Cormorants and discriminantmodelsof sex.J. Field Ornithol. 66:299-304. I-lys, C. 1984. The humboldt penguin in Peru. Oryx 18:92-95. KRy, K. R., D. J. AGNEW, R. CkARr, ANDG. D. ELSE.1992. Use of morphometricpaJ. rametersfor the determination of sex of Adlie penguins.Wildl. Res. 19:657-664. --, J. R. CkAP, ANDG. D. ELSE. 1993. Identification of sex of Adlie penguinsfrom r,E, observations incubating of birds.Wild. Res.20:725-732. LSHMAN, S. 1985. The comparative G. breedingbiologyof Adelie and ChinstrapPenguins, Pygoscelis adeliae and P antarctica Signyisland,SouthOrkney Islands. at Ibis 127:84-99. LORENTSEN, S-H., AND N. ROv. 1994. Sex determination of antarctic petrels Thalassoica antarctica discriminant by analysis morphometriccharacters. of Polar Biol 14:143-145.

112]

C. B. Zavalaga R. Paredes and

J.Field 1997 Ornithol. Winter

MCGILL,P.,)NDG. PERKINS. 1993. Humboldt PenguinSSPreport. Spheniscus PenguinNewsletter 6:2-4.

Mumv,J.O., L. S. D^vIs,D I. G. McLF_x. 1991. Identifyingthe sexof FiordlandCrested Penguins morphometriccharacters. by Notornis 38:233-238
MURPHY, R. C. 1936. Oceanic birds of South America, Vol. 1. The American Museum of

Natural History, New York. NoRusIs,M.J. 1990. SPSS advanced statistics user'sguide. SPSS Inc., Chicago. SCHOLTEN, C.J. 1987. Breedingbiologyof the Humbolt penguin,Spheniscus humboldti, at
Emmen Zoo. International Zoo Yearbook 26:198-204

1992. Choiceof nest-site and mate in Humboldt Penguins(Spheniscus humboldti). Spheniscus Penguin Newsletter5:3-13 ScOIRO,J. A., M. A. I-MLL,ANDI. M. XIMINEZ.1983. The MagellanicPenguin (Spheniscus magellanicus): sexingadultsby discrininant analysis morphometriccharacters. of Auk
100:221-224

SLaDEN, W.J.L. 1978. Sexingpenguins cloacascope. by InternationalZoo Yearbook 18:77-80. STFWSON, 1993. 1993meetingof the EEP PenguinTAG. Spheniscus M. PenguinNewsletter
6:10-12.

VAN FP,qFKR, A., D J.

Fulmarine Petrels from

C. J. F TF BmK. 1993. A generalizeddiscriminantfor sexing

external measurements. Auk 110:492-502.

YAMAZAKI,Y., A. YAMATO, A. YAMADA, AND K. NISHIWAKI. 1994. Sex determination

Conservation 7:7-11.

of Hum-

boldt Penguins(Spheniscus humboldti) using an original designedrestraint.Penguin

WIIKNSON, 1991. SYSTAT:the system statistics. L. for Systat, Inc., Evanston,Illinois. WILLLXMS, D. 1990. Annualvariationin breedingbiologyof the Gentoopenguin,Pygoscells T. papua,at Bird Island,SouthGeorgia Zool. Lond. 222:247-258 J. 1995. The Penguins: Spheniscidae. Oxford University Press, New York.

Received 23 Jan. 1996; accepted 24 Apr. 1996.

SEVENTY-FIFTH ASSOCIATION

ANNIVERSARY OF HELD

MEETING

ORNITHOLOGISTS

The Association Field Ornithologists hold its 75th anniversary of will meetingjointly with the AmericanBirding Association SanJosfi,CostaRica, 21-28 Jul. 1997, hostedby the in Associacitn Ornitoltgica Costa de Rica.Scientic paper sessions, symposia, workshops, field and tripsare planned.There will be a focuson conservation research CostaRicaand postand in meeting trips to major field stations researchers for and educators. Travel and per diem funds are being soughtfor Latin American participants.Likely symposium topicsinclude Conservation New World Psitticines, of Bird Sound Recording,Raptor Migration between the Americas, Bird Observatories, and Birds in Environmental Education for Indigenous Peoples.Registration will begin 10 Feb. 1997. Call for papers:ScottK. Robinson, Illinois Natural History Survey, 607 E. Peabody Dr., Champaign, IL 61820 (217-333-6857;
scottr@mail.inhs.uiuc.edu), and Rafael Campos Ramierez (fax 506-494-3346). For informa-

tion aboutsymposia (suggestions additional for topics welcome): Charles Duncan,Institute D. for Field Ornithology,Universityof Maine at Machias,Machias,ME 04654 (207-255-1358); cduncan@aced.umm.maine.edu). Registrar:Carol Wallace,American Birding Association, P.O. Box 6599, ColoradoSprings, CO 80934-6599(800-850-2473 719-578-1614; 719or fax 578-1480; member@aba.org).