David C. Dulfy (visiting professor' in the Programa de Vida Silves tre, Uni~ versidad Nacional, Her'edia, Costa Rica) was formerly at the Percy Fitz­Patrick Institute of African Ornithol~ ogv, University of Cape Town, South Africa. He has worked on seabirds in Peru, the Galapagos, Chile, and South Aft'ica, and is currently study~ ing the scar'let macaw in Costa Rica. Rory P. Wilson, r'esearcher a t the Institut far Meer'eskunde an de,. Uni~ versitat, Kiel, Federal Republic of Germany, obtained his Ph.D. at the University of Cape Town where he studied the Joraging ecology oj Aji'i­can penguins and their prey. His present work concerns the for-aging of European auks and Antarctic pen­guins. Robe,·t E. Ricklefs, professor' ofbiologJJ a t the University of Penn­sylvania, has worked extensively on the growth ofbir'ds. More recently, he has stwliedJoraging str'ategies oJGa­lapagos seabirds. Stephen C. Broni earned his M.Sc, from the University of Cape Town, studying African pen~ guins. Helen Veldhuis (M.Sc., bot­any) is coauthor of a book on endangered South African plants.

480

David C. Duffy, Rory P. Wilson, Robert E. R{{;Kle{s, Stephen C. Broni, and Helen Veldhuis

Penguins and Purse Seiners: Competition or Coexistence?

Competition with a commercial purse-seine fIShery has been suggested as the main cause of the decrease in the population of African penguins Spheniscus demersus off the coast of southern Africa. Comparison of fIShery landings with growth and diet of African penguins at Marcus Is­land, South Africa, suggests that direct competition during the breeding season is limited because penguins, beingflightless, are constrained to nearshore waters where purse seiners cannot fISh. Penguins also take smaller fISh than does the flShery,jurther reducing the potentialfor di­rect competition. In contrast, competition away from the breeding is­lands may be much stronger. Juvenile penguins from Marcus Island spend their first year at sea in an area of intense fIShery activity. Their survival is very low compared with that of juveniles from a colony that spend their first year in a bay closed to commercial fIShing. Results of this study show that interactions between penguins and fISheries are complex, and that competition between the seabirds and fISheries can­not be invoked to explain changes in seabird numbers until the compet­itive mechanism has been examined.

Competitive interactions between seabirds and commercial fisheries have been documented ill the North Sea (Furness 1978) and in the Peru­vian (Duffy 1983, Jordan {jp Fuentes 1966, Schaefer 1970), Californian (Ainley {jp Lewis 1974) , and Benguela upwelling systems (Abrams 1983; Burger {jp Cooper 1984; Crawford {jp Shelton 1978, 1981; Furness {jp Cooper 1982). Competition with a fishery may alter normal foraging behavior and diet of certain species (Fisher 1952), reduce reproductive rates (Crawford {jp Shelton 1978, Duffy 1983), and ultimately reduce populations (Crawford Ii» Shelton 1981, Nelson 1978, Valdivia 1978).

Effects,have also been beneficial for some species. Fishing fleets pro­vide food in the form of offal (Abrams 1983, Fisher 1952) or reduce com­petition from other predators such as whales or fishes (May et al. 1979, Schweigger 1964).

Most conclusions about seabird - fishery interactions have been based either on post hoc analyses or inferences. The former seek to explain past changes in bird numbers by relating them to changes in fishery land­ings. A common irtference is that - because seabirds take large amounts of particular species also exploited by commercial fisheries ­the two compete. The ideal approach to understanding seabird-fishery interactions would be to examine concurrent changes in seabird diets and reproductive performance with respect to changes in fishery land­ings. A change in diet, or decline in breeding success and population

NATIONAL GEOGRAPHIC RESEARCH 3(4):480-488 (1987)

size following intensified commercial exploitation of prey species, would suggest competition. The authors used such a method to examine the relationship between the African or jackass penguin Spheniscus de­mersus (Figure 1) and the South African purse-seine industry.

The African penguin population has decreased dramatically since 1900 (Cooper 1984, Crawford &; Shelton 1978, Frost et a!. 1976, Shelton et a!. 1984). Most of the penguin's prey are pelagic schooling fish (Wil­son 1985). Flightlessness requires the penguin to forage on a spacially and temporally predictable food supply (Frost et a!. 1976, Wilson 1985).

Cape pilchard Sardinops ocellata and maasbanker 01' Cape horse­mackerel Trachurus capensis stocks were heavily exploited during the late 1950s and early 1960s, resulting in the collapse of fisheries in 1964 and 1965 (Stander &; Le Roux 1968). In 1965, small-meshed nets (12.7 mm) were introduced to allow exploitation of Cape anchovy En.'jl·aulis japonicus capensis. Since then, anchovy has been the dominant species in pelagic catches - approximately 80% of the total landings (Butter­worth 1983). The collapse of the pilchard population during the mid-1960s may have forced the African penguin to switch to the anchovy, a spatially less predictable species (Crawford 1980a, b).

Methods Fieldwork was carried out on Marcus Island (33°03'S, 17°58'£) in Sal­danha Bay, off the southwest coast of South Africa, during July 1980 to June 1981, March to October 1982, and February 1983 toJuly 1986;

Penguin nests containing chicks were marked with numbered poles, . and chicks were individually marked with picric acid 01' flipper-banded if large enough. Culmen lengths of the chicks were determined using Vernier calipers (accurate to 0.1 mm), and mass was measured to ± 50 g using a Pesola spring balance. The chicks were remeasured between five and seven days later.

The study was divided into 41 nonsequential growih periods listed to the right. Chicks were grouped into three age-classes: young « 15 days), mid-age (15 to 42 days), and old (> 42 days). By employing three age-classes, the authors hoped to identify the stage at which chicks were most susceptible to fluctuations in fish availability.

Growih data were analyzed according to methods described by Rick­lefs et a!. (1984). For each of the age groups, a regression was calculated (WTI = a + b X age + c X age2; a, b, c are constants) which took into account the fact that initial weight (WTI) is related to age. The residuals about this regression (RWTI), that is, the adjusted initial weights, were used to calculate a second regression (DWT = a + b x-age + c X

age2 + d X RWTI; a, b, c, d are constants), where DWT is change in weight over a growih period. Residuals about this regression are the ad­justed weight increments (RDWT) for which the effects of age (within size group) and adjusted initial weight (indicating recent feeding histo­ry) have been removed statistically. The adjusted weight increments were then treated using standard analyses of variance to determine vari­ationamong growih periods. The GLM procedure of the StatisticafAnal­ysis System was used for the analyses.

Adult Diet Penguin diet was examined by sampling adult birds as they returned to the island in late afternoon or early evening. During each growth period, at least 10 samples were collected by stomach pump, using a water off-

PENGUINS VERSUS PURSE SEINERS

Chicl< Growth Periods

Hmo 1. n-14Junc 2. 24-:~O June

9-1GJuly :1. 4. 2:1-29 July

6-14 August 5.

1981 G. l1-ta Janumy 7. 20-28 Februmy 8. 4-14 March ~). lH-28 Mal'ch

10. 5-15 April 11. 21-28 April 12. 6-13 May 1:1. 18-28 May 14. G-12.June 15. 23-30Juue

1982 16. 1:1-17 March 17. 25 March-l April la, ::W October-4 November

198:1 19. 7-14 February 20. 1'5-28 Febnuuy 21. 1-14 March 22. 15-28 March 23, 29 March -11 April 24. 12-25 April 25, 2G April-~) May 26. 10-23 M'~' 27. 24 May-G June 28. 7-20June 29. }J-15 November

1984 :{O. 2GJanuary 31. 24 Fcbrw:lI)'-2 March 32. 21-26 March ;~3. 25-30 April :14. 21-2H May :~5, 2GJune-2July :lG. 21-27 July 37. 13-19 AUb'llst 38. 26 Octobcr-l November :1~}. 6-17 November

1985 40. 13-18Janw:uy 41. 7-14 Fcbmary

481

CCl Annie Griffi rhslDRK PHOTO

Figure 1. AJ7-ican penguins onshore in Namibia.

482

loading technique (Wilson 1984). Samples were analyzed in the field or frozen soon after collection and analyzed in the laboratory.

Each sample was weighed and the prey species identified where pos­sible. Intact fish were measured directly, whereas otoliths were removed from fish too digested to allow direct measurement oflength and were used to estimate body length and mass from regression equations devel­oped by the Sea Fisheries Research Institute (unpublished data) and Batchelor &> Ross (1984).

Anchovy was the main fishery target (Butterworth 1983) and the main prey of penguins (Wilson 1985), so it was used to examine interac­tions between the penguins and the fishery. Percentage of numerical abundance, mass of reconstituted anchovy, and total meal mass in stom­achs were used in analyses.

Fishery Landings Catch statistics supplied by Sea Fisheries Research Institute were used to examine the quantities and species compositions of all commercial catches made in waters off Saldanha Bay, for each month during which a growth period occurred. The percentage composition of catches and catch per unit effort (CPUE) of anchovy and total catches for each growth period were plotted. Species composition and caudal-length fre­quency of fish taken by the penguins and the fishery were examined.

The Sea Fisheries Research Institute data for 1983 and 1984 allowed examination of the distribution offishery effort (sets of purse seines) and catches with increasing distance offshore of Saldanha Bay.

Commercial catches of anchovy and of all species were compared with penguin chick growth-increments for each growth period. Spear­man rank correlations were used to test for relationships between fishery landings and growth increments. Diet and weight gain during growth periods were compared, but data from 1980 and 1981 were excluded because sampling was done throughout the month, in small batches, rather than concentrated over short periods as in subsequent years. Per­centage data were arc-sine transformed for analysis.

Results Penguin Diets and Fishery Landings During 27 months, 667 penguins were sampled with sufficient stomach contents for analysis. These contained 10 354 prey objects. Diets varied greatly, but the predominant prey was anchovy (Figure 2) (83.3%, SD = 6.7% , n = 27). Anchovy tended to be least abundant, relative to other prey, during the fitst four months ofthe year. There did not appear to be any annual trend in the relative abundance of anchovy in penguin diets (year versus anchovy percentage: r = -0.07, n = 27) .

Fishery catches consisted of anchovy (59.5%, SD = 32.9%, n = 27) (Figure 1) to a lesser extent than did penguin diets (t = 2.10, P < 0.05). The proportion of anchovy in fishelY landings was also much more vari­able (SD = 32.9% versus 6.7%) than in penguin diets. Similarly, fishery landings of anchovy were also lower in the firstfour months ofthe year, and there was no animal trend in the importance of anchovy in fishery landings (r = 0.03, n = 27).

Purse seiners did most oftheir fishing and made most of their catches 15 to 25 km from the mouth of Saldanha Bay (Figure 3). In 1983, only 15% of sets (fishing activity) and 13% of catches were made within 15 km of the coast; in 1984, 15% of sets, but only 3% of catches, occurred

DUFFYET AL.

100

a 100-r-r-r-r-r-r-r-r-r-r-r-r~~~.-'-'-'-"r-~r-r-r-

~

" i5 50 c ·s ..,

c ., 0..

o

within 15 km of the coast. Sets and catches were strongly correlated (1983: r = 0.97, n = 7, P < 0.05; 1984: r = 0.94, n = 6, P < 0.05) .

There was no correlation between the relative monthly abundance of anchovy in fishery landings and in penguin diets (r = -0.09, n = 27 months when growth was measured), or between monthly anchovy CPUE in the fishery and anchovy mass in penguin stomachs (r = 0.04, n = 26) (Figure 4). Monthly fishery CPUE for all target species and total mass in penguin stomachs were also not correlated (r = 0.18,'n = 25).

The proportions of anchovy in each size-class taken by the fishery and by penguins were significantly correlated within 17 of24 months when anchovy were taken by both predators (Figure 5), suggesting that the two were generally exploiting the same anchovy population. However, the months when the fishery and the penguins took different sizes of an­chovy were not randomly distributed within the year; six of 12 correla­tions were significant in the first four months of the year, while 11 of12 were significant in the last eight months (Fisher exact probability test; p = 0.07). African penguins exploit small anchovy before the fishery does, during the first half of the year (DuffY et al. 1985); this produces the size-discrepancy in anchovy taken during this period.

Growth of Young Penguins Variation in adjusted initial mass and adjusted growth increment is pre­sented in Figure 6. Proportions of anchovy in penguin diets and average growth increments were not significantly correlated (p > 0.05) for small (r = -0.14, n = 18), medium (r = 0.13, n = 21), or large (r = -0.29, n = 18) young. In adult stomachs, anchovy mass (Figure 4) was not sig­nificantly related (p > 0.05) to chick growth for young (r = 0.022,

PENGUINS VERSUS PURSE SEINERS

4 5 6 7 1984

Figure 2. Relative propor-tions of anchovy and other' pr'cy in the diet of African penguins and in fIShery landings, by month, 1980 to 1985.

483

~ :ll

..c u

100-;-;;s:;-----;:::===;_ 1983 • net-sets Dcatches

1;j 50--­~

~ ;; Z

o 0-5 16- 25 36-45 · 56-65

Distance from Shore (km)

1 00 - :;0;;-;,-----;::======;-1984

• net-sets Dcatches

o 0-5 16-25 36-45 56-65

Distance from Shore (km)

Figure 3 (above). The distribution a/net-sets and catches by distance oJJshor'e from the mouth of Saldan­ha Bay, for 1983 and 1984. 4 (right). Catch per unit effor-t and total mass offlShery landings, and meal mass and ,'econstructed ancho­J1l mass in penguin stomachs, by chick growth periods, 1980 to 1985.

484

60-IIF~ffiY~~~-----======----­LAND I

40

20

o 6

4

2

FISHERY LANDINGS

IN PENGUIN STOMACHS

0-- -,------

nother • anchovy

n other I anchovy -.

B :;: v

800 ----- ------------------IN PENGUIN

Cii i 600

E o = 400 >-

i ~ 200 <t

o

STOMACHS

n = 13), mid-age (r = 0.027, n = 17), or old (r = -0.275, n = 15) chicks; total meal mass (Figure 4) was significantly related to growth for young (r = 0.622, n = 20, P < 0.01), but not (p > 0.05) mid-age (r = 0.217, n = 24) or old (r = 0.275, n = 20) chicks.

Comparing fishery landings and growth of African penguins (Table 1) showed no evidence that the fishery catch affected penguin growth. Only three of24 correlations (12.5%) were significant (p < 0.05), and all three were positive, rather than negative as would be expected ifin­creased fishery landings were depressing penguin growth. .

Discussion Several worker'S have suggested that African penguins and the fishery compete for pelagic fishes (e.g., Crawford &> Shelton 1978, 1981; Shelt6n et al. 1984), but'the spatial and temporal aspects ofthe interaction have remained a mystery. Did past overfishing ofthe pilchard and consequent replacement by anchovy lead to less productive foraging by penguins and reduced nesting success? Has current overfishing so reduced the

DUFFYET AL.

July 1980 0 .95, p<O.Ol

January 1981 p<0.05

February 1981

10.33

++ i c", ",,,.,t June 1981

E 0.74, p<O.05 May 1982 0.93, p< O.Ol

February 1983 0.32

'"' oJ: ;c 5 c ~

1 9 u c « 13

I March 1981 0.88, p<0.01

, March 1983 10.78, p<O .OI

I Apri l 1981 0.94, p<O.Ol

I April 1983 0.9 1, p< O.Ol ,

I

May 1981 0.88 , p<O.Ol

May 1983 0.40

June 1983 0.78, p<0.05

November 1983 I February 1984 0-:-96, p<O.Ol 0.68, p<0.05

I March 1984 1 Apri I 1984 May 1984 0 .88, p<O.Ol I't rt

t

June 1984 0.69, p<0.05

Ju ly 1984 0.95, p<O.Ol

October 1984 0.98, p<0.05

November 1984 January 1985 February 1985 0.34 0.96, p<O.Ol 0.16

i , , , , , ,

15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15 15 0 15

Percentage of Anchovy Size

overall anchovy population that too little remains for penguins? Or do penguins and purse seiners compete day by day and school by school? If they do compete, does the competition occur most strongly during, or outside, the breeding season?

This study examined the effects of commercial fishing only on the breeding diet and growth of the penguin. Results suggest that competi­tion from the commercial fishery does not affect penguins that breed at Marcus Island. Breeding success on Marcus is similar to that at a colony where commercial fishing does not occur (La Cock et al. in press). The same study showed that juvenile penguins from Marcus Island, which spend most of their time at sea, have a very low survival rate. Dufty {!p

Cooper (in press) found that most of these birds occur north of Marcus Island, in St, Helena Bay, an area heavily exploited by the commercial fishe,y (cf. Crawford 1981), The present results suggest that competition

PENGUINS VERSUS PURSE SEINERS

Figur-c 5. Con'elation of size distri­butions of anchovy taken by African penguins (red) and South Afhcan pur-sc seine,.s W'ay J.

485

300

o

§ :l ~300 ~

" E ~ u .E

'" ~ -600

• I I ~I

~ -820

I

OLD CHICKS

.J -- ~ u"it '1

LJ

~ initial mass growth increment

I I I I e 300--------------------------------------------------------------------------------\.l "0 ~ .. ~ ~ .. ~

~ o---H~~n..w~~~~~~W1~~nr~~~~~~~~ :5 ] u -200 11 ~ ~

E

M ID-AGE CH ICKS

i _4oo----------,------------------,----,--------------------,------------------~-5~50 § 400 ------------------------------~6~60~----------------------------------------------:~ YOUNG CHICKS ~ Cl

200

o

-200--~2L-~4~-h6~~8~~1~0~~1'2L-~174~r.7~~~2vO~~2~2~~2~4L-~276L-~28~~30~~32~~3~4~~3u6L-~3~8L-~40~--

1980 1981 1983 1984 1985

Figure 6. Deviations from the expect­ed initial mass and from growth in­cr'ements for old, mid-age, and young African penguin chicks by gmwth pe­riods, 1980 to 1985.

486

with the commercial fishelY occurs in the SI. Helena region, which af­fects nonbreeding penguins, rather than off Marcus Island.

Both purse seiners and penguins rely on the same species and size­classes of fishes, but the effective foraging range of breeding African penguins at Marcus Island restricts direct interactions with purse sein­ers. African penguins must return to the nest each day to feed their young, so they are primarily inshore foragers (Siegfried et al. 1975, wil­son 1985). In the 13 or more hours available for feeding outside Sal­danha Bay, a penguin traveling 2.8 km/h has a maximum range of approximately 18 km (Wilson 1985, Wilson et al. 1987). In contrast, purse seiners are primarily offshore foragers, unable to fish close in­shore over rocky or irregular bottoms for fear of damaging their nets. On the other hand, nonbreeding penguins are essentially pelagic, so they can travel into areas where intense commercial fishing activity may re­duce the food supply and penguin survival (Dufty (j,> Cooper in press, Wilson et al. 1987).

This study suggests that competition between African penguins and

OUFFY ET AL.

fishermen is complex and dependent on time of year and the birds' breeding state. Putative competition betvveen birds and man in other marine ecosystems is likely to prove equally complicated (ef. MacCall 1984). Inferring competition from estimates of food consumption or from post hoc changes in bird numbers and fish stocks may mislead, un­less the competitive mechanism is understood.

Table 1. Spearman Ranl<: COlTelations Between Penguin Chick Growth and Fishery Catch pel' Unit Effort (CPUE)

Young Chicles Mid-age Chicl{.~_

FishelY J' n p

MASS INCREMENT (g\ SAME WEEK

Total catch Anchovy catch Total CPUE

0.16 -0.09

0.11 Anchovy CPUE 0.28 INITIAL CHICK MASS Cg), PREVIOUS WEEK

Total catch O.OG Anchovy catch Total CPUE Anchovy CPUE

Acknowledgments

0.13 -0.09 -0.04

26 22 26 20

25 17 20 13

I' n

0.07 30 0.04 20 0.13 30 0.19 21

0.16 29 0.02 21 0.34 25 0.70 15

We thank A Berruti and A. Maceall for discussions, and the Marine Development Branch for access to Marcus Island; the Sea Fisheries Research Institute for fishery data; and the Percy FitzPatrick Institute staff who assisted with fieldwork and data analysis. This project was funded by the National Geographic Society, the South African Nature Foundation, the Endangered Wildlife Trust, and the Benguela Ecology Programme (sponsored by the South African National Committee for Oceanographic Research).

Bibliography Abrams, R. W.

1983. Pelagic seabirds and trawl-fisheries in the southern Benh'llela Current re­gion. Marine EcoloS)' Pmp;ress Series 11:151-156.

Ainley, D. G.; &i Lewis, P. G. 1974. The histOlY of the Farallon Island marine bird populations, 1854~ 1972. Condor 76:432-446.

Batchelor, A.; &; Ross, GJ.B. 1984. The diet and implications of dietmy change of Cape gannets on Bird Island, Algoa Bay. Ostrich 55:45~63.

Burger, A. E.; &; Cooper,]. 1984. The effects of fisheries on seabirds in South Africa and Namibia. Nettle­ship, D. N., Sanger, G. A., &; Springer, P. E, editors: Mal'ine Birds: Their Feeding; Ecologv and Commercial Fisheries Relationships, special publication. Canadian Wildlife Service, Ottawa.

But1elworth, D. S. 1983. Assessment and management of pelagic stocks in the southern Benh'llela re­gion. Proceedings of the expert consultation to examine changes in abundance and species composition of neritic fish resources, San Jose, Costa Rica, 18~29 April 1983. Fish and Agriculture Organization Fisheries Repor·t 291:329~406.

Cooper,]. 1984. Changes in resource division among foul' breeding seabirds in the Benguela upwelling system, 1953~78. PmceedingsJ Fifth Pan African Ornithological Con­gress. Pan African Ornithological Congress,]ohannesbuI'g, 217~230.

Crawford, RJ. M. 1980a. Seasonal patterns in South Afl'ica's western Cape purse-seine fishery. Journal of Fisherp Biolo,(5») 16:649-664. 1980h. Catch pel' standard-boat-day and deplo,Ylnent of effort in the South African purse-seine fishery, 1964-1976. Investigational Report of the Division of Sea Fisheries of South Afdca 122:1 ~24.

Crawford, RJ.M.; &; Shelton, P. A. 1978. Pelagic fish and seabird interrelationships off the coasts of South and South West Africa. Biological Conservation 14:85~ 109. 1981. Population trends for some southern African seabirds related to fish availa-

PENGUINS VERSUS PURSE SEINERS

p

0.05 0.05

Old Chicloo -~---.~-

I' n p

-0.13 25 -0.25 21

0.01 25 0.05 20

0.42 24 0.05 -0.08 18

0.24 22 0.42 1:l

457

488

bility. Cooper,J., editor: Proceedings ofSymposiwn onBirds of the Sea and Shore. South African Seabird Group, Cape Town, 15-41.

Duffy, D. C. 1983. Environmental uncertainty and commercial fishing: cHects on PCl'l1vian guano birds. Biological Conservation 26:227~238.

Duffy, D. c.; {};; Cooperd. In press. Distribution and survival of juvenile jackass penguins at sea: an explana­tion of population trends. South AfricanJournal of Science 83.

Duffy, D. C.; Wilson, R. P.; {};; Berl'l1ti, A. 1985. Anchovy in the diet of Dyer Island penguins: toward a test of two models of anchovy distribution. South African}oumal of Science 81:552-554.

Fishcr,J. 1952. The Fulmar. Collins, London.

Frost, P.G.H.; Siegfried, W. R.; {};; Cooper,]. 1976. The conservation of the jackass penguin (Spheniscus derner'sus [Linnae­us]). Biological Conservation 9:79-99.

Ful'l1ess, R. W. 1978. Encl'b'Y requirements of seabird communities: a bioenergetics model. joumal of Animal EcologJ' 47:39-53.

Furness, R. W.; {};; Cooper,]. 1982. Interactions between seabirds and pelagic fish populations in the southern Benguela region. Marine Ecologv Progress Series 8:243-250.

,Jordan, R.; {};; Fuentes, H. 1966. Las poblaciones de aves guaneras y su situacion actual. Bolet{n del Insti­tuto del Mar' del Pent, Callao 10:1-31.

La Cock, G. D.; DuffY, D. C.; {};; Cooper,]. In press. Population dynamics of the African penguin <5pheniscus demer'sus at Marcus Island in the Benguela upwelling ecosystem: 1979-1985. Biolo.~rical Con­sel1Jation 30.

May, R. M.; Eeddinb>lon,]. R; Clark, C. W.; Holt, S.].; {};; Laws, R. M. 1979. Management of multi species fisheries. Bcience 205:267-277.

MacCall, A. 1984. Seabird-fishery trophic interactions in eastern Pacific boundary currents: California and Peru. Nettleship, D. N., Sanger, G. A., {};; Springer, P. F., editors: Marine Birds: their Feeding Ecolo,'{Y and Commer'cial Fisheries Relationships, spe­cial publication. Canadian Wildlife Service, Ottawa.

Nelson, J. B. 1978. The Sulidae: Gannet and Boobies. Oxford University Press, Oxford.

Ricklefs, R. E.; Dun", D. C.; {};; Coulter, M. 1984. Weight'gain of blue-footed booby chicks: an indicator of marine resources. Or-nis Scandinavica 15:162-166.

Schacfer, M. B. 1970. Men, birds and anchovies in the Peru current - dynamic interactions. Transactions of the Amer"ican Fisheries Society 9:461-4G7.

Schweigger, E. 1964. El Litoral Pm'uano. Universidad Nacional "Federico Villaneal," Lima.

Shelton, P. A.; Crawford, RJ.M.; Cooper,].; {};; Brooke, R. K. 1984. Distribution, population size and conservation of the jackass penguin (Spheniscus demer'sus). South Africanjournal of Marine Science 2:217 -257.

Siegfried, W. R; Frost, P.G.H.; Kinahan,J. B.; {};; Cooper,]. 1975. Social behaviour_of jackass penguins at sea. ZoologicaAfrica 10:87-100.

Stander, G .. H.; {};; Le Roux, P.]. 1968. Notes on fluctuations of the commercial catch of the South African pilchard (Sardinops ocellata) 1950-19G5. Investigational Report. of the Division of Sea FislwT'ies Union of South Africa G5:1-14.

Valdivia,J. E. 1978. The anchoveta and El Nino. Rappor'ts et Proces-verbau:r des Reunions du C071seil International pour l'E:rploration de la Mer 173:196-202.

Wilson, R. P. 1984. An improved stomach pump for penguins and other seabirds. Journal of Field Ornithologv 55:109-112. 1985. Thejackass pengllin{Spi1eniscus demersus) as a pelab>ic predator. Marine Ecologv Progr'ess Series 25:219-227.

Wilson, R. P.; Wilson, M. -P.; {};; DuffY, D. c. 1987. African Penguins at Sea, ms. document no. 5456. Centro de DOCllmenta­cion, Pl'ograma de Vida Silvestre, Universidad Nacional, Heredia, Costa Rica.

Acc(pted 25 February 1.987.

DUFFY ETAL.