Abstract

Chiaradia, A., Forero, M. G., Hobson, K. A., and Cullen, J. M. 2010. Changes in diet and trophic position of a top predator 10 years after a mass mortality of a key prey. – ICES Journal of Marine Science, 67: 1710–1720.

After the disappearance of primary prey, seabirds exhibit gradually decreased breeding performance, and eventually the population size drops. Results are presented of an investigation into the diet of little penguins (Eudyptula minor) at Phillip Island, Australia, during a period when their key prey, pilchard (Sardinops sagax), declined dramatically. Data from stomach flushing (1982–2006) were used, supported by stable isotope (δ15N, δ13C) analyses of blood samples (2003, 2004, and 2006). The effect of the pilchard mortality on penguin diet was immediate, the birds shifting to a diet almost devoid of pilchard, and this was followed by 2 years of low breeding success, with considerably fewer penguins coming ashore. During periods when pilchard was not part of the diet, penguins consumed prey of a higher trophic level, e.g. higher values of δ15N. Variability in penguin blood δ15N coincided with years of low prey diversity. The disappearance of pilchard resulted in a decrease in prey diversity and led penguins to “fish up” the foodweb, possibly because of the simplified trophic structure. After 1998, however, breeding success re-attained average levels and the numbers of penguins coming ashore increased, probably because of increased abundance of prey other than pilchard after a 3-year period of food scarcity. Although little penguins apparently compensated over time, a less-flexible diet could make them ultimately vulnerable to further changes in their foodweb.

Introduction

The natural or human-induced removal of key species in marine ecosystems has profound effects on foodwebs, leading to trophic cascades or ecological regime shifts (Österblom et al., 2007; Baum and Worm, 2009). Seabird data have been used to detect such changes in marine ecosystems. Seabirds are highly sensitive to changes in food abundance and diversity, whether related to climate, exploitation, or both (Crawford and Dyer, 1995; Reid and Croxall, 2001; Crawford, 2007; Frederiksen et al., 2007; Ainley and Blight, 2009). Several studies have shown that after the disappearance of a primary prey, whether through climate-related change or commercial fisheries, seabirds exhibit decreased breeding performance and eventually lower population size (Frank et al., 2005; Myers et al., 2007; Baum and Worm, 2009). However, resulting changes in seabird populations are often gradual, being detected over several years, and are often non-linear (Furness and Tasker, 2000; Montevecchi, 2007; Piatt et al., 2007). Rarely have examples of carry-over effects of an acute alteration of the availability of a key prey of seabirds been documented (Österblom et al., 2006, 2007; Crawford, 2007).

In 1995, there was a mass mortality of pilchard Sardinops sagax throughout southern Australia and New Zealand (Griffin et al., 1997), the largest mortality of a single marine fish species ever recorded (Jones et al., 1997; Whittington et al., 1997). Deaths were reported over 5000 km along the southern Australian coast, which represents the total range of the species in Australia (Griffin et al., 1997; Jones et al., 1997). The mortality, apparently caused by a herpesvirus spreading at 30 km d−1, decimated ∼70% of the population, seriously depleting the southern Australian stocks of pilchard in a very short time (Jones et al., 1997; Gaughan et al., 2000; Gaughan, 2002; Murray and Gaughan, 2003). Further pilchard mortality was recorded in 1998, but on a much smaller scale (Gaughan et al., 2000; Ward et al., 2001b).

The pilchard was a key component of southern Australia's marine foodweb, present in the diet of several top predators (Kailola et al., 1993), and one of the most abundant commercial species in southern Australia (Gaughan et al., 2000; Ward et al., 2001b). However, despite the potential impact on the ecosystem of the disappearance of pilchard from southern Australia, few studies monitored the ecological effects of the mortality in the marine system (Bernoth, 2007; Whittington et al., 2008).

Little penguins Eudyptula minor, as top predators, may be good monitors of change in their marine system. They are central-place foragers with one of the smallest foraging ranges among seabirds (<20 km; Collins et al., 1999; Hoskins et al., 2008), and their breeding success is strongly related to food availability (Chiaradia and Nisbet, 2006). At Phillip Island, Victoria, Australia, during the 1980s, the pilchard was a key prey in the diet of little penguins (Montague and Cullen, 1988; Cullen et al., 1992). The breeding onset of penguins seemed to be related to an increased presence of pilchard in their diet, and its absence delayed breeding onset and coincided with poor breeding success (Cullen et al., 1992). Anchovy Engraulis australis were equally important in penguin diet (Cullen et al., 1992) and their patterns of abundance in the area were similar to those of pilchard (Hobday, 1992); however, anchovy mortality was not recorded in 1995 (Jones et al., 1997). Therefore, the pilchard mortality provided a natural experiment in which the trophic response of a top predator such as the little penguin to changes in its marine ecosystem could be examined.

Previous studies of the diet of little penguins used flushed-stomach analyses, which yield prey-species composition, but such data can be complemented with stable isotope analyses to inform on trophic interactions between predator and prey (Barrett et al., 2007; Inger and Bearhop, 2008). Measurements of the stable isotope ratios of carbon (δ13C) and nitrogen (δ15N) can provide trophic-level and space-use information of consumers in marine foodwebs (Hobson and Welch, 1992; Forero et al., 2002; Hobson et al., 2002). The use of δ13C analyses can additionally provide information on feeding on inshore or benthically linked prey vs. offshore or pelagic prey (Hobson and Welch, 1992; France, 1995). In contrast to conventional dietary approaches, the stable-isotope method can trace assimilated food and integrate dietary information in the blood over longer periods (about a month for seabirds of medium size; Hobson and Wassenaar, 2008) than stomach analyses (often less than a day), which represent a single or at most few feeding events.

Predicting the diet changes of a predator in response to major reductions in its primary prey is difficult because it will depend on the relative abundance of alternative prey and the foraging limitations of the predator, in little penguin a range of 20 km (Collins et al., 1999). However, penguins could have shifted to a more diverse diet, as removal of the then highly abundant pilchard facilitated the expansion of other less abundant species in other areas (Ward et al. 2001a). Within that context, we examined not only shifts in diet, but also possible changes in the foodweb of little penguins, when the availability of a key prey varied dramatically.

Here, long-term data from flushed stomachs (1982–2006), supported by δ13C and δ15N analyses (2003, 2004, 2006) of penguin blood and of their prey, were used to investigate possible changes in diet and trophic position in little penguins at Phillip Island, Australia, following the mortality of pilchard in 1995.

Methods

First, we examined whether the initial change in penguin diet after the pilchard mortality in 1995 (Chiaradia et al., 2003) had persisted 10 years after the event, which we then related to changes in breeding success and population parameters. Second, the trophic positions of penguins and their prey were determined over 3 years of below- and above-average breeding success, a parameter considered a proxy for food availability (Chiaradia and Nisbet, 2006). Third, the isotopic foodweb of penguins was constructed before and after the pilchard mortality, using prey isotope values and prey proportion from the long-term conventional stomach-flushed dataset to determine whether penguins moved up or down in trophic position. Finally, penguin diet diversity was investigated, because shifting the balance between a predator and its main prey can increase or decrease the diversity of the predator's diet (Croxall et al., 1999).

Two sites at Summerland Peninsula, Phillip Island, Victoria, Australia (38°15′S 145°30′E) were used in this study—Parade and Radio-Tracking Bay, which are about 1 km apart. Both sites have about 100 nests marked individually, with an annual nest occupancy of ∼60%. All birds were weighed and sexed, then injected with transponders (Trovan and Allflex, Australia) for individual identification. Bill depth was used to determine the sex of adults (Gales, 1988; Arnould et al., 2004) and DNA analysis to sex chicks in 2003 and 2004, using polymerase chain reaction amplification of the CHD gene (Fridolfsson and Ellegren, 1999). This technique for determining the sex of little penguins has been validated (Overeem et al., 2006). A custom-made hand-held transponder reader recorded the presence of birds with transponders in the colony (Chiaradia and Kerry, 1999). Parental status, such as pre-laying, incubation, chick guard, and chick post-guard, was determined during regular visits to both colonies (details on the breeding stages are provided by Chiaradia and Kerry, 1999). The Parade site was visited three times per week (Chiaradia and Nisbet, 2006) and the Radio-Tracking Bay every 2 weeks (Chiaradia et al., 2003).

Breeding success was determined as the number of chicks fledged per pair (cpp). Data were standardized as relative values in relation to the long-term mean. For the years when diet data were available, years were grouped in below-average (≤0.7 cpp, years 1985, 1987, 1995, and 2004) and above-average (>0.7 cpp, 1986, 1996, 1998, 2000, 2001, 2002, 2003, 2005, and 2006) breeding success, for a comparison with diet diversity.

Daily beach counts of penguins were used as a proxy for population change. Little penguins always return to shore after sunset, most crossing the beach during the first 2 h after sunset (Stahel and Gales, 1987; Daniel et al., 2007), providing a good indication of daily attendance. Birds were counted at the Parade colony every day for 50 min after the first group of penguins crossed the beach. Data were standardized as relative mean values in relation to the long-term mean.

Stomach-flush samples were collected from 1982 to 2006 at Phillip Island. Data for 1982 and 1983 were from Montague and Cullen (1988, their Table 5). Although shown in Figure 1, they were not used in statistical analyses, because information on individual samples was not available. Data collected between 1985 and 1987 are from Cullen et al. (1992) and represent the period before the pilchard mortality. Although those samples were collected during the breeding period, the breeding status of individual birds was unknown. Therefore, all analyses performed with these data were grouped per year and not examined at the level of breeding stage. No data were available by stage between 1988 and 1994. Data from 1995 to 2006 (but no samples in 1997 and 1999) were collected at pre-laying, incubation, guard, and post-guard stages. Adults were trapped as soon as they arrived on the beach and food samples collected from marked birds with known recent breeding histories. Stomach samples were collected by stomach-flushing, according to the method of Gales (1987), later modified by Chiaradia et al. (2003). Samples were frozen at −20°C for later analysis (see details in Chiaradia et al., 2003).

Figure 1.

Mean relative prey abundance in the diet of little penguins from 1982 to 2006. Note that diet data are discontinuous and not available for the years 1984, 1988–1994, 1997, and 1999. Data for 1982 and 1983 are grouped, because the values were reconstructed from Montague and Cullen (1988), and data were not separated by years. Relative prey abundance does not add to 100% because zero values from empty stomachs are included (see the text). Therefore, decreased values of relative prey abundance indicate overall low prey abundance in the stomachs, most noticeably in 1995, 1996, and 1998.

Figure 1.

Mean relative prey abundance in the diet of little penguins from 1982 to 2006. Note that diet data are discontinuous and not available for the years 1984, 1988–1994, 1997, and 1999. Data for 1982 and 1983 are grouped, because the values were reconstructed from Montague and Cullen (1988), and data were not separated by years. Relative prey abundance does not add to 100% because zero values from empty stomachs are included (see the text). Therefore, decreased values of relative prey abundance indicate overall low prey abundance in the stomachs, most noticeably in 1995, 1996, and 1998.

Blood and prey sampling

Blood samples were collected from the Parade and Radio-Tracking Bay sites in 2003 and 2004 and from Radio-Tracking Bay only in 2006. At each colony, adults were sampled during the pre-laying, incubation, guard (chicks up to 3 weeks old), and post-guard stages [chicks between 6 and 7 weeks old; details on breeding stage in Chiaradia and Kerry (1999)]. Additionally, young (guard) and old chicks (post-guard) were sampled at each site. About 0.08 ml of blood was extracted from the brachial or tarsal vein (see Hobson et al., 1997). Muscle samples of prey species (Table 1) were from catches of size similar to the penguins' prey (standard length ∼14 cm; Cullen et al., 1992), obtained from fishing boats operating inside Port Phillip Bay and in areas of the Bass Strait within 100 km of the study site. Samples of prey and penguin blood were stored in 70% ethanol at room temperature (20–25°C) until isotope analysis.

Table 1.

Diet composition of little penguins during the breeding season from 1982 to 2006 at Phillip Island, Australia.

 Species 1982–1983a (321) 1985b (24) 1986b (164) 1987b (160) 1995c (6) 1996c (69) 1998 (61) 2000 (68) 2001 (95) 2002 (108) 2003 (80) 2004 (50) 2005 (60) 2006 (26) 
Fish Anchovy (Engraulis australis) 36 (2.1) 23 (7.3) 14 (1.9) 4 (1.0) 24 (12.6) 6 (2.1) <1 57 (5.0) 4 (1.1) 8 (1.8) 9 (2.4) 12 (3.2) 18 (3.8) 20 (7.3) 
Pilchard (Sardinops sagax) 18 (1.6) 10 (5.1) 22 (2.6) 1 (0.4)  <1 8 (3.5) <1 1 (1.1) 20 (3.5) <1 2 (1.2) <1  
Red Cod (Pseudophysis bachus)  <1 2 (0.5) <1 7 (9.8) 19 (3.7) 4 (2.0) 7 (2.5) 2 (0.7) 17 (2.7) 6 (2.1) 9 (3.0) 2 (1.2) 2 (1.9) 
Barracouta (Thyrsites atun) 1 (0.3) 3 (2.0) 4 (1.1) 3 (0.9) 11 (7.5) 7 (2.3) 2 (1.5) 11 (3.0) 41 (4.3) 21 (2.8) 30 (4.0) 26 (4.2) 29 (4.8) <1 
Blue Warehou (Seriolella brama)  4 (3.1)  1 (0.6)  7 (2.6) 15 (4.4) <1 8 (2.3) 12 (1.7) 3 (1.2) 9 (2.6) <1  
Leatherjackets (Monocanthidae) 2 (0.5) 3 (1.7) 3 (0.7) 4 (0.8)  <1  4 (1.1) 2 (0.9) 1 (0.3) <1 2 (0.8) 1 (0.6) 8 (4.0) 
Seahorses (Hippocampus sp.)  <1 2 (0.5) 1 (0.6) 3 (4.1) 4 (1.9) 1 (1.2) <1 <1 <1 <1 <1 <1 <1 
Sandy Sprat (Hyperlochus vittatus <1 <1 <1  <1  1 (0.8) <1 <1 <1 <1   
Hardyheads (Atherinason sp.)  <1  1 (0.8)  <1  <1 1 (1.1) <1 2 (1.4)    
Red Bait (Emmelichthys nitidus        <1 3 (1.4) 2 (1.1)    
Jack Mackerel (Trachurus declivis       <1 <1 <1 5 (1.4) <1  <1 
Garfish (Hemiramphus far 1 (1.2) <1 2 (0.6)   2 (1.6)   <1 <1  <1  
Gurnards (Triglidae)  1 (1.0) 2 (0.7) 3 (0.8)     <1      
Red Mullet (Upeneichthys porosus    5 (4.4) <1  <1 <1 <1     
Pink Ling (Genypterus blacodes     <1    <1  3 (2.1)   
Silver Warehou (Seriolella punctata       <1  <1  <1   
Trevallies (Carangidae)  <1 <1 <1    1 (1.5)   <1    
Fish postlarvae unknown  5 (3.4) 2 (0.9) 5 (1.2)  5 (2.2) 15 (4.1) 7 (2.7) <1 1 (0.8) 8 (2.5) 1 (1.1) 2 (1.4) 6 (4.2) 
Cephalopods Arrow squid (Nototodarus gouldi) 2 (0.3) 23 (6.0) 20 (2.2) 28 (2.5)  2 (1.1) <1 2 (0.6) 16 (3.0) 7 (1.4) 7 (2.0) 13 (3.4) 23 (3.9) 25 (6.7) 
Loliolus noctiluca  2 (1.6) <1 4 (1.2)    <1   <1    
Sepioteutis australis  <1 <1 1 (0.7)   <1  <1 <1 <1 <1   
Argonauta nodosa  2 (1.0) 5 (1.0) 9 (1.6)   <1  2 (1.3) 1 (0.6) <1 <1 <1 1 (1.0) 
Octopodidae       <1 <1  <1 <1    
Post Larvae Cephalopod        <1  <1  <1 <1  
Unknown Cephalopod   <1 2 (0.6)    <1 <1 <1 6 (2.5)  5 (2.3) 5 (3.9) 
Crustaceans Krill (Nyctiphanes australis <1 10 (2.1) 12 (1.9)   10 (3.4) <1 13 (3.2) <1 14 (3.8) 4 (2.5) 2 (1.0) <1 
Stomatopoda  <1  2 (0.7)   <1 <1 1 (1.1) 2 (0.9) <1 2 (1.8) 1 (1.0) <1 
Amphipoda   <1 1 (0.5)   <1 <1 <1 <1 <1 <1 <1 <1 
Brachyura       <1 <1  <1   <1  
Megalopa  3 (3.0) <1 1 (0.5)           
Unknown Crustacea  <1  <1    <1    <1  <1 
Other 18 (1.7)              
 Breeding success (ccp) – 0.7 0.9 0.5 0.3 1.2 1.0 1.0 0.8 1.3 1.1 0.7 1.3 0.8 
 Diversity index – 1.1 0.8 0.8 0.6 0.5 0.2 0.4 0.4 0.7 0.5 0.6 0.6 0.4 
 Species 1982–1983a (321) 1985b (24) 1986b (164) 1987b (160) 1995c (6) 1996c (69) 1998 (61) 2000 (68) 2001 (95) 2002 (108) 2003 (80) 2004 (50) 2005 (60) 2006 (26) 
Fish Anchovy (Engraulis australis) 36 (2.1) 23 (7.3) 14 (1.9) 4 (1.0) 24 (12.6) 6 (2.1) <1 57 (5.0) 4 (1.1) 8 (1.8) 9 (2.4) 12 (3.2) 18 (3.8) 20 (7.3) 
Pilchard (Sardinops sagax) 18 (1.6) 10 (5.1) 22 (2.6) 1 (0.4)  <1 8 (3.5) <1 1 (1.1) 20 (3.5) <1 2 (1.2) <1  
Red Cod (Pseudophysis bachus)  <1 2 (0.5) <1 7 (9.8) 19 (3.7) 4 (2.0) 7 (2.5) 2 (0.7) 17 (2.7) 6 (2.1) 9 (3.0) 2 (1.2) 2 (1.9) 
Barracouta (Thyrsites atun) 1 (0.3) 3 (2.0) 4 (1.1) 3 (0.9) 11 (7.5) 7 (2.3) 2 (1.5) 11 (3.0) 41 (4.3) 21 (2.8) 30 (4.0) 26 (4.2) 29 (4.8) <1 
Blue Warehou (Seriolella brama)  4 (3.1)  1 (0.6)  7 (2.6) 15 (4.4) <1 8 (2.3) 12 (1.7) 3 (1.2) 9 (2.6) <1  
Leatherjackets (Monocanthidae) 2 (0.5) 3 (1.7) 3 (0.7) 4 (0.8)  <1  4 (1.1) 2 (0.9) 1 (0.3) <1 2 (0.8) 1 (0.6) 8 (4.0) 
Seahorses (Hippocampus sp.)  <1 2 (0.5) 1 (0.6) 3 (4.1) 4 (1.9) 1 (1.2) <1 <1 <1 <1 <1 <1 <1 
Sandy Sprat (Hyperlochus vittatus <1 <1 <1  <1  1 (0.8) <1 <1 <1 <1   
Hardyheads (Atherinason sp.)  <1  1 (0.8)  <1  <1 1 (1.1) <1 2 (1.4)    
Red Bait (Emmelichthys nitidus        <1 3 (1.4) 2 (1.1)    
Jack Mackerel (Trachurus declivis       <1 <1 <1 5 (1.4) <1  <1 
Garfish (Hemiramphus far 1 (1.2) <1 2 (0.6)   2 (1.6)   <1 <1  <1  
Gurnards (Triglidae)  1 (1.0) 2 (0.7) 3 (0.8)     <1      
Red Mullet (Upeneichthys porosus    5 (4.4) <1  <1 <1 <1     
Pink Ling (Genypterus blacodes     <1    <1  3 (2.1)   
Silver Warehou (Seriolella punctata       <1  <1  <1   
Trevallies (Carangidae)  <1 <1 <1    1 (1.5)   <1    
Fish postlarvae unknown  5 (3.4) 2 (0.9) 5 (1.2)  5 (2.2) 15 (4.1) 7 (2.7) <1 1 (0.8) 8 (2.5) 1 (1.1) 2 (1.4) 6 (4.2) 
Cephalopods Arrow squid (Nototodarus gouldi) 2 (0.3) 23 (6.0) 20 (2.2) 28 (2.5)  2 (1.1) <1 2 (0.6) 16 (3.0) 7 (1.4) 7 (2.0) 13 (3.4) 23 (3.9) 25 (6.7) 
Loliolus noctiluca  2 (1.6) <1 4 (1.2)    <1   <1    
Sepioteutis australis  <1 <1 1 (0.7)   <1  <1 <1 <1 <1   
Argonauta nodosa  2 (1.0) 5 (1.0) 9 (1.6)   <1  2 (1.3) 1 (0.6) <1 <1 <1 1 (1.0) 
Octopodidae       <1 <1  <1 <1    
Post Larvae Cephalopod        <1  <1  <1 <1  
Unknown Cephalopod   <1 2 (0.6)    <1 <1 <1 6 (2.5)  5 (2.3) 5 (3.9) 
Crustaceans Krill (Nyctiphanes australis <1 10 (2.1) 12 (1.9)   10 (3.4) <1 13 (3.2) <1 14 (3.8) 4 (2.5) 2 (1.0) <1 
Stomatopoda  <1  2 (0.7)   <1 <1 1 (1.1) 2 (0.9) <1 2 (1.8) 1 (1.0) <1 
Amphipoda   <1 1 (0.5)   <1 <1 <1 <1 <1 <1 <1 <1 
Brachyura       <1 <1  <1   <1  
Megalopa  3 (3.0) <1 1 (0.5)           
Unknown Crustacea  <1  <1    <1    <1  <1 
Other 18 (1.7)              
 Breeding success (ccp) – 0.7 0.9 0.5 0.3 1.2 1.0 1.0 0.8 1.3 1.1 0.7 1.3 0.8 
 Diversity index – 1.1 0.8 0.8 0.6 0.5 0.2 0.4 0.4 0.7 0.5 0.6 0.6 0.4 

Values are mean (s.e.) of the relative abundance (%) of prey species. For each year, relative prey abundance <1% is represented as <1, and the absence of prey as no entry. Data from 1998 are unpublished (P. Dann, pers. comm.). The yearly sample size is given in parenthesis at the top of each column. The main species in the diet that were used in the stable isotope analysis are in bold. Breeding success is expressed as the number of chicks fledged per pair. Diversity index is a non-unit value based on the Shannon–Weaver index (Shannon and Weaver, 1949).

aRelative prey abundance values were reconstructed from Montague and Cullen (1988).

bOriginal data from Cullen et al. (1992).

cOriginal data from Chiaradia et al. (2003).

Stable-isotope analyses

Before stable-isotope analysis, ethanol was removed from the samples by decanting superfluous ethanol, then freeze-drying the balance. Prey tissues were treated for lipid extraction using a 2:1 chloroform:methanol solvent, then dried at 60°C for 24 h to remove residual solvent. The extraction of lipids was considered unnecessary for blood samples, because the lipid component in blood is generally very low (Deuel, 1955). Whole blood was freeze-dried, then powdered. Stable-carbon- and nitrogen-isotope assays were performed on 1-mg subsamples of homogenized materials by loading the material into tin cups and combusting at 1200°C in a Robo-Prep elemental analyser. The resultant carbon dioxide and nitrogen were then analysed using an interfaced Europa 20:20 continuous-flow isotope ratio mass spectrometer, with every five unknowns separated by two (baleen and egg albumen) laboratory standards. Based on replicate measurements on within-run standards, measurement error was estimated to be ±0.3 and ±0.1‰ for δ15N and δ13C measurements, respectively.

Analysis of relative prey

Diet composition was calculated by recording the proportional contribution of each species to the total mass of each stomach sample, hereafter referred to as relative prey abundance. The absence of a prey in the sample was recorded as a zero value, including empty stomachs, before the mean percentage value was calculated per prey item per year. As a consequence, the sum of the relative contributions of all species did not reach 100% in all years (Figure 1). Using relative prey abundance, each sample was weighted equally regardless of the mass of stomach contents. This method prevents biases towards the chick-rearing period, when stomach-content mass is several times heavier than at other stages of the penguin life cycle (Montague and Cullen, 1988; Cullen et al., 1992; Chiaradia et al., 2003).

Data analysis

Multivariate analyses were performed separately for adults and chicks to analyse intra-population variability on δ15N and δ13C values. The effect of sex on δ15N and δ13C was also tested for both age classes. The effect of breeding stage and year was tested on stable-isotope values of adults using generalized linear mixed models (GLMMs; Littell et al., 1996). The response variables were the δ15N and δ13C values of blood samples, and the three categorical variables were year (2003, 2004, and 2006), breeding stage (pre-laying, incubation, guard, and post-guard), and sex. Repeated measures from the same adult between years or breeding stages were treated as a random term in the GLMM using the SAS Macro program GLIMMIX (Littell et al., 1996). Changes in δ15N and δ13C values of male and female chicks were examined between guard and post-guard periods, and differences between chicks and their parents at guard and post-guard stages using generalized linear models (GLMs; Littell et al., 1996), applying normal-error and identity-link functions of PROC GENMOD in SAS (Littell et al., 1996).

Segregation among stable-isotope values of prey was analysed by a multivariate analysis of variance (ANOVA) test, and post hoc comparisons were tested with Wilks' lambda test. Prey species were classified into different trophic levels based on their δ15N values. Isotope values and stomach-flushed data were combined to infer trophic-level change of penguins from before to after the pilchard mortality. Krill were not included in the analysis because their abundance changed only slightly for the whole period.

Prey were grouped into three trophic-level categories in relation to their δ15N values and classified as low [pilchard and arrow squid (Nototodarus gouldi), <13‰], medium [red cod (Pseudophycis bachus) and blue warehou (Seriolella brama), >13‰ and <15‰], or high [anchovy and barracouta (Thyrsites atun), >15‰].

A two-isotope mixing-model analysis (Phillips and Gregg, 2003) was used to determine abundance of prey species in the penguin's isotope mixture, using diet-tissue isotope-discrimination values from Caut et al. (2009).

Diversity index

An annual index of diversity was calculated for each stomach sample using the Shannon–Weaver index (Shannon and Weaver, 1949), and it was compared before and after the pilchard mortality in 1995 according to the breeding stage, and between years of below- or above-average breeding success. For the within-year analysis, only data for years post-1995 were used when samples were from penguins of known breeding stage (incubation, guard, post-guard; n = 9 years). First, an ANOVA was performed with diversity index as the dependent variable, and the response was the period before and after 1995 (pilchard mortality). Second, the effect of breeding stage (incubation, guard, post-guard) on breeding success was tested with an ANOVA that included the year effect and the interaction between year and stage.

Results

Diet composition

In all, 20 species of fish, 6 of cephalopods, and 6 of crustaceans were identified in the diet of little penguins from 1292 samples (Table 1). The overall composition of fish, crustaceans, and cephalopods was similar to the results of previous studies (Cullen et al., 1992; Chiaradia et al., 2003). As stated above, relative prey abundance did not always add up to 100% because zero values from empty stomachs were included in the analysis. Therefore, low values of relative prey abundance indicated low prey abundance in the stomachs, most noticeably in 1995, 1996, and 1998 (Figure 1).

There was an effect of year on the combined proportion of main prey consumed (MANOVA; Wilks' λ = 0.005, F7,952 = 28 624, 98, p < 0.0001). The effect of year was also significant for each of the main prey items consumed (ANOVA; all p < 0.0001; Figure 1). There was an effect of pilchard mortality on the combined relative abundance of the main prey consumed (Wilks' λ = 0.805, F7,963 = 33.34, p < 0.0001). The relative abundance of barracouta, anchovy, red cod, and blue warehou increased after the pilchard mortality in 1995 (ANOVA; all p < 0.0001). In contrast, pilchard and arrow squid were less frequent in the diet after the pilchard mortality (all p < 0.0001). Penguin consumption of krill was unchanged throughout the study period (F1,971 = 1.12, p = 0.29). In 1998, pilchard presence increased slightly in the diet, but another pilchard mortality in southern Australia may have further depleted the stocks (Ward et al., 2001a). Since then, the proportion of pilchard in the diet has been <2%, except for a value of 20% in 2002.

Breeding success

Breeding success varied from 0.07 (1997) to 2 (1981) cpp. Breeding success was below average in the 1980s and recovered to above average in the 1990s until 1995, the year of the pilchard mortality (Figure 2b). Breeding success was very low for 2 years just after the mortality (1995 and 1997), but surprisingly high in 1996, though there was no significant difference in the breeding success before and after the mortality (t37 = 0.47, p = 0.64). However, there was a significantly lower variance in breeding success after the pilchard mortality (0.04 from 1998 to 2006 vs. 0.19 before the mortality, F24 = 5.14, p = 0.006).

Figure 2.

Commercial fishery landing and penguin parameters from 1980 to 2006. (a) Annual commercial catch of pilchard and anchovy near the little penguin colony (2006 not available for pilchard), showing the decline in pilchard landings after the mortality in 1995 and the relatively constant values for anchovy, except 1996 [data from Department of Primary Industries (2008)]. (b) Penguin breeding success expressed as the number of chicks fledged per breeding pair relative to the long-term mean. (c) Annual mean number of penguins crossing the beach daily as an index of penguin population change relative to the long-term mean. Note that zero on the x-axis represents the long-term mean value in (b) and (c).

Figure 2.

Commercial fishery landing and penguin parameters from 1980 to 2006. (a) Annual commercial catch of pilchard and anchovy near the little penguin colony (2006 not available for pilchard), showing the decline in pilchard landings after the mortality in 1995 and the relatively constant values for anchovy, except 1996 [data from Department of Primary Industries (2008)]. (b) Penguin breeding success expressed as the number of chicks fledged per breeding pair relative to the long-term mean. (c) Annual mean number of penguins crossing the beach daily as an index of penguin population change relative to the long-term mean. Note that zero on the x-axis represents the long-term mean value in (b) and (c).

Beach counts

The annual mean number of penguins counted crossing the beach showed a pattern similar to that of breeding success (Figure 2c). Small numbers were ashore in the 1980s, followed by an increase in the early 1990s and a drop to below average in the years following the pilchard mortality in 1995. However, there were no differences in the mean or the variance of the numbers of penguins ashore before and after the pilchard mortality (p > 0.5).

Diet diversity

Penguins had a more diverse diet before the pilchard mortality (ANOVA; F1,896 = 96.30, p < 0.001). Males and females did not show differences in diet diversity (sex * year: F1,896 = 0.49, p = 0.48). Within years, diet diversity increased as the season progressed from incubation to post-guard stages (ANOVA; breeding stage, F3,529 = 4.03, p = 0.02). There was no difference in diet diversity between years of below- and above-average breeding success (year: F1,529 = 2.35, p = 0.13; year * breeding stage: F2,529 = 0.22, p = 0.81), but the variability in diet diversity was greater in years of poor breeding success (Figure 3). In the 3 years for which stable isotope data were collected (2003, 2004, and 2006), diet diversity was higher in the year of least breeding success (2004; F1,156 = 0.3.14, p = 0.05).

Figure 3.

Diversity index of prey (mean ± s.e.) in the diet of little penguins from 1985 to 2006, grouped according to breeding stage and years of below- or above-average breeding success. The diversity index is a non-unit value based on the Shannon–Weaver index (Shannon and Weaver, 1949).

Figure 3.

Diversity index of prey (mean ± s.e.) in the diet of little penguins from 1985 to 2006, grouped according to breeding stage and years of below- or above-average breeding success. The diversity index is a non-unit value based on the Shannon–Weaver index (Shannon and Weaver, 1949).

Prey isotope values

There was a significant effect of species on the combined stable isotope values of the main prey of little penguins (MANOVA; Wilks' λ = 0.28, F10,182 = 16.1, p < 0.0001). As expected, variance was much greater in δ15N (3.6‰) than in δ13C values (0.43‰) from the prey (Table 2). Anchovy and barracouta had the highest and similar δ15N values (Tukey's DHS test, p = 1.00) and were segregated from the other few prey species (all p < 0.02). There were no significant differences in the δ13C values among the four other dominant prey species (all p > 0.126). Arrow squid and barracouta had the highest mean values of δ13C and differed from other prey species (all p < 0.03), which had more negative and similar δ13C values (all p > 0.99; Figure 4; Table 2).

Table 2.

Mean isotope values of nitrogen (δ15N) and carbon (δ13C) of the dominant prey of little penguins at Phillip Island, Australia, in 2003, 2004, and 2006.

Prey species δ15N(‰)
 
δ13C (‰)
 
na 
Mean s.d. Mean s.d. 
Anchovy 15.9 1.5 −19.9 0.5 14 
Pilchard 12.8 1.4 −20.0 0.9 16 
Barracouta 15.9 1.5 −19.4 0.5 21 
Red cod 13.8 0.4 −20.1 0.3 18 
Blue warehou 14.1 2.0 −20.0 0.3 10 
Arrow squid 12.8 0.9 −19.9 0.4 19 
Prey species δ15N(‰)
 
δ13C (‰)
 
na 
Mean s.d. Mean s.d. 
Anchovy 15.9 1.5 −19.9 0.5 14 
Pilchard 12.8 1.4 −20.0 0.9 16 
Barracouta 15.9 1.5 −19.4 0.5 21 
Red cod 13.8 0.4 −20.1 0.3 18 
Blue warehou 14.1 2.0 −20.0 0.3 10 
Arrow squid 12.8 0.9 −19.9 0.4 19 

aSample size was the same for both isotopes.

Figure 4.

Trophic position (mean ± s.e.) of little penguin prey (open circles).

Figure 4.

Trophic position (mean ± s.e.) of little penguin prey (open circles).

Penguin isotope values

Totals of 350 blood samples from adults and 175 from chicks were collected in 2003, 2004, and 2006 (Table 3). For adults, the GLMMs for δ15N and δ13C included the effects of breeding stage, year, and their interaction, after controlling for the random effect of individual identity (Table 4). Males and females did not differ in stable isotope values. The GLM for δ15N explained a higher percentage (up to 89%) of the variance than δ13C (up to 60%). During incubation and pre-laying, adults showed higher δ13C values than during chick-rearing stages. Adults feeding large chicks at post-guard had the lowest values of δ15N (Figure 5).

Table 3.

Mean (and s.d.) isotope values of nitrogen (δ15N) and carbon (δ13C) from the blood of little penguin adults and chicks in 2003, 2004, and 2006.

Parameter 2003
 
2004
 
2005
 
Adults Chicks Adults Chicks Adults Chicks 
δ15N(‰) 15.2 (0.7) 14.7 (0.3) 14.8 (0.4) 14.6 (0.3) 15.9 (1.3) 14.9 (0.6) 
δ13C(‰) −19.6 (0.4) −20.0 (0.2) −19.4 (0.6) −20.0 (0.3) −20.0 (0.4) −20.3 (0.2) 
n 234 119 68 44 48 12 
Parameter 2003
 
2004
 
2005
 
Adults Chicks Adults Chicks Adults Chicks 
δ15N(‰) 15.2 (0.7) 14.7 (0.3) 14.8 (0.4) 14.6 (0.3) 15.9 (1.3) 14.9 (0.6) 
δ13C(‰) −19.6 (0.4) −20.0 (0.2) −19.4 (0.6) −20.0 (0.3) −20.0 (0.4) −20.3 (0.2) 
n 234 119 68 44 48 12 
Table 4.

GLMMs for the effect of year, season, and age on stable isotope signatures from the blood of little penguins at Phillip Island, Australia, in 2003, 2004, and 2006.

Response variable GLM Variables Statistic p-value Explained deviance (%) 
δ15Adults Individual identity Z = 6.03 <0.0001 89 
Breeding stage F3,74 = 45.45 <0.0001 
Year F2,74 = 44.63 <0.0001 
Breeding stage * year F6,74 = 4.67 0.0004 
Chicks Age forumla= 38.36 <0.0001 34 
Year forumla = 11.86 0.0027 
Age * year forumla = 22.44 <0.0001 
Small chick-adult segregation Year forumla= 86.88 <0.0001 64 
Age forumla = 79.05 <0.0001 
Large chick-adult segregation Year * age* forumla = 10.42 0.005 10 

 
δ13Adults Individual identity Z = 0.42 0.34 60 
Breeding stage F3,74 = 35.05 <0.0001 
Year F2,74 = 45.71 <0.0001 
Breeding stage * year F6,74 = 14.39 <0.0001 
Chicks Year forumla= 18.97 <0.001 22 
Age forumla = 14.86 0.0001 
Year * age forumla = 9,88 0.007 
Small chick-adult segregation Age forumla= 68.46 <0.0001 49 
Year forumla = 30.37 <0.0001 
Age * year forumla = 17.50 0.0002 
Large chick-adult segregation Year forumla= 23.79 <0.001 19 
Age forumla = 5.00 0.02 
Year * age forumla = 7.92 0.02 
Response variable GLM Variables Statistic p-value Explained deviance (%) 
δ15Adults Individual identity Z = 6.03 <0.0001 89 
Breeding stage F3,74 = 45.45 <0.0001 
Year F2,74 = 44.63 <0.0001 
Breeding stage * year F6,74 = 4.67 0.0004 
Chicks Age forumla= 38.36 <0.0001 34 
Year forumla = 11.86 0.0027 
Age * year forumla = 22.44 <0.0001 
Small chick-adult segregation Year forumla= 86.88 <0.0001 64 
Age forumla = 79.05 <0.0001 
Large chick-adult segregation Year * age* forumla = 10.42 0.005 10 

 
δ13Adults Individual identity Z = 0.42 0.34 60 
Breeding stage F3,74 = 35.05 <0.0001 
Year F2,74 = 45.71 <0.0001 
Breeding stage * year F6,74 = 14.39 <0.0001 
Chicks Year forumla= 18.97 <0.001 22 
Age forumla = 14.86 0.0001 
Year * age forumla = 9,88 0.007 
Small chick-adult segregation Age forumla= 68.46 <0.0001 49 
Year forumla = 30.37 <0.0001 
Age * year forumla = 17.50 0.0002 
Large chick-adult segregation Year forumla= 23.79 <0.001 19 
Age forumla = 5.00 0.02 
Year * age forumla = 7.92 0.02 
Figure 5.

Stable isotope values (mean ± s.e.) for little penguin chicks (squares) and adults (dots) during different breeding stages and years, 2003, 2004, and 2006.

Figure 5.

Stable isotope values (mean ± s.e.) for little penguin chicks (squares) and adults (dots) during different breeding stages and years, 2003, 2004, and 2006.

For δ15N and δ13C values for chicks, there was a significant effect of year, age, and their interaction, but the effect of sex was not significant. For any given year of the study, small young chicks showed higher values of δ15N than large older chicks, with 2006 the year with the greatest difference (Figure 5). The mixing-model analyses did not converge (Phillips and Gregg, 2003).

When analysing inter-year variance in stable isotope values and its relationship with diet diversity and breeding success, inverse trends for δ15N and δ13C were found over the 3 years studied. In 2004, when breeding success was lowest (0.7 chicks per pair), variance in δ15N (0.13‰) was the lowest and variance in δ13C (0.31‰) the highest. 2004 also had the highest diversity index calculated from regurgitations (0.6). Results for 2006 were opposite, with variance highest in δ15N (1.55‰) and lowest in δ13C (0.15‰), diversity of the diet lowest (0.4), and breeding success higher (0.8 chicks per pair) than in 2004.

Penguin trophic position from stomach-flushed data

Trophic diet was reconstructed using the relative abundance of prey from 1985 to 2006 for the prey grouped in the three trophic-level categories in relation to low, medium, or high δ15N values. Combined proportions of consumed prey of different trophic levels varied over the study period (MANOVA; Wilks' λ = 0.054, F36,2787 = 18.18, p < 0.0001), and this year effect was significant for each trophic-level category (ANOVA; all p < 0.001). After the pilchard mortality in 1995, little penguins' trophic position rose, as they consumed more prey from medium and high trophic levels and less from low trophic levels (Wilks' λ = 0.812, F3,954 = 73.68, p < 0.001; Figure 6).

Figure 6.

Mean proportion of prey of each trophic level consumed by little penguins before and after the pilchard mortality in 1995.

Figure 6.

Mean proportion of prey of each trophic level consumed by little penguins before and after the pilchard mortality in 1995.

Discussion

There was a massive mortality of pilchard in 1995 in southern Australia which led to a greatly reduced biomass of this key marine prey species (Jones et al., 1997; Gaughan et al., 2000; Gaughan, 2002; Murray and Gaughan, 2003). The effect of the mortality was immediate on little penguins at Phillip Island. In the first 2 years (1995 and 1996), their diet changed to similar proportions of barracouta, red cod, anchovy, and blue warehou, with pilchard completely absent (Chiaradia et al., 2003), followed by a high mortality of penguins in winter 1995 (Dann et al., 2000) and 2 years of very poor breeding success in 1995 and 1997 (the latter is the lowest ever recorded). The breeding success in 1996 was above average, but the penguins showed high diet diversity and produce lighter chicks (Chiaradia and Nisbet, 2006), an indicator of poor food availability. Also from 1995 to 1997, considerably fewer penguins were counted coming ashore than in the years 1992–94. Overall, it seemed that food was scarce from 1995 to 1997 and that the penguins' low food supply and small chick size (Chiaradia and Nisbet, 2006) reflected this shortage. Then, in 1998, the situation for penguins seemed to improve. Breeding success re-attained average levels and the numbers of penguins coming ashore increased, likely because there was some improvement in feeding conditions after a 2–3-year scarcity; there was probably increased abundance of prey other than pilchard. However, the sudden removal of pilchard from the penguin diet caused a long-term increase in the trophic position of little penguins, coinciding with a less-diversified diet within a simpler foodweb, but with little long-term effect on breeding output and population parameters.

Little penguins, like all seabirds, obtain all their food at sea, but they must return to land to breed. This life strategy poses constraints both on land and at sea. In the 1980s, the breeding success of little penguins was below average and the population was in decline, partly because of an overall shortage of food at sea (Dann and Cullen, 1990; Cullen et al., 1992; Dann, 1992). These declining patterns started to reverse at the start of the 1990s (Figure 2b and c). As decreases in breeding success were partially attributed to low availability of pilchard and anchovy (Cullen et al., 1992; Hobday, 1992), the massive pilchard mortality in 1995 was expected to trigger a negative effect on food supply and breeding output similar to the years when pilchard were absent from the penguin diet in the 1980s.

Pilchard was probably unavailable to the penguins, as suggested by the crash in the commercial catch at Port Phillip Bay, Victoria, from 1339 t in 1995 to virtually zero in 2006 (Figure 2a; Department of Primary Industries, 2008), where the fishery operates within the foraging zone of penguins. Pilchard did not recover to pre-mortality levels in the penguin diet during the 10 years after the mortality in 1995 (with two exceptions, in 1998 and 2002), and the post-mortality absence of pilchard agrees with models simulating its recovery rate, which showed that it would take 15 years for pilchard to recover following the mortality (Murray and Gaughan, 2003).

Since 2000, little penguins have fed mainly on three species: barracouta, anchovy, and arrow squid. They may target pilchard as one of the optimal prey but, in their absence, they may target the next-most-profitable prey and can still breed successfully, as shown here. Therefore, the abundance of prey rather than the presence/absence of a particular prey species may be most important to the breeding success of little penguins.

Stable-isotope analyses during the 3 years of study (2003, 2004, and 2006) provided further detail of foraging behaviour and complemented the long-term stomach-flush data. Isotope values showed a noticeable shift within and between breeding seasons. Inter-year variability in stable-isotope values was high over the 3 years of intensive study. During pre-laying and incubation, adults had higher values of δ15N and δ13C, coinciding with inshore feeding within Port Phillip Bay (Collins et al., 1999). Stable isotopes also showed that the two main prey items of penguins, barracouta and anchovy, have the highest values of δ15N among the prey species.

Initially, it was thought that anchovy could increase in penguin diet to occupy the ecological niche left after the pilchard mortality. A decrease in pilchard abundance is usually followed by an increase in anchovy abundance in productive temperate ecosystems (Chavez et al., 2003). However, anchovy schools behave differently (Barange et al., 2009), feed on larger sizes of zooplankton (Espinoza et al., 2009), and exploit a slightly higher trophic level than pilchard (van der Lingen, 1994; van der Lingen et al., 2009). In this study, however, anchovy had δ15N values as high as predators such as barracouta, but this could be explained if anchovy diet was dominated by fish eggs or larvae, placing them in the same trophic level of piscivorous species as anchovy off South Africa's east coast (Mketsu, 2008). This high trophic position of anchovy off Australia suggests that they are less abundant in the foodweb than pilchard. The local commercial landings support this notion, showing that there has not been a substantial increase in the anchovy fishery following the pilchard mortality (Figure 2a).

The relative abundance of arrow squid in the diet of penguins was similar before and after the pilchard mortality. In the absence of pilchard, arrow squid became the only key prey at low trophic position in the penguins' diet. Therefore, it could potentially be a low trophic and abundant substitute in the absence of pilchard. However, it has a much lower nutritional value than pilchard (Kailola et al., 1993), and although squid species constitute main food resources for several species of seabird (Shealer, 2001), its dominance in the diet of little penguins has been related to poor breeding success (Cullen et al., 1992).

Lower diversity in seabird diet is often associated with a decrease in breeding performance (Österblom et al., 2006; Ainley and Blight, 2009), but in some cases, low diversity is related to an abundance of prey (Croxall et al., 1999). In our study, high diet diversity coincided with negative outcomes for little penguin foraging. Diet diversity increased towards the end of the breeding season when foraging conditions usually deteriorate (Chiaradia and Nisbet, 2006). At that time, penguins would increase their diving effort, diving deeper owing to a decrease in the rate of prey encounter (I. Zimmer, pers. comm.). In years of poor breeding success, diet diversity was significantly higher. Previously, little penguins made longer trips to deliver the same amount of food to their chicks in years of low breeding success (Chiaradia and Nisbet, 2006). Therefore, they may have taken a greater diversity of species to compensate for the poor availability of food overall. This hypothesis is supported by greater variance in penguin blood δ13C values during 2004 and lower values in 2006, the years of poorer breeding success than 2003. However, the lowest variance in penguin blood δ15N values for 2004 suggests that penguins fed at the same trophic position even if the diversity of prey species was greater. Notably, diet diversity was less after the pilchard mortality, indicating a greater abundance of fewer prey species. This suggests that generalist birds such as little penguins can be more specialized when prey density is greater (Krebs et al., 1977; Woo et al., 2008).

The large change in biomass and spatial scale of the pilchard mortality could have triggered an ecological regime shift in the southern Australian marine ecosystem, as seen elsewhere (Österblom et al., 2006, 2007). Although the mortality facilitated the expansion of other species off southern Australia (Ward et al., 2001a), it does not seem to have prompted a pilchard/anchovy regime shift (Lluch-Belda et al., 1989; Shannon et al., 2004). However, no long-term studies have been conducted to determine the ecological impact of the mortality event in the marine foodweb of southern Australia, and more research is needed to understand the system better, such as in studies of predator–prey interactions in the Baltic Sea, southern Benguela, North Atlantic, California Current, and Southern Ocean (Christensen et al., 2003; Shannon et al., 2004; Österblom et al., 2006, 2007; Ainley and Blight, 2009; Baum and Worm, 2009). In this context, an increase in the trophic position of little penguin prey reported here, probably a consequence of the pilchard mortality, is one of only a few studies on a predator response to that unique event (see Bunce and Norman, 2000; Ward et al., 2001a; Taylor and Roe, 2004). Fishing up the foodweb would decrease prey abundance of the order of tenfold per trophic level (Christensen, 2000), but such a shift seems to have buffered, if not had a positive impact on, little penguins. However, the simpler foodweb and a less-flexible penguin diet (Srivastava and Bell, 2009) may make little penguins ultimately more vulnerable to further changes in their foodweb.

Acknowledgements

We thank all those involved in collecting and processing data for this study, in particular P. Dann for the 1998 diet data and providing long-term data on breeding success, L. Renwick for her hard work assisting in the field, F. Hume for her meticulous sorting of the stomach samples, M. Healy for providing messy, hard-to-blend fish-shake, T. Montague and C. Hull for collecting data from before the pilchard mortality, T. Shaw for updating the diet database, P. Coutin for the fish samples, J McInnes for the DNA sexing of chicks, all the Penguin Parade® rangers for the daily count of penguins crossing the beach over decades, and C. D. van der Lingen for useful discussion on anchovy trophic levels. Stable isotope samples were prepared in the Environment Canada laboratory of KAH and analysed at the Department of Soil Science, University of Saskatchewan, by M. Stocki. We are also appreciative of the grants received from the Penguin Foundation and Junta Andalucía to AC, Ministério de Educación y Ciencia de España CGL 2006-278999-E/BOS for MGF, and additional funding provided by an operating grant to KAH from Environment Canada. The project was approved by the Phillip Island Nature Park Animal Experimentation Ethics Committee and by the Department of Sustainability and Environment of Victoria, Australia.

References

Ainley
D. G.
Blight
L. K.
Ecological repercussions of historical fish extraction from the Southern Ocean
Fish and Fisheries
 , 
2009
, vol. 
10
 (pg. 
13
-
38
)
Arnould
J. P. Y.
Dann
P.
Cullen
J. M.
Determining the sex of little penguins (Eudyptula minor) in northern Bass Strait using morphometric measurements
Emu
 , 
2004
, vol. 
104
 (pg. 
261
-
265
)
Barange
M.
Coetzee
J.
Takasuka
A.
Hill
K.
Gutierrez
M.
Oozeki
Y.
van der Lingen
C.
, et al.  . 
Habitat expansion and contraction in anchovy and sardine populations
Progress in Oceanography
 , 
2009
, vol. 
83
 (pg. 
251
-
260
)
Barrett
R. T.
Camphuysen
K. C. J.
Anker-Nilssen
T.
Chardine
J. W.
Furness
R. W.
Garthe
S.
Huppop
O.
, et al.  . 
Diet studies of seabirds: a review and recommendations
ICES Journal of Marine Science
 , 
2007
, vol. 
64
 (pg. 
1675
-
1691
)
Baum
J. K.
Worm
B.
Cascading top-down effects of changing oceanic predator abundances
Journal of Animal Ecology
 , 
2009
, vol. 
78
 (pg. 
699
-
714
)
Bernoth
M.
Structural arrangements in Australia for managing aquatic animal disease emergencies
Developments in Biologicals
 , 
2007
, vol. 
129
 (pg. 
55
-
64
)
[PubMed]
Bunce
A.
Norman
F. I.
Changes in the diet of the Australian gannets Morus serrator in response to the 1998 mortality of pilchards Sardinops sagax
Marine and Freshwater Research
 , 
2000
, vol. 
51
 (pg. 
349
-
352
)
Caut
S.
Angulo
E.
Courchamp
F.
Variation in discrimination factors (Δ15N and Δ13C): the effect of diet isotopic values and applications for diet reconstruction
Journal of Applied Ecology
 , 
2009
, vol. 
46
 (pg. 
443
-
753
)
Chavez
F. P.
Ryan
J.
Lluch-Cota
S. E.
Niquen
M. C.
From anchovies to sardines and back: multidecadal change in the Pacific Ocean
Science
 , 
2003
, vol. 
299
 (pg. 
217
-
221
)
Chiaradia
A.
Costalunga
A.
Kerry
K.
The diet of little penguins Eudyptula minor at Phillip Island, Victoria, following the 1995 mass mortality of one of their main prey, the pilchard Sardinops sagax
Emu
 , 
2003
, vol. 
103
 (pg. 
43
-
48
)
Chiaradia
A.
Nisbet
I. C. T.
Plasticity in parental provisioning and chick growth in little penguins Eudyptula minor in years of high and low breeding success
Ardea
 , 
2006
, vol. 
94
 (pg. 
257
-
270
)
Chiaradia
A. F.
Kerry
K. R.
Nest attendance and breeding success in the little penguins Eudyptula minor at Phillip Island, Australia
Marine Ornithology
 , 
1999
, vol. 
27
 (pg. 
13
-
20
)
Christensen
V.
Indicators for marine ecosystems affected by fisheries
Marine and Freshwater Research
 , 
2000
, vol. 
51
 (pg. 
447
-
450
)
Christensen
V.
Guénette
S.
Heymans
J. J.
Walters
C. J.
Zeller
D.
Pauly
D.
Hundred-year decline of North Atlantic predatory fishes
Fish and Fisheries
 , 
2003
, vol. 
4
 (pg. 
1
-
24
)
Collins
M.
Cullen
J. M.
Dann
P.
Seasonal and annual foraging movements of little penguins from Phillip Island, Victoria
Wildlife Research
 , 
1999
, vol. 
26
 (pg. 
705
-
721
)
Crawford
R. J. M.
Food, fishing and seabirds in the Benguela upwelling system
Journal of Ornithology
 , 
2007
, vol. 
148
 (pg. 
S253
-
S260
)
Crawford
R. J. M.
Dyer
B. M.
Responses by four seabird species to a fluctuating availability of Cape anchovy Engraulis capensis off South Africa
Ibis
 , 
1995
, vol. 
137
 (pg. 
329
-
339
)
Croxall
J. P.
Reid
K.
Prince
P. A.
Diet, provisioning and productivity responses of marine predators to differences in availability of Antarctic krill
Marine Ecology Progress Series
 , 
1999
, vol. 
177
 (pg. 
115
-
131
)
Cullen
J. M.
Montague
T. L.
Hull
C.
Food of little penguins Eudyptula minor in Victoria: comparison of three localities between 1985 and 1988
Emu
 , 
1992
, vol. 
91
 (pg. 
318
-
341
)
Daniel
T.
Chiaradia
A.
Logan
M.
Quinn
G. P.
Reina
R. D.
Synchronized group association in little penguins Eudyptula minor
Animal Behaviour
 , 
2007
, vol. 
74
 (pg. 
1241
-
1248
)
Dann
P.
Distribution, population trends and factors influencing the population size of little penguins Eudyptula minor on Phillip Island, Victoria
Emu
 , 
1992
, vol. 
91
 (pg. 
263
-
272
)
Dann
P.
Cullen
J. M.
Davis
L. S.
Darby
J. T.
Survival, patterns of reproduction, and lifetime reproductive output in little blue penguins Eudyptula minor on Phillip Island, Victoria, Australia
Penguin Biology
 , 
1990
Sydney
Academic Press
(pg. 
63
-
84
)
Dann
P.
Norman
F. I.
Cullen
J. M.
Neira
F. J.
Chiaradia
A.
Mortality and breeding failure of little penguins Eudyptula minor in Victoria, 1995–6 following a widespread mortality of pilchard Sardinops sagax
Marine and Freshwater Research
 , 
2000
, vol. 
51
 (pg. 
355
-
362
)
Department of Primary Industries
Fisheries Victoria Commercial Fish Production Information Bulletin 2007
2008
Victoria, Australia
Fisheries Victoria, Queenscliff
Deuel
H. J.
The Lipids. 2. Biochemistry, Digestion, Absorption, Transport, and Storage
1955
New York
Interscience
Espinoza
P.
Bertrand
A.
van der Lingen
C. D.
Garrido
S.
Mendiola
B. R.
Diet of sardine (Sardinops sagax) in the northern Humboldt Current system and comparison with the diets of clupeoids in this and other eastern boundary upwelling systems
Progress in Oceanography
 , 
2009
, vol. 
83
 (pg. 
242
-
250
)
Forero
M. G.
Hobson
K. A.
Bortolotti
G. R.
Donázar
J. A.
Bertellotti
M.
Blanco
G.
Food resource utilisation by the Magellanic penguin evaluated through stable-isotope analysis: segregation by sex and age and influence on offspring quality
Marine Ecology Progress Series
 , 
2002
, vol. 
234
 (pg. 
289
-
299
)
France
R. L.
Carbon-13 enrichment in benthic compared to planktonic algae: foodweb implications
Marine Ecology Progress Series
 , 
1995
, vol. 
124
 (pg. 
307
-
312
)
Frank
K. T.
Petrie
B.
Choi
J. S.
Leggett
W. C.
Trophic cascades in a formerly cod-dominated ecosystem
Science
 , 
2005
, vol. 
308
 (pg. 
1621
-
1623
)
Frederiksen
M.
Mavor
R. A.
Wanless
S.
Seabirds as environmental indicators: the advantages of combining data sets
Marine Ecology Progress Series
 , 
2007
, vol. 
352
 (pg. 
205
-
211
)
Fridolfsson
A-K.
Ellegren
H.
A simple and universal method for molecular sexing of non-ratite birds
Journal of Avian Biology
 , 
1999
, vol. 
30
 (pg. 
116
-
121
)
Furness
R. W.
Tasker
M. L.
Seabird–fishery interactions: quantifying the sensitivity of seabirds to reductions in sandeel abundance, and identification of key areas for sensitive seabirds in the North Sea
Marine Ecology Progress Series
 , 
2000
, vol. 
202
 (pg. 
253
-
264
)
Gales
R. P.
Validation of the stomach-flushing technique for obtaining stomach contents of penguins
Ibis
 , 
1987
, vol. 
129
 (pg. 
335
-
343
)
Gales
R. P.
Sexing adult blue penguins by external measurements
Notornis
 , 
1988
, vol. 
35
 (pg. 
71
-
75
)
Gaughan
D. J.
Disease-translocation across geographic boundaries must be recognized as a risk even in the absence of disease identification: the case with Australian Sardinops
Reviews in Fish Biology and Fisheries
 , 
2002
, vol. 
11
 (pg. 
113
-
123
)
Gaughan
D. J.
Mitchell
R. W.
Blight
S. J.
Impact of mortality, possibly due to herpesvirus, on pilchard Sardinops sagax stocks along the south coast of Western Australia in 1998–99
Marine and Freshwater Research
 , 
2000
, vol. 
51
 (pg. 
601
-
612
)
Griffin
D. A.
Thompson
P. A.
Bax
N. J.
Bradford
R. W.
Hallegraeff
G. M.
The 1995 mass mortality of pilchard: no role found for physical or biological oceanographic factors in Australia
Marine and Freshwater Research
 , 
1997
, vol. 
48
 (pg. 
27
-
42
)
Hobday
D. K.
Abundance and distribution of pilchard and Australian anchovy as prey species for the little penguin Eudyptula minor at Phillip Island, Victoria
Emu
 , 
1992
, vol. 
91
 (pg. 
342
-
354
)
Hobson
K. A.
Fisk
A.
Karnovsky
N.
Holst
M.
Gagnon
J-M.
Fortier
M.
A stable isotope (13C, 15N) model for the North Water food web: implications for evaluating trophodynamics and the flow of energy and contaminants
Deep-Sea Research II
 , 
2002
, vol. 
49
 (pg. 
5131
-
5150
)
Hobson
K. A.
Gibbs
H. L.
Gloutney
M. L.
Preservation of blood and tissue samples for stable-carbon and stable-nitrogen isotope analysis
Canadian Journal of Zoology
 , 
1997
, vol. 
75
 (pg. 
1720
-
1723
)
Hobson
K. A.
Wassenaar
L. I.
Tracking Animal Migration with Stable Isotopes
2008
New York
Academic Press
Hobson
K. A.
Welch
H. E.
Determination of trophic relationships within a high Arctic marine food web using 13C and 15N analysis
Marine Ecology Progress Series
 , 
1992
, vol. 
84
 (pg. 
9
-
18
)
Hoskins
A. J.
Dann
P.
Ropert-Coudert
Y.
Kato
A.
Chiaradia
A.
Costa
D. P.
Arnould
J. P. Y.
Foraging behaviour and habitat selection of the little penguin Eudyptula minor during early chick rearing in Bass Strait, Australia
Marine Ecology Progress Series
 , 
2008
, vol. 
366
 (pg. 
293
-
303
)
Inger
R.
Bearhop
S.
Applications of stable isotope analyses to avian ecology
Ibis
 , 
2008
, vol. 
150
 (pg. 
447
-
461
)
Jones
J. B.
Hyatt
A. D.
Hine
P. M.
Whittington
R. J.
Griffin
D. A.
Bax
N. J.
Special topic review: Australasian pilchard mortalities
World Journal of Microbiology and Biotechnology
 , 
1997
, vol. 
13
 (pg. 
383
-
392
)
Kailola
P. J.
Williams
M. J.
Stewart
P. C.
Reichelt
R. E.
McNee
A.
Grieve
C.
Australian Fisheries Resources
1993
Canberra
Fisheries Research and Development Corporation
Krebs
J. R.
Erichsen
J. T.
Webber
M. I.
Charnov
E. L.
Optimal prey selection in the great tit (Parus major)
Animal Behaviour
 , 
1977
, vol. 
25
 (pg. 
30
-
38
)
Littell
R. C.
Milliken
G. A.
Stroup
W. W.
Wolfinger
R. D.
SAS System for Mixed Models
1996
Cary, NC
SAS Institute Inc.
Lluch-Belda
D.
Crawford
R. J. M.
Kawasaki
T.
MacCall
A. D.
Parrish
R. H.
Schwartzlose
R. A.
Smith
P. E.
World-wide fluctuations of sardine and anchovy stocks: the regime problem
South African Journal of Marine Science
 , 
1989
, vol. 
8
 (pg. 
195
-
205
)
Mketsu
Q. K.
Comparative dietary analysis of small pelagic fishes from presumed mixed shoals off South Africa's east coast
2008
MSc thesis, University of Cape Town
pg. 
74
 
Montague
T. L.
Cullen
J. M.
The diet of the little penguin Eudyptula minor at Phillip Island, Victoria
Emu
 , 
1988
, vol. 
88
 (pg. 
139
-
149
)
Montevecchi
W. A.
Binary dietary responses of northern gannets Sula bassana indicate changing food web and oceanographic conditions
Marine Ecology Progress Series
 , 
2007
, vol. 
352
 (pg. 
213
-
220
)
Murray
A. G.
Gaughan
D. J.
Using an age-structured model to simulate the recovery of the Australian pilchard (Sardinops sagax) population following epidemic mass mortality
Fisheries Research
 , 
2003
, vol. 
60
 (pg. 
415
-
426
)
Myers
R. A.
Baum
J. K.
Shepherd
T. D.
Powers
S. P.
Peterson
C. H.
Cascading effects of the loss of apex predatory sharks from a coastal ocean
Science
 , 
2007
, vol. 
315
 (pg. 
1846
-
1850
)
Österblom
H.
Casini
M.
Olsson
O.
Bignert
A.
Fish, seabirds and trophic cascades in the Baltic Sea
Marine Ecology Progress Series
 , 
2006
, vol. 
323
 (pg. 
233
-
238
)
Österblom
H.
Hansson
S.
Larsson
U.
Hjerne
O.
Wulff
F.
Elmgren
R.
Folke
C.
Human-induced trophic cascades and ecological regime shifts in the Baltic Sea
Ecosystems
 , 
2007
, vol. 
10
 (pg. 
877
-
889
)
Overeem
R.
Wallis
R.
Salzman
S.
Sexing little penguins Eudyptula minor using bill measurements
Victorian Naturalist
 , 
2006
, vol. 
123
 (pg. 
390
-
395
)
Phillips
D. L.
Gregg
J. W.
Source partitioning using stable isotope: coping with too many sources
Oecologia
 , 
2003
, vol. 
136
 (pg. 
261
-
269
)
Piatt
J. F.
Harding
A. M. A.
Shultz
M.
Speckman
S. G.
van Pelt
T. I.
Drew
G. S.
Kettle
A. B.
Seabirds as indicators of marine food supplies: Cairns revisited
Marine Ecology Progress Series
 , 
2007
, vol. 
352
 (pg. 
221
-
234
)
Reid
K.
Croxall
J. P.
Environmental response of upper trophic-level predators reveals a system change in an Antarctic marine ecosystem
Proceedings of the Royal Society, Series B
 , 
2001
, vol. 
268
 (pg. 
377
-
384
)
Shannon
C. E.
Weaver
W.
The Mathematical Theory of Communication
1949
Urbana
University of Illinois Press
Shannon
L. J.
Field
J. G.
Moloney
C. L.
Simulating anchovy–sardine regime shifts in the southern Benguela ecosystem
Ecological Modelling
 , 
2004
, vol. 
172
 (pg. 
269
-
281
)
Shealer
D. A.
Schreiber
E. A.
Burger
J.
Foraging behavior and food of seabirds
Biology of Marine Birds
 , 
2001
Boca Raton, FL
CRC Press
(pg. 
137
-
177
)
Srivastava
D. S.
Bell
T.
Reducing horizontal and vertical diversity in a foodweb triggers extinctions and impacts functions
Ecology Letters
 , 
2009
, vol. 
12
 (pg. 
1016
-
1028
)
Stahel
C.
Gales
R.
Little Penguin, Fairy Penguins in Australia
1987
Kensington
New South Wales University Press
Taylor
I. R.
Roe
E. L.
Feeding ecology of little terns Sterna albifrons sinensis in south-eastern Australia and the effects of pilchard mass mortality on breeding success and population size
Marine and Freshwater Research
 , 
2004
, vol. 
55
 (pg. 
799
-
808
)
van der Lingen
C. D.
Effect of particle size and concentration on the feeding behaviour of adult pilchard Sardinops sagax
Marine Ecology Progress Series
 , 
1994
, vol. 
109
 (pg. 
1
-
13
)
van der Lingen
C. D.
Bertrand
A.
Bode
A.
Brodeur
R.
Cubillos
L. A.
Espinoza
P.
Friedland
K.
, et al.  . 
Checkley
D.
Alheit
J.
Oozeki
Y.
Roy
C.
Trophic dynamics
Climate Change and Small Pelagic Fish
 , 
2009
Cambridge, UK
Cambridge University Press
(pg. 
112
-
157
)
Ward
T. M.
Hoedt
F.
McLeay
L.
Dimmlich
W. F.
Jackson
G.
Rogers
P. J.
Jones
K.
Have recent mass mortalities of the sardine Sardinops sagax facilitated an expansion in the distribution and abundance of the anchovy Engraulis australis in South Australia?
Marine Ecology Progress Series
 , 
2001
, vol. 
220
 (pg. 
241
-
251
)
Ward
T. M.
Hoedt
F.
McLeay
L.
Dimmlich
W. F.
Kinloch
M.
Jackson
G.
McGarvey
R.
, et al.  . 
Effects of the 1995 and 1998 mass mortality events on the spawning biomass of sardine, Sardinops sagax, in South Australian waters
ICES Journal of Marine Science
 , 
2001
, vol. 
58
 (pg. 
865
-
875
)
Whittington
R. J.
Crockford
M.
Jordan
D.
Jones
B.
Herpesvirus that caused epizootic mortality in 1995 and 1998 in pilchard, Sardinops sagax neopilchardus (Steindachner), in Australia is now endemic
Journal of Fish Diseases
 , 
2008
, vol. 
31
 (pg. 
97
-
105
)
[PubMed]
Whittington
R. J.
Jones
J. B.
Hine
P. M.
Hyatt
A. D.
Epizootic mortality in the pilchard Sardinops sagax in Australia and New Zealand in 1995. Pathology and epizootiology
Diseases of Aquatic Organisms
 , 
1997
, vol. 
28
 (pg. 
1
-
16
)
Woo
K. J.
Elliott
K. H.
Davidson
M.
Gaston
A. J.
Davoren
G. K.
Individual specialization in diet by a generalist marine predator reflects specialization in foraging behaviour
Journal of Animal Ecology
 , 
2008
, vol. 
77
 (pg. 
1082
-
1091
)

Author notes

Deceased.