Abstract

Although seabirds frequently aggregate with feeding delphinids, the benefits to seabirds of feeding with dolphins have been rarely reported. We examined how dusky dolphins (Lagenorhynchus obscurus) influenced prey accessibility for seabirds in Admiralty Bay, New Zealand. Interactions of dusky dolphins and seabirds were characterized during 335 feeding bouts of dusky dolphins (52 video-recorded underwater). Dolphins increased prey accessibility for seabirds because they swam under the bottom half of prey balls for 59% of passes that were within 2 m of the prey ball. During feeding bouts by dolphins, 51% of prey balls ascended, whereas only 13% descended. Dolphins also influenced prey mobility; only 24% of stationary feeding bouts became mobile after dolphins began feeding, and 17% subsequently became stationary again. Significantly more Australasian gannets (Morus serrator) were near mobile than stationary prey balls after feeding, but not during feeding bouts. This suggests that feeding gannets increase mobility of prey balls, but that feeding dolphins counteract this effect. Seabirds also used dusky dolphins to locate prey. Numbers of gannets, shearwaters (Puffinus), and gulls (Larus) increased during the first 2 min of dolphin feeding, even when other seabirds were not present. Gannets fed with dolphins for 40% of gannet feeding observations and shearwaters fed with dolphins for 24% of shearwater feeding observations.

Interspecies feeding aggregations that include fishes, sea-birds, pinnipeds, and cetaceans are common in marine environments (Au 1991; Ballance et al. 2006; Bearzi 2006; Lukoschek and McCormick 2002), but the costs and benefits of feeding in aggregate are often unknown (Stensland et al. 2003). Presumably, the primary benefits of feeding in aggregate with cetaceans include reduced predation (Bertram 1978; Norris and Schilt 1988) and increased foraging success. For example, prey in marine environments are often patchy and unpredictable (Ballance et al. 2006; Wells et al. 1999), and feeding in aggregate can increase a predator's ability to find prey (Bearzi 2006; Würsig and Würsig 1980). Feeding in aggregate with cetaceans also may increase prey accessibility for predators (Bräger 1998; Martin 1986; Würsig 1986). Increased competition for food is a primary cost of feeding in aggregate (Acevedo-Gutiérrez 2002a; Pierotti 1988).

The degree that seabirds benefit from feeding with cetaceans depends on cetacean feeding tactics. Cetaceans that feed at depth do not usually attract seabirds (Evans 1982). Conversely, seabirds benefit from feeding with cetaceans when they feed near the surface, and when they feed on stationary prey (i.e., prey that is not moving horizontally—Bräger 1998; Ridoux 1987). For example, feeding behaviors of humpback whales (Megaptera novaeangliae) include using bubbles and lunging to concentrate prey balls into tight masses near the surface, which increases prey accessibility for seabirds (Evans 1982). Similarly, delphinids such as Atlantic spotted dolphins (Stenella frontalis) that herd prey to the surface make it easier for seabirds to capture prey (Martin 1986).

Dusky dolphins (Lagenorhynchus obscurus) are an excellent species with which to examine the potential costs and benefits to seabirds feeding with cetaceans, because they feed at or near the surface in several nearshore environments. Dusky dolphins occur in the Southern Hemisphere off South Africa, Argentina, Chile, Peru, several temperate to subantarctic islands, and New Zealand (Van Waerebeek and Würsig 2002). In Admiralty Bay, New Zealand, they are generally present only in winter and spring, and approximately 220 dolphins are present during any given week in winter (Markowitz et al. 2004). They feed during the day on small schooling fishes (including anchovy [Engraulis australis], garfish [Hyporhamphus ihi], pilchard [Sardinops neopilchardus], sprats [Sprattus], and yellow-eyed mullet [Aldrichetta forsteri]—Duffy and Brown 1994; McFadden 2003; Vaughn et al. 2007), and at times herd prey balls toward the surface (Markowitz 2004, Markowitz et al. 2004; McFadden 2003). New Zealand fur seals (Arctocephalus forsten), seabirds, spiny dogfish (Squalus acanthias), and thresher sharks (Alopias vulpinus) aggregate with dusky dolphins during feeding (McFadden 2003; Vaughn et al. 2007). Seabirds that most commonly feed with dusky dolphins include Australasian gannets (Morus serrator), shearwaters (Puffinus), white-fronted terns (Sterna striata), gulls (Larus), and spotted shags (Phalacrocorax punctatusMarkowitz et al. 2004; McFadden 2003). Fluttering shearwaters (P. gavia) are common in Admiralty Bay in winter (McFadden 2003), whereas flesh-footed shearwaters (P. carneipes) and sooty shearwaters (P. griseus) occur after September (Heather and Robertson 2005). For the remainder of this manuscript, “dolphin” refers to dusky dolphins.

The objective of this study was to investigate how dolphins influence prey accessibility for those seabirds that most commonly feed with them (gannets, shearwaters, terns, gulls, and shags) in Admiralty Bay, New Zealand. We determined the influence of dolphins on depth, mobility, and shape of prey balls. We also investigated if seabirds use dolphins as cues to initially locate prey. Finally, to examine the importance of foraging tactics of dolphins for the above seabirds, we determined how frequently they fed with versus without dolphins.

Materials and Methods

Study area.—Admiralty Bay (40°56′S, 173°53′E) is located in the Marlborough Sounds at the north part of the South Island of New Zealand and is bounded by D'Urville Island to the north and west (Fig. 1). It is a shallow-water environment with typical depth of 30–50 m, maximum depth of 105 m near French Pass, and mud substrate. Current Basin is connected to Admiralty Bay via French Pass. Typical May–November water temperatures are 10–14°C (http://www.niwascience.co.nz). During this study, water clarity ranged from 5 to 18 m ( =10 m), as measured by Secchi disk approximately hourly.

Fig. 1

Location of study area and survey routes followed in Admiralty Bay and Current Basin. Map was created using ArcView GIS version 3.3 (Environmental Systems Research Institute, Inc., Redlands, California).

Fig. 1

Location of study area and survey routes followed in Admiralty Bay and Current Basin. Map was created using ArcView GIS version 3.3 (Environmental Systems Research Institute, Inc., Redlands, California).

Above-water data collection and analyses.—Data were collected from 5 August to 4 November 2005, and from 22 May to 28 August 2006, from a 5.5-m rigid-hulled inflatable boat with an 85 hp 2-stroke engine (2005), or an 80 hp 4-stroke engine (2006; Xamaha Motor Co., Iwata-shi, Shizuoka-ken, Japan). Groups of dolphins were located either by driving on transect lines at 10–12 knots on 2 predetermined routes in Admiralty Bay and Current Basin (Fig. 1), or opportunistically while driving to or from a transect line. Survey routes were similar to those of previous studies in the area (Markowitz 2004; Markowitz et al. 2004).

Transect data included the recording of all dolphins, plus feeding gannets, shearwaters, terns, and gulls that were within 400 m. We did not record number of feeding shags because we could not determine if shags dived because they were feeding or because of proximity of the research boat. Seabirds and dolphins were sighted without using binoculars. For feeding seabirds sighted from transects, numbers of seabirds and location (using a handheld Garmin model 76 global positioning system unit; Garmin Ltd., Olathe, Kansas) were recorded. Number of seabirds in a group (if > 1 seabird was present) was determined by including all seabirds that were within 10 m of another seabird, and the group was considered feeding if at least 1 seabird was feeding. Indicators of feeding were diving underwater for gannets and shearwaters, and swooping down to the water's surface from the air for gulls and terns. For dolphins sighted from transects, location, approximate distance from the transect line, number of dolphins, initial dolphin behavior (see below for dolphin behavior categories), and numbers of aggregated seabirds were recorded. The same person estimated distance for all survey sightings. Number of dolphins in a group (if >1 dolphin was present) was determined by including all dolphins within 10 m of another dolphin (Smolker et al. 1992). A group of seabirds was defined as aggregated with dolphins if any seabirds were within 10 m of dolphins.

We recorded data while following focal groups of dolphins at a distance of 25–50 m. Data were recorded at 2-min intervals by 1 or 2 observers. Above-water data included location, numbers of dolphins and all seabirds, and dolphin behavior. Dolphin behavior was defined using predominant group activity sampling (Altmann 1974; Mann 1999). Five categories were used: travel, rest, social, foraging, and feeding. Traveling behavior was defined as horizontal movement that was predominantly in 1 direction, and at least 3 knots. Resting behavior was defined as movement that was less than 3 knots. Social behavior was defined as acrobatic or touching behavior. Foraging behavior was defined as low-level activity that was widely scattered and that included occasional burst swims or leaping. Feeding behavior was defined as high-level activity that changed direction frequently and included leaping or burst swims. Three types of feeding behaviors were specified: fish balling—defined as stationary feeding (i.e., feeding that is not moving horizontally); deep feeding—defined as mobile feeding characterized by long surface intervals or clean (headfirst reentry) leaps; and shallow feeding—defined as mobile feeding characterized by short surface intervals or burst swims. A feeding bout was defined as a continuous, discrete period of feeding.

SPSS version 11.0.3 for Mac OS X statistical software (SPSS Inc., Chicago, Illinois) was used for data analyses. The 2-tailed alpha level was set at 0.05. Descriptive and nonparametric (Spearman's rank order correlation, Kruskal–Wallis, Mann–Whitney, and Wilcoxon tests) statistics were used to analyze data because of low sample sizes and distributions that were not normal. Means ± SD are presented. This study was conducted under Texas A&M Animal Use Protocol 2005–48, and followed guidelines approved by the American Society of Mammalogists (Gannon et al. 2007).

Underwater data collection.—During stationary feeding bouts by dusky dolphins, 1 or 2 observers collected underwater data via surface swimming and shallow breath-hold diving at a distance of 3–10 m. Stationary feeding bouts occurred from August through October 2005, and in August 2006. Under-water data collection occurred simultaneously with above-water data collection.

Prey depth, defined as the distance from the water's surface to the middle of the prey ball, was estimated at 2-min intervals. The mean length of an adult New Zealand dolphin (1.73 m, sensu Cipriano [1992]) was used for reference. Additionally, during feeding bouts by dolphins and after feeding, we recorded if prey balls remained stationary or moved out of sight. Underwater video was recorded using a Sony DCR-HC1000 video camera (Sony Corporation of America, New York, New York) in an Amphibico Invader electronic underwater housing (Amphibico, Montreal, Quebec, Canada). Focal length was 3.6 mm, shutter speed 1/500 s, and speed of frame capture 30 frames/s.

Underwater video analyses.—From underwater footage, the 2-dimensional area of prey balls was measured with Image J software (http://rsb.info.nih.gov/ij/), using the mean length of an adult dusky dolphin from New Zealand for reference.

The frequency with which dolphins swam near the bottom half versus the top half of prey balls during feeding bouts was determined. For this analysis, we considered only the half of the prey ball facing the video camera. The statistical criteria for the analysis included determining frequency of dolphin swims that were within 2 m of the prey ball. Additionally, the analysis only included prey balls for which the top of the prey ball was at least 2 m below the surface of the water for 1 or more 2-min intervals, and feeding intervals that included >4 occurrences of dolphins swimming by the prey ball.

The shape of the prey ball was determined by measuring the width : height ratio using Image J software. Video frames were captured at 30-s intervals, then the height of each prey ball was set at 1 m as a reference from which to determine relative width of prey balls. The correlation between shape of the prey ball and activity level of dolphins was then determined. Feeding activity level of dolphins was categorized as high (>10 prey-capture attempts/2-min interval), medium (3–9 prey-capture attempts/2-min interval), or low (<2 prey-capture attempts/ 2-min interval). A prey-capture attempt was defined as a dolphin swimming adjacent to a prey ball and accompanied by fish spraying away from the dolphin, tilting of the dolphin's head toward the prey ball, or a slight burst of speed by the dolphin.

For analysis of the correlation between shape of the prey ball and feeding activity level of dolphins, we used only the activity level that occurred with the greatest frequency for each prey ball. Size of the group of dolphins was not included in this analysis because we were interested in assessing the active influence of feeding dolphins, not the passive influence of the group size of dolphins. To determine if we were biasing our analysis by not including group size (recorded above-water), we used a Spearman's rank order correlation to assess the relationship between size of the group of dolphins and size of the prey ball.

Results

Dolphin feeding bouts.—Mean length of time that focal groups were followed was 89 min (SD = 58 min, range = 2–244 min, n = 171). During follows of focal groups, we observed 335 dolphin feeding bouts above-water and 52 dolphin feeding bouts underwater. Mean length of feeding bouts was 4.9 min (SD = 6.2 min, range = 1–42 min, n = 221 feeding bouts for which we observed beginning and end of feeding), and there was a mean of 8.3 dolphins (SD = 5.0 dolphins, range = 1–30 dolphins, n = 268) present during feeding bouts. The presence of researchers in the water appeared to influence dolphin behavior minimally, because dolphins continued feeding when researchers were in the water.

Do dolphins influence depth of prey balls?—Mean depth of prey during feeding bouts was 3.6 m (SD = 2.7 m, range = 0.5–18 m, n = 52). Twelve prey balls were >10 m below the surface during a dolphin feeding bout. Underwater data were recorded for 2 consecutive dolphin feeding intervals for 37 of 52 prey balls. Of these 37 prey balls, 19 (51%) ascended, 5 (13%) descended, and 13 (35%) stayed at a constant depth. The mean distance that prey balls ascended was 5.0 m (SD = 4.7 m, range = 1–18 m, n = 19; Fig. 2A), or 1.4 m/2-min interval (SD = 1.5 m/interval, range = 0.1–7.0 m/interval, n = 19). The mean distance that prey balls descended was 2.5 m (SD =2.5 m, range = 0.5–5.5 m, n = 5; Fig. 2A), or 0.4 m/interval (SD = 0.3 m/interval, range = 0.1–0.8 m/interval, n = 5). The mean depth at which ascending prey balls were initially observed was 7.6 m (SD =4.4 m, range = 2.0–20.0, n = 19; Fig. 2B). The mean depth at which descending prey balls were initially observed was 1.5 m (SD = 2.0 m, range = 0.5–5.0, n =5; Fig. 2B).

Fig. 2

Prey balls that were observed underwater for at least 2 consecutive feeding intervals of dusky dolphins (Lagenorhynchus obscurus) that ascended (n = 19, gray) or descended (n = 5, white), showing A) change in depth of prey balls, and B) depth at which these prey balls were initially observed.

Fig. 2

Prey balls that were observed underwater for at least 2 consecutive feeding intervals of dusky dolphins (Lagenorhynchus obscurus) that ascended (n = 19, gray) or descended (n = 5, white), showing A) change in depth of prey balls, and B) depth at which these prey balls were initially observed.

Overall, dolphins swam near the bottom half of the prey ball for 693 (59%) of 1,174 passes that were within 2 m of the prey ball. Dolphins swam near the bottom half of the prey ball more frequently than they swam near the top half of the prey ball for 24 (89%) of 27 two-minute intervals, and 6 (67%) of 9 feeding bouts (Fig. 3). Total number of passes by the prey ball per 2-min interval ranged from 12 to 123.

Fig. 3

Frequency per feeding bout (n = 9) with which dolphins (Lagenorhynchus obscurus) swam within 2 m of the prey ball near the bottom half of the prey ball (gray) versus the top half of the prey ball (white), for the half of the prey ball facing the video camera.

Fig. 3

Frequency per feeding bout (n = 9) with which dolphins (Lagenorhynchus obscurus) swam within 2 m of the prey ball near the bottom half of the prey ball (gray) versus the top half of the prey ball (white), for the half of the prey ball facing the video camera.

Do dolphins influence mobility of prey balls?—During above-water observations of dolphin feeding behaviors, herding behaviors of dolphins appeared to influence mobility of prey. Stationary feeding occurred during 110 (33%) of 335 feeding bouts. Of these 110 stationary feeding bouts, 26 (24%) became mobile while feeding by dolphins was still occurring; of these 26 mobile feeding bouts, 4 (17%) subsequently became stationary again. The start of feeding by dolphins was observed for 82 stationary feeding bouts; of these, 20 (24%) were mobile for >1 min before they became stationary.

Prey balls were more often stationary after than during feeding bouts. During feeding bouts, 18 of 52 prey balls remained stationary until dolphins finished feeding whereas 34 of 52 prey balls became mobile (Fig. 4). Of these 34 mobile prey balls, 31 (91%) moved out of sight laterally, whereas 3 (9%) descended. After feeding bouts, 12 of 18 prey balls remained stationary, whereas 6 of 18 prey balls became mobile (Fig. 4). Of these 6 mobile prey balls, 4 (67%) moved out of sight laterally, whereas 2 (33%) descended.

Fig. 4

Percentage of stationary prey balls observed underwater that became mobile compared to those that remained stationary during (n = 52) and after (n = 18) feeding bouts by dusky dolphins (Lagenorhynchus obscurus). Sizes of circles approximate number of prey balls.

Fig. 4

Percentage of stationary prey balls observed underwater that became mobile compared to those that remained stationary during (n = 52) and after (n = 18) feeding bouts by dusky dolphins (Lagenorhynchus obscurus). Sizes of circles approximate number of prey balls.

During feeding bouts, the number of prey balls that remained stationary or that became mobile was not related to the number of dolphins (Mann–Whitney test; Z = −1.223, P =0.221), gannets (Z = −0.733, P =0.463), shearwaters (Z = −1.185, P =0.236), terns (Z = −1.858, P =0.063), gulls (Z = −0.831, P =0.406), or shags (Z = −0.764, P =0.445; Table 1). After feeding bouts, there were significantly more gannets present (Mann–Whitney test; Z = −2.759, P = 0.006) near mobile prey balls than near prey balls that remained stationary, but there was no difference in number of shearwaters (Z = −0.132, P = 0.895), or gulls (Z = −0.552, P =0.581).

Table 1

Mean numbers of dusky dolphins (Lagenorhynchus obscurus) and seabirds present during and after dolphin feeding bouts, for prey balls that remained stationary compared to prey balls that became mobile. NA = not applicable.

 During feeding bouts After feeding bouts 
Stationary prey balls Mobile prey balls Stationary prey balls Mobile prey balls 
X̄ ± SD n X̄ ± SD n X̄ ± SD n X ± SD n 
Dolphins 8.5 ± 3.2 18 10.2 ± 4.6 33 NA NA NA NA 
Gannets 4.1 ± 7.8 16 5.9 ± 11.7 31 0.4 ± 0.8 10 9.0 ± 15.3 
Shearwaters 18.6 ± 40.9 16 5.7 ± 16.4 31 4.0 ± 7.0 10 5.8 ± 12.0 
Terns 16 0.2 ± 0.7 31 10 
Gulls 4.1 ± 4.4 16 7.0 ± 9.9 31 2.8 ± 3.1 10 5.7 ± 7.3 
Shags 0.2 ± 0.3 16 0.3 ± 0.5 31 10 0.3 ± 0.2 
 During feeding bouts After feeding bouts 
Stationary prey balls Mobile prey balls Stationary prey balls Mobile prey balls 
X̄ ± SD n X̄ ± SD n X̄ ± SD n X ± SD n 
Dolphins 8.5 ± 3.2 18 10.2 ± 4.6 33 NA NA NA NA 
Gannets 4.1 ± 7.8 16 5.9 ± 11.7 31 0.4 ± 0.8 10 9.0 ± 15.3 
Shearwaters 18.6 ± 40.9 16 5.7 ± 16.4 31 4.0 ± 7.0 10 5.8 ± 12.0 
Terns 16 0.2 ± 0.7 31 10 
Gulls 4.1 ± 4.4 16 7.0 ± 9.9 31 2.8 ± 3.1 10 5.7 ± 7.3 
Shags 0.2 ± 0.3 16 0.3 ± 0.5 31 10 0.3 ± 0.2 

Do dolphins influence shape of prey balls?—Shape of prey balls was measured for 31 feeding bouts. Mean width:height ratios were 2.4 (SD =0.8, range = 1.2–3.7, n =12), 2.8 (SD =1.8, range = 1.0–5.5, n =9), and 2.1 (SD = 0.8, range = 1.5– 4.1, n = 10) for high, medium, and low feeding activity levels of dolphins, respectively. There was no difference in width : -height ratio between high, medium, and low activity levels (Kruskal–Wallis test; χ2 = 0.454, d.f. = 2, P =0.797). There was no correlation between size of groups of dolphins and size of prey balls (Spearman's R =−0.182, P = 0.336, n =30).

Do dolphins make it easier for seabirds to locate prey?— There was a significant increase in number of gannets (Wilcoxon test; Z = −4.351, P< 0.001), shearwaters (Z = −4.113, P< 0.001), and gulls (Z = −3.656, P< 0.001) during the first 2-min of feeding (Table 2). However, there was not a significant increase in number of terns (Z = −1.175, P =0.240) or shags (Z = −1.780, P =0.075).

Table 2

Mean number of seabirds present at the start of feeding bouts by dusky dolphins (Lagenorhynchus obscurus) compared to at the end of the first 2-min interval.

 Start of dolphin feeding bout 2-min later 
X̄ ± SD Range n X̄ ± SD Range n 
Gannets Shearwaters Terns Gulls Shags 1.4 ± 3.9 23.3 ± 55.1 0.1 ± 0.5 0.7 ± 4.0 0.1 ± 0.6 0–30.0 0–400.0 0–5.0 0–40.0 0–7.0 178 176 178 178 174 3.0 ± 7.2 33.9 ± 66.4 0.1 ± 0.6 1.5 ± 6.0 0.2 ± 0.7 0–50.0 0–400.0 0–5.0 0–50.0 0–7.0 178 176 178 178 174 
 Start of dolphin feeding bout 2-min later 
X̄ ± SD Range n X̄ ± SD Range n 
Gannets Shearwaters Terns Gulls Shags 1.4 ± 3.9 23.3 ± 55.1 0.1 ± 0.5 0.7 ± 4.0 0.1 ± 0.6 0–30.0 0–400.0 0–5.0 0–40.0 0–7.0 178 176 178 178 174 3.0 ± 7.2 33.9 ± 66.4 0.1 ± 0.6 1.5 ± 6.0 0.2 ± 0.7 0–50.0 0–400.0 0–5.0 0–50.0 0–7.0 178 176 178 178 174 

For 59 dolphin feeding bouts, there were no seabirds present when feeding started. At the end of the first 2-min interval, 15 (25%) of these feeding bouts included gannets, 20 (34%) included shearwaters, 4 (7%) included terns, 10 (17%) included gulls, and 5 (8%) included shags.

Survey sightings of feeding dolphins and seabirds.—While driving on transects, data were collected for 22 groups of feeding dolphins and 255 groups of feeding seabirds. Seabirds were observed feeding more frequently without dolphins than with dolphins (Fig. 5). Diving gannets and shearwaters fed with dolphins more frequently than did surface-feeding terns and gulls. For gannets, mean group sizes were significantly larger when feeding with dolphins (Mann–Whitney test; Z = −5.856, P =<0.001; Table 3). However, group sizes were not different for shearwaters (Z = −1.170, P = 0.242), gulls (Z = −0.265, P =0.791), or terns (Z = −0.226, P = 0.821).

Fig. 5

Seasonal and overall frequencies with which an individual seabird fed with (white) versus without (gray) dolphins. Data are initial sightings of groups of feeding seabirds and dolphins within 400 m of transects (n = 22 feeding dolphin groups; n = 255 feeding seabird groups). In 2005, data are from 24 August to 4 November; in 2006, data are from 25 May to 8 August.

Fig. 5

Seasonal and overall frequencies with which an individual seabird fed with (white) versus without (gray) dolphins. Data are initial sightings of groups of feeding seabirds and dolphins within 400 m of transects (n = 22 feeding dolphin groups; n = 255 feeding seabird groups). In 2005, data are from 24 August to 4 November; in 2006, data are from 25 May to 8 August.

Table 3

Mean sizes of groups of seabirds when feeding with versus without dolphins. Data are based on initial sightings from transect lines.

 With dolphins Without dolphins 
X̄ ± SD Range n X̄ ± SD Range n 
Gannets 23.3 ± 12.5 8–40 1.5 ± 1.5 1–10 143 
Shearwaters 43.7 ± 38.3 3–150 15 36.0 ± 42.9 1–260 57 
Terns 2.7 ± 1.2 2–4 4.4 ± 4.7 1–20 79 
Gulls 6.7 ± 4.9 1–10 11.7 ± 12.4 1–36 
 With dolphins Without dolphins 
X̄ ± SD Range n X̄ ± SD Range n 
Gannets 23.3 ± 12.5 8–40 1.5 ± 1.5 1–10 143 
Shearwaters 43.7 ± 38.3 3–150 15 36.0 ± 42.9 1–260 57 
Terns 2.7 ± 1.2 2–4 4.4 ± 4.7 1–20 79 
Gulls 6.7 ± 4.9 1–10 11.7 ± 12.4 1–36 

There were seasonal differences in the frequencies with which seabirds fed with dolphins, compared to feeding without dolphins (Fig. 5). Gannets and gulls seldom fed with dolphins from May to July, but frequently fed with dolphins from August to November. Shearwaters frequently fed with dolphins from May to November. However, fluttering shearwaters occurred before October, and flesh-footed and sooty shearwaters took their place in early October.

Discussion

Prey have evolved effective antipredation tactics. For example, fishes from the order Clupeiformes such as anchovy, herring (Clupea), and pilchard are able to hear ultrasonic sounds (Mann et al. 1997, 1998). This may allow them to reduce predation by cetaceans. In response to simulated odontocete echolocation sounds, clupeoids drop in the water column, actively school, and increase swimming speed (Wilson and Dill 2002). In Norway, herring (C. harengus) have been observed swimming from the surface to the bottom of a fjord (100 m depth) when killer whales (Orcinus orea) entered the fjord (Similä and Ugarte 1993). In response to predation by aerial predators, schooling fishes typically descend, actively school, and compress vertically (Litvak 1993). This vertical compression probably reduces predation via the confusion effect (Bertram 1978; Norris and Schilt 1988), because it increases the number of fishes that are in the predator's visual field and thus makes it harder for a seabird that is above-water to focus on a single fish. Thus, feeding tactics by dolphins that counter antipredation tactics of clupeoid fishes by changing depth, mobility, or shape of prey balls not only increase prey accessibility for dolphins, but also increase prey accessibility for other predators, such as seabirds (Bräger 1998; Evans 1982; Harrison et al. 1991; Martin 1986).

Dolphins in Admiralty Bay increased prey accessibility for gannets, shearwaters, terns, gulls, and shags by decreasing depth of prey. During feeding bouts, 51% of prey balls ascended. Because dolphins swam under prey balls more frequently than they swam over prey balls, their herding tactics appeared to influence this change in depth of prey balls. The most common herding tactic of dolphins was to swim under and around prey balls while turning their white ventral sides toward the prey just before capture. Dolphins in Argentina also herd schooling fishes (Würsig and Würsig 1980), as do other delphinids (Acevedo-Gutiérrez 2002b;,Heithaus and Dill 2002; Wells et al. 1999). In the Marlborough Sounds, New Zealand, dolphins have been observed herding pilchard against the shore; herding anchovy, yellow-eyed mullet, and garfish against the hull of a boat; and herding fishes against a point of land (Duffy and Brown 1994). These descriptions suggest that herding prey against a surface facilitates feeding by dolphins.

Herding behaviors of dolphins appeared to be most important to gulls, gannets, and shearwaters, because these seabirds fed with dolphins most often. Gulls only feed at the surface (Heather and Robertson 2005); thus, when dolphins herded prey to the surface, they made prey accessible to gulls. Although gannets and shearwaters dive deep to catch prey, herding behaviors of dolphins also increased prey accessibility for these seabirds by decreasing dive depth. Shallow diving requires less energy than deep dives (Page et al. 2006; Peck and Congdon 2006; Ropert-Coudert et al. 2004). Data loggers indicate that gannets are capable of diving up to 34 m, but they typically feed within 10–20 m of the surface (Adams and Walter 1993; Brierley and Fernandes 2001; Ropert-Coudert et al. 2004). Similarly, shearwaters are capable of diving up to 35–66 m, but they also typically feed within 10–20 m of the surface (Burger 2001). In Admiralty Bay, mean water clarity was 10 m. Thus, prey deeper than 10 m was probably not visible to gannets and shearwaters unless they were using feeding dolphins as cues to locate prey.

Dolphins also increased prey accessibility for gannets by decreasing prey mobility. Underwater observations of prey balls indicated that gannets increased prey mobility after feeding, but not during dolphin feeding bouts. Thus, prey balls with feeding dolphins and gannets were more likely to be stationary than prey balls with only feeding gannets. Because schooling fishes typically increase their swimming speed when they are being attacked (Wilson and Dill 2002), the decrease in prey mobility during increased predation was probably due to herding behaviors by dolphins.

Dolphins did not change the shape of prey balls in terms of width : height ratio during feeding bouts. Overall shape of prey balls was probably determined by diverse factors including depth or size of prey ball; presence of seabirds, fur seals, or sharks; and many others.

Other factors also might influence prey accessibility for gannets, shearwaters, terns, gulls, and shags. Dolphins may have made it easier for these seabirds to capture prey by increasing compaction of prey balls, similar to the effect of killer whales on herring prey balls in Norway (Domenici et al. 2000). Video resolution in this study was not sufficiently clear for an analysis of the density of prey balls (e.g., via a nearest neighbor analysis). Nonetheless, prey balls appeared to be compacted during dolphin feeding bouts. Scraps of fish in the water during dolphin feeding bouts also increased food accessibility for the above seabirds (Pitman and Ballance 1992), because they were observed eating these scraps.

Dolphins also made it easier for gannets, shearwaters, terns, gulls, and shags to initially locate prey. Prey is often sparse and patchily distributed in marine environments (Ballance et al. 2006; Wells et al. 1999). Echolocation allows delphinids to be efficient at locating prey (Norris 1969). Thus, it may be more efficient for the above seabirds to locate prey using cues from feeding delphinids than to locate prey on their own. Indeed, seabirds sometimes appear to use delphinids to locate prey (Harrison et al. 1991; McFadden 2003; Würsig and Würsig 1980). Numbers of gannets, shearwaters, and gulls increased during the first 2 min of dolphin feeding, even when other seabirds were not present. Additionally, shearwaters followed dolphins before dolphins started feeding. While following dolphins, shearwaters flew to locations where dolphins surfaced or leaped, periodically sticking their heads underwater, apparently to see if any fishes were present. Shearwaters then began diving underwater when dolphin feeding commenced.

Feeding in aggregate with dolphins appears to be most important to gannets and shearwaters, because they fed with dolphins for 25–50% of feeding observations. Additionally, mean group size for gannets was greater when they fed with versus without dolphins. This suggests that gannets feed for longer periods when they feed with dolphins, allowing more individuals to aggregate over time. Foraging benefits may be particularly important during the August–November breeding season, when dolphins most commonly herd prey to the surface (Vaughn et al. 2007).

In summary, dolphins increased prey accessibility for gannets, shearwaters, terns, gulls, and shags by decreasing depth and mobility of prey, and the dolphins made it easier for these seabirds to locate prey. The frequency with which gannets and shearwaters fed with dolphins suggests that feeding in aggregate with dolphins is important to these apex predators.

Acknowledgments

We thank H. and C. Pearson, J. Weir, N. Duprey, H. Amin, S. Tirapelle, and Earthwatch volunteers for feedback, assistance, and technical expertise during this research. Additionally, D., L., and A. Boulton, R. and V. McDonald, B. Lloyd, and the New Zealand Department of Conservation provided logistical support in the field. L. Griffing and C. Marshall contributed many helpful ideas during the development of this study, and regarding analyses. C. Marshall, A. Dahood, H. Pearson, T. Reuland, and 2 anonymous reviewers provided valuable comments that much improved this manuscript, and A. Whitley assisted with analyses. Funding for this research was provided by the Marlborough District Council, New Zealand Department of Conservation, National Geographic, Earthwatch, Texas A&M University Departments of Wildlife and Fisheries Sciences and Marine Biology, Erma Lee and Luke Mooney Graduate Student Travel Grants, and Melany Würsig. Underwater camcorder equipment was purchased with a grant from Texas A&M University Department of Marine Biology, with much help from G. Rowe, S. Arms, and J. Maxwell. Background map of New Zealand is courtesy of Eagle Technologies, Wellington, New Zealand.

Literature Cited

Acevedo-Gutiérrez
A
.
2002a
.
Interactions between marine predators: dolphin food intake is related to number of sharks
.
Marine Ecology Progress Series
 
240
:
267
271
.
Acevedo-Gutiérrez
A
.
2002b
.
Group behavior
. Pp.
537
544
in
Encyclopedia of marine mammals
  (
Perrin
W. F.
Würsig
B.
Thewissen
J. G. M.
, eds.).
Academic Press
,
San Diego, California
.
Adams
N. J.
Walter
C. B.
.
1993
.
Maximum diving depths of Cape gannets
.
Condor
 
95
:
734
736
.
Altmann
J.
1974
.
Observational study of behavior: sampling methods
.
Behaviour
 
49
:
227
267
.
Au
D. W.
1991
.
Polyspecific nature of tuna schools: shark, dolphin, and seabird associates
.
Fishery Bulletin
 
89
:
343
354
.
Ballance
L. T.
Pitman
R. L.
Fiedler
P. C.
.
2006
.
Oceanographic influences on seabirds and cetaceans of the eastern tropical Pacific: a review
.
Progress in Oceanography
 
69
:
360
390
.
Bearzi
M.
2006
.
California sea lions use dolphins to locate food
.
Journal of Mammalogy
 
87
:
606
617
.
Bertram
B. C. R.
1978
.
Living in groups: predators and prey
. Pp.
64
96
in
Behavioral ecology: an evolutionary approach
  (
Krebs
J. R.
Davies
N. B.
, eds.).
Blackwell Scientific Publications
,
Oxford, United Kingdom
.
Bräger
S.
.
1998
.
Feeding associations between white-fronted terns and Hector's dolphins in New Zealand
.
Condor
 
100
:
560
562
.
Brierley
A. S.
Fernandes
P. J.
.
2001
.
Diving depths of northern gannets: acoustic observations of Sula bassana from an autonomous underwater vehicle
.
Auk
 
118
:
529
534
.
Burger
A. E.
2001
.
Diving depths of shearwaters
.
Auk
 
118
:
755
759
.
Cipriano
F. W.
1992
.
Behavior and occurrence patterns, feeding ecology and life history of dusky dolphins (Lagenorhynchus obscurus) off Kaikoura, New Zealand
.
Ph.D. dissertation
 ,
University of Arizona
,
Tucson
.
Domenici
P.
Batty
R. S.
Similä
T.
.
2000
.
Spacing of wild schooling herring while encircled by killer whales
.
Journal of Fish Biology
 
57
:
831
836
.
Duffy
C. A. J.
Brown
D. A.
.
1994
.
Recent observations of marine mammals and a leatherback turtle (Dermochelys coriacea) in the Marlborough Sounds, New Zealand, 1981–1990
 .
Department of Conservation
,
Nelson, New Zealand
,
Occasional Publication 9
:
1
52
.
Evans
P. G. H.
1982
.
Associations between seabirds and cetaceans: a review
.
Mammal Review
 
12
:
187
206
.
Gannon
W. L.
Sikes
R. S.
the Animal Care,Use Committee of the American Society of Mammalogists
.
2007
.
Guidelines of the American Society of Mammalogists for the use of wild mammals in research
.
Journal of Mammalogy
 
88
:
809
823
.
Harrison
N. M.
Whitehouse
M. J.
Heinemann
D.
Prince
P. A.
Hunt
G. L.
Jr.
Veit
R. R.
.
1991
.
Observations of multispecies seabird flocks around South Georgia
.
Auk
 
108
:
801
810
.
Heather
B. D.
Robertson
H. A.
.
2005
.
Field guide to the birds of New Zealand
 .
Viking Press
,
Auckland, New Zealand
.
Heithaus
M. R.
Dill
L. M.
.
2002
.
Feeding strategies and tactics
. Pp.
412
422
in
Encyclopedia of marine mammals
  (
Perrin
W. F.
Würsig
B.
Thewissen
J. G. M.
, eds.).
Academic Press
,
San Diego, California
.
Litvak
M. K.
1993
.
Response of shoaling fish to the threat of aerial predation
.
Environmental Biology of Fishes
 
36
:
183
192
.
Lukoschek
V.
McCormick
M. I.
.
2002
.
A review of multispecies foraging associations in fishes and their ecological significance
.
Proceedings of the 9th International Coral Reef Symposium
 
1
:
467
474
.
Mann
D. A.
Lu
Z.
Hastings
M. C.
Popper
A. N.
.
1998
.
Detection of ultrasonic tones and simulated dolphin echolocation click by a teleost fish, the American shad (Alosa sapidissima)
.
Journal of the Acoustical Society of America
 
104
:
562
568
.
Mann
D. A.
Lu
Z.
Popper
A. N.
.
1997
.
A clupeid fish can detect ultrasound
.
Nature
 
389
:
341
.
Mann
J.
1999
.
Behavioral sampling methods for cetaceans: a review and critique
.
Marine Mammal Science
 
15
:
102
122
.
Markowitz
T. M.
2004
.
Social organization of the New Zealand dusky dolphin
.
Ph.D. dissertation
 ,
Texas A&M University
,
College Station
.
Markowitz
T. M.
Harlin
A. D.
Würsig
B.
McFadden
C. J.
.
2004
.
Dusky dolphin foraging habitat: overlap with aquaculture in New Zealand
.
Aquatic Conservation: Marine and Freshwater Ecosystems
 
14
:
133
149
.
Martin
A. R.
1986
.
Feeding associations between dolphins and shearwaters around the Azores Islands
.
Canadian Journal of Zoology
 
64
:
1372
1374
.
McFadden
C. J.
2003
.
Behavioral flexibility of feeding dusky dolphins (Lagenorhynchus obscurus) in Admiralty Bay, New Zealand
.
M.S. thesis
 ,
Texas A&M University
,
College Station
.
Norris
K. S.
1969
.
The echolocation of marine mammals
. Pp.
391
423
in
The biology of marine mammals
  (
Anderson
T. H.
, ed.).
Academic Press
,
New York
.
Norris
K. S.
Schilt
C. R.
.
1988
.
Cooperative societies in three-dimensional space: on the origin of aggregations, flocks, and schools, with special reference to dolphins and fish
.
Ethology and Sociobiology
 
9
:
149
179
.
Page
B.
McKenzie
J.
Sumner
M. D.
Coyne
M.
Goldsworthy
S. D.
.
2006
.
Spatial segregation of foraging habitats among New Zealand fur seals
.
Marine Ecology Progress Series
 
323
:
263
279
.
Peck
D. R.
Congdon
B. C.
.
2006
.
Sex-specific chick provisioning and diving behaviour in the wedge-tailed shearwater Puffinus pacificus
.
Journal of Avian Biology
 
37
:
245
251
.
Pierotti
R.
1988
.
Associations between marine birds and mammals in the northwest Atlantic Ocean
. Pp.
31
58
in
Seabirds and other marine vertebrates: competition, predation, and other interactions
  (
Burger
J.
, ed.).
Columbia University Press
,
New York
.
Pitman
R. L.
Ballance
L. T.
.
1992
.
Parkinson's petrel distribution and foraging ecology in the eastern Pacific: aspects of an exclusive feeding relationship with dolphins
.
Condor
 
94
:
825
835
.
Ridoux
V.
1987
.
Feeding association between seabirds and killer whales, Orcinus orea, around subantarctic Crozet Islands
.
Canadian Journal of Zoology
 
65
:
2113
2115
.
Ropert-Coudert
Y.
Grémillet
D.
Ryan
P.
Kato
A.
Naito
Y.
Maho
Y. L.
.
2004
.
Between air and water: the plunge dive of the Cape gannet Morus capensis
.
Ibis
 
146
:
281
290
.
Similä
T.
Ugarte
F.
.
1993
.
Surface and underwater observations of cooperatively feeding killer whales in northern Norway
.
Canadian Journal of Zoology
 
71
:
1494
1499
.
Smolker
R. A.
Richard
A. F.
Connor
R. C.
Pepper
J. W.
.
1992
.
Sex differences in patterns of association among Indian Ocean bottlenose dolphins
.
Behaviour
 
123
:
38
69
.
Stensland
E.
Angerbjörn
A.
Berggren
P.
.
2003
.
Mixed species groups in mammals
.
Mammal Review
 
33
:
205
223
.
Van Waerebeek
K.
Würsig
B.
.
2002
.
Pacific white-sided dolphin and dusky dolphin: Lagenorhynchus obliquidens and L. obscurus
. Pp.
859
861
in
Encyclopedia of marine mammals
  (
Perrin
W. F.
Würsig
B.
Thewissen
J. G. M.
, eds.).
Academic Press
,
San Diego, California
.
Vaughn
R. L.
Shelton
D. E.
Timm
L. L.
Watson
L. A.
Würsig
B.
.
2007
.
Dusky dolphin (Lagenorhynchus obscurus) feeding tactics and multi-species associations
.
New Zealand Journal of Marine and Freshwater Research
 
41
:
391
400
.
Wells
R. S.
Boness
D. J.
Rathbun
G. B.
.
1999
.
Behavior
. Pp.
324
422
in
Biology of marine mammals
  (
Reynolds
J. E.
III
Rommel
S. A.
, eds.).
Smithsonian Institution
,
Washington, D.C.
Wilson
B.
Dill
L. M.
.
2002
.
Pacific herring respond to simulated odontocete echolocation sounds
.
Canadian Journal of Fisheries and Aquatic Sciences
 
59
:
542
553
.
Würsig
B.
1986
.
Delphinid foraging strategies. 1986
. Pp.
347
359
in
Dolphin cognition and behavior: a comparative approach
  (
Schusterman
R. J.
Thomas
J. A.
Wood
F. G.
, eds.).
Lawrence Erlbaum Associates
,
Hillsdale, New Jersey
.
Würsig
B.
Würsig
M.
.
1980
.
Behavior and ecology of the dusky dolphin, Lagenorhynchus obscurus, in the south Atlantic
.
Fishery Bulletin
 
77
:
871
890
.

Author notes

Present address of DSS: University of Arizona, Department of Ecology and Evolutionary Biology, P.O. Box 210088, Tucson, AZ 85721, USA
Present address of LAW: Institute of Antarctic and South Ocean Studies, University of Tasmania, Hobart Campus, Centenary Building Room 204, Private Bag 77, Hobart, Tasmania 7001, Australia
Associate Editor was Gerardo Ceballos.