Foraging Behavior of American White Pelicans Pelecanus erythrorhyncos

Foraging behavior by American White Pelicans (Pelecanus erythrorhyncos was studied in the .Lahontan Basin in western Nevada. Pelicans engaged in cooperative fish herding and in kleptoparasitism upon Double Crested cormorants (Phalacrocorax auritus). Pelicans in groups of size 2 through 6 caught more fish than single birds. Mean strike number increased initially with increasing flock size but leveled off at a flock size of between 3 and 4. strike efficiency (captures/bird/strike) declined with flock size, reaching an asymptote at a flock size of 4. Analysis of the regurgitate of young birds revealed that the pelicans' primary food source consisted of Carp (Cyprinus carpio) and Tui Chub (Gila bicolor). Analysis of flocks of pelicans arriving and departing from the colony on Anaho Island revealed a peak in total arrivals and departures between 1100 and 1300 hrs. This peak appeared to be constant throughout the season although the total number of birds arriving and departing increased into July. Mean flock size increased from April to July. Thermal flocks departing and arriving at higher altitudes were generally larger than low level counterparts. The evolutionary significance of cooperative foraging is briefly discussed. ACKNOWLEDGEMENTS The studies leading to this dissertation have benefited enormously from the criticism, help, and support of a number of individuals and institutions. First of all I would like to thank the Tribal Council and members of the Pyramid Lake Paiute Tribe for allowing me to live and work at Pyramid Lake over the past seven years, for pulling me out of sand dunes when I had ventured too far off the beaten track, and for answering my many questions whenever our paths crossed. I would also like to thank Morris LeFever of the U.S. Fish and Wildlife Service for allowing me to work within the Stillwater Refuge, providing me ·with gate combinations and file access, and for dealing with me courteously and thoughtfully whenever I showed up unannounced out of the wild. I would also like to thank the staff of the Stillwater Refuge for helping me find addresses and permits and pelicans. I owe a debt to my Major Professor, Frank Heppner, who has sponsored my research over the past five years and provided me with many stimulating conversations on a wide variety of topics. I also thank the members of my Committee Ors. Jim Heltshe, Bill Krueger, and Bob Shoop, each of whom put up with my flights of fancy, provided good council in time of need, and helped me gain new perspectives on the ways of Science. I owe a special debt of gratitude to Dr. Sally Levings for provision of countless reprints, revision of countless

. At the present time the basis for claims of cooperative behavior rests on anecdotal accounts of fish herding (Goldsmith 1840, Goss 1888, Mills 1925, Cottam et al. 1942, Low et al. 1950, and although several authors (Behle 1958, Hall 1925, Woodbury 1966, Knopf and Kennedy 1980, for American White Pelicans (~ erythrorhyncos) and Din andEltringham 1974a, 1974b for Great White Pelicans (P. onocrotalus) have made reference to pelican foraging behavior, no study has been directed specifically at feeding. It has not been demonstrated that cooperation occurs, or that group feeding results in a benefit to individual pelicans.
White Pelicans are good subjects for foraging studies because they are conspicuous, relatively tame birds, found throughout central and western North America (Palmer 1962). White Pelicans give a characteristic "head toss" upon capturing prey, similar to that observed in the Brown Pelican (Pelecanus occidentalis) by Orians (1969).
It is thus possible to obtain an accurate count of prey captures. All previous ethological studies of the Anaho pelicans (Hall 1925, Marshall and Giles 1953, Woodbury 1966, Anderson 1982 have focused on pelican behavior either on or in the immediate vicinity of the colony. Knopf and Kennedy (1980) provide valuable data on pelican foraging and loafing sites in western Nevada. Their study was conducted however immediately prior to the rise and subsequent decline of water levels in the Lahontan drainage system, and several of the areas that they report as suitable pelican habitat have been significantly 2 altered in the interim.
In addition Knopf and Kennedy conducted the bulk of their observations from the air and therefore were unable to obtain precise information on actual pelican foraging patterns. The regurgitation response of pelicans is well known 6 and has been used in a number of studies (Hall 1925, Gromme 1930, Marshall and Giles 1953, Behle 1958  I constructed a number of pelican decoys using commercial White-fronted Goose (Anser albifrons) floater decoys as a base. The bill of each decoy was replaced with a beak/pouch assembly carved from blocks of styrofoam, and the entire decoy was then painted to resemble a pelican in 8 breeding plumage. In the latter portion of the season to August) I repainted the decoys' white crowns (June black to simulate birds in the post-nuptial molt (Knopf 1975 ). unmodified goose decoys served as controls for each experiment. During static tests each decoy was anchored by a 1 kg concrete weight attached to a 2-3 m monofilament tether. Decoys were deployed in groups of 1 to 7, and distributed in both crescent and haphazard patterns.
Each test lasted for 45 minutes. The closest approach by pelicans to the decoys was noted as was any activity by other birds in the vicinity of the decoys. A test was considered over if a pelican approached to within 2 m of a decoy, because at that point the "flock" might be regarded as decoys plus real pelican, thus affecting its possible attractiveness. Control and experimental decoys were alternated in successive trials to remove possible temporal effects on sociality, and equal numbers and patterns of both controls and experimentals were used in each test. A total of 30 experimental tests of anchored decoys consisting of 3-5 replicates of 1 to 7 decoys were run during the 1985 season. A positive response to a decoy group was recorded in the event of an approach by a pelican to within 2 m of a decoy. In addition I noted if pelicans executed a tight Wheeling flight low over the decoys in an apparent prelude to landing. Similar criteria are described in Barnard and 9 Thompson (1985).
The anchors were removed for herding simulations and the decoys were linked above the waterline by single monofilament lines. The outside decoys were then linked to additional lines held by assistants on either side of a slough or stream. An array of 1 to 3 decoys could be drawn through the water in linear or crescent formations.
TWO additional tests of a single moving decoy were conducted in 1986.
A series of three simulations using one, two, and three decoys connected above the waterline by 30 cm of monofilament line was conducted to test the response of fish to a herding situation. Each set of decoys was drawn through the water by assistants standing approximately 5 m on either side of a slough in the Stillwater Refuge. As a control measure observers walked along both sides of the slough at equivalent distances to those maintained during the herding simulation. In each case the responses of fish were recorded by an observer standing on the embankment above the slough.
The carcasses of 5 adult pelicans found dead at Pyramid Lake and in the Stillwater Marshes were obtained for morphological data and analysis of stomach contents.
Measurements of neck and bill length were made using dial calipers and a meter stick. The time, direction of departure or arrival, and the number of birds seen approaching or leaving the colony were recorded during one min intervals.
In addition note was made whether the birds arrived or departed below or above the observation point, which was located approximately 200 m above the surface of Pyramid Lake.
Evidence of the use of thermal vortices in soaring was recorded.
Statistical analysis was performed using the SYSTAT (Systat Inc. 1986) SAS (SAS Institute 1987) and MINITAB (Ryan et al. 1986) statistical packages. In cases where variances differed significantly and/or populations were not normally distributed appropriate transforms  were performed and tests of significance were performed on the transformed data. Levels of significance for statistical tests were obtained from Rohlf and Sokal ( 1981) .

COMPOSITION OF PREY SPECIES AND FOOD REQUIREMENTS
All studies to date agree that the major food source for White Pelicans in the Pyramid Lake region consists of the Asiatic Carp Cyprinus carpio and the Lahontan Tui Chub ~ bicolor (Table 1). Although estimates are based on food fed to birds-of-the-year it seems reasonable to suggest that this is also representative of the adults' diet.
Analysis of the regurgitate of near-fledging young pelicans on Anaho Island immediately after feeding revealed that these birds had been fed a mean of 9.9 fish  gh mid-May was concentrated in the southeastern thrOU portion of the Carson Sink, and along the lower reaches of the Humboldt River (Fig 1). 87% (n= 4762) of all pelicans observed foraging during this period were seen in the Stillwater region. Although this estimate is biased in favor of areas with easy ground-access, it is in good accord with earlier, more systematic surveys (Knopf and Kennedy 1980). Pelicans used low mud islands for loafing areas between feeding bouts. Several of these islands appeared to serve as assembly areas, both for feeding groups and for flights assembling for the return trip to Anaho Island. Groups of up to several hundred birds might gather on islets in the Stillwater Marshes before taking off together and spiralling slowly off in the direction of the colony.

FORAGING BEHAVIOR
No evidence of diving activity was recorded during the course of the study. Groups of pelicans feeding in the Stillwater Marshes and along the Humboldt River foraged almost exclusively in water less than 2 m deep.
Measurements of dead pelicans found at Stillwater revealed a mean neck plus bill length of 85.5 cm (15.2 cm s.o. n = 5). Partial submersion duri ng a vigorous strike 16 adds several centimeters to this effective length, thus White pelicans appear to be restricted to foraging on prey in the upper 1.25 m of the water column.
Group foraging behavior fell into two general forms.
In the first ( Foraging behavior by large (usually > 20) groups Fiq·eiicans in open water. Ovals represent individual 0~ ~s numbers identify the same bird throughout, Letters bir 5 C) refer to three distinct phases of the entrapment ~~~r~tion, arrows indicate direction of travel. "flock" of three decoys to steer the school into such the water that individual carp were exposed to the air shallow as they pushed past each other. These responses were not _,ed when observers simply walked parallel to the obse~" slough without decoys. small numbers of pelicans (<400 out of a colony population of over 8000) were seen in the delta of the Truck~e River and near the site of Old Popcorn Beach during the first three months of the breeding season.

20
Many of these birds appeared to be loafing on the sandbars at the mouth of the river and Knopf (pers. comm.) has suggested that this area serves as a way-station between the colonies and the foraging areas to the south.
Scattered groups totalling less than 100 pelicans in any given day were observed foraging along the north-west and eastern shores of Anaho Island and in the Truckee Delta area as early as April 2. Group foraging at Pyramid Lake did not begin in earnest until early June however, when groups of pelicans moved into areas along the South and West shore of the lake and th.e East shore of Anaho Island.
Mixed pelican and cormorant flocks were observed herding schools of chub in towards . the shelving lake shorelines, beginning their drive in water over 7 m deep.
Other flocks performed apparent encirclement maneuvers at the southeastern end of the lake in water over 15 m deep.
Pel' icans were never observed foraging along the 22 th ern half of the West side of Anaho Island. sou Ur ements with an echo sounder revealed that the island Meas steeply away to depths of over 70 m along this shore drops . making the area unsuitable as a spawning ground and rendering herding-to-shore Impractical.
Herding groups generally consisted of less than 10 birds ( Fig. 4) (Table 2). There was no evidence that a particular position within a flock affected foraging success and all members of a given flock appeared to have an equal probability of catching a fish during a given time interval. Single birds did significantly worse than members of groups of sizes 3-6 (F = 4.89 Fisher's LSD test p < 0.05). Di 'stribution of flock numbers and total bi'rds i'n Fiq. 4 · · · flocks of given size seen feeding in the Lahontan Basin during the course of the study. Solid bars indicate the number ~f flo~ks in a given size-class (left vertical axis) while hatched bars represent the total number of birds seen in that size class (right axis) .

Number of flocks of given size class
Total birds in flock size class 25 able 2. Analysis ,of pelic~n foraging success. Numbers T r to head-tosses per bird per minute in flocks of a r~f:n size. The data were analyzed under the SAS General qtvear Models procedure (SAS Institute, 1987). Analysis Lfnvariance revealed a significant difference within the ~ta set (F = 4.89, P < 0.0001). Means separation using F~sher's Least Significant Difference test revealed iqnif icant differences between values for single birds :nd flocks.consisting ~f two, three, four, five, and six birds (indicated ~y ~ ~n the table). Values for flocks of three were also significantly larger than those for flocks of two.   Pelican responses to decoys. A flyover was Tal:>leco~ded if one or more birds banked low over the decoys ret did not land. A "close approach" consisted of a bUlican landing near the decoys and/or approaching to P~thin 2 m. "Pelican" refers to modified goose decoys w ipped with bill/pouch structures and painted to ~~emble White Pelicans. "Goose" refers to unmodified qoose decoys. Large flocks (> 20 birds) feeding at Pyramid Lake in July were clearly attractors, with birds often leaving loafing areas to join in a foraging session. After dark however the focus of pelican foraging activity may shift to the creeks and sloughs along the roadways. Groups of 20 to 300 pelicans were observed aoving upstream, driving fish ahead of them until they.  Table 4. In each case flocks conformed to Heppner's (1974)  Arrivals and departures to and from the colony followed a similar pattern throughout the season (Fig. 8 a-f). A disproportionately large number of birds arrived and departed between 1100 hours and 1300 hours (chi-square = 402 33, p < 0.001 assuming an equal number of departures or arrivals during any given 2 hour period) . The total number of birds arriving and departing both overall and during the peak 1100 to 1300 hours period increased markedly between April and July (Table 5).

DISCUSSION
White Pelicans rarely dive for fish. Gunter (1958) states that despite extensive observations he never saw a dive, and cites Bent (1924) in asserting that White Pelicans seldom dive. Hall (1925) mentions seeing a pelican executing an aerial plunge on one occasion. Skinner (1917)            . summary statistics on the number of pelicans Ta observed arriving at and departing from Anaho Island during both the "peak" 1100-1300 hour period and throughout a given observation period.  (Anderson 1982) and that both parents participate. in feeding (Hall 1925) , and that adult birds require at least as much food as fledglings, an adult pelican must capture somewhere between 9 and 20 138 g fish or the equivalent per day during the last month of the breeding season.
It seemed possible that some birds might attempt to reduce the number of captures required per day by taking a few large rather than many small fish. Although the negative slope of the regression of number of fish fed/weight of individual fish is consistent with a capture minimization strategy, the small r-squared value suggests that the pelicans may be opportunists, taking as many fish as Possible during a feeding bout regardless of the size 55 of prey items. That capture of larger fish may on occasion be in vain is shown by the mummified remains of extremely large (480-500 mm length) carp found abandoned in the breeding colonies after the pelicans have departed.
The location of these carcasses·makes it extremely unlikely that they were transported by predators other than pelicans. Anaho Island is closed to . the general public, thus eliminating human fishermen as a probable source of fish remains.
Applying the lengths of the mummified fish carcasses to a regression curve based on the known weights and lengths of carp captured at the Stillwater Marshes produces estimated weights of 1690 to 2003 g for these fish. While it is hard to imagine even a large adult pelican capturing so monstrous a fish, Koonz (1981) reports that pelicans in Saskatchewan have been known to feed on whole fish and fish scraps left by sportsmen.
Bowhunting for carp is a popular sport in the Lahontan Basin and hunters discartl all but the largest fish along the edges of sloughs in the Stillwater Marshes. It is possible that a foraging pelican might come across a dead or dying carp and attempt to bring it back to the colony.
Although pelicans were frequently seen feeding near piles of abandoned carp scavenging behavior was never observed.
FORAGING LOCATIONS Vigg (1978Vigg ( , 1981  The importance of food availability in pelican colony establishment and regulation is discussed by Brown and Urban (1969), Tait et al. (1978) and Smith et al. (1984).
As the water levels within the Lahontan Basin declined to pre-flood levels large numbers of fish were trapped and concentrated in isolated pools along the edge of the Humboldt River and Carson Sinks. These fish proved relatively easy to capture, and this concentration of food provides a possible cause of pelican overwintering in the Sinks duri ng 1986-87.

57
Honey Lake, Lassen County, California has been identified as a historical foraging area for pelicans breeding at Pyramid Lake (Knopf and Kennedy 1980 Because the area has supported a pelican colony in the past Honey Lake must be regarded as an important site in any long-term planning for pelican management. The result is a "zero-sum game" in which the pre a · the prey group as a whole may b~ more susceptible to the predator than if each individual simply scattered on its own. More recently Gottmark et al. (1986)  The response of fish to pelicans feeding in the deeper waters of Pyramid Lake is something of a mystery.
water clarity in the lake is mucq high~r than that in the sinks and it would seem relatively easy for fish to dive below maximal pelican reach as soon as a feeding flock was detected.
Larqe flocks of pelicans were frequently observed foraging in water over 20 m deep from June through August.
The bulk of these birds engaged in variations of the "surround and strike" technique depicted in Fig. 3. The apparent failure of fish to avoid the pelicans by diving may be the result of some form of "Selfish Herd" behavior.
Other groups of pelicans and cormorants were seen herding fish in to shore (Fig. 2) along the south and west sides of Pyramid Lake. Chub use these areas as spawning grounds (Vigg, 1978)  Evidence for this hypothesis is incomplete at best.
Members of groups of two through six birds did catch significantly more fish than did birds foraging alone.
The small sample size available for larger groups makes it impossible to determine whether there was an eventual point of diminishing returns as group size increased further. It may be that a combination of cooperative herding and social facilitation.results in all groups of pelicans doing better than single individuals, but this cannot be stated categorically at this time.
In addition to any increase in foraging success it has been suggested by several authors (Krebs 1974, caraco et al. 1980, Rubenstein 1982) that group foraging may be a way of minimizing the variance in food intake by individuals. As shown in Fig probability of a premature strike with increased flock size.
once again the small sample sizes available for flocks of 7 or more pelicans makes it impossible to assess differences among these larger groups.

FORAGING GROUP FORMATION
The results of the decoy experiments showed that pelicans are attracted to an area by the presence of other pelicans. Although it is tempting to suggest further that smaller groups were more attractive than larger ones the small sample of tests conducted makes such a suggestion premature. The failure of the control (goose) decoys to attract any pelicans rules out site characteristics alone as an attractant.
Observed success on the part of an individual or group does not appear to be a major source of attraction. The decoys obviously catch no fish, yet Pelicans would land in their immediate vicinity, and often remain near them until approached by an observer. 63 pIRACY piracy or kleptoparasitism has been reported in a nwnber of bird species (Brockman and Barnard 1979).
Generally the kleptoparasite is a smaller, more agile bird (but see Barnard and Thompson 1985) that takes advantage of its victim's slower speed, or inability to swallow a food item rapidly. For example, members of the Pelecaniformes must surface to swallow prey items and are vulnerable to attacks by gulls (Bent 1921, Baldwin 1946Schnell et al. 1983, Carroll and Cramer 1985. Large birds rarely use their superior size and strength to obtain prey forcibly from a smaller individual. The only previously published reference to this activity was by Skinner (1917) who mentions instances of kleptoparasitism by White Pelicans on "fish ducks" feeding along the Yellowstone River.
The importance of piracy or kleptoparasitism in either the pelicans' or cormorants' biology · is probably minimal.
Although the majority of observed attacks by pelicans on cormorants resulted in either a pelican taking the fish or the fish being lost to all birds, the probability of an individual cormorant losing even one fish is very low.
A more interesting question is perhaps that given the small number of fish obtained by the pirates, why does the behavior continue at all? Pelicans at Pyramid Lake that are successful at kleptoparasitism make available a resource relatively close to the nest site (the Truckee 64 I 1 River delta is only 15 km from Anaho Island, whereas the Humboldt and Carson sinks are over 100 km away). At the same time however, kleptoparasites are investing both time and energy that could be spent in foraging for themselves.
In addition kleptoparasites run the risk of injury both from their intended victims and from other pelicans.
Throughout the season large numbers of pelicans congregate along the sand bars at the mouth of the Truckee River. Knopf (pers. comm.) has informed me that he has correlated these assemblages with patterns of bad weather over the Carson Sinks. A few pelicans also engage in apparent foraging activity along the lower stretches of the Truckee River as early as the beginning of April. It is likely that the primary reason that the pelicans are in the vicinity of foraging cormorants at the beginning of the season is that the cormorants are feeding near pelican assembly grounds. Kleptoparasitism may thus be an opportunistic response to a given situation rather than a major facet of the pelicans' life history. In any case the impact of the pelicans upon the cormorants is insufficient to make the cormorants change breeding sites or foraging areas. Hall (1940) has shown that pelicans and cormorants have bred in close proximity within the Lahontan Basin since at least the Pleistocene and cormorants and pelicans overlap in geographic distribution throughout their range.
Kleptoparasitism is not limited to the Pyramid Lake  Hall (1925) makes a brief reference to hearing sounds that he assumed were caused by pelicans feeding at Pyramid Lake "into the early hours of the night". Low et al. -(1950), working at the Great Salt Lake in Utah, state that "Most feeding activities take place at nig:t:it or early morning, although there have been notable exceptions to this." The low levels of prey capture reported here for daylight feeding suggest that. the pelicans must be doing a sizable proportion of their feeding at night. Vigg (1981)  they move from place to place in discrete units of two to several hundred birds. Because pelicans nest in large colonies and typically occupy open habitat (Bent 1924) the possibility for information exchange and following behavior between widely separated groups is high. Armstrong (1971)  Leaders may thus be affecting followers at distances greater than that supposed by a human observer. Heppner's (1974) definition of a flock when applied to other than aerodynamic characteristics is thus probably quite conservative. The failure of time series analysis to reveal any consistent pattern in arrivals and depart~res suggests that either no following behavior is occurring or the birds are cueing in on more extended visual flocks.
Previous observations suggested that pelicans flying to foraging areas near the mouth of the Truckee River or along the western shore of the lake flew low, close to the water surface. Pelicans traveling to foraging sites at a distance from the lake often soared in thermals over Anaho Island before departing at high altitudes. Similar behavior was observed at the foraging sites themselves.
Because thermal soaring and low flight represent two 68 I : discrete forms of behavior I present data for each separately.
observations of birds foraging at the mouth of the Truckee River and at the South East end of Pyramid Lake suggest that it is unlikely that birds that had initially gained altitude over the islan~ would descend to feed at the river. Thermal soaring, ·although energy efficient (Pennycuick 1972), is costly in terms of time, and birds commuting to and from the South end of the lake appeared to be taking advantage of a ground-effect similar to that observed in Skimmers (Rhynchops niger) (Withers et al. 1977) rather than soaring to a high altitude only to descend after covering a short linear distance.  Evans (1982a) and in Pelecanus onocrotalus, which also feeds at some distance from breeding colonies, by Brown and Urban (1969). This suggests the possibility that a variety·of factors may be influencing the pelicans' behavior.
One possible explanation for the observed distribution of arrivals and departures lies in the interplay between the adults' foraging behavior and the behavior of pre-fledging young. If the greatest part of the pelicans' foraging is done at night or in the early morning, departure times from the feeding areas would begin within the period 0800-1100. Ross (1933)  If young are fed at 1200 hrs there is plenty of time for the adults to return from the island to the feeding areas for afternoon and evening fishing. Guillet and Crowe (1983) report that Carp (Cyprinus carpio) move into shallow water as the water temperature increases during the day. The "evening rise" exhibited by many fish 72 I species has been known to generations of human fishermen and I suggest that the pelicans may be tailoring their commute times to take advantage of this phenomenon.
As the season progresses an increasing number of adult birds can be seen fishing in Pyramid Lake, often within 0.5 km of Anaho Island. Vigg (1978) has demonstrated that chub are most common in the upper portions of the water column from approximately 1600 to 0800 hours. Thus, although pelicans feeding at the lake are not constrained by commute time, there is still an advantage to conducting non-feeding business during the middle portion of the day.
Once the young pelicans develop a cover of protective feathers and are large enough to defend themselves from would-be predators both parents engage in feeding activity away from the colony, returning only to provision their young. The young birds wander around the island either singly or in pods of several birds often congregating near the island's shoreline which is up to 1 km from nesting areas. Feeding of young takes place on the original nest site. Adult birds feed only their own young (Hall 1925).
I observed . adults that were not greeted by a fledgling at the nest scrape depart from the island following a brief wait.
Young pelicans return to the nesting areas during the middle of the day, gathering in dense clusters in any shaded spot near the nest scrapes. Adult pelicans 73 arriving high over the island dive on the colonies at steep angles, producing a pronounced whistling tone that can be heard at some distance. As the adults begin to arrive young birds at ·a distance from the nesting areas hurry back to the colonies to receive nourishment.
If adult arrivals at the colony were randomly distributed young pelicans would be unable to disperse far from their nest sites for fear of missing a day's feed.
This daily wandering may be important for both muscle development and water balance. The young pelicans spend several hours a day running along "runways" away from the breeding areas flapping their wings in an apparent prelude to flight. Apart from the moisture in the food brought by the adults the only source of water available to the young is the lake, and dehydration in the intense desert heat is a real possibility. Adults would also benefit from having a set time of return to their offspring. Birds that had to search for chicks over the 300 hectare expanse of Anaho Island would reduce the amount of time that they had available for feeding.
The peak in arrivals between 1100 and 1300 hrs therefore may be a compromise between the need of the young birds for exercise and for water from the lake, and the need of the adults to minimize the amount of time spent at the colony and away from the foraging grounds.

THEORETICAL IMPLICATIONS AND CONCLUSIONS
The role of social behavior in foraging by birds has 74 I been the subject of a number of theoretical and experimental studies (Ward and Zahavi 1973, Krebs 1974, pulliam and Millikan 1982, Caldwell 1981, Barnard and Thompson 1985, Gotmark et al. 1986. A possible evolutionary pathway to cooperation has been proposed by Axelrod and Hamilton (1981) and discussed at some length by Axelrod (1984) and Maynard Smith (1982 New arrivals to a feeding -group are certainly competitors and there is justification in expecting that they would be resisted unless they provided some benefit to group members. Although non-resistance is consistent with a cooperative advantage from increased group size it would also be indicated in cases where the cost of resistance is greater than that of increased competition. Pelicans do occasionally dive, diving is a relatively rare phenomenon. Given that the birds are restricted to fish in the upper levels of the water column cooperative fishherding is one mechanism of ensuring access to food. Pelican flocks observed departing from Anaho Island tend to be much larger than those that eventually engage in fishing ( Fig. 4 and Table 4). Much of the bir~s' time away from the colony is spent on loafing grounds in the immediate vicinity of foraging sites, and it is here that the feeding flocks form. It is critically important that these loafing areas are preserved in any management scheme.
For much of this study increased water levels in the indicates that this overlap has been going on since at least the Pleistocene.
Refuge personnel should be encouraged to establish and maintain graded sides to impoundments as post-flood repairs continue. The pelicans' use of sloughs as fish traps can be enhanced by elevating culvert mouths slightly above the stream-bed to slow fish passage upstream.
Because pelican use of the Stillwater region is greatest at the beginning and end of the season it would be advisable to regulate water levels in a number of impoundments such that appropriate water depths for foraging are maintained. The critical period of pelican use extends from mid February to May and mid July through September.
As the flood waters continue to recede there will be a gradual reduction in available foraging habitat.
Initially we may expect this to have a positive effect on pelican numbers as schools of fish become concentrated in drying pools. Reports from the Fallon region during the Winter of 1986-1987 indicate that a number of pelicans may have over-wintered in the basin, presumably in part to take advantage of the flush of food.
An inherent danger to this concentration of food is that it will also lead to a concentration of waterfowl, and this in turn may lead to an eventual increase in mortality due to predation and disease. Newspaper accounts of a bird die-off in the Carson Sinks have suggested that avian cholera and botulism may be taking a toll of the pelican population. Prompt removal of dead and dying fish and bird carcasse~ when practical may reduce the possibilities of an epidemic.
A further source of concern engendered by the declining water-levels in the Basin is the inevitable concentration of pesticide residues, industrial wastes, and heavy metals that are the inevitable by-products of the use of much of the terminal stage of the Carson/Truckee/Humboldt watersheds as a dumping ground.
Continual monitoring of levels of these toxins is vitally important to the health of the entire Lahontan ecosystem.
Dead birds should be analyzed for the presence of pesticides, and a program of water-quality monitoring throughout the Basin should be encouraged.
As fish populations decline with the receding water we may expect a corresponding ~ecline in the number of birds breeding at Anaho. Offsetting . this assumption however is the fact that the Anaho colony appeared to be increasing in size prior to the increase in foraging habitat.
It is likely that a number of factors may be affecting the western population of White Pelicans as a whole. Close monitoring of breeding success at a number of colonies would provide much useful information as to the general trend; The third point relating directly to management programs is the importance of nocturnal feeding to overall 89 pelican foraging success. In view of the low capture rates recorded for pelicans feeding during daylight hours it seems certain that a sizable proportion of the birds total catch must come at night. Nocturnally foraging pelicans made extensive use of the creeks and sloughs in the Stillwater region, in some cases feeding near roadways that are in heavy use by humans during daylight hours. River and twice at Pyramid Lake-was "buzzed" by attack aircraft engaged in simulated strafing runs. Given the large number of pelicans flying through these areas this activity amounts to an accident waiting to happen.