Many birds with colored dorsal plumage have white underparts and among waterbirds extensive white plumage is common. Feathers are made of keratin, which is naturally transparent and colorless. They appear white because keratin foam or fibers scatter all frequencies of the visible spectrum. Colored feathers have keratin structures that reflect light selectively or mechanisms for depositing pigment within feathers. Color is not essential for flight, streamlining, insulation or waterproofing; some of the largest birds fly and breed successfully with all-white plumage. Adaptive scenarios concerned with camouflage, communication and thermoregulation have been based upon the visual properties of white feathers. These are discussed, but few examples are convincing enough to falsify the hypothesis that feathers may be white by default. Arctic land birds provide evidence for cryptic white plumage, but some white feathers are retained after they cease to be cryptic. White plumage also occurs in some tropical land birds.

We surveyed the shoreline of a large terminal lake in southern Saskatchewan (Old Wives Lake) for sick or dead birds and estimated shorebird numbers during avian botulism outbreaks. Nearly double the number of shorebirds were present in 1999 compared with 1998. The most commonly observed shorebirds in both years were the Short-billed Dowitcher (Limnodromus griseus) and Long-billed Dowitcher (L. scolopaceus). Water levels were monitored, revealing a greater drawdown effect in 1998. Small species (e.g. Least Sandpiper (Calidris minutilla), Semipalmated Sandpiper (C. pusilla), and Sanderling (C. alba)) were much more common in 1998. Much of the data that exist on shorebird mortality during botulism kills is in the form of unpublished annual agency clean-up reports. These reports suggest that the American Avocet (Recurvirostra americana), yellowlegs (Tringa melanoleuca and T. flavipes), and the small Calidris sandpipers are the most commonly affected species. We suggest some biases inherent in these cleanup efforts that may unintentionally underestimate mortality due to botulism in some species of shorebirds. At Old Wives Lake, the most frequently affected shorebird was the Semipalmated Sandpiper (14) in 1998, and the American Avocet (5) in 1999. This study suggests that feeding habits may influence a species’ susceptibility to botulism, namely those species that feed near the surface of the substrate are more prone to ingesting the botulinal toxin than those species which probe deeply in the sediment. There is also evidence suggesting that the intensity of the avian botulism outbreak affecting waterfowl may dictate the intensity of the shorebird botulism die-off.

Social foraging in waterbirds in Ghanaian coastal lagoons was studied during October and November 1994. Two types of foraging were social: directionally synchronized flocks (often involving distinctive feeding methods used in unison) and dense pecking aggregations. Social flocks were typically larger and denser than non-social flocks, and primarily involved piscivorous birds, especially the longer-legged shorebirds and egrets. It is probable that the flocks concentrate fish into temporarily high densities through herding or confusing escape reactions. We suggest that these flocks, rather than simply forming in areas of existing high prey density, may actively enhance their foraging success. There may be a true social element to either the initiation or persistence of waterbird social foraging flocks in coastal Ghana.

The reproductive ecology of the Least Tern (Sterna antillarum) was examined at four nesting colonies along the Lower Mississippi River during the 1995-1997 breeding seasons. Nest success, hatching success, and reproductive success were calculated in order to estimate productivity and evaluate reproductive status. Nesting colonies varied in size from 172 to 550 nests, with average clutch sizes ranging from 1.9-2.7 eggs/nest. Both nest success, the proportion of nests from which at minimum one egg hatched, and hatching success, the proportion of eggs that hatched, were significantly higher in 1995 (97% and 94% respectively) as compared to 1996 and 1997 (∼40%). Reproductive success averaged between 0.28-1.27 fledglings/clutch. Predation and untimely flooding were the largest negative impacts upon successful tern reproduction.

During 1987-2000, the numbers of Swift Terns (Sterna bergii) breeding in South Africa’s Western Cape varied between 1,449 and 5,668 pairs, distributed at 13 localities. Numbers were significantly related to the combined biomass of Anchovy (Engraulis capensis) and Sardine (Sardinops sagax), two of the main food items. In years when food was scarce, numbers breeding decreased by up to 67%. The sizes of breeding aggregations of Swift Terns were significantly related to the combined biomass of Anchovy and Sardine. Numbers of Swift Terns breeding at a locality ranged from 1-4,192 pairs, with an average of 750. Numbers breeding in discrete colonies ranged from 1-3,000 pairs, with an average of 513. Swift Terns showed low fidelity to breeding areas, 61% of which were used just once. This is attributed to their requirement to breed near to fish prey, and its variable distribution from year to year. Breeding by other seabirds probably often indicated safe sites for nesting. Swift Terns bred by themselves in 17%, with Hartlaub’s Gulls (Larus hartlaubii) in 77%, and with Crowned Cormorants (Phalacrocorax coronatus) in 18% of breeding attempts.

Wintering roosts of Great Cormorants (Phalacrocorax carbo) on Lake Como, northern Italy, are often inaccessible. We describe an easy-to-build, cheap and efficient technique to collect pellets at inaccessible roosts, which enabled us to study the diet of cormorants without causing disturbance, or killing birds. A floating platform was built using iron bars, sea-plywood panels and plastic tanks which was placed on the lake’s surface, to collect pellets ejected by cormorants using a roost on the eastern shore of the western basin of Lake Como. During two winter periods (February 2001-March 2001 and December 2001-March 2002) 106 samples were retrieved from the platform, measuring 17.5 m².

We assessed reliability of plumage as an indicator of age class in the Surf Scoter (Melanitta perspicillata) for demographic and behavioral studies. Three age classes were distinguished among male Surf Scoters, based on the degree of concordance between plumage characteristics and known age-related features (bursal depth and tail feather notching). Males in their first year (1Y) were distinguishable from older males (>1Y) with nearly total accuracy. Discriminating between second year (2Y) and after second year (>2Y) males had an error rate of 11%. Female Surf Scoters could not be reliably aged using plumage characteristics. Field observations suggested the timing of feather changes is an important variable affecting accurate age class determination. First year male Surf Scoter plumage is brown and female-like at the time of fledging, and gradually becomes more adult male-like during the first year. Observations of plumage changes throughout the annual cycle on wild birds suggested that females and 1Y males may be confused in early autumn, and that 1Y males and 2Y males may be misidentified during late spring and summer. Further, variation in timing and speed of pre- and post-breeding molt among 2Y and older males is uncertain. Therefore, mid-January until the end of March is the period when age class determinations based on plumage are most reliable.

We studied nightly movement rates, behavior, flight directions, and flight altitudes of Hawaiian Petrel (Pterodroma sandwichensis) and possibly Newell’s Shearwater (Puffinus auricularis) at 15 sites on Maui, Hawaiian Islands, in June 2001. We observed no Newell’s Shearwaters visually, but saw Hawaiian Petrels at all of the nine sites around eastern Maui and at two of the six sites around western Maui. Mean nightly movement rates on radar generally were higher in eastern Maui than western Maui, although movement rates in the northeastern part of western Maui were comparable to those at several sites in eastern Maui. The highest movement rates occurred at the Ke’anae Valley, Mokuia Point, Nu’u Bay, and Kaupo sites, all of which are located on the northeastern or southeastern slopes of Haleakala in eastern Maui. Hawaiian Petrels (identified visually) flew inland primarily between 10 min after sunset and the point of complete darkness (about 30 min after sunset), with no movements observed beyond 50 min after sunset. Radar movements peaked just before the point of complete darkness, but a substantial number of radar targets also flew inland at 30-50 min after sunset, with some movement occurring even after that period. These later movements suggest that small numbers of Newell’s Shearwaters are flying inland at several locations. Further, there was a higher proportion of these late flights in western Maui than in eastern Maui, suggesting that Newell’s Shearwaters, if present, form a higher proportion of the two species in western Maui than in eastern Maui. Mean flight directions at all sites were in an inland direction, with little variation in flight directions at most sites. Mean flight altitudes of the Hawaiian Petrel were 190 m above ground level, similar to those at Kaua’i Island. The data suggest that the number of Hawaiian Petrels on Maui may be higher than the current estimate of 1,800 birds.

We evaluate sexual adult size dimorphism and provide a sex-discriminating function for the Balearic Shearwater (Puffinus mauretanicus). Blood samples taken from adults were used to sex birds by the amplification of the CHD gene. Linear discriminant analysis was applied to eight morphometric characters to determine whether any single variable or combination of them could provide reliable sex determination. Females were significantly smaller than males for most tested variables. Analysis indicated that head plus bill length and minimum bill depth were the most accurate variables in a discriminant function model, predicting sex with about 90% accuracy.

A voluntary waterfowl avoidance area (VWAA) was established on Lake Onalaska in Navigation Pool 7 of the Upper Mississippi River, Wisconsin, USA, in 1986, to reduce boating disturbance to migratory waterfowl. We monitored boater compliance with the VWAA program in 1993 and 1997. Of 1,664 “boating events” observed on Lake Onalaska, boats intruded into the VWAA on 127 occasions. Boating events have increased from 1.82 boating events/h in 1986-88 to 1.97 in 1993 and 2.58 in 1997. Despite a 60% increase in boating traffic, the lake-wide disturbance rates in 1997 were comparable to that in 1981. We attribute this to a significant reduction in the proportion of lake-wide boating events that resulted in disturbance, a direct consequence of the VWAA program. Rate of intrusion into the VWAA was 0.11 per boating event in 1997 compared to 0.18 per boating event in 1986-88. Boating disturbances to waterfowl within the VWAA occurred at about half the rate (0.24 to 0.28 disturbances·hr^-1) observed prior to establishment of the program (0.48 disturbances·hr^-1). We also identified access points used by boaters and boating activities that were most likely to result in intrusion into the VWAA and associated disturbance to waterfowl. Results of these analyses have provided useful information to resource managers for targeting public education efforts. The VWAA program has contributed to the value of Lake Onalaska as a waterfowl refuge and demonstrates an effective collaboration among government agencies and non-governmental organizations.

We document timing and location of wing molt in the Horned Grebe (Podiceps auritus), Red-necked Grebe (Podiceps grisegena), and Western Grebe (Aechmophorus occidentalis). Horned Grebes left breeding ponds in late May to August and were observed on large ponds and lakes (near breeding locations) where they replaced remiges before progressing to wintering areas in September and October. Red-necked Grebes moved to much larger bodies of water such as the Great Lakes and the coast following breeding and prior to molt. On the Pacific Coast, Boundary Bay, British Columbia was identified as a major molt site. Freshwater molting areas were identified around Manitoulin Island in northern Lake Huron. On the Atlantic coast, molt sites were located in the Gulf of St. Lawrence. Western Grebes were found in wing molt at both large freshwater lakes and coastal locations (including Boundary Bay, British Columbia). Like the Great Crested Grebe (P. cristatus) and Eared Grebe (P. nigricollis), Horned, Red-necked and Western Grebes move to special molt locations following breeding.

The Red-necked Grebe (Podiceps grisegena) is a species in which breeding pairs use both overt aggression and ritualistic behavior to defend territories for breeding and feeding. However in some areas, they also breed in colonial groups. The behavior of grebes breeding on a single large lake (2,537 ha), where some breed solitarily and others breed in colonies, was examined. Specifically, I compared the frequency and spatial distribution of behavioral interactions between pairs that had dispersed nests with those breeding in close proximity. Grebes breeding in a colony engaged in more overt behavioral interactions and spent more time conducting platform behavior during pre-nesting than their solitary counterparts. In addition, the locations of ritualized and overt interactions from the nest were greater for solitary breeders. However, grebes breeding in the colony tolerated conspecifics at closer distances which, suggests that those breeding in higher aggregations may expend more energy when dealing with conspecifics.

In 1998 and 1999, the costs of experimentally extending incubation among female Common Eiders (Somateria mollissima) were examined in relation to changes in female body mass. Clutches were switched between nests to extend incubation by an average of five days to compare with control clutches. In the four days prior to hatch, females with manipulated clutches incubated less per day, and lost more body mass than control females. The probability of nest failure was similar for control and manipulated females. Manipulated females who lost their clutches to predation or abandonment appeared to have greater body mass than those who hatched their eggs successfully. Manipulated females who were unsuccessful also took more frequent incubation recesses than those who hatched eggs successfully. The probability of return to the colony in the following year did not differ significantly between control (66%) and manipulated females (57%). These results suggest that some female Common Eiders breeding in the arctic can physically increase their incubation effort and decrease their body condition at the end of the breeding season with no major consequences to long-term survival.

The assumption that brood patches identify incubating birds is a pervasive one in avian literature, and as a result, brood patches are often used to infer breeding status. Although the developmental stages of the brood patch with specific reproductive stages in passerines have been described, this information for seabirds is not often reported. Thus, for birds whose breeding activities are not easily observed, it is difficult to confirm (1) that it is valid to assume that a bird which has some stage of brood patch is a nester or putative nester, and (2) whether specific stages of brood patch development reflect specific stages of the breeding cycle. We tested the utility of brood patch scores to infer breeding status in a non-colonial seabird, the Marbled Murrelet (Brachyramphus marmoratus), a species always captured away from the nest site. We confirmed the breeding status of murrelets with brood patches, and assessed the specific stages of brood patch development to the timing of egg-production (using a physiological analysis) and the onset of incubation (using radio telemetry). Murrelets with brood patches were not always nesters or putative nesters (58% of birds with brood patches were producing eggs, and 56% of radio-tagged birds with brood patches began incubation), and brood patch score did not predict which birds were more likely to become egg-producers or incubators. Specific brood patch stages did not always correlate with specific breeding stages (e.g., the brood patch of egg-producers ranged from absent to fully-developed). Birds with fully developed brood patches took from 3-30 days to start incubation. Brood patch development accurately depicted the average population incubation time, but we caution against using brood patches to predict the timing of an individual breeding attempt, and suggest that when possible, researchers should try to confirm breeding activities using other methods.

The impact of anthropogenically-caused salinization on wildlife is a conservation issue of global concern. Increasing salinization of inland wetlands is especially problematic for breeding and migrating birds. In the arid American west, inland wetlands are centers of avian biodiversity, providing habitat for millions of birds. We experimentally examined the effects of different salinity levels on the growth, health and behavior of American Avocets (Recurvirostra americana) chicks. Birds raised in highly saline conditions exhibited significant changes in behavior, including increased activity and decreased feeding and preening. High salinity resulted in significant weight loss of 4.4 g by day 3. Blood measures of chick health indicated that hypersaline-raised chicks suffered from dehydration (e.g., osmolality of chicks reared in brackish water was 305, compared with 381 among chicks reared in hypersaline water). Results of this study, and similar studies on ducks, geese and egrets, indicate that waterbirds in general suffer adverse effects when living in highly saline environments, without access to freshwater. We recommend that water managers ensure the maintenance of freshwater inflows to wetlands throughout the breeding season.