General field methods

3.2.1.1. Study sites and study period

Fieldwork was conducted in a population of avocets breeding on alkaline lakes in the Kiskunság National Park (KNP) in central Hungary, from 1998 to 2000. Alkaline lakes, varying in size from 2-3 ha to 500 ha, are natural habitats of avocets. Fieldwork was also conducted in artificial habitats (fishpond and reconstructed wetland) because of their importance in avocet nesting.

Avocets do not use every site every year and fieldwork was concentrated in areas of high nesting density and good accessibility in each year. Data from six alkaline lakes and two artificial sites are included in this study. More details on habitat and study sites can be found in Chapter 2.

3.2.1.2. Model species

The Avocet is a middle-sized, pied-plumaged shorebird breeding mostly in coastal areas in Europe and near alkaline lakes to the east of the Black Sea. Its main breeding habitats are salt marshes, salt evaporating ponds on the coast and large alkaline or saline lakes inland (Cramp & Simmons

1983; Hagemeijer & Blair 1997). Avocets nest in colonies, which are often formed on small islands. The species is monogamous and both parents care for young. The chicks are fully precocial and are able to walk and feed on their own a few hours post-hatch. Details on the breeding biology of avocets are given in Chapter 2.

3.2.1.3. General fieldwork

Nests were located by searching areas used by avocets for nesting. Every nest was numbered, marked and recorded on a map. The length, breadth and mass of the eggs was measured and eggs were floated to determine incubation stage, from which the expected date of hatching was estimated (Chapter 2).

Colonies that contained hatching nests were searched for young chicks at least once per day and the chicks were considered to belong to the nest in which they were found. Chicks were measured (culmen length, tarsus length and body mass) and banded with a brood-specific combination of two plastic color-bands and a metal band above the tarso-metatarsus less than 24 hours after hatching. Hatching order, if known, was indicated by different colored tapes attached to the metal band of the chicks. Colony visits were limited to 1 hr in cold weather and 0.5 hr in warm weather. A more detailed description of field methods can be found in Chapter 2.

Broods were monitored by locating families during their movement from the colony and in the chick-rearing areas. Searches were conducted every two or three days using a telescope to find vigilant adults and identify their chicks. During observations, I used hunting blinds or a car as observation points positioned at large distances from families to minimize disturbance. The location of territories and composition of avocet broods were recorded on habitat vegetation maps.

A chick was considered dead if it was not seen on at least three consecutive observation bouts.

Fledging was defined as living to 35 days, the age when juveniles are able to fly.

3.2.1.4. Behavioral observations

Behavioral observations were conducted to estimate the time budgets of both adults and young in as many broods as was possible. Behavioral observations were made when a brood was first seen to occupy a territory in the chick-rearing areas. If more than one brood was available for

behavioral observation on any given occasion, I randomly chose the order in which observations were made. Broods were observed only once per day for either adult or chick behavior with the latter enjoying a priority because chick behavior was more difficult to observe than adult behavior.

Behavioral observations lasted for 30 minutes per family group for both adults and young.

Adult behavior was recorded continuously by measuring the time spent with each behavior using a stopwatch. Adult behavior types included self-maintenance behavior (feeding, preening, resting, walking) and parental behavior (brooding, calling to the chicks, defense against predators, leading, territorial aggression against conspecifics and non-predators, vigilance).

Chick behavior was recorded instantaneously for every chick once every minute for 30 minutes. Chick behavior included being brooded, crouching, feeding, preening, resting and vigilance. In every 5th minute, I also estimated the distance between the vigilant parent and every chick using the height of the vigilant parent (ca. 0.4 m) as a template.

3.2.1.5. Sampling of food availability

I quantified territory quality by measuring the abundance of potential food items (aquatic invertebrates: waterbugs Heteroptera: Corixidae, Notonectidae and Gerridae, dragonfly larvae Odonata: Anisoptera and Zygoptera, water beetle larvae Coleoptera: Dytiscidae, Gyrinidae, chironomid larvae Chironomidae and annelid worms Oligochaeta) on the territory of as many broods as was possible. Sampling the food base was conducted once in 1998 and twice in 1999 and

in 2000. Territories were selected for sampling only if the chicks were young (< 2 wk old;

mortality of chicks is highest during the first week after hatching) and if the brood was observed on the same territory at least twice in 4-5 days.

The sampling location was standardized in the center of each territory in water depths of 3-4 cm. I enclosed a known area of water using an open-ended plastic cylinder (diameter 0.3-45 m) open on both ends and collected all aquatic invertebrates from within the known volume of water using a sweepnet (mesh size 0.2 mm). Macroscopic (>0.5 mm) aquatic invertebrates were stored in jars, preserved in 70% ethanol, and identified in the laboratory. The abundance of invertebrates was calculated by dividing the total number of invertebrates collected in a sample by the volume of water from which it was collected.

3.2.2. Variables and statistical analyses

Egg volume (cm³) was calculated and fresh egg mass (g) was estimated according to formulae by Hoyt (1979) and parameterized with coefficients derived from actual measurements of avocet eggs.

The estimated values did not differ from measured values (paired t-test, p > 0.3).

To quantify adult quality, I used indirect measures that usually positively correlate with adult quality in shorebirds (laying date, clutch size, egg size, number of young hatched, territory quality, proportion of time spent with parental behavior). Direct measures of adult quality (e.g.

body size) were not available because adults were not captured and measured in this study.

Chick body size variables were highly correlated, and were reduced to a single variable by a principal component analysis. Body condition of chicks at hatch was estimated by the residuals from a linear regression of body mass on tarsus length. This variable is important for precocial chicks because they rely on their yolk reserves for survival during the first days after hatch (Starck

1993). To avoid pseudoreplication, I averaged chick body sizes for broods and used broods as data points in most analyses. The difference between the brood average and the body size of individual chicks was used to quantify individual chick quality.

Pseudoreplication may be present in the data combined from three years if some pairs or adults nested in more than one year. However, this was unlikely because only one of the marked adults that bred in at least one of the study years (n = 25) was observed to breed in more than one year. However, I treated the years separately in relevant analyses.

Sample sizes differ among statistical tests because complete information was not available for every brood. Data from three years were pooled. However, I controlled for yearly variation where it was appropriate (Chapter 2). Parametric tests were used only if the data met the assumptions of such tests. Homoscedasticity was tested by Fmax and Bartlett-tests and residual plots, and normality was checked using the Kolmogorov-Smirnov test (Zar 1984). If the data did not meet the requirements of parametric tests even after transformations I used non-parametric tests. Means and  1 S.D.s are reported in the text and means  1 S.E.s are shown on graphs. I used two-tailed probabilities and  = 0.05 in statistical tests.