480 
THE WILSON JOURNAL OF ORNITHOLOGY • VoL 123, No. 3. September 2011 
b) reported the variable clutch size of Mandarin 
Duck was a function of CBP. and provided 
information about egg-laying, clutch size, and 
incubation for the introduced population in 
Britain. The literature on Mandarin Ducks pro¬ 
vides information on breeding dates, descriptions 
of eggs, clutch size, nest-site, and parental care 
(Clawson et al. 1979, Shurtleff and Savage 1996. 
Li et al. 2009, Jin et al. 2010). However, no 
studies of CBP and updated nesting biology of the 
Mandarin Duck are available for native popula¬ 
tions. 
The objectives of our study were to: (I) provide 
information on aspects of CBP and nesting 
biology of the Mandarin Duck, (2) identify 
conditions which facilitate occurrence of CBP, 
and (3) provide a better understanding of the 
behavioral traits and their effects on productivity 
of the species. 
METHODS 
Study Area .—We studied Mandarin Ducks for 
six breeding seasons (2004-2009) at Zuojia 
Nature Reserve in northeast China. The Zuojia 
Nature Reserve ranges from the eastern Chang 
Bai Mountains to the western plain (126' 1'-] 27 
2' N, 44 6'-45 5' E), and ranges from 200 to 
530 m asl. The region is subject to an eastern 
monsoon climate, characterized by hot, dry 
summers and cold, snowy winters. The forest 
type within the reserve is secondary growth: trees 
were ~40 to 50 years of age. 
Field Procedures .—No nests of Mandarin 
Ducks were found in Zuojia prior to this study 
although the species was recorded as a summer 
visitor. We installed nest boxes in the study area 
in 2004 and Mandarin Ducks began to use them. 
Nest boxes were constructed from rough-cut 
boards and assembled with exterior nails or deck 
screws. The internal dimensions of all boxes were: 
50 cm deep with a 35 X 35-cnr floor and a 12.5- 
or 15-cm diameter entrance hole near the top. The 
roof of each nest box was designed to be 
waterproof and the bottom was designed to permit 
drainage. The roof and body of each nest box 
were placed together with hinges, and a hook and 
ring were used to prevent the roof from being 
opened unintentionally. A slat of wood was nailed 
to the back of every nest box to provide a ledge 
for nest placement. Nest boxes were attached to 
trees with four spikes and a section of wire. Nest 
~ 8 “ 13 m ubove ground with 
50m be ‘ween adjacent boxes. We replaced 
nest boxes when they were destroyed or unusable 
Numbers of boxes monitored were 52, 56.61.6S. 
77. and 80 for the years from 2004 to 2009. 
respectively. 
We checked nest boxes at intervals of 5-7 days 
while attempting to first locate Mandarin Duck 
eggs in boxes, and then at 1-2 day intervals 
during the remaining egg laying period. Eggs 
were numbered consecutively with indelible 
marker to monitor egg losses as well as newly 
deposited eggs. We prolonged the visiting interval 
to 6-7 days during incubation to avoid female 
disturbance and. on days of expected hatching, we 
shortened the visiting interval to 1-2 days. We 
minimized disturbance caused by researchers and 
did not capture incubating females to mark them 
in the study. We designated nests in which two or 
more eggs were added within a day or clutch size 
more than 12 eggs (using the published clutch size 
of the species as criteria) as parasitized (Semel 
and Sherman 1992, Lyon 2003). Nests with ^12 
eggs were considered un-parasitized. We estimat¬ 
ed clutch initiation dates of nests that contained 
several eggs from the earliest known point in the 
nesting sequence. We back-dated the nest assum¬ 
ing a laying rate of one egg per day for nests first 
located with fewer eggs than the days elapsed 
since the previous nest box check (when the box 
was still empty). We assumed the nest was 
parasitized and was initiated on the first day after 
boxes were last known to be empty for nests first 
located with more eggs than days elapsed since 
the previous nest box check. 
A nest was classified as successful if nestlings 
were observed, or egg shell fragments w'ere found 
in box on day of expected hatching. We classified 
a nest as abandoned if it contained undamaged 
eggs and adults were absent for >5 days. We 
collected unhatched eggs for analysis to ascertain 
tertilization. We recorded nesting chronology for 
all active nest boxes and attempted to identify the 
causes ol failed nests using a combination of clues 
whenever possible. 
Statistical Analysis. —We defined incomplete 
clutches as those that did not enter the incubation 
stage, and a final clutch size could not be 
ascertained. We defined ‘combined clutch size’ 
as the number of eggs laid into any single nest box 
as a completed clutch: ‘clutch size of un¬ 
parasitized implied that all eggs were laid by a 
single female, and ‘clutch size of parasitized' 
implied that eggs in a completed clutch were laid 
by two or more females. We assumed clutches 
