RESPONSE OF MALLARD DUCKLINGS TO VARIATION IN HABITAT QUALITY: AN EXPERIMENT OF FOOD LIMITATION.
HANNU POYSA 
JOHAN ELMBERG 
PETRI NUMMI 
Abstract. Occurrence of Mallard (Anas platyrhynchos) pairs and broods was studied on 86 boreal lakes in two areas in south Finland during 1988--1997 (35 lakes) and 1989--1996 (51 lakes), and field experiments were conducted to determine whether food limitation is an important factor regulating population densities. In general, pairs and broods used and avoided the same lakes, but in both study areas, the proportion of lakes that were unoccupied every year was higher for broods (71% and 69%) than for pairs (26% and 31%). We hypothesized that lakes without breeding Mallards, and especially broods, were too poor to raise young in, and we tested that in field experiments on 22 lakes in boreal Sweden. Based on concentration of total phosphorus in the water, these lakes were divided into two groups: 11 "poor" lakes and 11 "rich" lakes. Mallard ducklings imprinted on humans were used to address brood-stage food limitation by studying mass change. Thirty-one individually marked ducklings were divided into three expe rimental groups: 12 ducklings were used in poor lakes, 12 ducklings in rich lakes, and 7 ducklings were used as controls receiving no experimental treatment in the field. Each group was studied in 11 daytime trials on 11 different lakes, except the control group which was kept in a pen with free access to food. Ducklings of all groups spent nights together as a big 31-individual group in the same pen and with free access to food. Ducklings foraging on poor lakes gained significantly less body mass than those feeding on rich lakes. At night, when ducklings of both groups had equal and free access to food, ducklings that had been feeding in a poor lake that day gained more mass than those that had been feeding in a rich lake. Overall, [tilde]95% of daily mass gain was accounted for by daytime gain in the control group, the corresponding percentage was 35% in the rich-lake experimental group and 11% in the poor-lake experimental group. Thus, we have observational, as well as experimental, evidence to conclude th at many boreal lakes are inferior breeding habitat for Mallards, especially during the brood stage. Our results suggest (1) that it is important to recognize at what stage of the yearly reproductive cycle food limitation may occur, (2) that brood stage may be more important than nesting stage for possible population limitation and regulation in breeding Mallards, and (3) that the reason why so many lakes in the boreal region lack breeding Mallards may be food limitation.
Key words: Anas platyrhynchos; Finland; food limitation; habitat saturation; Mallard; mass gain; population limitation; Sweden.
Food limitation may influence individual reproductive success and survival (review in Martin 1987), limit number of breeding individuals and population densities (reviews in Newton 1980, 1998, Boutin 1990), and affect species diversity through interspecific competition (review in Wiens 1989). However, as is evident in all reviews mentioned, it is difficult to demonstrate food limitation directly in the field, though both food addition experiments (Boutin 1990) and density-perturbation experiments (Harrison and Cappuccino 1995) done on a variety of animal taxa have detected resource-based bottom-up regulation.
Some of our earlier studies of assemblages of breeding dabbling ducks (Anas spp.) indicate that food limitation may affect breeding density and species diversity (Elmberg et al. 1993, Poysa et al. 1994). Introduction experiments showed that a notable proportion of our study lakes lacked ducks in some years, and that lakes remained unoccupied despite the fact that introduced ducks can attract wild birds (see Elmberg et al. 1997, Poysa et al. 1998).
Those results concern the Mallard (Anas platyrhynchos), the most common and widespread species of dabbling duck in the world. One explanation of our findings may be that the boreal forest biome, with its many lakes, is not saturated with Mallards, i.e., that number of birds returning from wintering areas is not enough to fill up all suitable breeding lakes (see also Jarvinen and Ulfstrand 1980, Elmberg et al. 1997:134). On the other hand, some lakes may produce too little food to meet dietary demands of Mallard ducklings. In other words, food resource limitation may operate in some years or at some stage of the breeding cycle, possibly influencing habitat selection of nesting pairs via their capability to raise young. Further, as indicated by differences in habitat distribution and lake use between pairs and broods of Mallards within a region (Danell and Sjoberg 1978, Pehrsson 1984, Nummi and Poysa 1993), it may be a common pattern that lakes do have breeding pairs but not broods. In such lakes, food resourc es may not be abundant enough for raising young.
Here we address the question of food resource limitation in breeding Mallard populations focusing on the brood stage. First, we use long-term data on lake occupancy by breeding pairs and broods from two study areas to show that, indeed, a notable proportion of lakes are unoccupied year after year and that this pattern is more pronounced for broods than for pairs. Second, we report on an experiment testing the hypothesis that lakes without breeding Mallards, and especially broods, are too poor in invertebrate food to raise young. By using Mallard ducklings imprinted on humans in controlled field experiments, we examined daily changes in body mass in relation to lake quality.
Observational study of lake occupancy
We collected data on lake occupancy of Mallard pairs and broods in Hame (61[degrees] 10' N, 25[degrees]05' E) and Karjala (61[degrees]30' N, 29[degrees]30' E) in south Finland (study regions number 3 and 4, respectively, in Fig. 1 in Elmberg et al. 1993). Every year from 1989 to 1996 in Hame and from 1988 to 1997 in Karjala, number of breeding pairs and broods was recorded on 51 (Hame mean size 5.8 ha, range 0.1--49.5 ha) and 35 lakes (Karjala; mean size 3.7 ha, range 0.2-24.0 ha), respectively. Pairs were censused with a standard waterfowl point-count method (Koskimies and Vaisanen 1991) and, according to recommendations about optimal census times in Poysa (1996), presence-absence in each year on a given lake was determined from a census conducted in the middle of May. Five (Hame) or three (Karjala) brood surveys were made each year between mid-June and early August using the point-count method. If a brood was observed on a given lake on any survey, the lake scored as hosting Mallard broods in that particul ar year; if no broods were seen in any census, the lake was considered to be unoccupied by broods.
We selected 22 lakes in the Umea region in north-central Sweden (63[degrees]40' N, 20[degrees] E), an area situated in the mid-boreal zone, and rich in lakes and wetlands. Experimental lakes are typical for the region, i.e., they are surrounded by mixed coniferous forest or wetlands, and they freeze over in winter. In total, 33 potential experimental lakes were selected. Primarily we looked for lakes [tilde]10 ha in size and on which Mallard pairs had been observed during previous breeding seasons, but the following selection criteria were added: (1) no cabins or other buildings or frequent human activities around the lakes that could influence breeding results, (2) within convenient distance from where the experimental birds were kept, and short distance to carry ducklings from a road to the lake, (3) tall vegetation should be sparse to minimize risk of losing experimental ducklings (thus excluding lakes at the eutrophic end of the local productivity gradient), and (4) having a well-defined shoreline. In th e region Mallard broods can be found in lakes down to 0.05--0.10 ha in size (Danell and Sjoberg 1978).
To get an objective value of the productivity level of each lake, we measured the concentration of total P in the water, a good indicator of trophic status of lakes (e.g., Kerekes et al. 1990, Staicer et al. 1994). In a closely related species, the Black Duck (Anas rubripes), total phosphorus has been found to be a good predictor of lake utilization by broods (Staicer et al. 1994). Water samples were taken in winter when lakes were frozen over and ice effectively reduced turbulence. Based on phosphorus values we selected two groups of lakes as experimental lakes, (1) the 11 "richest" (mean size 5.3 ha, median 4.1 ha, range 1.7- 12.6 ha), and (2) the 11 "poorest" (mean size 4.3 ha, median 3.0 ha, range 1.5-13.4 ha) among the 33 lakes ("rich" lakes, 19-30 [micro]g/L total-P; "poor" lakes, 8-18 [micro]g/L total-P; note that the term rich is used here only in a relative sense and not to refer to the locally most productive lakes, see above). This classification corresponds well to results from counts of wild bro ods in the same lakes in 1995 and 1996 (two brood censuses each year per lake): one brood was observed in the poor lakes and 14 in the rich, the number of lakes without broods was 10 (out of 11) among the poor and 5 (out of 11) among the rich lakes (Fisher's exact test, P = 0.032). Each day experiments were run (see below) we used one lake from each of these groups, the within-group order of usage being random.
Thirty-one Mallard ducklings from a hatchery in Sodermanland (central Sweden), hatched on 11 June 1996, were delivered to the experimental area the same day. An outdoor pen supplied with a heating bulb served as a "nest." Initially ducklings had free and direct daytime access to a fenced area along the shore of Umealven river, permitting them to search for food in water in a natural manner. They were also fed ad libitum with a diet consisting of 5% animal protein and 95% barley. To obtain effective imprinting one person spent several hours per day with the ducklings. The day before experiments started, a short test was made by releasing some ducklings along the shoreline. They reacted as expected by responding to the activities of the person following them. For example, they left the water when the person walked away from the shore at the same time calling to them by whistling.
Experiments started when ducklings were 4 d old. We randomly picked birds and put them in two baskets until there were 12 in each, i.e., about the number in a large wild-Mallard brood. One group of ducklings was assigned to always forage on rich lakes, the other on poor. The seven remaining ducklings were used as a reference group not subjected to daytime experimental treatment. All ducklings (including the reference group) were individually color-marked so that we could tell them apart from a distance even when they were swimming. This made it possible to identify individual birds in the field experiment. During the experimental period, food was taken away from the pen at 0800 each morning, and the two groups of experimental ducklings were collected. After that, reference ducklings got access to food again. When experimental birds were brought back to the pen in the afternoon (normally around 1700), they too had free access to food. In other words, all 31 ducklings spent roughly two-thirds of each day as on e big "super brood," and 2 x 12 ducklings were taken out on experiments and were absent for 8-9 h each day. Experiments were run between 15 and 27 June with two interruptions (one rain day and one holiday; all ducklings kept in the pen the whole day).
Upon arrival to an experimental lake, each duckling was weighed (nearest milligram) on a battery-driven balance. Thereafter they were released as a group on the shoreline, and the experimental trial started. Our aim was to start trials at 1000, but some variation due to different travel time to lakes was unavoidable. Each experimental trial lasted 6 h, after which ducklings were collected as quickly as possible and then again weighed individually, transported to the pen and the super brood was restored.
Two persons accompanied ducklings, but remained on the shore, while the latter searched for food along the shoreline and in aquatic vegetation (observers followed ducklings during the experiment, ducklings did not follow them). During each experimental period, all broods traveled around the lake at least once, so that all sections of shoreline were used for food searching. After [tilde] 1 h of foraging, ducklings normally selected a place along the shore to rest together. Roughly half an hour later they usually started looking for food again. This behavior was surprisingly independent of weather and not induced in any way by the observer.
Our experiment had a repeated-measures design with 11 trials for each of the three (body mass) groups of ducklings. Repeated measures were obtained for changes in body mass for each duckling from each experimental trial ("day-time change" is the mass after a trial minus the mass before) from (1) 11 poor lakes (12 ducklings in each), 11 rich lakes (12 ducklings in each), and 11 controls (7 ducklings kept in the pen; timing and weighing procedure as in experimental ducklings), and (2) the night following each of eleven trials ("night-time change" is the body mass before the trial in a given day minus mass after the trial the day before). We used absolute mass change in the analyses because: (1) duckling body mass did not differ among the three groups when the experiments started (ANOVA, F = 1.096, df = 2, 28, P = 0.348), (2) all ducklings were the same age, and (3) parallel trials were run on the same days in all three groups. Since individuals within each treatment group did not belong to the same natal brood, but were randomly selected from a large group of ducklings, and because they showed reasonable between-individual variation in terms of foraging place and body-mass change in each trial, we consider individuals within treatment groups as independent observations. Repeated-measures analysis of variance had one grouping factor (three levels: control, poor lakes, rich lakes) and one 11-level, within-subject factor (trial). This was done separately for daytime and nighttime changes in body mass. After a linear transformation of the mass changes (i.e., because some values were negative, the origin was translated; see Zar 1996:27), data were log(y + 1) transformed to improve normality and homoscedasticity. As mass changes in the control group appeared to differ drastically from those of the experimental groups (see Results), we first tested differences between the latter excluding the control group. We then ran analyses with all groups, and the only difference was that results were even more significant. Adopting a conservative approach we thus give results only with the experimental groups included.
All statistical analyses were run with SYSTAT procedures (Wilkinson 1992). When interpreting the output of repeated analysis of variance we followed instructions given in Wilkinson (1992:307-308, 336-339). All probability values are two-tailed, according to SYSTAT procedures or Zar (1996), as appropriate.
Observational study of lake occupancy
In both study areas, frequency of unoccupied lakes was high for pairs as well as for broods (Fig. 1). In general, pairs and broods used and avoided the same lakes; their lake-specific occurrence frequencies were correlated (Spearman rank correlations, Karjala region, [r.sub.s] = 0.707, N = 35, P [less than] 0.001; Hame region, [r.sub.s] = 0.429, N 51, P [less than] 0.01). However, there was a clear difference between pairs and broods in distribution of occurrence frequency, especially in number of unoccupied lakes. In Karjala, proportion of empty lakes was 26% in pairs and 71% in broods (Kolmogorov-Smirnov two-sample test, D = 0.486, P [less than] 0.001). The corresponding proportions were 31% and 69%, respectively, in Hame (Kolmogorov-Smirnov two-sample test, D = 0.373, P [less than] 0.01).
Experiments on duckling body-mass change
There were clear differences in daytime and nighttime mass change among the three groups of ducklings. In daytime experimental trials, ducklings feeding on poor lakes gained less mass than those feeding on rich lakes (Table 1), and highest mass gain was recorded in the control group (Fig. 2; means ([plus or minus]1 SE) of the eleven trial means: control, 11.0 [plus or minus] 1.8 g; poor lakes, 1.9 [plus or minus] 1.4 g; rich lakes, 5.5 [plus or minus] 0.8 g). By contrast, at night, when ducklings of all three groups had equal and free access to food, the reverse was the case: ducklings that had been feeding on a poor lake that day gained more mass than those that had been feeding on a rich lake (Table 2), and the control group gained the least (Fig. 3; means ([plus or minus]1 SE) of the eleven trial means: control, 0.6 [plus or minus] 0.8 g; poor lakes, 15.2 [plus or minus] 2.7 g; rich lakes, 10.3 [plus or minus] 2.1 g). Duckling age (increasing with trial number, see Figs. 2 and 3) also had a significant ef fect on mass change, the interaction between age (trial number) and treatment type being significant. Based on overall daytime and nighttime mass changes averaged over the 11 trials (see above), [tilde]95% of daily mass gain was accounted for by daytime gain in the control group, the corresponding percentages being 35% in the rich-lake experimental group and 11% in the poor-lake experimental group.
The less mass ducklings gained in experimental trials or in the pen (control group) during the day, the more they seemed to gain during the following night, but, interestingly, the correlation was significant only for ducklings feeding in poor lakes (Table 3). In all groups, the correlation between daytime mass gain and mass gain during the previous night was also negative but not significant (Table 3). These results, together with those in the paragraph above, indicate that food resources in poor lakes did not meet energy demands of ducklings as well as did food resources in rich lakes. Hence, ducklings feeding on the former compensated by feeding more the night after.
Considering lake occupancy, proportion of unoccupied lakes was high for pairs as well as broods, but considerably higher for the latter on both Finnish study areas. Because those patterns were derived from long-term data gathered in intensive brood-survey programs, we can exclude chance effects, and we thus consider absence of broods in those lakes a genuine pattern. Furthermore, among lakes used for duckling experiments in Sweden, the proportion without wild broods in the 1995 and 1996 censuses was 68% (see Methods: Duckling experiments), which is remarkably close to figures from our Finnish study areas (69% and 71%).
What might explain the difference in lake occupancy between pairs and broods? Some of it may be due to nesting failure owing to predation on nests and adults. However, because the pattern was so consistent among years within study areas, between different study areas differing in terms of nest predation pressure, at least for hole-nesting ducks (Karjala vs. Hame, see Poysa et al. 1997, study areas Evo and Intsila corresponding with Karjala and Hame, respectively), and between different studies (citations in Introduction), we suggest that predation explains only a minor part of this difference.
In ducks, breeding period can be divided into at least two parts, pair phase and brood phase, and those may be rather different with respect to food limitation (e.g., Pehrsson 1984, Nummi and Poysa 1993; see also Rohwer 1992). Breeding females carry fat reserves to breeding grounds, reserves which they may use while searching for protein- and calcium-rich invertebrates needed for egg production (Krapu 1981, Krapu and Reinecke 1992). Adult Mallards are further capable of foraging both on and under the water surface (to a depth of 40 cm; Poysa 1983), giving access to a wide range of food resources. On the other hand, newly hatched ducklings are very dependent on emerging insects encountered on the water surface or on emergent plants (Chura 1961, Pehrsson 1979), availability of which may be weather dependent (Sjoberg and Danell 1982). As found in studies with radio-marked females in North America, movements between ponds are frequent in Mallard broods during the first weeks after hatching (Talent et al. 1982, R otella and Ratti 1992, Mauser et al. 1994, Dzus and Clark 1997). Those movements may be important in locating habitats rich in food and cover (Talent et al. 1982, Rotella and Ratti 1992; but see Dzus and Clark 1997). Furthermore, broods that move frequently may suffer greater mortality (Rotella and Ratti 1992; but see Talent et al. 1983, Dzus and Clark 1997) and fledging success may be lower in nonpreferred than in preferred habitats (Mauser et al. 1994). Hence, observations from two continents together suggest a general pattern that some ponds or lakes are inferior brood habitats, but the reason why remains unclear.
Our experimental results of changes in duckling mass give further evidence that many boreal lakes are indeed inferior duckling habitats. Ducklings foraging on poor lakes gained less mass than ducklings foraging on rich lakes despite the fact that our contrast was conservative (i.e., excluding eutrophic lakes). Furthermore, at night when ducklings from both lake types had equal and free access to food, those from poor lakes gained more body mass than those from rich lakes, indicating that ducklings feeding on poor lakes during the day compensated for inferior habitat quality by feeding more the following night. This dynamic response was clearest in ducklings foraging on the poor lakes.
Because we were able to exclude all other factors potentially affecting mass changes, except trophic status of lakes, we conclude that poorer performance of ducklings on poor lakes was because of food limitation. This conclusion is consistent with earlier studies addressing possible food competition between ducklings and fish. Mallard and American Black Duck ducklings feeding on ponds with fish-reduced abundance of invertebrates obtained less food and gained less body mass than ducklings feeding on ponds with a high invertebrate abundance (Hunter et al. 1984, 1986, Pehrsson 1984, Hill et al. 1987).
Our experimental results, together with earlier findings show that mass change of Mallard ducklings in response to food limitation is a general phenomenon. How does this relate to population limitation and regulation? Because the proportion of lakes lacking wild Mallard broods was higher among poor lakes than among rich ones, and because proportion of unoccupied lakes in general was higher for broods than for pairs, we suggest that, if present, food limitation at the population level is likely to operate at the brood stage, rather than during the preceding phases of lake selection and nesting. As proportion of lakes lacking wild Mallard broods was similar and quite high in all three study areas (see above), food limitation operating at the brood stage may be a frequent and general phenomenon on boreal lakes. In another experiment we (Poysa et al. 1998) manipulated occupation of lakes by introducing extra Mallard pairs to study habitat selection rules used by later arriving, wild, Mallard pairs. Many (29%) la kes lacked wild Mallard pairs following manipulation, even though, based on the conspecific attraction hypothesis (Stamps 1988), we expected that introduced birds should attract wild conspecifics (see Poysa et al. 1998 for details). However, the reason why lakes lacked wild pairs in that experiment remains unclear. The present results suggest that many unoccupied boreal lakes are inferior Mallard habitats and that foodbased limitation of population density may occur during the brood stage. Thus, the fact that many boreal lakes are unoccupied does not necessarily support the hypothesis that boreal bird communities are unsaturated (cf. Jarvinen and Ulfstrand 1980).
However, the connection between food limitation found here, at the brood stage, and the actual limitation and regulation of breeding Mallard populations remains open for further studies. We have indeed shown that ducklings on poor lakes gain less mass, but we do not know whether this reduction is great enough to actually affect survival of individuals and broods, let alone to limit or regulate populations. There is evidence that slow growth rate, due to late hatching, affects survival and recruitment of goslings in Lesser Snow Geese (Anser c. caerulescens) (Cooke et al. 1984, Cooch et al. 1991) and Black Brant (Branta bernicla) (Sedinger and Flint 1991, Sedinger et al. 1995). In a review of the effects of intraspecific variation in egg size in birds, Williams (1994) stressed that slower growth of smaller chicks (hatched from smaller eggs) immediately after hatching could determine high levels of early chick mortality observed in many species. In addition, lower survival of young that fledge at lower masses h as been demonstrated in many species with altricial young (Martin 1987). Finally, even though Mallard females can take their broods to lakes with ample food, movements between poor and rich lakes may increase duckling mortality (see above), thus directly influencing fitness and population regulation.
Studying mechanisms of population dynamics and regulation with controlled field experiments remains a challenging task for the future even in the most wellknown and frequently studied taxonomic groups (Boutin 1990, Harrison and Cappuccino 1995, Myers and Rothman 1995). With respect to Mallards, the body of evidence accumulated so far suggests that (1) it is important to recognize at what stage of the yearly reproductive cycle food limitation may occur, (2) brood stage may be more important than nesting stage for possible population limitation and regulation, and (3) absence of Mallards from a high proportion of lakes in the boreal region of Fennoscandia may be a result of low production of ducklings due to food limitation, rather than a result of unsaturated populations.
This study was supported by a grant from the Swedish Environment Protection Agency. Anna Lindqvist took care of the ducklings, including the imprinting procedure. Urban Friberg, Julia Grims, Rebecka Gunnarsson, Anna Lindqvist, Thomas Sundstrom, and Per Wedholm provided invaluable help during the field work. We received helpful comments on the manuscript from two anonymous reviewers. We are grateful to all of them.
(1.) Department of Animal Ecology, Swedish University of Agricultural Sciences, S-901 83 Umea, Sweden
(2.) Finnish Game and Fisheries Research Institute, Evo Game Research Station, Kaitalammintie 75, FiN-16970 Evo, Finland
(3.) Department of Applied Zoology, University of Helsinki, P.O. Box 27 (Viikki, C), FIN-00014 Helsinki, Finland
(4.) E-mail: firstname.lastname@example.org
Boutin, S. A. 1990. Food supplementation experiments with terrestrial vertebrates: patterns, problems, and the future. Canadian Journal of Zoology 68:203-220.
Chura, N. J. 1961. Food availability and preference in juvenile mallards. Transactions of the North American Wildlife and Natural Resources Conference 26:121-134.
Cooch, E. G., D. B. Lank, A. Dzubin, R. F. Rockwell, and F. Cooke. 1991. Body size variation in Lesser Snow Geese: environmental plasticity in gosling growth rates. Ecology 72:503-512.
Cooke, F., C. S. Findlay, and R. F Rockwell. 1984. Recruitment and the timing of reproduction in Lesser Snow Geese (Chen caerulescens caerulescens). Auk 101:451-458.
Danell, K., and K. Sjoberg. 1978. Habitat selection by breeding ducks in boreal lakes in northern Sweden. Swedish Wildlife Research 10:161-190,
Dzus, E. H., and R. G. Clark. 1997. Overland travel, food abundance, and wetland use by mallards: relationships with offspring survival. Wilson Bulletin 109:504-515.
Elmberg J., P. Nummi, H. Poysa and K. Sjoberg. 1993. Factors affecting species number and density of dabbling duck guilds in North Europe. Ecography 16:25 1-260.
Elmberg, J., H. Poysa, K. Sjoberg, and P. Nummi. 1997. Interspecific interactions and co-existence in dabbling ducks: observations and an experiment. Oecologia 111:129-136.
Harrison, S., and N. Cappuccino. 1995. Using density-manipulation experiments to study population regulation. Pages 131-147 in N. Cappuccino and P. F. Price, editors. Population dynamics: new approaches and synthesis. Academic Press, San Diego, California, USA.
Hill, D., R. Wright, and M. Street. 1987. Survival of Mallard ducklings Anas platyrhynchos and competition with fish for invertebrates on a flooded gravel quarry in England. Ibis 129:159-167.
Hunter, M. L., Jr., J. J. Jones, K. E. Gibbs, and J. R. Moring. 1986. Duckling responses to lake acidification: do Black Ducks and fish compete? Oikos 47:26-32.
Hunter, M. L., Jr., J. W. Witham, and H. Dow. 1984. Effects of a carbaryl-induced depression in invertebrate abundance on the growth and behavior of American Black Duck and Mallard ducklings. Canadian Journal of Zoology 62:452-456.
Jarvinen, O., and S. Ulfstrand. 1980. Species turnover of a continental bird fauna: northern Europe, 1850-1970. Oecologia 46:186-195.
Kerekes, J. J., A. C. Blouin, and S. T. Beauchamp. 1990. Trophic response to phosphorous in acidic and non-acidic lakes in Nova Scotia, Canada. Hydrobiologia 191:105-110.
Koskimies, P., and R. A. Vaisanen 1991. Monitoring bird populations, Zoological Museum, Finnish Museum of Natural History, Helsinki, Finland.
Krapu, G. L. 1981. The role of nutrient reserves in Mallard reproduction. Auk 98:29-38.
Krapu, G. L., and K. J. Reinecke. 1992. Foraging ecology and nutrition. Pages 1-29 in B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, Minnesota, USA.
Martin, T. E. 1987. Food as a limit on breeding birds: a life-history perspective. Annual Review of Ecology and Systematics 18:453-487.
Mauser, D. M., R. L. Jarvis, and D. S. Gilmer. 1994. Movements and habitat use of Mallard broods in northeastern California. Journal of Wildlife Management 58:88-94.
Myers, J. H., and L. D. Rothman. 1995. Field experiments to study regulation of fluctuating populations. Pages 229-250 in N. Cappuccino and P. F. Price, editors. Population dynamics: new approaches and synthesis. Academic Press, San Diego, California, USA.
Newton, I. 1980. The role of food in limiting bird numbers. Ardea 68:11-30.
Newton, I. 1998. Population limitation in birds. Academic Press, San Diego, California, USA.
Nummi, P., and H. Poysa. 1993. Habitat associations of ducks during different phases of the breeding season. Ecography 16:319-328.
Pehrsson, O. 1979. Feeding behaviour, feeding habitat utilization, and feeding efficiency of Mallard ducklings (Anas platyrhynchos L.) as guided by a domestic duck. Swedish Wildlife Research 10:193-218.
Pehrsson, O. 1984. Relationships of food to spatial and temporal breeding strategies of Mallards in Sweden. Journal of Wildlife Management 48:322-339.
Poysa, H. 1983. Morphology-mediated niche organization in a guild of dabbling ducks. Ornis Scandinavica 14:317-326.
Poysa, H. 1996. Population estimates and the timing of waterfowl censuses. Ornis Fennica 73:60-68.
Poysa H., J. Elmberg, P. Nummi, and K. Sjoberg. 1994. Species composition of dabbling duck assemblages: ecomorphological patterns compared with null models. Oecologia 98:193-200.
Poysa H., J. Elmberg, K. Sjoberg, and P. Nummi. 1998. Habitat selection rules in breeding Mallards (Anas platyrhynchos): a test of two competing hypotheses. Oecologia 114:283-287.
Poysa H., M. Milonoff, and J. Virtanen. 1997. Nest predation in hole-nesting birds in relation to habitat edge: an experiment. Ecography 20:329-335.
Rohwer, F. C. 1992. The evolution of reproductive patterns in waterfowl. Pages 486-539 in B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, Minnesota, USA.
Rotella, J. J., and J. T. Ratti. 1992. Mallard brood movements and wetland selection in southwestern Manitoba. Journal of Wildlife Management 56:508-515.
Sedinger, J. S., and P. L. Flint. 1991. Growth rate is negatively correlated with hatch date in Black Brant. Ecology 72:49-59.
Sedinger, J. S., P. L. Flint, and M. S. Lindberg. 1995. Environmental influence on life-history traits: growth, survival, and fecundity in Black Brant (Branta bernicla). Ecology 76:2404-2414.
Sjoberg, K., and K. Danell. 1982. Feeding activity of ducks in relation to diel emergence of chironomids. Canadian Journal of Zoology 60:1383-1387.
Staicer, C. A., B. Freedman, D. Srivastava, N. Dowd, J. Kilgar, J. Hayden, F. Payne, and T. Pollock. 1994. Use of lakes by black duck broods in relation to biological, chemical, and physical features. Hydrobiologia 279/280:185-199.
Stamps, J. A. 1988. Conspecific attraction and aggregation in territorial species. American Naturalist 131:329-347.
Talent, L. G., R. L. Jarvis, and G. L. Krapu. 1983. Survival of Mallard broods in south-central North Dakota. Condor 85:74-78.
Talent, L. G., G. L. Krapu, and R. L. Jarvis. 1982. Habitat use by Mallard broods in south central North Dakota. Journal of Wildlife Management 46:629-635.
Wiens, J. A. 1989. The ecology of bird communities. Volume 2. Processes and variations. Cambridge University Press, Cambridge, UK.
Wilkinson, L. 1992. SYSTAT: the system for statistics. SYSTAT, Evanston, Illinois, USA.
Williams, T. D. 1994. Intraspecific variation in egg size and egg composition in birds: effects on offspring fitness. Biological Review 68:35-59.
Zar, J. H. 1996. Biostatistical analysis. Third edition. Prentice-Hall, Englewood Cliffs, New Jersey, USA.
Repeated-measures analysis of variance of day-time mass change of Mallard ducklings in poor-lake and rich-lake experimental groups. Source ss df MS F P Between Subjects Group 3.713 1 3.713 29.984 [less than]0.001 Error 2.725 22 0.124 Within subjects [+] Trial 11.470 10 1.147 22.622 [less than]0.001 Trial X group 5.516 10 0.552 10.879 [less than]0.001 Error 11.155 220 0.051 Notes: The analysis is based on data presented in Fig. 2, excluding the control group, using log (y + 1) transformed data (see Methods). The repeated factor was trial. (+.)Multivariate tests based on Wilks' lambda also showed that all effects were highly significant. Repeated-measures analysis of variance of night-time mass change of Mallard ducklings in poor-lake and rich-lake experimental groups. Source ss df MS F P Between subjects Group 1.338 1 1.338 11.600 0.003 Error 2.538 22 0.115 Within subjects [+] Trial 10.913 10 1.091 85.461 [less than]0.001 Trail X group 0.806 10 0.081 6.314 [less than]0.001 Error 2.809 220 0.013 Notes: The analysis is based on data presented in Fig. 2, excluding the control group, using log (y + 1) transformed data (see Methods). The repeated factor was trial. (+.)Multivariate tests based on Wilks' lambda also showed that all effects were highly significant. Pearson correlations between daytime mass change and mass change the following night, and between daytime mass change and mass change the previous night in three groups (control, poor-lake experiment, and rich- lake experiment). Experimental group N r Daytime mass change vs. mass change the next night Control 11 -0.555 Poor-lake experiment 11 -0.745 Rich-lake experiment 11 -0.378 Daytime mass change vs. mass change the previous night Control 8 -0.736 Poor-lake experiment 8 -0.461 Rich-lake experiment 8 -0.535 Experimental group P Daytime mass change vs. mass change the next night Control [greater than]0.05 Poor-lake experiment [less than]0.02 Rich-lake experiment [greater than]0.20 Daytime mass change vs. mass change the previous night Control [greater than]0.05 Poor-lake experiment [greater than]0.20 Rich-lake experiment [greater than]0.10 Notes: Analyses are based on trial-specific averages given in Figs. 2 and 3. Sample sizes for the second set of correlations are smaller because the experimental series were interrupted two times.
|Printer friendly Cite/link Email Feedback|
|Author:||SJOBERG, KJELL; POYSA, HANNU; ELMBERG, JOHAN; NUMMI, PETRI|
|Date:||Feb 1, 2000|
|Previous Article:||PATTERNS OF NEST PREDATION CONTRIBUTE TO POLYGYNY IN THE GREAT REED WARBLER.|
|Next Article:||STRESS AND DEVELOPMENTAL STABILITY: VEGETATION REMOVAL CAUSES INCREASED FLUCTUATING ASYMMETRY IN SHREWS.|