Freshwater Biology (2011) 56, 13901402

doi:10.1111/j.1365-2427.2011.02576.x

Do birds of a feather disperse plants together?

ELISA RAULINGS, KAY MORRIS, ROSS THOMPSON AND RALPH MAC NALLY Australian Centre for Biodiversity, School of Biological Sciences, Monash University, Clayton, Vic., Australia

SU M M A R Y 1. Dispersal of propagules by waterbirds is thought to be important for wetland plants because of the abundance of birds and their frequent movements among aquatic habitats. Differences in bird characteristics (size, movement, feeding ecology) were expected to lead to different outcomes for plant dispersal. 2. We investigated heterogeneity in plant dispersal by ducks (Anas superciliosa, Anas gracilis, Anas castanea). We calculated the probability of transport of viable seeds by germinating propagules retrieved from feathers and feet (epizoochory) and the contents of the oesophagus, gizzard and lower gut (endozoochory). 3. The abundance and richness of seeds carried internally and externally did not differ among sympatric bird species. We used estimates from the literature of movements of Anas species to approximate dispersal kernels for the transport of plant propagules. 4. Heterogeneity in the abundance and movement ecology of disperser species will result in differing patterns and degrees of connectivity for wetland plant metacommunities. Sedentary waterfowl are likely to have an important role in replenishing propagules and connecting aquatic metacommunities over small distances. Nomadic waterfowl may facilitate long-distance dispersal. We discuss the implications of differences between duck species in movement patterns for connectivity of aquatic plant metacommunities across landscapes.
Keywords: aquatic macrophytes, biotic connectivity, metacommunities, waterbird dispersal

Introduction
Aquatic habitats such as rivers, oodplains and wetlands are connected biotically across landscapes by dispersal of aquatic plants (Amezaga, Santamaria & Green, 2002). Dispersal via waterbirds may be crucial for the exchange of seeds and vegetative fragments (collectively, plant propagules) because of the abundance, widespread distribution and high frequency of movement of waterbirds within and among habitats (Amezaga et al., 2002). Waterbirds disperse plant propagules in three ways: rst, by attachment to feathers, feet and bill (epizoochory or external dispersal), second, in the crop or oesophagus
Correspondence: Elisa Raulings, Australian Centre for Biodiversity, School of Biological Sciences, Monash University, Clayton, Vic. 3800, Australia. E-mail: [email protected]

coupled with regurgitation (endozoochory or internal dispersal) and third, through gut passage and subsequent faecal deposition (also endozoochory). Differences in foraging mode and movement patterns among waterbirds may inuence the plant species dispersed and the frequency, distance and direction of their dispersal (Green, Figuerola & Sanchez, 2002). As such, different bird species may produce different patterns of biotic connectivity across landscapes. Several factors suggest that waterbird species may disperse different plant taxa or disperse similar taxa but with different frequencies. Co-occurring waterbird species may forage in different vegetation communities and consume different food sources (Marchant & Higgins, 1990; Green et al., 2002). Duck species with heavier gizzards may destroy a higher proportion of ingested seeds (Figuerola, Green & Santamaria, 2002), and differences in mechanical or chemical digestion among waterbird species may
2011 Blackwell Publishing Ltd

1390

Plant dispersal by ducks 1391

affect seed viability (Proctor, 1968; Charalamabidou, a & Langevoord, 2003; Brochet et al., 2010a). Santamar Information on external transport is relatively scarce (but see Vivian-Smith & Stiles, 1994; Figuerola & Green, 2002b; Brochet et al., 2010b) but the frequency and type of plant propagules carried also may differ with bird size and foraging mode (Ridley, 1930; Figuerola & Green, 2002a,b). Assessing patterns of connectivity for aquatic plant populations requires knowledge of the movement and behaviour of birds (Haig, Mehlman & Oring, 1998; Holt Mueller & Van Der Valk, 2002), which often involves complex foraging patterns (Roshier, Asmus & Klaassen, 2008). Many studies have attempted to predict the maximum distances over which plant populations are connected by waterbird dispersal, although shorter events are also important in determining metacommunity dynamics of aquatic plants. Dispersal kernels, which describe the probability of seed deposition as a function of distance from a source (Nathan & Muller-Landau, 2000; Tackenberg, Posschlod & Bonn, 2003), differ among disperser species because of different movement patterns (Sun et al., 1997) and also because of differences in retention time and, hence, effective viability (Brochet et al., 2010a; Figuerola et al., 2010). Nomadic waterbirds make rapid and long-distance movements in response to spatial uctuations in wetland availability, while more sedentary species move frequently between feeding and resting areas. Seeds transported by these species are likely to be deposited in different locations. Here, we investigated heterogeneity in the dispersal of plant propagules by three duck species (Pacic Black Duck Anas superciliosa Gmelin, Grey Teal Anas gracilis Buller and Chestnut Teal Anas castanea Eyton) and considered implications for patterns of connectivity of aquatic plant populations across landscapes. We tested whether waterbirdmediated dispersal of aquatic plant propagules differs among dispersers, because of differences in the plant species carried, the probability of transport and movement behaviour. We attempted to germinate propagules in the gut contents and propagules borne externally on ducks. We anticipated that there would be differences among dispersers in the types of plants dispersed, the frequency with which they are transported, and the relative importance of internal and external dispersal. We also obtained estimates of the
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

movements of waterbird species to construct approximate dispersal kernels for propagules transported by these ducks. We hypothesized that interspecic differences in dispersal ecology of waterbird species would result in differences in the dispersal kernels for aquatic plants.

Methods Study site

The Gippsland Lakes Ramsar site (380802.43S, 1471341.94E: Fig. 1) contains a diversity of wetland types from deep freshwater marshes to semi-permanent saline wetlands (Corrick & Norman, 1980). Deep, freshwater wetlands are dominated by the woody Swamp Paperbark (Melaleuca ericifolia Sm.) and Common Reed [Phragmites australis (Cav.) Steud.], with rushes, sedges, herbs and submerged species (e.g. Vallisneria australis S.W.L. Jacobs & Les) scattered throughout. Saline wetlands have diverse saltmarsh vegetation including Sea Rush (Juncus kraussii Hochst.), Australian Salt Grass (Distichlis distichophylla Labill. Fassett) and Beaded Glasswort [Sarcocornia quinqueora (Ung.-Sternb.) A. J. Scott]. The area surrounding the Gippsland Lakes is predominantly cleared agricultural land dominated by introduced grasses and herbs.

Foraging behaviour and dispersal of waterbirds

Waterfowl in eastern Australia display a regular breeding peak in the late Austral winter and spring (AugustNovember; Frith, 1968). Flocks of the Pacic Black Duck co-occur with the small dabbling ducks, Grey Teal and Chestnut Teal, in wetlands of the Gippsland Lakes Ramsar site. Pacic Black Duck males weigh almost twice as much (up to 1300 g; Johnsgard, 1978) as males of Grey Teal (670 g) and Chestnut Teal (708 g), and these species have different foraging behaviour in the same habitat (Frith, 1959, 1968; Norman, 1983; Marchant & Higgins, 1990). Pacic Black Duck are among the most abundant ducks in Australia and occur in a range of habitats from deep, permanent, reed-dominated freshwater wetlands to rivers, creeks and brackish or saline wetlands (Frith, 1968; Norris et al., 1983). They rarely feed far from water, dredging mud in shallow water, up-ending to feed from the bottom, stripping the

1392

E. Raulings et al.

Fig. 1 Map of the study area showing the collection sites for gut samples (Dowd Morass, Heart Morass, Clydebank Morass, Lake Wellington and Lake Victoria) and the collection site for external transport (Dowd Morass).

seeds from plants growing in or overhanging the water and ltering seeds from the surface (Frith, 1968; Goodrick, 1979). They occasionally make long journeys (e.g. between New Zealand and Australia), although they are typically more local and sedentary than Grey Teal and can remain in the same waterbody for long periods (Frith, 1968; Marchant & Higgins, 1990). Grey Teal are highly nomadic, move in response to local conditions and occasionally move from Australia to neighbouring parts of Asia (Frith, 1968; Norris et al., 1983; Tracey et al., 2004). They are common and widely distributed across Australia in a range of habitats from mangrove estuaries on the coast to deep cold lakes in the highlands and inland Australia. Grey Teal consume fallen seeds, although some animal material is often also taken (Goodrick, 1979). Grey Teal rarely stay in the same wetland for extended periods of time (Frith, 1968) and can move hundreds of kilometres overnight (Slater, Slater & Slater, 2002; Roshier et al., 2008). Chestnut Teal are most common in brackish coastal lagoons, saltwater estuaries and the lower reaches of creeks around coastal areas. In contrast to the highly nomadic Grey Teal, Chestnut Teal are more or less sedentary with some summer movement (Norris et al., 1983). They are common in south-eastern and south-western Australia, but vagrants may occur elsewhere. They feed in and around the margins of

wetlands and tend to be non-selective and opportunistic in their feeding preferences (Norman & Mumford, 1982).

Sampling
We sampled Pacic Black Duck, Chestnut Teal and Grey Teal harvested by licensed Field and Game Australia hunters on the opening morning of the legal duck-hunting season (21 March 2009). Ducks were collected in early morning as they returned from feeding areas. Waterbirds were shot during ight over Dowd Morass and Heart Morass State Game reserves, with occasional samples from Lake Wellington and Clydebank Morass (Fig. 1). Birds were shot in-ight, so propagules recovered from these birds were in the process of dispersal. Further sampling throughout duck hunting season was not feasible because sample sizes would have been too small to compare among waterbird species. To assess external transport, seeds were removed from the feathers and feet of three Black Duck, 22 Chestnut Teal and one Grey Teal harvested at Dowd Morass. Waterbirds were retrieved immediately and placed by hunters into clean plastic bags. If ducks had been placed on the ground or on other surfaces, they were excluded from external propagule sampling. The feet were placed in water to the ankle and scrubbed with a toothbrush. The bird then was held
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

Plant dispersal by ducks 1393

over a tray, and the feathers on the head and body of the bird were brushed continuously for 3 min using a brush. To assess internal dispersal of aquatic plant propagules by ducks, the gut contents of 21 Black Duck, 24 Chestnut Teal and 21 Grey Teal were removed by hunters in the eld and placed in sealed bags on ice. The crop is located near the throat and remained in the head cavity when gut contents were removed by hunters. Where present, the oesophageal sample included the region below the crop and above the proventriculus and is likely to underestimate the number of seeds present in this part of the tract. We separated samples into oesophagus, gizzard and lower gut. To distinguish whether waterbirds differed in diets or in the digestion of plant propagules, we sampled the gizzard and the lower intestine. We sampled the lower two-thirds of the gut because viable propagules in this part of the digestive system are considered to have survived the harsh digestive juices and mechanical forces of the gizzard and upper intestine, and are most representative of faecal matea, 2005). The conrial (Charalambidou & Santamar tents of the oesophagus, gizzard and lower gut were obtained by squeezing, and the remaining tissue was thoroughly rinsed with water to ensure propagules had been removed. We used a seedling-emergence trial to estimate the number of viable seeds transported by waterbirds. We preferred this approach to visual estimates for two reasons. First, common aquatic plants, such as Juncus sp., have seeds <200 lm and visual estimates are biased towards seeds with a maximum axial dimension of 0.5 mm (Cottrell, 2004). Second, we were interested not only in dispersal per se but also in which plant species could germinate in a wetland habitat after dispersal. Containers (0.08 0.08 m) were lled with potting compost and placed in water-lled trays (0.11 m diameter) that maintained the water level just below the soil surface. This method facilitates germination of most wetland plant species and species associated with moist habitats (Boedeltje, Ter Heerdt & Bakker, 2002). Seedlings that failed to grow or showed signs of senescence under the waterlogged conditions were transferred to irrigated pots and the water level altered. External and internal duck samples were washed through 125 lm mesh, spread across the soil and placed in a temperature-controlled glasshouse. Twenty additional soil-only trays were
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

used to control for any seeds blown into the glasshouse. Germination was followed for 4 months. Seedlings were checked daily and removed from trays when they could be identied and prior to owering. Eleocharis pusilla R.Br. rapidly produces clonal growth after germinating, and it was difcult to distinguish whether new emergents were a consequence of seed germination or subsequent clonal growth. Where possible, clonally produced plants were identied from connecting rhizomes and counted as one individual. Plants were formally identied by staff at the Royal Botanic Gardens (Melbourne) and taxonomic nomenclature follows Ross & Walsh (2003).

Statistical analyses
Grey Teal were eliminated from the analysis of external transport because there was just one individual. Mann Whitney U-tests were used to compare differences among Black Duck and Chestnut Teal in the abundance and richness of seedlings emerging from feather and feet samples. KruskalWallis tests were used to compare differences in the abundance and richness of seedlings emerging from different gut regions and among all three bird species. Data were pooled across localities because birds were harvested in-ight and had not necessarily been feeding at the location from which the samples were obtained. We used logistic regression with a binomial error distribution to determine the probability that a species of bird could disperse at least one viable seed (i.e. presence absence) internally (in its lower gut) or externally (on its feet and feathers). We assumed internal and external carriage of viable propagules were not mutually exclusive. Therefore, we calculated the probability that a single bird (i.e. the dispersal unit) would transport 1 viable propagule (Pt) internally (i) or externally (e) or both as Pi [ e Pi Pe Pi \ e We did not collect externally and internally carried propagules from the same waterbirds, so we do not have an estimate of P(i \ e). Therefore, we report the minimum probability as the larger value of P(i) or P(e) and the maximum probability as P(i) + P(e). Data were analysed in S Y S T A T version 10 (SPSS, 2000). To test whether there were interspecic differences in the composition and abundance of plant taxa

1394

E. Raulings et al.
Figuerola et al., 2010) to 60 h (Pollux, Santamaria & Ouborg, 2005), but up to 94% of seeds are evacuated within 12 h (Brochet et al., 2010a). Satellite tracking of Grey Teal showed that they regularly make longdistance movements; one individual moved 502 km in 6.6 h and was travelling at 99 km h)1 for the rst half of this period (Roshier et al., 2008). To calculate the range of distances Grey Teal may travel before propagules may be egested, we used the average ight speed (76 km h)1) rather than the maximum ight speed (99 km h)1; Roshier et al., 2008) because individual birds may not be able to sustain this speed over longer distances. Reliable ight speeds were not available for Chestnut Teal or Pacic Black Duck.

in the gizzards of waterbirds, a one-way A N O S I M analysis was undertaken using P R I M E R version 5 (Clarke & Gorley, 2001).

Waterbird dispersal and construction of dispersal kernels

To determine the range of distances over which propagules may be carried by Black Duck and Grey Teal, we combined data on the probability that a bird (the dispersal unit) carried 1 viable propagule(s) (i.e. the joint probability of internal or external transport) with movement data from two studies. Between 31 December 1957 and 31 March 1958, Frith (1959) banded 5535 Black Duck at 10 locations and 10 333 Grey Teal at 12 locations in Tasmania, New South Wales, Northern Territory and South Australia. Distances travelled by recovered or recaptured birds from the original banding locality were provided. Norman (1971) banded 2041 Black Duck and 1915 Grey Teal at ve locations in South Australia over a 10 year period (195363). He tallied band returns with distance from the banding site for species recaptured from 1953 to 1963. Quantitative information on the mobility of Chestnut Teal is not available. The distance that a propagule will be deposited from its source (i.e. dispersal kernel) also depends on propagule retention time and the ight speed of birds. Propagules may be carried long distances on feathers and feet if they are carried in positions where they cannot be removed by the bird, or fall off easily (e.g. on the head; Sorensen, 1986), but germination rates may decline as a consequence of desiccation over long periods. Thus, the dispersal curves presented here assume 100% retention of externally carried seed. Gut retention time also has important effects on the percentage of seeds that germinate and the germination rate, and varies depending on both the seed species and bird vectors involved (Figuerola et al., 2010). Thus, it is not yet possible to modify the dispersal curve as a function of gut retention time. To determine the distances over which internally carried propagules are likely to be egested, we combined data from studies of faecal matter retention with ight speeds for Grey Teal. Feeding studies of other Anas species suggest the retention time of seeds in the gut varies from 1 h (Brochet et al., 2010a;

Results External transport of plants by waterbirds

Few seedlings emerged from feather and feet samples (Table 1). The native grass Lachnagrostis liformis Trin., three species of introduced pasture weeds and one native herb germinated from feathers and feet of Black Duck and Chestnut Teal (Table 1). The probability that a bird would carry one or more seeds externally, as measured by the emergence of 1 seedling from feathers or feet, was 0.22 (i.e. approximately one in ve birds carried one or more seed on their feathers or feet). The median ranks of abundance (MannWhitney U-test = 172, 1 d.f., P = 0.099) and richness (MannWhitney U-test = 171, 1 d.f., P = 0.11) of plant taxa emerging from external samples did not differ between Black Duck and Chestnut Teal. However, the small sample of Black Duck reduced the statistical power for these tests. There was little evidence that feet and feather samples of Chestnut Teal and Black Duck differed in having 1 seedling (v2 = 1.1, 1 d.f., P = 0.30), and the median ranks of abundance (MannWhitney U-test = 360, 1 d.f., P = 0.162) and richness (Mann Whitney U-test = 354, 1 d.f., P = 0.125) of plant taxa were greater in feathers than on feet, but not signicantly so.

Internal transport of plants by waterbirds

Overall, 22 taxa germinated from gut samples. The median ranks of seedling abundance (KruskalWallis
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

Plant dispersal by ducks 1395

Table 1 List of plant taxa that germinated from the feathers and feet of Pacic Black Duck and Chestnut Teal. Seeds did not germinate from the Grey Teal sample (n = 1). Values given are the number of birds with viable seeds (total number of viable seeds) Total number of birds carrying germinants Chestnut Teal (n = 22)

Feathers Pacic Black Duck (n = 3) Grasses Lachnagrostis liformis (G. Forst.) Trin. Herbs Trifolium cf. glomeratum L.* Plantago coronopus L.* Conyza bonariensis (L.) Cronquist* Senecio glomeratus Desf. ex Poir. Unidentied taxa Species 1 Total number of germinants *An introduced species. Chestnut Teal (n = 22)

Feet Pacic Black Duck (n = 3)

Feathers

Feet

1 (3)

4 (12)

1 (1)

1 (1) 1 (1) 2 (2) 1 (1) 1 1

1 2

2 (2) 15

2 1 3

H test = 38.1, 2 d.f., P < 0.001) and seedling richness (KruskalWallis H test = 37.9, 2 d.f., P < 0.001) differed signicantly among oesophagus, gizzard and lower gut samples. At least 60% of plant taxa were carried by two or more duck species. However, six of the 22 species were unique to different disperser species; in each of these cases, only one individual carried seed. The oesophagus was often missing from gut samples (n = 25) or, where present, was empty (i.e. digestive contents were absent and presumably lost during gut removal from the body cavity, n = 23). Overall, 45 seedlings from ve plant species emerged from 18 oesophageal samples. Most of the seedlings (40) were E. pusilla, which were from a single Grey Teal (Table 2). Four birds carried one or more seeds and the probability of carrying 1 seed in the oesophagus also was 0.22, and did not differ among bird species (v2 = 3.175, 2 d.f., P = 0.20). One Grey Teal had two species (E. pusilla and Isolepis sp.) emerge from the oesophagus sample. Seedling emergence from gizzards showed that waterbirds consumed a varied diet of grasses, sedges, rushes and herbs (Table 2). Some 396 seedlings emerged from gizzard samples from 19 plant taxa (Table 2). The probability of 1 seedling emerging
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

from a gizzard sample was 0.61, and seven gizzard samples had the maximum of three plant species germinate. Eleocharis pusilla was the most abundant taxon, representing 70% of all seedlings. This species was present in 19% of Black Duck gizzards and 24% of Grey Teal gizzards, but was not found in Chestnut Teal gizzards. After germination, E. pusilla rapidly produced clonal growth and it was difcult to distinguish vegetative spread from seedlings, but at least 125 seeds germinated from the gizzard of one Black Duck. The introduced plant Water Buttons (Cotula coronopifolia) was abundant in the gizzards of all bird species, with several other introduced species (Table 2). The species composition of the seedlings germinating from the contents of duck gizzards was similar among bird species based on abundance (A N O S I M global R = 0.011, P = 0.16) and diversity (A N O S I M global R = 0.019, P = 0.08). That is, the ducks consumed similar plants. Thirteen plants, from seven plant taxa, germinated from the lower gut contents (Table 2). These taxa were native sedges and introduced herbs, and were a mixture of aquatic and terrestrial species. The number of seedlings emerging from the lower gut contents was typically one per bird, but one Black Duck carried four seeds of E. pusilla. Only a few individuals of

Table 2 List of plant taxa that germinated from the oesophagus (O), gizzard (G) and lower gut (LG) samples of Pacic Black Duck, Grey Teal and Chestnut Teal. Values given are the number of birds with germinants (total number of germinants). Some birds carried one or more plant species Total number of birds carrying germinants

1396

Pacic Black Duck

Chestnut Teal

Grey Teal

E. Raulings et al.

Plant taxa

O (n = 7)

G (n = 21)

LG (n = 21)

O (n = 6)

G (n = 24)

LG (n = 22)

O (n = 5)

G (n = 21)

LG (n = 20)

LG

Grasses Poaceae 1 (4) 1 (1) 4 (129 + ) 3 (5) 1 (1) 1 (1) 1 (2) 1 (14) 1 (1) 1 (4) 1 (2) 1 (40) 1 (1)

Lachnagrostis liformis (G. Forst.) Trin. Unidentied grass

2 1 5 (148 + ) 4 (6) 1 (1) 1 1 9 8 3 2 7 1 1

Sedges Cyperaceae

Eleocharis pusilla R.Br. Bolboschoenus medianus Sojak Eleocharis acuta R. Br. Isolepis sp. Schoenoplectus tabernaemontani Palla 1 (1) 1 (2) 6 (8) 1(5) 1(1) 1 (1) 2 (3) 1 (1) 1 (1) 1 (1) 1 (1) 1 (1) 2 (4) 6 (24) 1 (1) 1 (1)

Rushes Juncaceae

Juncus gregiorus L.A.S. Johnson Juncus pallidus R. Br.

1 1 1 (1) 1 (1) 1 (2) 4 (14) 1 (1) 1 (1) 1 1 2 1 2 2 1 (1) 1 (1) 12 1 1 1 4

Herbs Asteraceae Asteraceae Asteraceae Asteraceae Asteraceae Polygonaceae Chenopodiaceae Fabaceae 2 (7) 1 (1)

Aster subulatus Michx. Conyza sp. Cotula coronopifolia* L. Lactuca serriola* L. Sonchus sp.* Persicaria prostrata (R.Br.) Sojak Atriplex prostrata DC. Trifolium glomeratum* L. Trifolium cernuum* Brot. Triglochin striata Ruiz & Pav.

1 1 (4) 1 (1) 1 (1) 1 1 (1) 1 1 1 4

Juncaginaceae Unidentied taxa Species 1 Species 2 Species 3 Total number of germinants 1

167+

1 (1) 54

44

175+

2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

*An introduced species. Eleocharis pusilla species were extensively and rapidly clonal; individual seedlings were counted, but where they could not be distinguished they are indicated with a +.

Plant dispersal by ducks 1397

several plant species survived gut passage. The probability of gut samples having 1 seedling that germinate from the lower gut was 0.14, and there was no difference among waterbird species (v2 = 0.035, 2 d.f., P = 0.98). Similarly, there was no evidence of interspecic differences in median ranks of abundance (KruskalWallis H test = 2.413, P = 0.30) or richness (KruskalWallis H test = 1.380, P = 0.50) of seeds germinating from the lower gut. With the exception of Sonchus, the plant taxa germinating from lower gut samples differed among waterbirds (Table 2). We calculated the maximum probability that a waterbird would carry 1 viable propagule(s) internally or externally was 0.36. This value assumes that internal carriage and external carriage of seeds are not mutually exclusive events and does not include oesophageal data nor detachment from feathers and feet. Seeds did not germinate in control trays, and vegetative propagules were not observed in external or internal samples.
1.0
Pacific black duck Grey teal

Probability of seed carriage

0.8

0.6
1h 60 h

0.4

0.2

0.0

1100

101300

301500 5011000

1000+

Distance from source (km)

Dispersal of waterbirds and dispersal kernels for plant propagules

The dispersal kernels for propagules carried by Black Duck and Grey Teal suggest propagules have a comparatively high probability of dispersal within 100 km of the source, before declining with increasing distance (Fig. 2). Although these waterbird species have a similar maximum probability (c. 0.36) of carrying 1 viable propagule(s), the distances that propagules are carried, and hence the shape of the dispersal kernel differed among ducks because of different mobility (Fig. 2). Assuming 100% retention, propagules carried by Black Duck are almost twice as likely to arrive at locations within 100 km of the source plant (c. 0.23) than those carried by the nomadic Grey Teal (c. 0.13), and there is a c. 10% chance that propagules attached to, or ingested, by Grey Teal will be transported more than 100 km (Fig. 2). There is a c. 1% chance that Grey Teal will carry viable propagules 1000 km, while the equivalent gure for Black Duck is 5011000 km. If gut retention as a function of ight time is overlaid on the dispersal kernel for Grey Teal, we estimate that during migratory ights propagules carried internally and egested after 1 h would be deposited <100 km from the source and up to 1800 km from the source
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

Fig. 2 Dispersal kernels for seed carriage by Pacic Black Duck and Grey Teal. Dispersal kernels assume 100% retention of externally and internally transported seed. Externally transported seed may be detached anywhere along the dispersal kernel. Internally transported seeds will be egested at a given distance along this curve depending on gut retention time and ight speeds. The range of distances potentially travelled by Grey Teal between the minimum (c. 1 h) and maximum (60 h) retention times is illustrated here. Distances travelled by recovered Pacic Black Duck and Grey Teal individuals that were banded by Frith (1959) and Norman (1971) at locations across Australia are presented as a mean SE.

wetland after 24 h if the bird is able to y continuously during this period (Fig. 2).

Discussion
The duck species considered have the potential to spread large numbers of viable seeds, although the rates of transport both internally and externally were low, seeds remained viable in both means of transport. The large number of birds moving between habitats means that the total number of propagules moved among metacommunities is likely to be high, but disperser species did not differ in their capacity to transport seed. For these three closely related duck species, different patterns of connectivity of aquatic plant metacommunities arise from differences in the abundance and movement patterns of dispersers. Plant taxa dispersed by waterfowl were aquatic and terrestrial plants [e.g. sedges (Cyperaceae), grasses (Poaceae) and amphibious herbs (particularly Asteraceae)]. Relatively few individuals of each plant species germinated from external (feather and feet)

1398

E. Raulings et al.
conditions. In arid Australia, the abundance of plant propagules in faecal samples differed between Black Swan, Eurasian Coot, Grey Teal and Pelican (Green et al., 2008). The probability of a waterbird carrying 1 viable seed(s) in the lower gut at any time was low (0.14), although the likelihood of internally mediated dispersal is likely to have been underestimated. Ducks were hunted after early morning feeding and a large proportion of the previous days food would have been egested when they were hunted, or within 12 h of consumption (Charalamabidou et al., 2003; Pollux et al., 2005; Brochet et al., 2010a). The probability of a waterbird carrying 1 viable seed(s) in the gizzard was higher (0.61) and many seedlings emerged from some gizzard samples (up to 125 seedlings per gizzard). Many of these propagules were likely to have remained viable and subsequently egested had they not been hunted. In captive experiments, for example, 32% of diaspores ingested by teal were recovered intact (Brochet et al., 2010a). It is difcult to compare the proportion of samples in which one or more seeds survived digestion in our study with other studies because of small sample sizes and differences in sampling approaches (e.g. visual identication versus germination), seed viability and waterbirds sampled (Figuerola & Green, 2002a). However, estimates vary from 28% of faecal samples containing propagules collected from ve waterbird species in a, 2005), to 33% Europe (Charalambidou & Santamar of hunted ducks in Louisiana, U.S.A. (Powers, Noble & Chabreck, 1978), to 66% of faecal samples collected from six duck species in the prairie potholes of North America (Holt Mueller & Van Der Valk, 2002). We may have underestimated the probability of internally mediated dispersal if the large numbers of seeds in the gizzard pass through the lower gut intact and remain viable. Our study and others (Figuerola & Green, 2002b) have found no differences among waterbirds in numbers of seed attached externally, despite differences in waterbird size. The percentage of birds carrying 1 seeds on feathers or feet was 22% for Black Duck and Chestnut Teal. This is much lower than that found by Vivian-Smith & Stiles (1994), who reported seeds adhering to feathers and feet of 78% of Brant Geese Branta bernicla (L.) and ducks. The latter estimates are likely to be inated because viability was not measured, but differences in substratum
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

samples, which restricted our capacity to determine whether different ducks disperse a similar suite of plant species externally. The ducks consumed similar numbers and species of seeds, but only a few individuals of each species survived gut passage. It is plausible that differences in digestion or physiology among the duck species may result in different germination and subsequently, dispersal of plants species. However, ingestion rate may be the best predictor of seed viability in the lower gut, and all plant species present in the gizzard have dispersal potential (Brochet et al., 2009). Dispersers consumed similar numbers and species of seeds in this study, despite differences in body size and reportedly feeding in different habitats within a swamp. Therefore, these disperser species probably transport similar types of plant taxa between wetland metacommunities. Occasionally, plant taxa were represented by a single germinant; for these taxa, it is difcult to interpret whether this reects real differences among waterbird diets, the effects of ducks on seed viability, or incidental ingestion of propagules.

Quantity of waterbird-mediated dispersal

In our study, closely related duck species consumed, and potentially dispersed, similar quantities of seed on a per capita basis, despite having different foraging behaviour in the same habitat and variation in body size. Others have reported interspecic differences in the abundance of internally dispersed plant propagules (Guillemain et al., 2002; Figuerola, Green & Santamaria, 2003; Green et al., 2008; Brochet et al., 2010a; Figuerola et al., 2010). Seasonal differences in propagule availability may partly explain dietary overlap of duck species in our study (Green et al., 2002). In western France, Mallard and Teal had considerable food overlap in early winter, when food was abundant, but as food availability declined, diets diverged (Guillemain et al., 2002). Aquatic plant propagules were abundant in our study region at the time of sampling and reduced competition for dietary resources may partly explain dietary overlap among Anas species in this study. When propagules are abundant, these closely related disperser species may also have more overlap in habitat use and feeding ecology than previously reported. However, divergent feeding preferences of other types of waterfowl are likely to be maintained under these

Plant dispersal by ducks 1399

(e.g. mud versus sand) also may contribute towards variation in attachment capacity (Figuerola & Green, 2002b). There remains an unavoidable bias in eld studies such as this that seeds may have been added to, or lost from, feathers and feet between being shot and picked up. wetland and arrive in a greater variety of locations and habitats. Most aquatic plant propagules are likely to be deposited at short distances (<100 km) from the propagule source. Frequent movements from foraging to resting areas for all duck species suggest aquatic plant metacommunities in the study region are relatively well connected over short distances (<100 km) and are likely to occur in any direction. Such dispersal events will be of great importance in replenishing propagule banks in the rst instance and facilitating genetic exchange if plants subsequently establish. Local interactions (e.g. competition and habitat mismatch) may constrain species diversity in metacommunities over these scales more than the supply of propagules dispersed by waterbirds. Sedentary species such as Chestnut Teal may be more likely to deposit seeds over short distances than more nomadic species, such as Black Duck and Grey Teal, because of longer residence times of the birds within wetlands and shorter ying distances (Norman & Powell, 1981; Norman, 1983). Assuming equal retention times across species, sedentary ducks may be more inuential in facilitating arrival of propagules to suitable sites and inuencing the genetic structuring of plant populations over 10100s of km. The tail of the dispersal kernel was characterized by low(er) probability, long-distance dispersal events that occur when birds undertake long periods of ight at relatively high speeds. Nomadic Grey Teal have a higher probability of dispersing seeds furthest from the source wetland, as a greater proportion of the population y longer distances, migrating more frequently and possibly at greater speeds than Black Duck and Chestnut Teal. Aquatic plants may be more reliant on nomadic waterbird species for long-distance dispersal, but may be less likely to arrive in a suitable site, or in suitable condition, for germination and establishment. This is particularly so if such movements occur without any particular directionality (as opposed to directed migration; Figuerola & Green, 2002a) or between different habitat types. Using gut retention times of 160 h (Holt Mueller & Van Der Valk, 2002; Pollux et al., 2005; Wongsriphuek, Dugger & Bartuszevige, 2008; Brochet et al., 2010a; Figuerola et al., 2010) and an average ight speed of 76 km h)1 (Roshier et al., 2008), seeds transported by Grey Teal may be deposited 761800 km from the source during migratory ights. The lower limit

Waterbird density, movement and the connectivity of aquatic metacommunities

The probability of success or failure of dispersal for aquatic plant propagules depends more on the abundance and movement of the dispersers (i.e. dispersal quality) than on the quantity of seeds that any individual waterbird disperses. The quantity of seed dispersed by each duck species is likely to be positively related to the abundance of that species at any given time. Estimating the abundance of waterfowl is difcult, but conservative estimates suggest that c. 2400 Chestnut Teal, 3600 Grey Teal and 10 000 Black Duck (c. 16 000 total) were shot in the study catchment during the 7-week duck-hunting season (Gormley & Turnbull, 2009). We estimate conservatively that the minimum number of viable plant propagules being actively dispersed across the catchment by birds at any time during the 7-week duck hunting period was c. 5760. Presuming these ducks are equally likely to be harvested, Black Duck would have transported nearly half of these propagules over the 7-week period. The time between seed consumption or attachment and departure from the seed source will affect rates of deposition and thus distances over which plant populations are effectively connected (Green et al., 2002). The lag phase associated with internal transport (1 h minimum; Brochet et al., 2010a; Figuerola et al., 2010) suggests that seeds are unlikely to be deposited adjacent to the parent plant, but most seeds will be deposited within 1224 h. There are few data on how long plant propagules remain attached externally to birds, but seeds may be removed immediately by the ducks once detected, or deposited when they come into contact with vegetation or water (Sorensen, 1986). Propagules transported on feet are likely to be washed off relatively quickly, but propagules may be carried indenitely in plumage. Similarly, the rate and frequency with which seeds are regurgitated by waterbirds, if at all, is poorly known. Seeds retained for longer will be deposited further from the source
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

1400

E. Raulings et al.
manuscript and Alan Lill, Birgita Hansen, Zachary Powell, Gary Howard and Field and Game Australia members for their assistance. This paper is contribution number 223 from the Australian Centre for Biodiversity, Monash University.

concurs with estimated internal dispersal distances of Ruppia maritima L. seed by mallard (3001400 km; Holt Mueller & Van Der Valk, 2002) and invertebrate propagules by waterfowl (601800 km; Green & Figuerola, 2005). The maximum recorded distance own by a Grey Teal within the range of retention times was 978 km in 1.8 days in the Lake Eyre Basin (Roshier et al., 2008). Whether Grey Teal can actually y 1800 km without a stopover or in-ight egestion is unknown, and seeds are likely to fall in unsuitable places for germination and survival if deposited en route. At these broader spatial scales (1001000 km), dispersal limitation may constrain the geographical ranges of aquatic plants and inuence the structure of aquatic plant metacommunities. For the rst time, we have been able to provide an integrated assessment of the effects of waterbirds on propagule transport at landscape scales. Further research is required to reduce uncertainties associated with propagule retention and frequency of duck movements between habitats. The realized impact of dispersal by waterfowl on aquatic plant population dynamics is also difcult to evaluate without information on how dispersal subsequently translates into plant tness (Figuerola & Green, 2002a). We mimicked water regime conditions that simulate requirements for germination in moist or aquatic habitats, but whether species subsequently persist and reproduce in such environments depends on a suitable water regime and water quality over long periods. Our research provides the rst step in understanding how landscape scale connectivity of wetland plant communities may be mediated by animals.

References
Amezaga J.M., Santamaria L. & Green A.J. (2002) Biotic wetland connectivity supporting a new approach for wetland policy. Acta Oecologia, 23, 213222. Boedeltje G., Ter Heerdt G. & Bakker J.P. (2002) Applying the seedling-emergence method under waterlogged conditions to detect the seed bank of aquatic plants in submerged sediments. Aquatic Botany, 72, 121128. Brochet A.L., Guillemain M., Fritz H., Gauthier-Clerc M. & Green A.J. (2009) The role of migratory ducks in the long-distance dispersal of native plants and the spread of exotic plants in Europe. Ecography, 32, 919928. Brochet A.L., Guillemain M., Fritz H., Gauthier-Clerc M. & Green A.J. (2010a) Endozoochory of Mediterranean aquatic plant seeds by teal after a period of desiccation: determinants of seed survival and inuence of retention time on germinability and viability. Aquatic Botany, 93, 99106. Brochet A.L., Guillemain M., Fritz H., Gauthier-Clerc M. & Green A.J. (2010b) Plant dispersal by teal (Anas crecca) in the Camargue: duck guts are more important than their feet. Freshwater Biology, 55, 12621273. a L. & Langevoord O. Charalamabidou I., Santamar (2003) Effect of ingestion by ve avian dispersers on the retention time, retrieval and germination of Ruppia maritima seeds. Functional Ecology, 17, 747753. a L. (2005) Field evidence Charalambidou I. & Santamar for the potential of waterbirds as dispersers of aquatic organisms. Wetlands, 25, 252258. Clarke K.R. & Gorley R.N. (2001) Primer v5. Primer-E Ltd, Plymouth, U.K. Corrick A.H. & Norman F.I. (1980) Wetlands of Victoria I. Wetlands and waterbirds of the Snowy River and Gippsland Lakes catchment. Proceedings of the Royal Society of Victoria, 91, 115. Cottrell T. (2004) Seed rain traps for forest lands: considerations for trap construction and study design. BC Journal of Ecosystems and Management, 5, 16. Figuerola J. & Green A.J. (2002a) Dispersal of aquatic organisms by waterbirds: a review of past research and priorities for future studies. Freshwater Biology, 47, 483 494. Figuerola J. & Green A.J. (2002b) How frequent is external transport of seeds and invertebrates by
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

Acknowledgments
This work was funded by ARC grant no. LPO776494 with contributions from Parks Victoria, Department of Sustainability and Victoria, West Gippsland Catchment Management Authority, East Gippsland Catchment Management Authority, Gippsland Lakes Taskforce, Royal Botanic Gardens (Victoria), Australian Ecosystems and Field and Game Victoria. Research was conducted under the specications of the Department of Sustainability and Environment Wildlife Act Research Permit 10004314 and Monash University School of Biological Sciences Animal Ethics Committee application BSCI 2008 19. We thank Andy Green and an anonymous reviewer for their comments on the

Plant dispersal by ducks 1401

waterbirds? A study in Donana, SW Spain. Archiv fur Hydrobiologie, 155, 557565. Figuerola J., Green A.J. & Santamaria L. (2002) Comparative dispersal effectiveness of wigeongrass seeds by waterfowl wintering in south-west Spain: quantitative and qualitative aspects. Journal of Ecology, 90, 9891001. Figuerola J., Green A.J. & Santamaria L. (2003) Passive internal transport of aquatic organisms by waterfowl in Donana, south-west Spain. Global Ecology and Biogeography, 12, 427436. Figuerola J., Charalambidou I., Santamaria L. & Green A.J. (2010) Internal dispersal of seeds by waterfowl: effect of seed size on gut passage time and germination patterns. Naturwissenschaften, 97, 555565. Frith H.J. (1959) The ecology of wild ducks in Inland New South Wales, III. Food habits. Csiro Wildlife Research, 4, 131155. Frith H.J. (1968) Waterfowl in Australia. Angus and Robertson, Sydney. Goodrick G.N. (1979) Food of the Black Duck and Grey Teal in Coastal Northern New South Wales. Australian Wildlife Research, 6, 319324. Gormley A.M. & Turnbull J.D. (2009) Estimates of harvest for deer, duck and quail in Victoria: results from surveys of Victorian game license holders in 2009. Arthur Rylah Institute for Environmental Research Technical series No. 196. Victorian Government. Green A.J. & Figuerola J. (2005) Recent advances in the study of long-distance dispersal of aquatic invertebrates via birds. Diversity and Distributions, 11, 149 156. Green A.J., Figuerola J. & Sanchez M.I. (2002) Implications of waterbird ecology for the dispersal of aquatic organisms. Acta Oecologia, 23, 177189. Green A.J., Jenkins K.M., Bell D., Morris P.J. & Kingsford R.T. (2008) The potential role of waterbirds in dispersing invertebrates and plants in arid Australia. Freshwater Biology, 53, 380392. Guillemain M., Fritz H., Guillon N. & Simon G. (2002) Ecomorphology and coexistence in dabbling ducks: the role of lamellar density and body length in winter. Oikos, 98, 547551. Haig S.M., Mehlman D.W. & Oring L.W. (1998) Avian movements and wetland connectivity in landscape conservation. Conservation Biology, 12, 749758. Holt Mueller M.H. & Van Der Valk A.G. (2002) The potential role of ducks in wetland seed dispersal. Wetlands, 22, 170178. Johnsgard P.A. (1978) Ducks, Geese and Swans of the World. University of Nebraska Press, Lincoln. Marchant S. & Higgins P.J. (1990) Handbook of Australian, New Zealand and Antarctic Birds. Oxford University Press, Melbourne.
2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402

Nathan R. & Muller-Landau H. (2000) Spatial patterns of seed dispersal, their determinants and consequences for recruitment. Trends in Ecology & Evolution, 15, 278 284. Norman F.I. (1971) Movement and mortality of Black duck, Mountain duck and Grey Teal banded in South Australia, 19531963. Transactions of the Royal Society of South Australia, 95, 17. Norman F.I. (1983) Grey Teal, Chestnut Teal and Pacic Black Duck at a saline habitat in Victoria. Emu, 83, 262 270. Norman F.I. & Mumford L. (1982) Food of the Chestnut Teal, Anas castanea, in the Gippsland Lakes Region of Victoria. Australian Wildlife Research, 9, 151155. Norman F.I. & Powell D.G.M. (1981) Rates of recovery of bands, harvest patterns and estimates for Black Duck, Chestnut Teal, Grey Teal and Mountain Duck shot during Victorian open seasons, 1953-77. Australian Wildlife Research, 8, 659664. Norris K.C., Mansergh I.M., Ahern K.D., Belcher C.A., Temby I.D. & Walsh N.G. (1983) Vertebrate fauna of the Gippsland Lakes Catchment Victoria. In: Occasional Paper Series Number 1. Fisheries and Wildlife Division, Victoria, Ministry for Conservation, East Melbourne. Pollux B.J.A., Santamaria L. & Ouborg N.J. (2005) Differences in endozoochorous dispersal between aquatic plant species, with reference to plant population persistence in rivers. Freshwater Biology, 50, 232 242. Powers K.D., Noble R.E. & Chabreck R.H. (1978) Seed distribution by waterfowl in southwestern Louisiana. Journal of Wildlife Management, 42, 598605. Proctor V. (1968) Long-distance dispersal of seeds by retention in digestive tract of birds. Science, 160, 321 322. Ridley H.N. (1930) The Dispersal of Plants Throughout the World. Reeve and co., Ltd., Ashford, Kent. Roshier D.A., Asmus M. & Klaassen M. (2008) What drives long-distance movements in the nomadic Grey Teal Anas gracilis in Australia? Ibis, 150, 474484. Ross J.H. & Walsh N.G. (2003) A Census of the Vascular Plants of Victoria. National Herbarium of Victoria, South Yarra, Australia. Slater P., Slater P. & Slater R. (2002) The Slater Field Guide to Australian Birds. New Holland Publishers (Australia) Pty. Ltd., Singapore. Sorensen A.E. (1986) Seed dispersal by adhesion. Annual Review of Ecology and Systematics, 17, 443463. SPSS (2000) SYSTAT version 10. SPSS Inc., Chicago, IL. Sun C., Ives A.R., Kraeuter H.J. & Moermond T.C. (1997) Effectiveness of three turacos as seed dispersers in a tropical montane forest. Oecologia, 112, 94103.

1402

E. Raulings et al.
Wongsriphuek C., Dugger B.D. & Bartuszevige A.M. (2008) Dispersal of wetland plant seeds by Mallards: inuence of gut passage on recovery, retention and germination. Wetlands, 28, 290299. (Manuscript accepted 5 January 2011)

Tackenberg O., Posschlod P. & Bonn S. (2003) Assessment of wind dispersal potential in plant species. Ecological Monographs, 73, 191205. Tracey J.P., Woods R., Roshier D.A., West P. & Saunders G.R. (2004) The role of wild birds in the transmission of avian inuenza for Australia: an ecological perspective. Emu, 104, 109124. Vivian-Smith G. & Stiles E.W. (1994) Dispersal of salt marsh seeds on the feet and feathers of waterfowl. Wetlands, 14, 316319.

2011 Blackwell Publishing Ltd, Freshwater Biology, 56, 13901402