Endozoochory by mallard in New Zealand: what seeds are dispersed and how far?

Introduction

Endozoochory, the dispersal of propagules (e.g., seeds) through the gut by an animal, is a common dispersal strategy of many plant species (Reynolds & Cumming, 2016; Janzen, 1984; Wenny et al., 2016). In New Zealand, endozoochory is considered the second most frequent dispersal mechanism of plants and is reportedly used by 33% of native flora (Thorsen, Dickinson & Seddon, 2009). However, only 12% of native plant species produce fleshy fruits, and endozoochory of other plants depends mainly on granivory (Thorsen, Dickinson & Seddon, 2009). Granivory is often an over-looked pathway of seed dispersal (Costa et al., 2014; Green et al., 2016; Farmer et al., 2017), and although its significance has been recognized in New Zealand, no previous studies have documented seed dispersal by granivorous waterfowl there (Thorsen, Dickinson & Seddon, 2009).

Plants benefit from endozoochory because it allows them to colonise new sites, evade predators, and may provide a higher likelihood of germination (Wenny et al., 2016; Green et al., 2016). Often, plants coevolve with their dispersers, and plant populations may decline if their dispersers become endangered or extinct (Thorsen, Dickinson & Seddon, 2009). However, nowadays alien disperser species, which had no time to coevolve with the native plants, are widespread, particularly in New Zealand (Turbelin, Malamud & Francis, 2017). Some introduced animals can disperse seeds of native flora effectively and compensate for reductions in native vectors (Staddon, Compton & Portch, 2010). For instance, introduced birds in Hawaii have replaced the role of extinct frugivorous birds as the primary vectors of some native plants (Cox, 1983; Foster & Robinson, 2007). In New Zealand, introduced blackbirds (Turdus merula) and song thrushes (T. philomelos) dispersed more seeds than native frugivorous avifauna within secondary forested areas (Wyman & Kelly, 2017; but see Macfarlane, Kelly & Briskie, 2016). However, alien plants can also disperse by endozoochory and this could have negative consequences for native biodiversity and ecosystem functions (Reynolds, Miranda & Cumming, 2015; Green, 2016; Van Leeuwen, 2018).

In New Zealand, studies of avian endozoochory have largely focused on frugivorous avifauna in forested and native habitats (Wyman & Kelly, 2017; Clout & Hay, 1989). Nearly 60% of New Zealand is comprised of low-lying agricultural and pastoral land (MacLeod et al., 2008), which contains remnant patches of native forests, wetlands, rivers, and swamp areas, with a vast network of drainage ditches and ponds (e.g., Miskell, 1993; Buckland, 2008). While these modified habitats have become largely dominated by invasive plants (e.g., Salix cinerea, Juncus bufonius, Rubus fruticosus, Ulex europeaus), numerous native plant species such as raupō (Typha orientalis), harakeke (Phormium tenax), swamp kiokio (Blechnum novae-zelandiae), manuka (Leptospermum scoparium) and toetoe (Austroderia toetoe) still inhabit remaining peatlands and swamps. However, little research has considered seed dispersal in these ecosystems and the role avian vectors may have on the maintenance of indigenous flora, or the expansion of invasive species, in such highly fragmented landscapes. Mallards (Anas platyrhynchos) are one of the most abundant birds in these habitats.

Waterbirds such as ducks, rails, and shorebirds are major vectors of dispersal for aquatic and terrestrial plants by endozoochory (Green et al., 2008; Green et al., 2016; Lovas-Kiss et al., 2018b). However, many of the extant native waterbird species such as brown teal (Anas chlorotis), South Island takahe (Porphyrio hochstetteri), Australasian bittern (Botaurus poiciloptilus), and black stilt (Himantopus novaezelandiae) are endangered or threatened (Robertson et al., 2017), thus their role as vectors in New Zealand may be limited. Conversely, mallards were introduced to New Zealand in the late 1800s, and they became widespread following the combined release of ∼25,000 individuals during 1940–1960 and the genetic introgression between mallard and the native grey duck (Williams & Basse, 2006; Dyer & Williams, 2010; Guay et al., 2014). Today, mallards and their hybrids (hereafter mallards) are the most common and widespread waterfowl (Anatidae) species in the country, and although total population size is unknown, around 500,000 are hunted each year (Mcdougall & Amundson, 2017). Mallards use agriculturally dominated habitats and are major vectors of plant dispersal in Europe and North America (Soons et al., 2016; Green et al., 2013; Farmer et al., 2017; Lovas-Kiss et al., 2018b). During daily movements of mallards between foraging and roosting sites in autumn and winter (outside migration periods) in the Netherlands, 41% of egested seeds were dispersed away from the sites where they were ingested, over distances of up to 25 km away (Kleyheeg et al., 2017a). Given their abundance and widespread distribution, understanding endozoochory by mallards in New Zealand is important for conservation agencies who manage native and alien vegetation in areas where mallards reside.

Here, we evaluated seed dispersal of wild mallards and mallard-grey duck hybrids (A. platyrhynchos × superciliosa) in New Zealand. We combined mallards and their hybrids in our study because they exhibit similar behaviour and female mallards were difficult to distinguish from mallard-grey duck hybrids without the aid of genetic analysis. Specifically, we aimed to (i) identify plant species dispersed by mallards, (ii) determine the indigenous status and germinability of seeds egested by mallards, (iii) assess whether these plant species are recognized as having the capacity to disperse via birds, (iv) evaluate the likely dispersal distances for seeds by endozoochory, as determined by radio-tracking of mallards.

Methods

Results

Mallard movements

Within the 72 h seed dispersal period, we collected 3,887 consecutive observations (homing = 3,863, triangulation = 130) including 2,059 records during pre-breeding and 1828 records during breeding (including laying, incubation, renesting intervals, and brood-rearing). Mean error ellipse polygon of triangulations was 0.75 ha (SD = 1.50 ha; range = 0.0001–8.88 ha). We summarised pre-breeding and breeding periods separately because activities during breeding are typically constrained to a nesting site or breeding territory.

Within 72 h, the mean distance travelled between consecutive observations was 392 m (SD = 1,066 m), but birds moved an average of 258 m further during pre-breeding than during the breeding period (t =  − 7.70, p < 0.001; Table 3). Approximately 90% of relocations occurred within 1,000 m of the previous tracking location, and during both the pre-breeding and breeding periods most of these movements were less than 200 m (Fig. 1). The fraction of movements exceeding 500 m was greater during pre-breeding, but there were no obvious differences in the distances moved between the <24 h, 24–<48 h and 48–72 h intervals (Fig. 1). However, 10% of movements exceeded 1000 m, and all the movements recorded over 10 km occurred between 24–72 h (Fig. 2). Thus, although no movement was reported in 30% of consecutive observations (i.e., the bird was in the same location during the previous tracking event), the maximum distance between tracking events was 8,028 m within 24 h, 19,230 m within 24–48 h, and 16,640 m within 48–72 h (Fig. 2; Table 3).

Table 3:

Distance moved by mallards.

Sample size (n), mean ($\bar{x}$), standard deviation (sd), and maximum (max) distance (m) travelled within 24 h, between 24–48 h, and between 48–72 h during the pre-breeding and breeding periods of female mallards equipped with radio-transmitters in Southland and Waikato, New Zealand, 2014–2015.

	<24 h				Between 24–48 h				Between 48–72 h
	n	$\bar{x}$	sd	max (m)	n	$\bar{x}$	sd	max (m)	n	$\bar{x}$	sd	max (m)
Pre-breeding	122	393.7	705.6	3,969.6	876	451.7	1,206.4	19,230.7	1,061	578.3	1,177.1	16,017.2
Breeding	461	221.8	605.8	8,028.1.4	842	255.1	932.2	16,643.9	523	284.8	1,104.7	16,640.1

DOI: 10.7717/peerj.4811/table-3

Discussion

To our knowledge, this was the first study to investigate dispersal of seeds in New Zealand by waterfowl. Previous studies in the southern hemisphere have shown that seed dispersal by granivorous waterfowl is an important process, but did not include tracking the movements of the vectors (Green et al., 2008; Raulings et al., 2011; Reynolds & Cumming, 2016). In a large sedentary population of mallards and mallard-grey duck hybrids, we found that females dispersed intact seeds of at least 12 different plant taxa. These included nine alien species and three unidentified taxa which may have been of native origin. Many of the alien species we identified have been reported throughout New Zealand, but there have been few records of Agrostis stolonifera, Polygonum aviculare, and Persicaria maculosa occurring in the Southland region, which suggests that these species have recently extended their range southwards (AVH, 2017).

We also demonstrated that female mallards were capable of moving up to 8 km within 24 h, or up to 19 km within 48 h. Our telemetry data indicated that mallards frequently moved between various waterbodies or between pastoral land and wetlands (Figs. 3 and 4). Thus, there is high potential for mallard-mediated seed dispersal for a range of wetland or pastoral species in New Zealand, including native and alien flora. Given the abundance of mallards throughout the country (Robertson, 2007), they are clearly important vectors of plants, as has previously been demonstrated for mallards in agricultural landscapes in Europe and North America (Soons et al., 2016; Farmer et al., 2017; Kleyheeg et al., 2017a).

Our results also provided important insights into seed dispersal mechanisms. Of nine plant species identified, endozoochory was recognised as the dispersal mechanism only for Solanum nigrum (Table 2). Most of the other species are generally assumed to disperse via barochory (self-dispersal) or hydrochory, yet these dispersal mechanisms provide much shorter dispersal distances than those expected from endozoochory by mallards (Bullock et al., 2017). Of particular importance are the high maximum dispersal distances (MDDs) of up to 19 km provided by mallards, since MDD has a vital influence on plant migration and colonization rates, including rates of spread of alien species (Bullock et al., 2017). In highly fragmented areas, such as the agriculturally dominated regions of Southland and Waikato, plants which disperse via endozoochory are better able to exploit patchy and isolated habitats (Soons et al., 2016; Kleyheeg et al., 2017a). As such, waterbird-mediated dispersal of indigenous flora may be especially important for the conservation of threatened or endangered plants, whereas endozoochory of alien species may aid in their expansion (Green, 2016). Even for plants able to disperse via water, mallards are vital for dispersal between catchment areas (Figs. 3 and 4). Many of the alien plant species we identified are already ubiquitous throughout New Zealand, and waterfowl may have played an important historical role in their spread. Persicaria hydropiper and S. nigrum are both widespread throughout the country (Champion & Hofstra, 2013; AVH, 2017), and the other species we identified appear to have spread beyond their previously reported distributions.

Given the limited sample size and timing of our study, we probably only detected a small proportion of the plant taxa that mallards disperse in New Zealand. Sample size effects detected in other endozoochory studies on mallards elsewhere indicate that the number of plant taxa detected increases with sample size (Soons et al., 2016; Lovas-Kiss et al., 2018b). We collected faecal samples in June and July during the late Austral-winter when many indigenous plant species are dormant but invasive species are fruiting (Wotton & McAlpine, 2015). Hence, the availability and abundance of diaspores from native flora may have been limited during our collection period. Whether mallards play an active role in dispersing endemic plants during peak floral-fruiting in Austral-spring to autumn should be investigated.

We confirmed the germinability of two plant taxa following gut passage. Many species did not germinate during our tests, perhaps because the protocols we used were not ideal. The long delay between collection and germination trials in different continents and the changes in humidity and temperature of seeds during this period likely reduced germinability rates. Previous studies of mallards have shown that intact seeds tend to maintain higher germinability after gut passage (Garcia-Alvarez et al., 2015; Van Leeuwen et al., 2012; Farmer et al., 2017), so we suspect that under better conditions more seeds may have germinated successfully. Indeed, viability after gut passage has been demonstrated for some of these same species during previous studies of mallards elsewhere (Farmer et al., 2017). Conversely, approximately 55% of Solanum nigrum seeds germinated. S. nigrum is a noxious, invasive weed common in cropping soils in New Zealand (Shah, Falloon & Bulman, 2010) and can remain dormant for extended periods of time (Edmons, 1983). Possibly, extended dormancy of S. nigrum enabled it to withstand the storage and transport from New Zealand to Spain, and may explain why the species was so successful in our trials.

The mean distance moved by female mallards between consecutive observations was 392 m (0.39 km), yet the longest distance recorded in this study was 8.0 km within 24 h and 19.2 km within 48 h. Given that seed dispersal is highly likely within 24 h of ingestion (Garcia-Alvarez et al., 2015; Kleyheeg et al., 2017a), our results suggest that mallards were most likely to disperse seeds within 8.0 km of the core home range, but that movements up to 20 km also occurred. Our irregular tracking meant that we underestimated the frequency of movements, yet our results are generally similar to Kleyheeg et al. (2017b) who used continuous tracking and reported that average and maximum flight distances of mallards in the Netherlands was 0.4 km and 25.5 km, respectively. Our results demonstrate that mallards allow dispersal of plants in New Zealand over greater distances than those predicted based on seed morphology and associated syndromes (see Bullock et al., 2017 for distances for other dispersal mechanisms). However, they also illustrate how endozoochory by mallards can result in much shorter dispersal distances than those predicted by studies assuming that they fly continuously in a straight line during the interval between seed ingestion and egestion (e.g., Raulings et al., 2011; Farmer et al., 2017).

We also found that, on average, mallards moved slightly further per day during the pre-breeding period (early June–early August) relative to the breeding period. Prior to breeding, birds are feeding in flocks and beginning to establish breeding territories, and most movements occurred between feeding and roosting areas. However, some of the longer movements occurred during the breeding period and were related to nesting, incubation recesses and brood-rearing. Breeding movements may be greatest during breeding-phase transition periods (i.e., between nesting and brood-rearing areas, or renesting attempts). Only females were radio-marked, and it is likely that males exhibited fewer constraints and showed greater movement during the breeding period, as only females incubate eggs and attend broods. The non-breeding period coincides with peak fruiting of flora that are indigenous to New Zealand, and movements of mallards or other waterbirds during this time period are likely to be more extensive and should be further evaluated.

Conclusion

Since their initial introduction into New Zealand ∼150 years ago, mallards have become the dominant waterfowl, and now play an important role as vectors of seed dispersal by endozoochory as observed in their native range. Most of the plants dispersed in our study were aliens, including all nine taxa identified to species, such that an alien bird is dispersing alien plants. With one exception (S. nigrum), these are not plants that have a fleshy fruit and hence an “endozoochory syndrome”, but rather are dry achenes (Julve, 1998) that are attractive to granivorous birds such as ducks. In New Zealand, as elsewhere, seed morphology has been used to identify putative mechanisms of dispersal for the native flora (Thorsen, Dickinson & Seddon, 2009). Generally, fleshy fruits are diagnostic of endozoochory, and researchers often focus on frugivores as plant vectors rather than granivorous waterbirds (Green et al., 2016). Recent studies from Europe have demonstrated that a range of waterbirds act as regular vectors by endozoochory for a wide variety of plants assigned to different dispersal syndromes (Van Leeuwen et al., 2017; Lovas-Kiss et al., 2018b; Lovas-Kiss et al., 2018a). In the absence of empirical evidence based on data on dispersal in field situations, diaspore morphology should not be considered as a reliable indicator of dispersal mechanisms. Ours is only a preliminary study of seed dispersal by waterfowl in New Zealand, and future work is required to establish how important mallards and native waterbirds are for the conservation of native flora or the spread of alien species.

Supplemental Information