Endozoochory of the same community of plants lacking fleshy fruits by storks and gulls

1 Endozoochory of the same community of plants lacking fleshy 1 fruits by storks and gulls 2 short running title: seed dispersal by waterbirds 3 Víctor Martín-Vélez*1, Ádám Lovas-Kiss2, Marta I. Sánchez3, Andy J. Green1 4 *[email protected] 5 Víctor Martín-Vélez: https://orcid.org/0000-0002-4846-8177 6 Ádám Lovas-Kiss: https://orcid.org/0000-0002-8811-1623 7 Marta I. Sánchez: https://orcid.org/0000-0002-8349-5410 8 Andy J. Green: https://orcid.org/0000-0002-1268-4951 9 1 Department of Wetland Ecology, Estación Biológica de Doñana, CSIC, Américo Vespucio 10 26, 41092 Sevilla, Spain. 11 2 Wetland Ecology Research Group, MTA Centre for Ecological Research, Danube Research 12 Institute, Department of Tisza River Research, Debrecen, Bem tér 18/C. H-4026 Hungary. 13 3 Departamento de Biología Vegetal y Ecología, Facultad de Biología, Universidad de 14 Sevilla, 41012 Seville 15 16 Funding information: Becas “La Caixa-Severo Ochoa 2016” (VMV). Ministerio de 17 Economía, Industria y Competitividad project CGL2016-76067-P (AEI/FEDER, EU)(MIS, 18 AJG). János Bolyai Research Scholarship of the Hungarian Academy of Sciences, New 19 National Excellence Programme of the Ministry of Innovation and Technology ÚNKP-19-4- 20 DE-172. NKFIH OTKA FK-127939 and KH-129520 grants (ÁLK). 21 22 23 24 25 2 Abstract 26 Aims and methods: research into the dispersal of plants lacking a fleshy fruit by avian 27 endozoochory remains limited, particularly the different roles of specific vectors in the same 28 habitat. We compared plants dispersed by endozoochory between two migratory waterbirds 29 differing in body size: the lesser black-backed gull Larus fuscus, and the white stork Ciconia 30 ciconia. We collected faeces and pellets from roosting flocks on dykes in ricefields in 31 Doñana, SW Spain, and extracted intact seeds. 32 Results: We recovered 424 intact seeds from excreta, representing 21 plant taxa, 11 of which 33 germinated under laboratory conditions. Eight plant species are considered weeds, four of 34 them as alien species, and only two have a fleshy fruit. Seed abundance and species richness 35 per sample did not differ between storks and gulls. Toadrush Juncus bufonius was the 36 dominant species, accounting for 49% of seeds recovered. Permanova and mvabund analyses 37 revealed no differences in the proportions of each plant species dispersed by the two vectors, 38 and seasonal variation in abundance was absent. Overall, germinability was 19 %, and 39 declined with increasing delay between sample collection and processing. Transects along 40 dykes identified 52 plant taxa, only 18 of which were recorded in excreta. 41 Conclusions: overlap in the communities of non-fleshy fruited plants dispersed by two 42 unrelated birds of different size suggests that waterbird-plant dispersal networks are different 43 from frugivore-plant networks. Unlike for frugivores, decoupling between seed production 44 and ingestion reduces seasonal variation in endozoochory rates. For J. bufonius and other 45 plants, these avian vectors provide maximum dispersal distances several orders of magnitude 46 greater than predicted from their dispersal syndromes. Endozoochory by migratory 47 waterbirds has major implications for plant distributions in a rapidly changing world, and 48 more research is required before we can predict which plants disperse regularly via this 49 mechanism. 50 3 Key words: dispersal syndromes, weeds, Larus fuscus, Ciconia ciconia, Juncus bufonius, seeds, 51 pellets, faeces, ricefields. 52 53 54 4 Introduction 55 56 Dispersal is a crucial determinant of plant distribution, demography and genetic structure 57 (Vekemans and Hardy 2004; Caughlin et al., 2014), and therefore of plant responses to 58 environmental perturbations, including anthropogenic land use change and climate change 59 (Thuiller et al., 2008; Corlett and Westcott, 2013; Tamme et al., 2014). Moreover, dispersal 60 ecology is central to the spread and potential control of alien plants and weeds (Gosper et al., 61 2005). 62 There are still knowledge gaps in our understanding of plant dispersal, including a need for a 63 better characterization of dispersal vectors (Bullock et al., 2017). Many studies have relied on 64 the classifications of plant species into syndromes based on seed morphology to make 65 predictions about vectors, and about dispersal distance (Thomson et al., 2010; Tamme et al., 66 2014). Animal vectors generally provide the longest dispersal distances for angiosperms 67 (Bullock et al., 2017), yet dispersal syndromes assume that only plants with a fleshy fruit are 68 dispersed by endozoochory (i.e. gut passage). However, repeated empirical studies have 69 shown that this assumption is invalid (Costea et al., 2019; van Leeuwen et al., 2020). 70 Migratory waterbirds act as dispersal vectors for a broad variety of angiosperms (Green et al., 71 2016; Costea et al., 2019). In Europe, hundreds of non-fleshy fruited angiosperm species 72 previously assigned to other syndromes have now been shown to be dispersed regularly by 73 ducks and shorebirds via endozoochory (Soons et al., 2016; Lovas-Kiss et al., 2018a, 2019). 74 Nevertheless, only a handful of detailed studies of waterbird endozoochory exist, compared 75 to an extensive literature of plant dispersal by frugivorous birds (Wenny et al., 2016). 76 Consequently, basic questions remain unanswered, such as to whether there are specialised 77 dispersal relationships between specific waterbird and plant species, or how the considerable 78 5 range of body size and morphology amongst waterbird groups influences plant dispersal. It is 79 well established that larger frugivores disperse plants with larger fruits (Jordano, 1995; 80 Falcón et al., 2020) and that different bird species have different roles in plant-frugivore 81 interactions (Tsunamoto et al., 2020). In contrast, for non-fleshy fruited plants at a global 82 scale, larger animals tend to ingest smaller seeds, and a higher number of plant species (Chen 83 and Moles, 2015). 84 Over the past century, the extent of natural wetlands across the globe has been greatly 85 reduced, whereas that of artificial environments such as ricefields has greatly increased 86 (Davidson et al., 2018). Many waterbird species have shifted their habitat use and movement 87 patterns to take advantage of agricultural environments such as ricefields, which are now 88 important for waterbirds across the world (Rendón et al., 2008; Toral and Figuerola, 2010; 89 Sesser et al., 2018). During the harvest period, food availability peaks and ricefields can 90 support high numbers and diversity of waterbirds (Toral et al., 2011; Rendón et al., 2008; 91 Sesser et al., 2018), which can disperse plants from the seed bank (Powers et al., 1978), 92 which is particularly diverse for weeds (Chauhan et al., 2010). Within ricefields, it is 93 therefore possible to investigate the essential differences in seed dispersal between different 94 plant vectors feeding in the same habitat. 95 Studies regarding the role of waterbirds as plant vectors within ricefields are scarce (Powers 96 et al., 1978; Brochet et al., 2010). Waterbirds feed on the alien red swamp crayfish 97 (Procambarus clarkii) in Iberian ricefields, and there is evidence for secondary dispersal of 98 seeds carried on the outside of the crayfish by lesser black-backed gulls (Larus fuscus) 99 (Lovas-Kiss et al., 2018b). The white stork (Ciconia ciconia) is a much larger waterbird also 100 known to feed on crayfish within ricefields (Tablado et al., 2010), but its role in 101 endozoochory is unknown. Both these species are benefitting from the expansion of artificial 102 6 habitats, and can show high functional connectivity between different habitat types (Bécares 103 et al., 2019; Martín-Vélez et al., 2020), increasing their potential as plant vectors. 104 In this study, we compared endozoochory by these two omnivorous waterbirds in ricefields. 105 We identified and quantified intact seeds through faecal and pellet analyses, and evaluated 106 their germinability. Our specific objectives were: (1) To establish how plant dispersal 107 interactions differ between these two different birds, and determine the roles of diet and 108 seasonal variation. (2) To evaluate the traits and life history strategies of the plants dispersed, 109 including whether they were alien species or agricultural weeds. (3) To compare the plant 110 species dispersed by these birds with the vegetation where seeds are egested, by carrying out 111 transects along ricefield borders, and comparing traits between plants recorded and those 112 dispersed. 113 114 Material and methods 115 Study area and study species 116 The ricefields of the Guadalquivir delta (37° 7' 50" -6° 9' 54", SW Spain, Fig. 1), flooded 117 from May to January, are the largest ricefield complex (37 000 ha) in Spain and an important 118 part of the Doñana wetland complex (Green et al, 2018). These ricefields support a diverse 119 avifauna (Rendón et al. 2008; Toral and Figuerola, 2010). We selected two model bird 120 species owing to their high abundance, major difference in morphology, and the ease with 121 which their excreta could be collected. 122 The lesser black-blacked gull (LBBG) is a wintering migratory waterbird breeding in North 123 Europe (Baert et al., 2018). LBBG typically arrive in SW Spain in September and migrate 124 back to their breeding grounds in March (Klaassen et al., 2012; Rendón et al. 2008). The 125 white stork breeds from Northern Europe to West Africa with major differences in migration 126 7 patterns between populations, and the Doñana ricefields hold a mixture of residents and 127 winter migrants (Bécares et al., 2019; Flack et al., 2016). Both species have increased across 128 Europe and in the study area in recent decades (Ramo et al., 2013; Rendón et al., 2008; 129 Wetlands International, 2020). The increases in numbers are related with increased food 130 availability, largely from landfills and ricefields (Martín-Vélez et al., 2020; Massemin- 131 Challet et al., 2006; Ramo et al., 2013). Over 10 000 LBBG and over 1000 white storks were 132 present in the Doñana ricefields during our study. Mean body mass is 762 g for LBBG and 133 3345 g for white stork (Wilman et al., 2014). 134 135 Excreta collection 136 A total of 463 excreta samples were collected in 36 different locations around the ricefields to 137 the north-east of Doñana National park (Fig. 1). 183 samples (136 faeces and 47 regurgitated 138 pellets) were collected from white stork and 280 (183 faeces and 97 pellets) from LBBG 139 during two consecutive winters: (1) November 2016 and (2) September, October and 140 November 2017 (Table 1). Fresh faeces and pellets were collected from dykes that serve as 141 field borders and public access routes, where monospecific flocks were resting after feeding 142 in the fields (no feeding was observed on dykes). Samples were taken from points separated 143 by at least one-meter to ensure they were from different individuals. To avoid contamination, 144 we removed the surface in contact with the soil with a knife before storing the samples in 145 separate zip bags. We preserved the samples in the fridge at 4°C until analysis. Average 146 storage time was 35 days (range 4 to 80). 147 Description of local flora along the dykes 148 Twenty vegetation transects were selected opportunistically to determine the most 149 representative flora of the dykes in the ricefield complex (Fig. 1). Ten transects were carried 150 8 out during February 2017 and ten during September-October 2017 in order to account for 151 seasonal differences. We identified all taxa present in the transects along 100 meters in a 152 straight line along the dykes, including moist soil and aquatic plants along the ricefield 153 borders. We did not sample vegetation within the ricefields, where gulls and storks were 154 feeding, because we were unable to get permission to do so. 155 Sample processing 156 The fresh mass of pellet and faecal samples was first measured on a balance (Sartorius 157 MSE225P). Diet composition based on the main food items present was categorized as (1) 158 crayfish based (2) rice based or (3) mixed (presence of both rice and crayfish). Samples were 159 then sieved (100 µm mesh) and inspected under a stereomicroscope in Petri dishes. Plant 160 diaspores (seeds and oogonia; “seeds” from hereon) were then retrieved, counted, 161 photographed and measured (with ZEN 2-2.0 software). We identified them to the lowest 162 taxonomic level by comparing the shape, size and seed coat pattern with available literature 163 (Benedí and Orell, 1992; Bojnanský and Fargašová, 2007; Cappers et al., 2012; Castroviejo, 164 1998). When it was not possible to assign a morphotype to species level with certainty, genus 165 or family level was reported. We did not include rice grains (Oryza sativa) as seeds dispersed 166 because they were unlikely to be viable (Cummings et al., 2008). Immediately after retrieval, 167 intact seeds were placed in Petri dishes that contained bacteriological agar, and placed in 168 germination chambers with 12/12 photoperiod and 22°C/18°C temperature conditions. 169 Germination tests lasted for three months and seeds were checked every day for germination. 170 Once germinated, seeds were counted and removed from the Petri dish. Seeds infected with 171 fungi were also removed and considered not-germinated. 172 173 Statistical analyses 174 9 To evaluate the sampling effect on taxa richness within the samples, we carried out 175 rarefaction analyses for each study species and sample type (Sanders, 1968), using the R 176 package iNEXT for rarefaction analyses (Hsieh et al., 2016). We applied non-metric 177 multidimensional scaling (NMDS) and Permanova analyses (applying Bray-Curtis for 178 distance matrices) to identify differences in community composition between samples 179 through the metaNMDS and adonis functions in vegan R package (Oksanen et al., 2010). 180 Abundance and richness (per sample) of seeds were compared between sample types (faeces 181 or pellets), species (LBBG or white stork), period (November 2016, September 2017, 182 October 2017 and November 2017) and diet (crayfish based, rice based or mixed) as fixed 183 factors, using sample weight as a continuous variable and sampling location as a random 184 factor. We used Generalized Mixed Models (GLMM) with negative binomial error 185 distribution and log link function under the glmmTMB package (Magnusson et al., 2017) to 186 account for the many samples with zero values, and overdispersion. For the dominant species 187 Juncus bufonius, we carried out similar GLMM analyses for abundance, but results were the 188 same as for total seed abundance (details not shown). For J. bufonius, we also tested the 189 effect of bird species, sample type, diet, period and storage time on germinability (binomial 190 model) and time (days) until germination (linear model) with the package lme4 (Bates et al., 191 2014). 192 We carried out multivariate negative binomial tests for abundance of the remaining taxa, 193 which was much lower than for J. bufonius. These tests included the variables species, 194 sample type, period and diet and were perfomed with the manyglm function in mvabund 195 package (Wang et al., 2012). All analyses were performed with R (v3.3.4 R Core Team, 196 2018). 197 16 harder and/or large food items (including large seeds) are more likely to be egested in pellets 340 than in faeces (Sánchez et al., 2005; Lovas-Kiss et al. 2019), we found no difference in seed 341 size between these two forms of excreta, presumably owing to the generally small and similar 342 size of all seeds. 343 Per individual, the larger storks ingest more, egest more, and disperse more seeds a day than 344 gulls. However, LBBGs are about ten times more abundant than white storks within the 345 ricefields (Rendón et al., 2008). Bearing in mind the peak numbers of gulls and storks 346 counted (Estación Biologica de Doñana monitoring data), and estimates for daily production 347 of excreta (Martín-Vélez et al., 2019), in the order of 105 intact seeds per day are dispersed 348 within the ricefield complex (including both fields and dykes) by these two bird species alone 349 at peak periods. In the case of LBBG, about 8% of seeds are dispersed beyond the ricefield 350 complex into other habitats over distances of up to 150 km (Martín-Velez 2021). 351 The lack of difference between storks and gulls in plants dispersed suggests that other birds 352 of an intermediate size (e.g., herons, egrets, glossy ibis, other gulls) that are abundant in 353 ricefields and feed in a similar manner on crayfish (Tablado et al., 2010) may be vectors for 354 the same plant species. On the other hand, other birds such as ducks, shorebirds and greater 355 flamingos have different feeding strategies, and may disperse plants represented in ricefield 356 seed banks in different proportions. 357 358 Comparison between seeds dispersed and vegetation transects 359 Endozoochory by storks or gulls was only recorded for 26% of plant taxa from transects 360 along dykes. Most of the plant taxa in excreta have high water requirements (Fig. 2) and 361 occur within rice stands, but not along the dykes. Vegetation transects were conducted along 362 17 the dyke habitats where birds roosted, and thus recorded plants with low moisture 363 requirements (e.g. C. canadensis, absent from excreta) on the top of dykes (Fig. 2) and 364 generalist plants occurring at field edges. However, 73% of seeds from excreta were from 365 taxa recorded in transects, therefore seeds dispersed to dykes by birds may become 366 established. Some species recorded in excreta have low moisture requirements yet were not 367 recorded in dyke transects (Table S1, Appendix 2), possibly because they grow in ricefields 368 in their dry phase before they are flooded sometime in May. The vegetation growing in fields 369 during this period, and the composition of the seedbank, should be studied in future to clarify 370 which plant species are preferentially dispersed by waterbirds. 371 372 Seasonality 373 Different management practices (e.g. harvesting, tilling) may change the availability of 374 different species in the seed bank and of food resources for birds in ricefields (Toral et al., 375 2011; Li et al., 2012; Chauhan et al., 2006), but we recorded no variation in the species 376 richness and abundance of seeds dispersed by birds over a three month period. The only 377 seasonal effects were for Ranunculus sceleratus and Cyperus difformis, which showed less 378 abundance in September and October respectively. Both species grow within the rice stands, 379 and perhaps their seed dispersal may be favoured by tilling practices in November. Since 380 storks and gulls are generally dispersing seeds after they have left the mother plant and have 381 entered the seed bank, this decoupling between seed maturity and endozoochory means that 382 differences in phenology between plant species dispersed do not readily translate into 383 differences in the timing of dispersal. Similarly, Brochet et al., (2010) found no seasonal 384 changes in the frequency of endozoochory in teal Anas crecca wintering in the Camargue and 385 feeding partly in ricefields. In contrast, frugivorous birds can show major temporal 386 18 differences in the proportions of different plants dispersed (Carnicer et al., 2009; Vázquez et 387 al., 2009). 388 Possibly, much greater variation between months and bird species in plants dispersed would 389 be recorded if excreta were sampled within natural wetlands. Ricefields are more predictable 390 in their flooding patterns and food resources than natural wetlands in Doñana, and there are 391 important differences in the bird communities they hold (Rendón et al., 2008). Likewise, 392 storks resident in Doñana are likely to disperse different plants in different parts of the annual 393 cycle, when they mainly feed in other habitats (Ramo et al., 2013). 394 395 Importance of endozoochory for long-distance dispersal 396 Storks, gulls and other birds in ricefields regularly move into natural wetlands as well as into 397 different agricultural habitats, facilitating the dispersal of alien species, weeds and other 398 plants between habitats (Bécares et al., 2019; Ramo et al., 2013; Martín-Vélez et al., 2020). 399 The plant species dispersed in our study are found in many other natural and anthropogenic 400 habitats apart from ricefields. For example, J. bufonius occurs in various terrestrial habitats 401 such as grasslands (Milotic and Hoffmann, 2016), is a new arrival in Antartica (Cuba-Diaz et 402 al., 2013), and may provide an interesting model for the study of how zoochory influences 403 genetic patterns at different spatial scales. Many of the plants we recorded in excreta samples 404 are agricultural weeds, some of which have herbicide resistant populations (Table 2), and 405 waterbirds may facilitate their effective dispersal to other habitats beyond ricefields (Farmer 406 et al., 2017; Martín-Vélez, 2021). Interestingly, five species dispersed in our study were 407 previously reported in Polish agricultural landscapes during seed dispersal by storks into their 408 nests (Table 4, the mechanism could be transfer in the beak as nest material, or via excreta) in 409 a region lacking ricefields. This suggests there is a class of non-fleshy fruited plants (e.g. J. 410 19 bufonius, Amaranthus retroflexus), with an extreme ability to disperse via birds, which is 411 worthy of future research. 412 Tamme et al. (2014) considered J. bufonius, R. sceleratus and Spergularia marina to have a 413 maximum dispersal distance of 100 m, 35 m and 340 m, respectively (via wind dispersal). All 414 three species are dispersed by storks and gulls over much longer distances, illustrating how 415 studies that make macroecological predictions about plant dispersal based on syndromes 416 ignoring non-classical endozoochory (e.g. Thomson et al. 2010; Tamme et al., 2014) are 417 likely to be unreliable. Juncus bufonius is also dispersed by endozoochory by shorebirds 418 (Lovas-Kiss et al. 2019), at least five species of Anatidae (Lovas Kiss et al., unpublished) and 419 ungulates (Milotic & Hoffmann, 2016). Different authors assigned this taxon to anemochory, 420 hydrochory and epizoochory syndromes (Löve, 1963; Cope & Stace, 1978; Julve, 1998), 421 exemplifying the subjectivity when syndromes are assigned based on seed morphology. 422 Increasing numbers of both storks and LBBG in SW Spain in recent decades may have 423 facilitated range expansions of plant species and genotypes. LBBG and white stork move at 424 three spatial scales, enabling seed dispersal into a range of habitats: 1) daily movements of up 425 to 20 km between different feeding and roost sites within the extensive ricefield complex, 426 where they often stay for several days at a time (Bouten et al., 2013; Martín-Vélez et al., 427 2020); 2) between ricefields and other habitats in Andalusia, including other agricultural 428 lands and natural wetlands such as coastal marshes and inland shallow lakes, with direct 429 flights concentrated within a radius of 150 km (Sanz-Aguilar et al., 2015; Martín-Vélez et al., 430 2020; Martín-Vélez 2021); 3) long-distance migratory flights over hundreds of km to other 431 parts of Europe or Africa. Between September to November, many gulls and storks are on 432 passage to Africa (Baert et al., 2018; Flack et al., 2016). Gut retention times for seeds easily 433 allow endozoochory over such distances (Green et al., 2016). 434 20 Conclusions and future work 435 Even though 92% of European angiosperms in continental Europe lack a fleshy fruit (Heleno 436 and Vargas, 2015), avian endozoochory studies to date have concentrated on the remaining 437 8% (i.e. on frugivores). Our study illustrates the importance of avian endozoochory for other 438 angiosperms within and beyond a wetland landscape. Waterbirds provide maximum dispersal 439 distances for many angiosperms that greatly exceed those predicted from their dispersal 440 syndromes, with major implications for how plants respond to climate change, land use 441 transformation or introductions of alien species. 442 Studying waterbird-plant dispersal interactions can improve our understanding of community 443 structure, connectivity and distributions of plant species. Effective dispersal also requires that 444 seedlings become established in new habitats, and the potential for such establishment should 445 be investigated. Detailed studies of J. bufonius are required to establish how endozoochory 446 influences population genetics and phylogeography. More research is vital to address plant- 447 bird dispersal networks involving larger numbers of waterbird species and families, and in 448 natural habitats (Sebastián-González et al., 2020). 449 Author contribution- VMV collected and analysed the samples, performed data analyses and 450 figures and wrote the first draft; ALK identified the plant taxa and reviewed several drafts; 451 MIS reviewed several drafts; AJG contributed with sampling design, and co-wrote advanced 452 drafts. All authors gave their final approval of the document and declare no conflict of 453 interests. 454 Data accessibility.- datasets used for statistical analyses can be found in Digital CSIC 455 repository: http://hdl.handle.net/10261/221014 456 Acknowledgements 457 21 Support was provided by staff of the Aquatic Ecology Laboratory LEA-EBD and the Remote 458 Sensing Lab LAST-EBD. Census data were provided by Programa de Seguimiento de EBD- 459 CSIC. Logistic and technical support for fieldwork were provided by Doñana ICTS-RBD. 460 Many volunteers contributed to fieldwork. Funding information: Becas “La Caixa-Severo 461 Ochoa 2016” (VMV). Ministerio de Economía, Industria y Competitividad project CGL2016- 462 76067-P (AEI/FEDER, EU)(MIS, AJG). János Bolyai Research Scholarship of the Hungarian 463 Academy of Sciences, New National Excellence Programme of the Ministry of Innovation 464 and Technology ÚNKP-19-4-DE-172. NKFIH OTKA FK-127939 and KH-129520 grants 465 (ÁLK). 466 467 References 468 Baert, J. M., Stienen, E. W., Heylen, B. C., Kavelaars, M. M., Buijs, R. J., Shamoun-Baranes, 469 J., ... and Müller, W. (2018). High-resolution GPS tracking reveals sex differences in 470 migratory behaviour and stopover habitat use in the Lesser Black-backed Gull Larus fuscus. 471 Scientific reports, 8(1), 5391. https://doi.org/10.1038/s41598-018-23605-x. 472 Bates, D., Mächler, M., Bolker, B., and Walker, S. (2014). Fitting linear mixed-effects 473 models using lme4. arXiv preprint arXiv:1406.5823. 474 Bécares, J.; Blas, J.; López-López, P.; Schulz, H.; Torres-Medina, F.; Flack, A.; Enggist, P.; 475 Höfle, U.; Bermejo, A. y De la Puente, J. 2019. Migración y ecología espacial de la cigüeña 476 blanca en España. Monografía n.º 5 del programa Migra. SEO/BirdLife. Madrid. 477 https://doi.org/10.31170/0071. 478 Benedí, C., and Orell, J. J. (1992). Taxonomy of the genus Chamaesyce SF Gray 479 (Euphorbiaceae) in the Iberian Peninsula and the Balearic Islands. Collectanea Botanica, 480 21,9–55. https://doi.org/10.3989/collectbot.1992.v21.92 481 22 Brochet, A. L., Guillemain, M., Fritz, H., Gauthier‐Clerc, M., and Green, A. J. (2010). Plant 482 dispersal by teal (Anas crecca) in the Camargue: duck guts are more important than their feet. 483 Freshwater biology, 55(6), 1262-1273. https://doi.org/10.1111/j.1365-2427.2009.02350.x 484 Bojnanský, V., and Fargašová, A. (2007). Atlas of seeds and fruits of Central and East- 485 European flora: The Carpathian Mountains region. Springer Science & Business Media. 486 Bullock, J. M., González, L. M., Tamme, R., Götzenberger, L., White, S. M., Pärtel, M., and 487 Hooftman, D. A. (2017). A synthesis of empirical plant dispersal kernels. Journal of Ecology, 488 105(1), 6-19. https://doi.org/10.1111/1365-2745.12666 489 Burns, K.C. (2013). What causes size coupling in fruit–frugivore interaction webs? Ecology, 490 94, 295–300. https://doi.org/10.1890/12-1161.1 491 Calvino-Cancela, M. (2011). Gulls (Laridae) as frugivores and seed dispersers. Plant 492 Ecology, 212(7), 1149-1157. https://doi.org/10.1007/s11258-011-9894-2 493 Cappers, R. T., Bekker, R. M., and Jans, J. E. (2012). Digitale Zadenatlas van 494 Nederland/digital seed atlas of the Netherlands (Vol. 4). Eelde, The Netherlands: Barkhuis. 495 Carnicer, J., Jordano, P., and Melián, C. J. (2009). The temporal dynamics of resource use by 496 frugivorous birds: a network approach. Ecology, 90(7), 1958-1970. 497 https://doi.org/10.1890/07-1939.1 498 Castroviejo, S. (Ed.). (1998). Flora ibérica: plantas vasculares de la Península Ibérica e Islas 499 Baleares. Editorial CSIC-CSIC Press. 500 Caughlin, T. T., Ferguson, J. M., Lichstein, J. W., Bunyavejchewin, S., and Levey, D. J. 501 (2014). The importance of long‐distance seed dispersal for the demography and distribution 502 of a canopy tree species. Ecology, 95(4), 952-962. https://doi.org/10.1890/13-0580.1 503 23 Chauhan, B. S., Migo, T., Westerman, P. R., and Johnson, D. E. (2010). Post‐dispersal 504 predation of weed seeds in rice fields. Weed Research, 50(6), 553-560. 505 https://doi.org/10.1111/j.1365-3180.2010.00807.x 506 Chauhan, B. S., Gill, G. S., and Preston, C. (2006). Tillage system effects on weed ecology, 507 herbicide activity and persistence: a review. Australian Journal of Experimental 508 Agriculture, 46(12), 1557-1570. https://doi.org/10.1071/EA05291 509 Chen, S.C. and Moles, A.T., (2015). A mammoth mouthful? A test of the idea that larger 510 animals ingest larger seeds. Global Ecology and Biogeography, 24, 1269-1280. 511 https://doi.org/10.1111/geb.12346 512 Cope, T. A., and Stace, C. A. (1978). The Juncus bufonius L. aggregate in western Europe. 513 Watsonia, 12, 113-128. 514 Corlett, R. T., and Westcott, D. A. (2013). Will plant movements keep up with climate 515 change?. Trends in ecology & evolution, 28(8), 482-488. 516 https://doi.org/10.1016/j.tree.2013.04.003 517 Costa-Pereira, R., Lucas, C., Crossa, M., Anderson, J.T., Albuquerque, B.W., Dary, E.P., 518 Piedade, M.T.F., … and Correa, S.B., (2018). Defaunation shadow on mutualistic 519 interactions. Proceedings of the National Academy of Sciences of the United States of 520 America 115, E2673-E2675. https://doi.org/10.1073/pnas.1801106115 521 Costea, M., El Miari, H., Laczkó, L., Fekete, R., Molnár, A. V., Lovas-Kiss, Á., and Green, 522 A. J. (2019). The effect of gut passage by waterbirds on the seed coat and pericarp of 523 diaspores lacking “external flesh”: Evidence for widespread adaptation to endozoochory in 524 angiosperms. PLoS One, 14(12). 525 24 Cuba-Diaz, M., Troncoso, J. M., Cordero, C., Finot, V. L., and Rondanelli-Reyes, M. (2013). 526 Juncus bufonius, a new non-native vascular plant in King George Island, South Shetland 527 Islands. Antarctic Science, 25, 385-386. https://doi.org/10.1017/S0954102012000958 528 Cummings, J.L., Handley, L.W., MacBryde, B., Tupper, S.K., Werner, S.J. and Byram, Z.J., 529 (2008). Dispersal of viable row-crop seeds of commercial agriculture by farmland birds: 530 implication for genetically modified crops. Environ. Biosafety Res. 7, 241-252. 531 Czarnecka, J., and Kitowski, I. (2013). The white stork as an engineering species and seed 532 dispersal vector when nesting in Poland. Annales Botanici Fennici 50 (1-2), 1-13. Finnish 533 Zoological and Botanical Publishing Board. https://www.jstor.org/stable/23728170 534 Davidson, N.C., Fluet-Chouinard, E., and Finlayson, C.M. (2018). Global extent and 535 distribution of wetlands: trends and issues. Marine and Freshwater Research, 69, 620-627. 536 https://doi.org/10.1071/MF17019 537 Dinno, A., and Dinno, M. A. (2017). Package ‘dunn. test’. CRAN Repos. doi, 10. 538 Falcón, W., Moll, D., and Hansen, D. M. (2020). Frugivory and seed dispersal by chelonians: 539 A review and synthesis. Biological Reviews, 95(1), 142-166. 540 https://doi.org/10.1111/brv.12558 541 Farmer, J. A., Webb, E. B., Pierce, R. A., and Bradley, K. W. (2017). Evaluating the potential 542 for weed seed dispersal based on waterfowl consumption and seed viability. Pest 543 management science, 73(12), 2592-2603. https://doi.org/10.1002/ps.4710 544 Figuerola, J., Green, A. J., and Santamaría, L. (2003). Passive internal transport of aquatic 545 organisms by waterfowl in Doñana, south‐west Spain. Global Ecology and Biogeography, 546 12(5), 427-436. https://doi.org/10.1046/j.1466-822X.2003.00043.x 547 25 Flack, A., Fiedler, W., Blas, J., Pokrovsky, I., Kaatz, M., Mitropolsky, M., ... and Azafzaf, H. 548 (2016). Costs of migratory decisions: A comparison across eight white stork 549 populations. Science Advances, 2(1), e1500931. DOI: 10.1126/sciadv.1500931 550 Gilbert, N. I., Correia, R. A., Silva, J. P., Pacheco, C., Catry, I., Atkinson, P. W., ... and 551 Franco, A. M. (2016). Are white storks addicted to junk food? Impacts of landfill use on the 552 movement and behaviour of resident white storks (Ciconia ciconia) from a partially migratory 553 population. Movement Ecology, 4(1), 7. https://doi.org/10.1186/s40462-016-0070-0 554 Green, A. J., Soons, M., Brochet, A. L., and Kleyheeg, E. (2016). Dispersal of plants by 555 waterbirds. Why birds matter: Avian ecological function and ecosystem services, 147. 556 Green, A.J., Bustamante, J., Janss, G.F.E., Fernández-Zamudio, R. and Díaz-Paniagua C. 557 (2018). Doñana Wetlands (Spain). Pp 1123-1136 in: The Wetland Book: II: Distribution, 558 Description and Conservation. Eds. C.M. Finlayson, G.R. Milton, R.C. Prentice, N.C. 559 Davidson. Springer. 560 Gosper, C. R., Stansbury, C. D., and Vivian‐Smith, G. (2005). Seed dispersal of fleshy‐ 561 fruited invasive plants by birds: contributing factors and management options. Diversity and 562 distributions, 11(6), 549-558. https://doi.org/10.1111/j.1366-9516.2005.00195.x 563 Hagemeijer, W. J., and Blair, M. J. (1997). The EBCC atlas of European breeding birds. 564 Poyser, London, 479. 565 Heap, I. M. (2009). International survey of herbicide-resistant weeds. Retrieved from 566 http://weedscience.org/summary/species.aspx. 567 Heleno, R., and Vargas, P. (2015). How do islands become green?. Global Ecology and 568 Biogeography, 24(5), 518-526. https://doi.org/10.1111/geb.12273 569 32 Viana, D. S., et al. 2016. Migratory Birds as Global Dispersal Vectors. - Trends Ecol Evol 704 31: 763-775. 705 Wang, Y. I., Naumann, U., Wright, S. T., and Warton, D. I. (2012). mvabund–an R package 706 for model‐based analysis of multivariate abundance data. Methods in Ecology and 707 Evolution, 3(3), 471-474. https://doi.org/10.1111/j.2041-210X.2012.00190.x 708 Wenny, D. G., Sekercioglu, C., Cordeiro, N. J., Rogers, H. S., and Kelly, D. (2016). Seed 709 dispersal by fruit-eating birds. Why birds matter: avian ecological function and ecosystem 710 services, 107-145. 711 Wetlands International (2020). Waterbird Population Estimates. Retrieved from 712 wpe.wetlands.org. 713 Wilman, H., Belmaker, J., Simpson, J., De La Rosa, C., Rivadeneira, M. M., and Jetz, W. 714 (2014). EltonTraits 1.0: Species‐level foraging attributes of the world's birds and mammals: 715 Ecological Archives E095‐178. Ecology, 95(7), 2027-2027. https://doi.org/10.1890/13- 716 1917.1 717 33 Table 1. Numbers and fresh mass in grams (reported median values and interquartile 718 range (IQR) between brackets) of samples collected in ricefields from white stork and 719 LBBG. 720 721 Species Sample type Median mass (IQR) Nov. 2016 Sept. 2017 Oct. 2017 Nov. 2017 Total White stork Faeces 2.55 (0.54) 51 32 23 30 136 Pellets 8.0 (0.01) 22 - 14 11 47 LBBG Faeces 2.0 (0.55) 53 42 47 41 183 Pellets 8.2 (0.06) 25 15 27 30 97 34 Table 2. Details of intact seeds found in white stork and LBBG excreta samples from the winters of 2016-2017. Shown are whether or not a taxon was recorded within dyke transects (Dyke Occur.), total number of seeds found in excreta, the number of samples in which each taxon was recorded (NST), the combined number of seeds in those samples (N seeds) and the maximum number of seeds recorded in a single sample (Max. sample). Larus fuscus Ciconia ciconia Pellets (N=97) Faeces (N=183) Pellets (N=47) Faeces (N=136) Family Taxa Dyke Occur Tota l NST N seed Max. sample NST N seed Max. sample NST N seed Max. sample NST N seed Max. sample Amaranthaceae Amaranthus albus a, b, c Yes 10 1 1 1 1 3 3 4 4 1 1 2 2 Amaranthus retroflexus a, b, c No 1 - - - - - - 1 1 1 - - - Chenopodium cf. album - 2 - - - - - - - - - 1 2 2 Araceae Lemna minor Yes 3 1 1 1 - - - - - - 1 2 2 Lemna gibba Yes 2 - - - - - - 1 1 1 1 1 1 Caryophyllacea e Spergularia marina Yes 27 - - - - - - 1 24 24 1 1 1 Characeae Chara sp. - 35 3 3 1 9 11 3 3 12 9 8 9 2 Cyperaceae Cyperus difformis b, c No 38 2 2 1 8 11 3 7 10 4 10 15 4 Elatinaceae Bergia capensis a No 1 - - - 1 1 1 - - - - - - Fabaceae Trifolium repens Yes 1 - - - - - - 1 1 1 - - - Juncaceae Juncus subnodulosus Yes 15 1 1 1 3 3 1 4 5 2 5 6 2 Juncus bufoniusb Yes 210 14 20 4 35 52 4 20 42 10 45 96 9 Poaceae Polypogon monspeliensisb, c Yes 2 - - - - - - - - - 2 2 1 35 Sorghum cf. halepense a, b, c No 2 - - - - - - 1 2 2 - - - Panicum cf. miliaceum No 1 - - - - - - 1 1 1 - - - Unidentified - 2 - - - 1 1 1 1 1 1 - - - Portulacaceae Portulaca oleracea b, c Yes 4 1 1 1 - - - 1 1 1 1 2 2 Ranunculaceae Ranunculus sceleratus Yes 34 1 1 1 8 9 2 4 5 2 9 19 6 Ranunculus repens Yes 1 - - - - - - 1 1 1 - - - Solanaceae Solanum nigrum b, c Yes 14 1 5 5 2 6 3 - - - - - - Solanum dulcamara Yes 19 1 1 1 2 15 8 1 1 1 3 3 3 Total 424 26 36 5 70 112 8 52 111 10 88 158 9 a Taxa alien to Spain according to Agroatlas (2005), http://agroatlas.ru., b considered an agricultural weed, and c known to have herbicide resistant populations according to Heap (2009). 36 Table 3. Effects of bird species, period, diet, sample type and weight on (A) total abundance of seeds and (B) taxon richness per sample, from negative binomial mixed models. White stork, faecal samples, November 2016, and a diet of crayfish are absent from the table because these levels of the respective factors were aliased, and so effectively had estimates of zero. Sampling location (Figure 1) was included as a random factor. Shown for each term are the parameter estimates (β) and their standard errors, and the main effects for each predictor variable. (A) Seed abundance Level of effect β S.E. χ2 p Species LBBG -0.726 0.183 15.738 <0.001 Period Sept. 2017 0.110 0.322 4.127 0.248 Oct. 2017 -0.048 0.327 Nov. 2017 0.546 0.310 Diet Mixed 0.733 0.252 2.069 0.355 Rice 0.338 0.238 Sample mass 0.076 0.014 30.31 <0.001 Sample type Pellets -0.594 0.219 7.393 0.007 Random contribution (variance): location= 0.156 (B) Plant richness Level of effect β S.E. χ2 p Species LBBG -0.647 0.168 14.784 <0.001 Period Sep. 17 0.016 0.281 4.751 0.191 Oct. 17 -0.167 0.286 Nov. 17 0.444 0.264 Diet Mixed 0.096 0.233 0.586 0.746 Rice 0.164 0.220 37 Sample mass 0.069 0.012 30.312 <0.001 Sample type Pellets -0.459 0.203 5.108 0.024 Random contribution (variance): location= 0.0908 38 Table 4. Germination of seeds found in LBBG and white stork pellets and faeces. Larus fuscus Ciconia ciconia Pellets Faeces Pellets Faeces Plant family Plant taxa Total diaspores Total germinated N diaspores N germinated N diaspores N germinated N diaspores N germinated N diaspores N germinated Amaranthaceae Amaranthus albus 9 6 1 - 3 3 3 1 2 2 Amaranthus retroflexus 1,2,3 1 1 - - - - 1 1 - - Chenopodium cf. album1,3 3 3 - - - - 1 1 2 2 Araceae Lemna minor 3 0 1 - - - - - 2 - Lemna gibba 2 0 - - - - 1 - 1 - Caryophyllaceae Spergularia marina 27 11 - - - - 24 11 1 - Characeae Chara sp. 35 0 3 - 11 - 12 - 9 - Cyperaceae Cyperus difformis 1 38 4 2 1 11 - 10 2 15 1 Elatinaceae Bergia capensis 1 0 - - 1 - - - - - Fabaceae Trifolium repens1,3 1 1 - - - - 1 1 - - Juncaceae Juncus subnodulosus 1 15 1 - - - - 1 - - 1 Juncus bufonius 1,2,3 210 39 20 4 52 12 42 4 96 19 Poaceae Polypogon monspeliensis 1 2 0 - - - - - - 2 - Sorghum cf. halepense 2 0 - - - - 2 - - - Panicum cf. milleanum 1 0 - - - - 1 - - - Unidentified 2 0 - - 1 - 1 - - - Portulacaceae Portulaca oleraceae 4 2 1 - - - 1 1 2 1 Ranunculaceae Ranunculus sceleratus 1,2 34 0 1 - 9 - 5 - 19 - 39 Ranunculus repens3 1 0 - - - - 1 - - - Solanaceae Solanum nigrum1 14 2 5 - 6 2 - - - - Solanum dulcamara1 19 10 1 - 15 10 1 - 3 - Total 424 80 35 5 110 27 108 22 154 26 1endozoochory previously reported for intact seeds in gulls (Calvino-Cancela 2011; Lovas-Kiss et al., 2018b). 2germination after gut passage confirmed in Lovas-Kiss et al., 2018b. 3 seed dispersal previously reported within white stork nest material (Czarnecka & Kitowski, 2013). 40 Table 5. Effects of species, period, sample type, diet and storage time on germinability (A) and germination time (B) of Juncus bufonius based on binomial and linear models respectively. LBBG, faecal samples, November 2016, and a diet of crayfish were aliased. See Table 3 for further explanation. (A) Germinability Juncus bufonius Level of effect β S.E. χ2 p Species White stork -0.493 0.443 0.074 0.786 Period Sept. 2017 -0.213 0.560 8.416 0.038 Oct. 2017 -2.304 1.110 Nov. 2017 -0.226 0.548 Diet Mixed -0.434 0.703 1.319 0.517 Rice -0.167 0.564 Sample type Pellets -0.528 0.4644 2.405 0.121 Storage time -0.024 0.013 3.893 0.048 (B) Germination time Juncus bufonius Level of effect β S.E. χ2 p Species Ciconia ciconia -0.196 1.114 0.098 0.756 Period Sep. 2017 3.4923 1.434 3.522 0.028 Oct. 2017 -1.621 3.204 Nov. 2017 2.325 1.587 Diet Mixed -2.262 2.118 1.002 0.380 Rice -1.487 1.512 Sample type Pellets 0.672 1.428 0.048 0.827 Storage time -0.001 0.043 0.006 0.938 Adj. R2=0.119. 41 Figure 1. Location of sites within the ricefields of Doñana Biosphere Reserve where gull and stork samples (including pellets and faeces) were collected, and vegetation transects were monitored, in 2016 and 2017.