SYNTHESIS & INTEGRATION

The ecological significance of secondary dispersal by € € € 1, 1 1 1 ANNI HAMALAINEN , KATE BROADLEY, AMANDA DROGHINI, JESSICA A. HAINES, 1 1 1,2 CLAYTON T. LAMB, STAN BOUTIN, AND SOPHIE GILBERT

1Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2M9 Canada 2Department of Fish and Wildlife Sciences, University of Idaho, Moscow, Idaho 83843 USA

Citation: Ham€ al€ ainen,€ A., K. Broadley, A. Droghini, J. A. Haines, C. T. Lamb, S. Boutin, and S. Gilbert. 2017. The ecological significance of secondary by carnivores. Ecosphere 8(2):e01685. 10.1002/ecs2.1685

Abstract. Animals play an important role in the seed dispersal of many plants. It is increasingly recog- nized, however, that the actions of a single disperser rarely determine a seed’s fate and final location; rather, multiple abiotic or animal dispersal vectors are involved. Some carnivores act as secondary dis- persers by preying on primary seed dispersers or seed predators, inadvertently consuming contained in their prey’s digestive tracts and later depositing viable seeds, a process known as diploendozoochory. Carnivores occupy an array of ecological niches and thus range broadly on the landscape. Consequently, secondary seed dispersal by carnivores could have important consequences for plant dispersal outcomes, with implications for functioning under a changing climate and across disturbed landscapes where dispersal may be otherwise limited. For example, trophic downgrading through the loss of carni- vores may reduce or eliminate diploendozoochory and thus compromise population connectivity for lower trophic levels. We review the literature on diploendozoochory and conclude that the ecological impact of a secondary vs. primary seed disperser depends on the relative dispersal distances, success, and the proportion of seeds exposed to secondary dispersal by carnivores. None of the studies up to present day have been able to rigorously assess the ecological significance of this process. We provide a framework of the components that determine the significance of diploendozoochory across systems and identify the components that must be addressed in future studies attempting to assess the ecological importance of diploendozoochory.

Key words: ; ; diploendozoochory; fragmentation; indirect dispersal; invasive species; polychory; predator; secondary dispersal; seed dispersal; seed dispersal effectiveness; seed predation.

Received 19 December 2016; accepted 22 December 2016. Corresponding Editor: Debra P. C. Peters. Copyright: © 2017 Ham€ al€ ainen€ et al. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. E-mail: [email protected]

INTRODUCTION resources, reduce density-dependent seedling mortality, and may have better chances of encoun- Owing to their sessile adult lives, plants have tering suitable microhabitats for recruitment evolved numerous ways to ensure offspring dis- (Howe and Smallwood 1982). In addition to dis- perse away from the immediate vicinity of the persal by abiotic means such as wind and water, parent plant, typically via the movement of seeds. seed dispersal by animals (zoochory) is a common This movement is beneficial because seedlings dispersal mechanism that allows seeds to be that take root further away from their parent plant deposited some distance away from the parent avoid competing with related individuals for plant through animal vectors. While plants with

❖ www.esajournals.org 1 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL. high rates of seed predation have evolved adapta- been made to sufficiently establish the importance tions that minimize seed losses by animals (Hulme of the mechanism. A synthesis of the topic is there- and Benkman 2002), plants that rely on zoochory fore needed to enable broader predictions on the often produce seeds encased in fleshy fruits to prevalence and ecological role of diploendozoo- promote consumption by dispersers (endozoo- chory to help direct future research into the chory; Schaefer and Ruxton 2011). Estimates of phenomenon. the number of plants dispersed by animals vary Presumably, the effects of secondary dispersal widely, but endozoochory is estimated to be the depend strongly on the characteristics of the ani- primary dispersal mechanism in up to 94% of mal vectors and plants involved, as well as the woody plants, depending on the region (Jordano habitats they occupy. Therefore, we identified 2000, Buitron-Jurado and Ramırez 2014). Further- characteristics of the dispersal process that are more, many plants that do not seem to have adap- likely to influence plant fitness via germination tations for endozoochory can, nevertheless, be or recruitment success and access to suitable dispersed by animals in addition to dispersal by habitat. Using this information, we devised a other means such as wind, water, or gravity (Pake- framework to assess the importance of diploen- man et al. 2002, Orłowski et al. 2016). dozoochory in plant dispersal success under Despite such adaptations, there is increasing different conditions. Using this framework, we evidence that the final fate of seeds is not necessar- identified potential ecological consequences of ily determined by the animals that remove them secondary dispersal for ecosystem functioning. from the parent plant. Instead, multiple dispersal We then reviewed the existing literature on the vectors may be involved in taking seeds to their role of carnivores in multi-stage seed dispersal final destination or destruction (Ozinga et al. to assess the evidence in support of the identi- 2004, Vander Wall and Longland 2004); at the fied mechanisms and their implications for plant community level, plants were estimated to have dispersal. on average 2.15 dispersal vectors per species among Dutch (Ozinga et al. 2004). WHAT MAKES DIPLOENDOZOOCHORY AN There has been increasing interest in diplochory ECOLOGICALLY SIGNIFICANT SEED DISPERSAL (two-phase dispersal, also known as “secondary MECHANISM? dispersal” or “indirect dispersal”) involving a sec- ond dispersal phase by , dung beetles, or scat- The overall impact of the dispersal mechanism ter-hoarding that physically carry the on plant fitness is composed of the quantity of seeds to a new location (Vander Wall and Long- seeds processed by each dispersal vector, the via- land 2004), but relatively little attention has been bility of the seeds after handling and consump- paid to “diploendozoochory,” that is, seed disper- tion by each disperser, and the likelihood that sal that involves the ingestion of the seed by two the dispersed seeds will germinate and mature or more separate species of animals in sequence. into reproductive adult plants where they are Typically, this occurs when a carnivorous predator deposited (seed dispersal effectiveness frame- (hereafter referred to as carnivores) consumes a work; Schupp and Jordano 2010). In general, a primary disperser or a seed predator, along with prerequisite for successful endozoochory is that seeds in its prey’s digestive tract, and subsequently the digestive process of the animal vector does deposits the seeds in or in regurgitated pel- not damage the seed; thus, seeds should be swal- lets (Dean and Milton 1988, Nogales et al. 2007, lowed whole. Furthermore, the viable seed must 2012). Diploendozoochory was first documented be deposited by the final disperser at a site that by Darwin (1859), and opportunistic observations meets the minimum requirements for successful have since then been infrequently reported. germination: Deposition in a cave or a building Although diploendozoochory has recently been (Dean and Milton 1988), a highway or the ocean, approached more rigorously using experiments will usually impede seed germination. (Nogales 1999, Nogales et al. 2007, Padilla and For a to improve seed dispersal out- Nogales 2009, Padilla et al. 2012), the broader eco- comes, the seed dispersal effectiveness of diploen- logical significance of this phenomenon remains dozoochory must naturally be higher than that of largely unknown as few attempts have thus far dispersal by a single vector (Schupp and Jordano

❖ www.esajournals.org 2 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

2010). The types of plants and the primary and dispersers may thus deposit seeds in maladaptive secondary dispersers involved all influence dis- locations or out of their climate zone, but the persal effectiveness and the ecological significance plant could, nevertheless, benefit from the long- of the secondary dispersal phase, as detailed in distance dispersal if novel, suitable patches are the framework in Table 1. The involvement of a even occasionally reached and colonized via this carnivore in the second phase of the seed disper- process (Nathan et al. 2008, Caughlin and Fergu- sal process can influence plant fitness in three son 2013). ways: by transporting seeds, by altering the via- bility of transported seeds, and by changing the Effects on germination success quantity of seeds that are dispersed. Germination success may be altered via the treatment of a seed in a carnivore’s digestive tract Long-distance dispersal and can be differentially affected by various sec- Long-distance dispersal in plants, although ondary dispersers (Table 1). Improvement in ger- rare (Cain et al. 2000, Nathan et al. 2008), mination success follows if the seeds benefitfrom increases the rate of spread and colonization of a double digestion due to a longer gut retention new areas (Higgins and Richardson 1999, Lesser time (e.g., thick-coated seeds; Nogales et al. 2015) and Jackson 2013), with effects on plant popula- or, possibly, if carnivore feces is richer in nutrients tion dynamics (Cain et al. 2000, Nathan et al. or includes a lower number of competing seeds 2008, Caughlin and Ferguson 2013). As carni- than that of the primary disperser. Negative effects vores tend to range much farther than frugivores on germination can result from damage to thin- or (Carbone et al. 2005), the dispersal coated seeds due to coarse materials ingested distance and deposition site (i.e., location at alongside seeds (Traveset et al. 2007). For example, which the seed ends up after being processed by germination success was improved by a secondary a disperser, typically within feces or within dispersal phase by shrikes, but reduced by kestrels regurgitated pellets) may differ dramatically for (Nogales et al. 2002), and gray herons caused a seeds deposited by primary vs. secondary dis- complete loss of seed viability (Rodrıguez et al. persers (Dean and Milton 1988, Nogales et al. 2007, Tables 2 and 3). Variability in carnivore 2007, 2012). Secondary dispersal can thus con- effects on germination success may be related to tribute significantly to plant dispersal range and species-specific gut enzymatic activity and other population dynamics, especially when the pri- foods ingested with the seeds (Rodrıguez et al. mary disperser has a relatively small home range 2007, Traveset et al. 2007). Damage to seeds might size, is movement-restricted, or a habitat special- also be influenced by the evolutionary past of ist (Higgins and Richardson 1999, Nogales et al. coexistence of the plant and the secondary disper- 2012). For example, after consuming a prey item sal, and the potential for adaptation by the plant with a very small home range (such as a small to minimize such losses. Plant germination or ), a raptor pellet may take up to 22 h to recruitment is likely also impaired by a deposition form. During a migration, a can cover in unsuitable microhabitats (e.g., on a road, in 480 km in that time (Balgooyen and Moe 1973); poor soil, or in dense vegetation), or at a site that this distance could be even longer as some seeds elevates the risk of post-dispersal seed predation. can germinate after a retention time of up to 62 h These determinants of dispersal effectiveness are in of prey (Darwin 1859). not unique to carnivore-mediated dispersal, but Secondary dispersal by far-ranging carnivores their importance in diploendozoochory remains to can thus enable colonization of newly suitable be studied. habitats under climate change or remote areas such as islands (Nogales et al. 2012), or may Interruption of seed predation locally influence the number of seeds entering the Carnivores can also indirectly improve seed seed pool. An increase in dispersal distance may, viability by interrupting seed predation (Sarasola however, reduce the dispersal success of rare, or et al. 2016). Granivorous birds, for example, con- habitat-specialist species (Herrmann et al. 2016), sume large quantities of seeds that are not imme- unless the dispersers have similar, specialized diately destroyed but rather move to the gizzard habitat requirements as the plant. Secondary intact, and are only later broken down and

❖ www.esajournals.org 3 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Table 1. Mechanisms by which a plant’s dispersal success is influenced by a predator as a secondary disperser.

Effect of How is dispersal outcome altered? diploendozoochory Under which conditions is on plant dispersal the outcome expected to success Deposition site Germination success have ecological relevance?

Seed dispersal 1. Longer dispersal distance 1. Seeds benefits from a double 1. Habitat becomes avai- success improved provides access to newly digestive process (thick seed lable for colonization available habitat suitable for coat) within the secon- colonization or increases the 2. Fecal (or regurgitate) nutri- dary disperser’srange speed of spread ent content is higher or more via fragmentation or 2. Distance from parent can suited to the seed’s needs changing landscapes reduce kin competition, than that of primary dis- 2. The dispersed plant is post-dispersal seed preda- perser a pioneer species that tion, and disease mortality 3. Predation of the primary requires newly dis- 3. Travel between fragments consumer saves the seed turbed habitats for improves plant resilience by from destruction (seed pre- establishment maintaining gene flow dation) and thus increases 3. The plant is adapted between populations its viability to the alternative 4. Better or more varied posi- habitats made avail- tions on the landscape are able by the secondary reached because the preda- disperser and pro- tor reaches a wider range of duces seeds that sur- suitable sites vive or benefit from two-phase digestion 4. The predator inter- cepts a significant proportion of seed predation or seeds transported by an ineffective primary disperser

Seed dispersal 1. The predator tends to use 1. Seed is damaged by double 1. Predation on primary success lowered habitat that is unsuitable for digestion (thin seed coat) dispersers signifi- the plant and that is not or mechanical damage by cantly reduces the used by the primary dis- other, simultaneously inges- proportion of viable perser, or deposits seeds in ted food items seeds entering the poor locations (e.g., dense 2. Seeds from multiple prey seed pool due to vegetation with high compe- items may increase the num- reduced germination tition, or a site with high risk ber of seeds in a deposit, success or unsuitable of seed predation or post- leading to increased compe- deposition sites by germination consumption) tition or reduced viability the secondary dis- 2. Longer dispersal distances due to release of allelochem- perser take the seed of a specialized icals from certain seeds that 2. A significant propor- or rare plant outside of the prevent germination of other tion of seeds of rare species’ potential range and seeds (Traveset et al. 2007) or specialized plant thus reduce the plant’s effec- 3. Nutrient content is lower or species are removed tive population size less well suited to the seed’s from their suitable needs in the feces or regurgi- habitats by a sec- tate of the secondary than of ondary disperser the primary disperser

No significant effect 1. The secondary disperser 1. Processing of the seed by the on dispersal uses similar habitats and secondary disperser does success does not range significantly not change the germination wider than the plant’s pri- success of the seeds, relative mary dispersers to effects of the primary dis- 2. The plant is common and perser widespread with sufficient 2. The proportion of a plant’s population overlap that it seeds transported by a sec- does not suffer from reduced ondary disperser is very gene flow small compared to other means of dispersal

❖ www.esajournals.org 4 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Table 2. A review of the literature on diploendozoochory. See Table 3 for further detail.

Type of References Plant species being dispersed Primary consumer disperser † Secondary disperser

Balgooyen and Moe Calamagrostis Canadensis ? NA American kestrels (Falco (1973) sparverius) Darwin (1859) Oat, wheat, millet, canary, Pigeons; speculated: P/NA Hawks, owls, fishing- hemp, clover, beet freshwater fish, other eagles, storks, pelicans prey species Dean and Milton Various, most unidentified Various granivorous birds P Elanus caeruleus, Falco (1988) and biarmicus, Tyto alba; inferences for other raptors Grant et al. (1975) Chamaesyce amplexicaulis Finches (Geospiza) P Unidentified owl Green et al. (2008) Poaceae sp., Lemna disperma, Unidentified small fish NA Australian Pelican Myriophyllum crispatum, (Pelecanus conspicillatus) Nitella sp., Typha dominguensis, Typha orientalis Kurek and Holeksa Secale cereale, Avena sativa Speculated: granivorous P Foxes (Vulpes vulpes), (2015) birds martens (Martes sp.) Lopez-Darias and 39 (buzzard) and 62 (kestrel) European Rabbits NA Common Buzzards (Buteo Nogales (2016) species of mainly weeds, (Oryctolagus cuniculus), buteo), Eurasian Kestrels few fleshy-fruited plants Barbary Ground (Falco tinnunculus) Squirrels (Atlantoxerus getulus), house mice (Mus domesticus), (Gallotia atlantica), unidentified birds Nogales et al. (1996) Plocama pendula; Rubia fruticose Lizards (Gallotia galloti) D Cats (Felis catus) Nogales et al. (1998) Lycium intricatum Lizards (Gallotia atlantica) D Shrikes (Lanius excubitor) Nogales (1999) Lycium intricatum Lizards (Gallotia atlantica) D Shrikes (Lanius excubitor) Nogales et al. (2002) Lycium intricatum Lizards (Gallotia atlantica) D Shrikes (Lanius excubitor); kestrels (Falco tinnunculus) Nogales et al. (2007) Lycium intricatum, Rubia Lizards (Gallotia atlantica) D Shrikes (Lanius fruticosa, Asparagus nesiotes meridionalis); kestrels (Falco tinnunculus) Nogales et al. (2015) Plocama pendula, Rubia Lizards (Gallotia atlantica) D Cats (Felis catus) fruticosa, Juniperus turbinata, Opuntia dilleniid Padilla and Nogales Rubia fruticose Lizards (Gallotia galloti) D Eurasian kestrels (Falco (2009) tinnunculus) Padilla et al. (2012) 78 species (26 sp. in shrike Lizards (Gallotia spp.) D Southern grey shrikes pellets, 76 sp. in kestrel (Lanius meridionalis), pellets) Eurasian kestrels (Falco tinnunculus) Pearson and Ortega Centaurea maculosa Deer mice (Peromyscus D Great Homed Owls (Bubo (2001) maniculatus) virginianus) Rodrıguez et al. at least 12 taxa Speculated: lizards D Grey herons (Ardea cinerea) (2007) Sarasola et al. (2016) Chenopodium album; Panicum Eared doves (Zenaida P Cougars (Puma concolor) bergii; Sorgum bicolor auriculata) Twigg et al. (2009) speculated: Trifolium sp., European rabbits D Red foxes (Vulpes vulpes) Romulea rosea, Poaceae, (Oryctolagus cuniculus) Hypochaeri ssp., Erodium spp. † Primary consumer is D = a seed disperser, P = a seed predator, NA = role in seed dispersal uncertain. digested. Some such seeds survive the digestive destruction and subsequently deposited by the process and remain viable (van der Pijl 1982, carnivore. The effect of secondary consumption Orłowski et al. 2016), but consumption of seed by the carnivore depends on the number of seeds predators by carnivores can improve seed disper- rescued relative to the number of seeds that sal as a larger proportion of seeds are saved from would normally enter the seed bank without

❖ www.esajournals.org 5 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Table 3. Evidence of diploendozoochory (DZ) and its ecological significance. For further detail on the study systems, see Table 2.

Type of Evidence Ecological Relative effect of References study †, ‡ of DZ ‡ significance of DZ § secondary dispersal ‡, ¶

Balgooyen and Moe O; Eg Viable seeds in R2 LDD; SP- Not quantified (1973) Darwin (1859) O; Ef; Eg Viable seeds in F2 & R2 LDD; SP- Not quantified (F2, R2) Dean and Milton O; R; Eg Viable seeds F2 & R2 ADS; SP- Not quantified (1988) (F2, R2) Grant et al. (1975) O; E# Seeds in R2 LDD; SP- Seeds only found in 2 samples; most seeds tested do not float and have no other mechanism to move between islands Green et al. (2008) O; Eg (F2) Viable seeds in F2 LDD; CFH 4 times more diaspores in pelican feces than in samples collected from other waterbirds (primary dispersers) Kurek and Holeksa O; Eg (F2) Seeds in R2, some co- SP- Not quantified, but only 1.1% of (2015) occurring with samples contained seeds and GR of feathers, few seeds seeds was very low viable Lopez-Darias and O; Eg (R2) Co-occurring viable LDD Not quantified; low GR (overall 5%, Nogales (2016) seeds and animal range: 0-34.7%) remains in R2 Nogales et al. (1996) O; R Co-occurring seeds and LDD Low damage to seeds; 18.9 (Plocama animal matter in F2 pendula) and 4.8 (Rubia fruticosa) seeds/ dropping, only 0.2 and 0.01 in cats; 66.5% and 80.5% of F1 have seeds, only 7.5 and 3.5% of F2; number of seeds passing through cats is low; possible seed dehydration due to slow F2 breakdown Nogales et al. (1998) O, Eg Viable seeds in R2 LDD; ADS; GR+ Higher GR from R2 (64.3%) than (Pre, F1, R2) from F1 (49.5%) or Pre seeds (54.3%). Nogales (1999) O; Eg (Pre, Co-occurring viable LDD; ADS; GR+ Higher GR from R2 (64.3%) than F1, R2) seeds and lizard from F1 (49.5%) or Pre seeds remains in R2 (54.3%); other effects not quantified Nogales et al. (2002) O; Eg (Pre, Co-occurring viable LDD; GR- GR reduced by a third in kestrel R2 F1, R2) seeds and lizard relative to F1 and by 25% relative remains in R2 to Pre seeds; GR slightly higher in shrike R2 relative to F1 and similar to Pre seeds. Nogales et al. (2007) O; Eg (Pre, Co-occurring seeds and LDD; ADS; GR- 99% seeds undamaged by ingestion, F1, R2) animal matter in R2 similar GR for Pre and Lanius R2 seeds; 28.2-95.1% reduction in GR in Falco R2, significant differences in microhabitat and dispersal distance relative to primary disperser Nogales et al. (2015) Ef; Eg (Pre, Experimental GR-; SD- Seed thickness reduced (sometimes F1, F2) comparably to primary disperser effect), lower GR, disruption of native seed dispersal by an invasive predator Padilla and Nogales O; Ef; Eg Study of kestrel feeding LDD; ADS; GR- G1 discarded by carnivore (89% of (2009) (Pre, F1, behavior; seeds and seeds) have the same GR as Pre G1, R2) lizard remains in R2 seeds; seeds in F1 have significantly lower GR

❖ www.esajournals.org 6 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Table 3. Continued.

Type of Evidence Ecological Relative effect of References study †, ‡ of DZ ‡ significance of DZ § secondary dispersal ‡, ¶

Padilla et al. (2012) O; Eg (Pre, Co-occurring viable LDD; ADS; GR+ Shrikes improve GR of Rubia fruticosa F1, R2) seeds and lizard (95% in shrike R2 compared to remains in R2 80.5% in Pre and 83.0% in lizard F1); no effect on germination of most plant species Pearson and Ortega O; Eg (R2) Co-occurring seeds and IS+ Not quantified but very low GR (2001) animal matter in (<1%) from R2 feces, one seed viable Rodrıguez et al. O; Eg (R2) Co-occurring seeds and GR-; IS- Complete disruption of germination; (2007) animal matter in R2 number of seeds in pellets considered small Sarasola et al. (2016) O; Eg Co-occurring seeds and LDD; SP-; IS+ GR similar to seeds in dove (G, F2) animal matter in F2 gizzards, estimated that cougars could disseminate up to 5000 seeds/ km2/ year Twigg et al. (2009) O; Eg Viable seeds in F2, LDD; IS+ Not quantified (F1, F2) some co-occurring with fur † Type of study: O = observation, R = review, Ef = feeding experiment, Eg = germination experiment; ‡ The different dispersal stages: “Pre” = diaspores or seeds collected pre-dispersal from the plant or from the ground below the plant; “F1” = seeds recovered from feces of primary disperser, “G1” = seeds recovered from the gizzard or intestine of pri- mary disperser; “F2” seeds recovered from feces of secondary disperser; “R2” = seeds recovered from regurgitate/pellets of sec- ondary disperser. § Ecological significance of diploendozoochory (LDD = Long Distance Dispersal; ADS = Alternative deposition sites; CFH = Continuity in fragmented habitat; SP = Seed predation; SD = Seed dispersal; GR = Germination rate; IS = Spreading of invasive plant species; “À” = reduction in SP, SD, GR or IS; “+” = improvement in SP, SD, GR or IS). ¶ GR = Germination rate. # Flotation experiment. carnivore involvement (Fig. 1). Dean and Milton lower for carnivore-dispersed seeds than for (1988) extrapolated that a single raptor might seeds dispersed by the primary mechanism, disperse thousands of seeds annually, based given that near-complete seed destruction would on the prey intake rate and the number of be expected from the seed predator. Similarly, seeds in the guts of each prey item. Sarasola et al. positive effects are expected if a carnivore inter- (2016) estimated that cougars could disseminate cepts an ineffective disperser that has a negative 5000 seeds/km2 annually by intercepting seed or negligible effect on seed germination success, predation by their main granivorous prey, eared and treatment by the carnivore improves germi- doves. These numbers are only relevant, however, nation success. On the contrary, if a carnivore in the context of the total number of seeds pro- intercepts dispersal by an effective primary seed duced by plants in the area, and their fates without disperser, and/or causes a decline in germination carnivores (Fig. 1; Appendix S1). These proportions success, the dispersal success declines (Fig. 1, were not estimated in the above two studies and Table 1). have yet to be assessed in any study of diploendo- zoochory (see calculations for a potential scenario IMPLICATIONS OF SECONDARY SEED DISPERSAL in the Appendix S1). A step in this direction was BY CARNIVORES taken by Culot et al. (2015), who attempted to quantify the relative importance of secondary dis- Human effects: loss of habitat and species persal at the microscale by dung beetles, and a sim- Diploendozoochory may influence plant ilar approach might also prove useful for the study adaptability to human-altered landscapes and of the phases of diploendozoochory. resilience in the face of changing community Where the carnivore intercepts seed predation, structures through increased seed dispersal dis- positive effects would, nevertheless, be expected tance, reaching of alternative habitats, or seed even if overall germination success were slightly germination success. Long-distance dispersal is

❖ www.esajournals.org 7 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Fig. 1. Possible seed fates in systems with diploendozoochory. (A) A plant that relies on animal vectors for dis- persal (e.g., fruit-bearing plants) produces a set of seeds. These may remain unconsumed (and therefore less likely to germinate) or be consumed by the primary seed disperser. Seeds deposited by this seed disperser will germinate/ mature at some rate to produce new plants. Some primary seed dispersers will be consumed by the secondary seed disperser (carnivore). Seeds that enter this fate are deposited by the secondary seed disperser and germinate/mature at some rate into new plants. (B) A plant with no obvious adaptations for zoochory produces a set of seeds. Seeds that are not consumed spread via the primary dispersal mechanism (e.g., wind) and germinate/mature at some rate to produce new plants. Some seeds are consumed by a seed predator and destroyed with little chance of germina- tion. Through predation of the seed predator, the secondary disperser diverts some seeds from this fate and instead deposits them. These seeds germinate/mature at some rate into new plants. Illustrations by Kate Broadley. of increasing importance for ecosystem resilience flow among populations (Bacles et al. 2006). in the face of environmental changes such as Escalating habitat loss, fragmentation, and local and climate change by extinctions due to human actions can disrupt dis- facilitating rapid dispersal among disconnected persal pathways via declines in the abundance, areas (Nathan et al. 2008) and by boosting gene species richness, and shifting ranges of primary

❖ www.esajournals.org 8 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL. dispersers (Michalski and Peres 2005, Farwig and carnivores might thus have an impact on the Berens 2012, Beaune et al. 2013, Caplat et al. efficacy of diploendozoochory, although further 2016) and carnivores that act as secondary dis- research is needed to confirm the role of top car- persers (Crooks and Soule 1999). The importance nivores in seed dispersal. of carnivores as dispersal agents might increase Carnivores could also play a role in stabilizing in fragmented habitats because they may facili- changes in community structure via their prey tate plant species’ gene flow between fragments selection. A study conducted in the Democratic due to their larger ranges and often broader Republic of Congo found that all seed dispersers habitat use (Carlo and Morales 2016), although in the system were hunted by humans, whereas the relative importance of the carnivore depends very few seed predators experienced hunting on species-specific responses to fragmentation pressure (Beaune et al. 2013). Such bias in har- (Crooks 2002). Some carnivores could even be vest may result in reduced zoochory and thus creating seed corridors between habitat frag- potentially disrupt ecosystem function by alter- ments when they defecate more often on linear ing plant communities. Likewise, a study in Bra- features such as trails (Suarez-Esteban et al. zil found frugivores to be the most integral 2013). Seed dispersal by carnivores thus has the group for ecological network structure, yet also potential to influence the magnitude of detrimen- the most threatened by extinction (Vidal et al. tal effects of habitat loss and fragmentation on 2014). Processes such as these can alter the rela- plant communities. tive abundance of functional groups (e.g., higher Secondary dispersal by carnivores has the losses of seed dispersers vs. seed predators) and potential to effectively increase the potential speed density-dependent prey selection by carnivores of plant movements (see Naoe et al. 2016) and col- could buffer the impact on plant dispersal by tar- onization of vacant habitat. This may become an geting the more abundant seed predators over increasingly important process at the leading edge mutualistic seed dispersers. Further research is of a shifting range due to climate change. The needed to assess the effects of carnivores on the potential involvement of carnivores in seed dis- resilience of plant communities. persal processes could shift the predicted out- The structure and function of ecosystems can comes of distribution models (Thuiller 2004) and be influenced by top-down influences such as conservation plans due to their specific role in trophic cascades (Schmitz et al. 2004), and the intercepting seed predation and dispersal, as well involvement of carnivores in shaping the distri- as in transporting seeds to novel locations (see bution and abundance of plants through also Higgins and Richardson 1999, Caplat et al. diploendozoochory provides additional insights 2016, Estrada et al. 2016). As seed dispersal is a into complex community-level interactions and key mechanism determining whether plants will ecosystem functioning. Carnivores can also have be able to shift their ranges to match changing cli- indirect effects on seed dispersal when seed dis- mate conditions (Higgins and Richardson 1999, persal behavior of the primary disperser (or seed Chen et al. 2011, Corlett and Westcott 2013), a predator) is altered by carnivore presence (Sun- thorough understanding of dispersal mechanisms yer et al. 2013, Steele et al. 2015). is required to predict plant responses to climate change (Cain et al. 2000, Naoe et al. 2016). Invasive plants The structure of carnivore communities may Carnivores could facilitate invasions by allow- play a role in determining the utility of sec- ing invasive plants to disperse over a greater dis- ondary seed dispersal. With the exception of the tance than they would reach by other means, or cougar (Puma concolor), all documented cases of by transporting them to a novel location. It has diploendozoochory to date (Table 2) involved a been shown that long-distance dispersal can lead mesocarnivore. Mesocarnivore release has been to faster rates of spread (Higgins and Richardson well documented where top carnivores have 1996, 1999). A decreased time to germination been removed from the landscape by human resulting from a double digestive process may interventions (Prugh et al. 2009). Globally, a loss also promote rapid regeneration of invasive in North America (Ripple et al. 2014) and recolo- plants, which also tend to have adaptations that nization in Europe (Chapron et al. 2014) of top increase their rate of reproduction (Rejmanek and

❖ www.esajournals.org 9 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Richardson 1996, Kolar and Lodge 2001). In at LITERATURE REVIEW least two cases studied so far, invasive or pioneer plants’ viable seeds were indirectly dispersed by To assess the evidence from empirical research carnivores that consumed seed predators (Pearson on the significance of diploendozoochory on plant and Ortega 2001, Sarasola et al. 2016, see also dispersal and broader ecological and conservation Twigg et al. 2009). implications, we conducted a literature review to identify studies that have observed or empirically Potential for plant adaptation to diplochory tested aspects of diploendozoochory. The search Dean and Milton (1988) suggested that some was done in Google Scholar in June 2016 using plants may be selected to promote, or at least not the search terms diploendozoochory; diplochory; prevent, seed predation when secondary disper- polychory; double endozoochory; secondary seed sal by raptors functions as an effective dispersal dispersal; indirect seed dispersal; indirect disper- mechanism. They proposed that this may explain sal; two-phase dispersal; two-stage dispersal; and why many seeds consumed by granivorous birds multi-phase dispersal. From these articles and show no obvious adaptations to alternative references therein, we identified sources that means of dispersal. For instance, thick seed coats empirically addressed secondary seed dispersal might develop to facilitate long-distance disper- by carnivores. For those studies, we identified the sal via a long retention time in a far-ranging ani- plant, primary seed consumer, and carnivore mal vector, such as in migrating raptors. This involved. We also recorded any attempts to quan- hypothesis might be tested by comparing seed tify the relative significance of the second, carni- coat thickness or other relevant adaptations on vore-facilitated dispersal stage for the plant’s island populations, where thicker coats would be dispersal success. For each study, we collected expected on the islands due to increased disper- information on the effects of diploendozoochory sal requirements from the mainland. Indeed, on seed viability or germination success, dispersal some evidence suggests that island populations over long distances or to novel environments, and have more thick-coated seeds relative to main- whether the primary seed consumer was likely to land (van der Pijl 1982, Vargas et al. 2015). be a mutualistic seed disperser or a seed predator, Nathan et al. (2008) suggested that not only seed where these details were reported. We then morphology but also fruiting phenology might inferred any likely broader implications of each evolve to match migration schedules of potential study from these variables as well as other charac- dispersers. Plants are most likely to develop teristics of the plants, dispersers, and the ecosys- adaptations to certain dispersal pathways when tems involved. the dispersal vectors and dispersion pathways are quite fixed; that is, certain primary and sec- Evidence of secondary seed dispersal by ondary dispersers handle the majority of the carnivores plant’s seeds. If multiple vectors with different All but one of the 19 studies that reported retention times or habitat selection are involved, diploendozoochory (Table 2) inferred the mecha- a range of seed types might be expected to nism from the discovery of seeds in the feces or evolve (van der Pijl 1982, Cheptou et al. 2008, regurgitates of carnivores that were assumed to Nathan et al. 2008). Disruption of dispersal path- not purposely consume seeds (Table 3). The one ways due to extinctions may have negative exception took a purely experimental approach implications for such potentially evolved traits under laboratory conditions (Nogales et al. 2015). (Vander Wall and Longland 2004). If the costs of The co-occurrence of seeds and remains of poten- diplochory (Cheptou et al. 2008, Nathan et al. tial primary seed consumers in the secondary dis- 2008) are high, owing, for example, to poor seed perser’s excrements (11 studies) was generally deposition sites or dispersal out of suitable habi- accepted as evidence of diploendozoochory tat zone, counter-selection might be expected. (Table 3). The primary consumer of seeds was Modeling exercises on the potential for plant with certainty a seed predator (granivorous birds) adaptations to diplochory as well as empirical in four studies, and 10 studies involved a primary data on recruitment rates are needed to clarify seed disperser (lizards, a mouse, and a rabbit). the adaptive potential of plants to diplochory. The remaining studies were not able to identify

❖ www.esajournals.org 10 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL. the primary disperser with certainty, or it was disperser used alternative habitats or deposited unclear whether the primary seed consumer was a seedsinnovellocationsinsixstudies(Table3). seed predator or a mutualistic disperser (Table 2). Secondary dispersal by a carnivore also con- Studies of diploendozoochory thus far (Table 2) tributed to the spreading of an invasive or alien represent a limited range of taxonomic groups. plant species in three studies and disrupted their The number of plant species that disperse seeds spread in one study. Diploendozoochory resulted via diploendozoochory is unknown, as many in the disruption of seed predation in six studies studies identified only some of the species present and the maintenance of connectivity between habi- among the discovered seeds (range of identified tat patches in one study. The disruption of the nat- plant species: 0–78 species or 12 taxa; Table 2). ural dispersal by an invasive secondary disperser Several weeds, grains, and fleshy-fruited plants was found in one study (Table 3). Overall, diploen- were among the identified plant species. Of pri- dozoochory thus influenced seed dispersal efficacy mary seed consumers, only doves, finches, a rab- or potentially the broader plant distribution or bit, three rodent, and two lizard species have been communities, but the study designs and the studied in a natural setting (see also experiments emerging patterns are currently too diverse to per- with dead fish by Darwin 1859); in terms of sec- mit definitive conclusions about the exact signifi- ondary dispersal, two feline, one canine, and one cance of the phenomenon across systems. mustelid, as well as several raptors (11 avian spe- cies identified, but exact number of species is FUTURE DIRECTIONS unknown due to unspecified species in Darwin (1859) and Grant et al. (1975)), have been studied. While the circumstantial evidence and the pro- jected potential frequency of secondary dispersal Potential significance of diploendozoochory events suggest that diploendozoochory is not Overall, there is evidence from the majority of uncommon, systematic research into diploendo- these studies that secondary seed dispersal by zoochory is needed to better understand the phe- predators can have an influence on the compo- nomenon and its overall ecological significance. nents that influence seed dispersal efficiency: Ger- While carnivores may have a significant influence mination success, dispersal distance or the types on seed dispersal patterns under the conditions of habitats reached, and the ecological implica- we have described in Table 1, evidence should be tions of these studies varied broadly. Carnivores carefully weighed before assuming diploendozoo- improved seed germination success relative to chory is relevant ecologically or otherwise (e.g., seeds unprocessed by a secondary disperser in suggested ecosystem services, Sarasola et al. 2016, eight studies and reduced germination success in see also Appendix S1). Continuing developments three studies, while two studies observed no in methodology, such as the use of DNA barcod- change in germination success (Table 3). None of ing, molecular “log books,” stable isotopes, and the studies directly tested the germination of radio-active tags along with a mechanistic vector- seeds consumed by the primary consumer which centered approach to study seed dispersal (Bullock was subsequently consumed by the carnivore, et al. 2006, Nathan et al. 2008, Fordham et al. although an attempt at testing this was made by 2014, Gonzalez-varo et al. 2014, Culot et al. 2015, Nogales et al. (2015). Seven studies compared ger- Herrmann et al. 2016, Naoe et al. 2016), make mination rates between seeds that were not con- aspects of the phenomenon more readily testable. sumed by any animal, seeds that passed through Future studies should explicitly address the poten- a primary disperser (or were contained in their tial mechanisms and implications of the phe- intestines, Twigg et al. 2009, or gizzards, Sarasola nomenon among systems. In the face of the et al. 2016), and seeds that passed through a sec- changing landscapes, it would be especially ondary disperser (Table 3). None of the studies important to look for further evidence of the influ- followed germinated seeds through to maturity to ence of diploendozoochory on metapopulation assess recruitment rates. dynamics in fragmented environments, on the role Long-distance dispersalwasinferredasthe of carnivores in disrupting or facilitating the most significant ecological consequence of diploen- spread of rare or invasive species and recolonizing dozoochory in 14 studies, and the secondary degraded habitats.

❖ www.esajournals.org 11 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

To rigorously assess the significance of diplo- potential secondary (as well as primary) dis- chory, the relative proportions of seeds with vari- persers of up to 46 plant genera, although the ous fates must be determined (Culot et al. 2015, effects on seed viability is unknown and the exis- Fig. 1; Appendix S1). For a controlled experimen- tence and frequency of diplochory remains to be tal approach, known quantities of seeds from confirmed (Platt et al. 2013). Diploendozoochory known plants would need to be fed to primary may also occur in systems where consumers (dispersers and seed predators), and a and fish function as primary dispersers (Darwin proportion of these consumers then exposed to 1859, van der Pijl 1982, Pollux 2011). These ani- predation by secondary dispersers. The droppings mals are often eaten whole and consumed in and pellets of each disperser, as well as the intact, large quantities by birds and mammals, some- unconsumed seeds, would then be collected and times followed by long-distance movements by the viability of seeds therein assessed via germina- the secondary disperser (Green and Figuerola tion experiments (preferably on actual site of 2005). Some taxa, such as omnivorous , deposition). Furthermore, recruitment of the seed- may include species that serve as both a primary lings should be monitored because the survival of and a secondary disperser by consuming small the seedlings to maturity and their subsequent prey items, such as along with fruit (e.g., reproduction are relevant for assessing the evolu- mouse lemurs; Dammhahn and Kappeler 2008) tionary significance of the different dispersal syn- and falling prey to raptors, snakes, and mam- dromes (Schupp and Jordano 2010). malian carnivores (Rasoloarison et al. 1995). Effects of diploendozoochory on seed fates can be most readily quantified when a single animal CONCLUSIONS species is responsible for the majority of primary dispersal or seed removal of a given plant species, Several authors have suggested that polychory and a single carnivore is responsible for the major- is likely a much more common phenomenon ity of the mortality of the primary vector. In such than has been previously assumed (Ozinga et al. systems, the effect of the carnivore is also likely 2004, Vander Wall and Longland 2004) and can the highest, and plant adaptations to diploendo- be more beneficial for the dispersing plant than zoochory might therefore be expected to evolve. single-phase dispersal (Vander Wall and Long- Relatively simple systems with limited species land 2004). While these studies have largely con- interactions would thus likely prove most fruitful centrated on abiotic vectors and short-distance, for the study of the phenomenon (such as the rap- second-phase dispersal by invertebrates and tors, lizards, and Lycium fruit studied in an island scatter-hoarding rodents, the impact of carni- ecosystem; Nogales 1999) as seed fates of com- vores may be similarly important, particularly in plete seed cohorts could potentially be followed discontinuous habitats. Secondary dispersal by throughout the dispersal pathway. carnivores is by no means exclusive of the types primary dispersers with reasonably small-ranged of diplochory defined by Vander Wall and Long- carnivores, such as insectivorous , might land (2004); rather, it is very likely that further also prove useful because of their relatively short seed transport by ants, dung beetles, or scatter- dispersal distances. However, multiple taxa hoarding rodents often occurs after seeds are should eventually be studied to determine how deposited by the secondary disperser. widespread the phenomenon really is and its Our framework provides guidelines for future potential management and conservation conse- research, with predictions that should aid in tar- quences (Levey et al. 2002). geting systems that are likely to be most affected Although recent studies of two-phase dispersal by carnivore involvement in seed dispersal. In have focused on raptors and mammalian carni- addition to disrupting heavy seed predation pres- vores as well as their prey (typically small mam- sure, carnivores that intercept large proportions of mals, birds, and reptiles), numerous other taxa a plant population’s seeds and significantly alter disperse seeds and eat seed dispersers or seed the germination or recruitment success of seeds predators, suggesting multiple pathways of two- relative to the primary disperser will most likely phase dispersal have yet to be identified. For be an important ecological force for the plant example, crocodilians are thought to serve as species and, possibly, the community structure.

❖ www.esajournals.org 12 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Another important role for far-ranging secondary Buitron-Jurado, G., and N. Ramırez. 2014. Dispersal dispersers may involve long-distance dispersal or spectra, size and the importance of endo- gene flow between remote populations or habitat zoochory in the equatorial Andean montane for- fragments. While carnivore effects will likely be ests. Flora: Morphology, Distribution, Functional – small in most systems, such circumstances may Ecology of Plants 209:299 311. Bullock, J. M., K. Shea, and O. Skarpaas. 2006. Measur- indeed result in secondary seed dispersers signifi- ing plant dispersal: an introduction to field meth- cantly influencing plant range shifts, dispersal fi ods and experimental design. Plant Ecology success, tness, and potentially species viability. 186:217–234. It is currently unknown how important the Cain, M. L., B. G. Milligan, and A. E. Strand. 2000. phenomenon is ecologically, but given its poten- Long-distance seed dispersal in plant populations. tially vast prevalence and the possible implica- American Journal of 87:1217–1227. tions, it is possible that ignoring it could impair Caplat, P., et al. 2016. Looking beyond the mountain: the interpretation of broad ecological patterns or dispersal barriers in a changing world. Frontiers in hinder conservation efforts. Considering diploen- Ecology and the Environment 14:261–268. dozoochory as a part of the dispersal mechanism Carbone, C., G. Cowlishaw, N. J. B. Isaac, and J. M. of plants can potentially improve modeling out- Rowcliffe. 2005. How far do animals go? Determi- nants of day range in mammals. American Natu- comes for range shifts due to climate change, or ralist 165:290–297. help explain current plant distributions, as his- Carlo, T. A., and J. M. Morales. 2016. Generalist birds torical effects of carnivores (or other large-bodied promote tropical forest regeneration and increase fl animals; Pakeman 2001) may have in uenced plant diversity via rare-biased seed dispersal. Ecol- plant movement rates. Where the secondary dis- ogy 97:1819–1831. perser facilitates different dispersal processes Caughlin, T., and J. Ferguson. 2013. The importance of than are accomplished by other means of disper- long distance seed dispersal for the demography sal, carnivore involvement may have important and distribution of a canopy tree species. Ecology consequences for the spread of invasive plant 95:952–962. Chapron, G., P. Kaczensky, J. D. C. Linnell, M. von species, as well as the ability of plants to adapt to habitat loss and changing climatic conditions. Arx, D. Huber, H. Andren, J. V. Lopez-bao, and M. Adamec. 2014. Recovery of large carnivores in Where such relationships exist, the extinction or Europe’s modern human-dominated landscapes. decline of involved species can affect multiple Science 346:1517–1519. trophic levels and disrupt ecosystem functions. Chen, I., J. K. Hill, R. Ohlemuller,€ D. B. Roy, and C. D. Thomas. 2011. Rapid range shifts of species of ACKNOWLEDGMENTS climate warming. Science 333:1024–1026. Cheptou, P.-O., O. Carrue, S. Rouifed, and A. Cantarel. We thank Walter Carson and two anonymous 2008. Rapid evolution of seed dispersal in an urban reviewers for their comments on an earlier draft and environment in the weed Crepis sancta. Proceedings Richard Schneider, Eric Neilson, and Darcy Doran- of the National Academy of Sciences USA 105: Myers for their contributions to the early stages of the 3796–3799. paper. The authors have no conflicts of interest. Corlett, R. T., and D. A. Westcott. 2013. Will plant movements keep up with climate change? Trends LITERATURE CITED in Ecology and Evolution 28:482–488. Crooks, K. R. 2002. Relative sensitivities of mammalian Bacles, C. F. E., A. J. Lowe, and R. A. Ennos. 2006. carnivores to habitat fragmentation. Conservation Effective seed dispersal across a fragmented land- Biology 16:488–502. scape. Science 311:628. Crooks, K., and M. Soule. 1999. Mesopredator release Balgooyen, T. G., and L. M. Moe. 1973. Dispersal of and avifaunal extinctions in a fragmented system. grass fruits-an example of endornithochory. Ameri- Nature 400:563–566. can Midland Naturalist 90:454–455. Culot, L., M. C. Huynen, and E. W. Heymann. 2015. Beaune, D., F. Bretagnolle, L. Bollache, G. Hohmann, Partitioning the relative contribution of one-phase M. Surbeck, and B. Fruth. 2013. Seed dispersal and two-phase seed dispersal when evaluating strategies and the threat of defaunation in a Congo seed dispersal effectiveness. Methods in Ecology forest. and Conservation 22:225–238. and Evolution 6:178–186.

❖ www.esajournals.org 13 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Dammhahn, M., and P. M. Kappeler. 2008. Compara- editors. Plant–animal interactions: an evolutionary tive feeding ecology of sympatric Microcebus approach. Wiley-Blackwell, Oxford, UK. berthae and M. murinus. International Journal of Jordano, P. 2000. Fruits and frugivory. Pages 125–166 Primatology 29:1567–1589. in M. Fenner, editor. Seeds: the ecology of regener- Darwin, C. R. 1859. On the origin of species by means ation in plant communities. Second edition. CABI of natural selection, or the preservation of favoured Publishing, Wallingford, UK. races in the struggle for life. First edition. John Kolar, C. S., and D. M. Lodge. 2001. Progress in inva- Murray, London, UK. sion biology: predicting invaders. Trends in Ecol- Dean, W. R. J., and S. J. Milton. 1988. Dispersal of seeds ogy and Evolution 16:199–204. by raptors. African Journal of Ecology 26:173–176. Kurek, P., and J. Holeksa. 2015. Grains in the diets of Estrada, A., I. Morales-Castilla, P. Caplat, and R. Early. medium-sized carnivores – A case of diplochory? 2016. Usefulness of species traits in predicting range Polish Journal of Ecology 63:1–9. shifts. Trends in Ecology and Evolution 31:190–203. Lesser, M. R., and S. T. Jackson. 2013. Contributions of Farwig, N., and D. G. Berens. 2012. Imagine a world long-distance dispersal to population growth in without seed dispersers: a review of threats, conse- colonising Pinus ponderosa populations. Ecology quences and future directions. Basic and Applied Letters 16:380–389. Ecology 13:109–115. Levey, D. J., W. R. Silva, and M. Galetti. 2002. Seed Fordham, D. A., B. W. Brook, C. Moritz, and dispersal and frugivory: ecology, evolution and D. Nogues-Bravo. 2014. Better forecasts of range conservation. CABI Publishing, Wallingford, UK. dynamics using genetic data. Trends in Ecology Lopez-Darias, M., and M. Nogales. 2016. Raptors as and Evolution 29:436–443. legitimate secondary dispersers of weed seeds. Ibis Gonzalez-varo, J. P., J. M. Arroyo, and P. Jordano. 158:428–432. 2014. Who dispersed the seeds? The use of DNA Michalski, F., and C. A. Peres. 2005. Anthropogenic barcoding in frugivory and seeds dispersal studies. determinants of and carnivore local extinc- Methods in Ecology and Evolution 5:806–814. tions in a fragmented forest landscape of southern Grant, P. R., J. N. M. Smith, B. R. Grant, I. J. Abbott, Amazonia. Biological Conservation 124:383–396. and L. K. Abbott. 1975. Finch numbers, owl preda- Naoe, S., I. Tayasu, Y. Sakai, T. Masaki, K. Kobayashi, tion and plant dispersal on Isla Daphne Major, A. Nakajima, Y. Sato, K. Yamazaki, H. Kiyokawa, Galapagos. Oecologia 19:239–257. and S. Koike. 2016. Mountain-climbing bears Green, A. J., and J. Figuerola. 2005. Recent advances in protect cherry species from global warming the study of long-distance dispersal of aquatic through vertical seed dispersal. Current Biology invertebrates via birds. Diversity and Distributions 26:R315–R316. 11:149–156. Nathan, R., F. M. Schurr, O. Spiegel, O. Steinitz, Green, A. J., K. M. Jenkins, D. Bell, P. J. Morris, and A. Trakhtenbrot, and A. Tsoar. 2008. Mechanisms R. T. Kingsford. 2008. The potential role of water- of long-distance seed dispersal. Trends in Ecology birds in dispersing invertebrates and plants in arid and Evolution 23:638–647. Australia. Freshwater Biology 53:380–392. Nogales, M. 1999. Some ecological implications of the Herrmann, J. D., T. A. Carlo, L. A. Brudvig, E. I. broadening habitat and trophic niche of terrestrial Damschen, N. M. Haddad, D. J. Levey, J. L. Orrock, vertebrates in the Canary Islands. Ecologia Illes and J. J. Tewksbury. 2016. Connectivity from a 1:67–82. different perspective: comparing seed dispersal Nogales, M., I. Castaneda,~ M. Lopez-Darias, F. M. kernels in connected vs. unfragmented landscapes. Medina, and E. Bonnaud. 2015. The unnoticed effect Ecology 97:1274–1282. of a top predator on complex mutualistic ecological Higgins, S. I., and D. M. Richardson. 1996. A review of interactions. Biological Invasions 17:1655–1665. models of alien plant spread. Ecological Modelling Nogales, M., J. D. Delgado, and F. M. Medina. 1998. 87:249–265. Shrikes, lizards and Lycium intricatum (Solanaceae) Higgins, S. I., and D. M. Richardson. 1999. Predicting fruits: a case of indirect seed dispersal on an ocea- plant migration rates in a changing world: the role nic island (Alegranza, Canary Islands). Journal of of long-distance dispersal. American Naturalist Ecology 86:866–871. 153:464–475. Nogales, M., R. Heleno, A. Traveset, and P. Vargas. Howe, H. F., and J. Smallwood. 1982. Ecology of seed 2012. Evidence for overlooked mechanisms of dispersal. Annual Review of Ecology and System- long-distance seed dispersal to and between ocea- atics 13:201–228. nic islands. New Phytologist 194:313–317. Hulme, P. E., and C. W. Benkman. 2002. Granivory. Nogales, M., F. M. Medina, and A. Valido. 1996. Indi- Pages 132–154 in C. M. Herrera and O. Pellmyr, rect seed dispersal by the feral cats Felis catus in

❖ www.esajournals.org 14 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

island ecosystems (Canary Islands). Ecography Rejmanek, M., and D. M. Richardson. 1996. What attri- 19:3–6. butes make some plant species more invasive? Nogales, M., D. P. Padilla, C. Nieves, J. C. Illera, and Ecology 77:1655–1661. A. Traveset. 2007. Secondary seed dispersal sys- Ripple, W. J., et al. 2014. Status and ecological effects of tems, frugivorous lizards and predatory birds in the world’s largest carnivores. Science 343:1241484. insular volcanic badlands. Journal of Ecology Rodrıguez, A., B. Rodrıguez, B. Rumeu, and 95:1394–1403. M. Nogales. 2007. Seasonal diet of the Grey Heron Nogales, M., V. Quilis, F. M. Medina, J. L. Mora, and Ardea cinerea on an oceanic island (Tenerife, Canary L. S. Trigo. 2002. Are predatory birds effective Islands): indirect interaction with wild seed plants. secondary seed dispersers? Biological Journal of Acta Ornithologica 42:77–87. the Linnean Society 75:345–352. Sarasola, J. H., J. I. Zanon-Mart ınez, A. S. Costan, and Orłowski, G., J. Czarnecka, A. Goławski, J. Karg, and W. J. Ripple. 2016. Hypercarnivorous apex preda- M. Panek. 2016. The effectiveness of endozoochory tor could provide ecosystem services by dispersing in three avian seed predators. Journal of Ornithol- seeds. Scientific Reports 6:19647. ogy 157:61–73. Schaefer, H. M., and G. D. Ruxton. 2011. Plant-animal Ozinga, W. A., R. M. Bekker, J. J. Schaminee, and J. M. communication. OUP, Oxford, UK. van Groenendael. 2004. Dispersal potential in plant Schmitz, O. J., V. Krivan, and O. Ovadia. 2004. Trophic communities depends on environmental condi- cascades: the primacy of trait-mediated indirect tions. Journal of Ecology 92:767–777. interactions. Ecology Letters 7:153–163. Padilla, D. P., A. Gonzalez-Castro, and M. Nogales. Schupp, E. W., and P. Jordano. 2010. Seed dispersal 2012. Significance and extent of secondary seed effectiveness revisited: a conceptual review. New dispersal by predatory birds on oceanic islands: Phytologist 188:333–353. the case of the Canary archipelago. Journal of Ecol- Steele, M. A., G. Rompré, J. A. Stratford, H. Zhang, ogy 100:416–427. M. Suchocki, and S. Marino. 2015. Scatterhoarding Padilla, D. P., and M. Nogales. 2009. Behavior of kes- rodents favor higher predation risks for cache trels feeding on frugivorous lizards: implications sites: The potential for predators to influence the for secondary seed dispersal. Behavioral Ecology seed dispersal process. Integrative Zoology 10: 20:872–877. 257–266. Pakeman, R. J. 2001. Plant migration rate and seed dis- Suarez-Esteban, A., M. Delibes, and J. M. Fedriani. persal mechanisms. Journal of Biogeography 2013. Barriers or corridors? The overlooked role of 28:795–800. unpaved roads in endozoochorous seed dispersal. Pakeman, A. R. J., G. Digneffe, and J. L. Small. 2002. Journal of Applied Ecology 50:767–774. Ecological correlates of endozoochory by herbi- Sunyer, P., A. Muñoz, R. Bonal, and J. M. Espelta. 2013. vores. Functional Ecology 16:296–304. The ecology of seed dispersal by small rodents: A Pearson,D.E.,andY.K.Ortega.2001.Evidenceof role for predator and conspecific scents. Functional an indirect dispersal pathway for Spotted Ecology 27:1313–1321. Knapweed, Centaurea maculosa,seeds,viaDeer Thuiller, W. 2004. Patterns and uncertainties of species’ Mice, Peromyscus maniculatus, and Great Horned range shifts under climate change. Global Change Owls, Bubo Virginianus. Canadian Field-Natural- Biology 10:2020–2027. ist 115:354. Traveset, A., A. W. Robertson, and J. Rodrıguez-Perez. Platt, S. G., R. M. Elsey, H. Liu, T. R. Rainwater, J. C. 2007. A review on the role of endozoochory in seed Nifong, A. E. Rosenblatt, M. R. Heithaus, and F. J. germination. Pages 78–101 in A. J. Dennis, E. W. Mazzotti. 2013. Frugivory and seed dispersal by Schupp, R. J. Green, and D. A. Westcott, editors. crocodilians: An overlooked form of saurochory? Seed dispersal: theory and its application in a Journal of Zoology 291:87–99. changing world. CABI Publishing, Wallingford, UK. Pollux, B. J. A. 2011. The experimental study of seed Twigg, L. E., T. J. Lowel, and G. R. Martin. 2009. The dispersal by fish (ichthyochory). Freshwater Biol- presence and implications of viable seed in the ogy 56:197–212. faeces of invasive free-ranging European rabbits Prugh, L. R., C. J. Stoner, C. W. Epps, W. T. Bean, W. J. and red foxes. Pacific Conservation Biology 15: Ripple, A. S. Laliberte, and J. S. Brashares. 2009. 158–170. The rise of the mesopredator. BioScience 59:779–791. van der Pijl, L. 1982. Principles of dispersal in higher Rasoloarison, R., B. Rasolonandrasana, J. Ganzhorn, plants. Springer-Verlag, Berlin, Germany. and S. Goodman. 1995. Predation on vertebrates in Vander Wall, S. B., and W. S. Longland. 2004. Diplo- the Kirindy forest, western Madagascar. Ecotropica chory: Are two seed dispersers better than one? 1:59–65. Trends in Ecology and Evolution 19:155–161.

❖ www.esajournals.org 15 February 2017 ❖ Volume 8(2) ❖ Article e01685 € € € SYNTHESIS & INTEGRATION HAMALAINEN ET AL.

Vargas, P., Y. Arjona, M. Nogales, and R. H. Heleno. Vidal,M.M.,E.Hasui,M.A.Pizo,J.Y.Tamashiro, 2015. Long-distance dispersal to oceanic islands: W.R.Silva,andP.R.Guimaraes.~ 2014. Frugivores success of plants with multiple diaspore specializa- at higher risk of extinction are the key elements of tions. AoB Plants 7:1–9. a mutualistic network. Ecology 95:3440–3447.

SUPPORTING INFORMATION

Additional Supporting Information may be found online at: http://onlinelibrary.wiley.com/doi/10.1002/ecs2. 1685/full

❖ www.esajournals.org 16 February 2017 ❖ Volume 8(2) ❖ Article e01685