Factors influencing repeated seed movements by scatter-hoarding rodents in an alpine forest

Effects of Seed Size and Frequency on Seed Dispersal and Predation by Small Mammals

Frequency-dependent predation is common in predator-prey interactions. Size is an important characteristic of seeds and is crucial in the regeneration stage of plant seeds. However, the frequency dependence of animal predation on seed size has not been reported. In this study, we conducted a field experiment and used different sizes of Liaodong oak (Quercus wutaishanica) seeds to test the frequency dependence of intraspecific seed size selection in rodents. We used the number ratio of large to small seeds as the frequency. The results show that the rate of small seeds being eaten in situ was significantly higher than that of large seeds (p < 0.05). The rates of different-sized seeds being eaten after removal decreased with increasing frequencies, and there was no significant difference between frequencies except for 1:9 and 9:1. The rates of large seeds being scatter-hoarded were significantly higher than those of small seeds at different frequencies (p < 0.05). The eating distances after removal of large seeds were significantly longer than those of small seeds at the same frequencies (p < 0.05). Furthermore, the scatter-hoarding distances of large seeds were significantly longer than those of small seeds at three frequencies (1:9, 3:7, and 9:1) (p < 0.05). That is, rodents consumed more small seeds in situ, dispersed and scatter-hoarded more large seeds, and dispersed large seeds over longer distances. Rodents exhibited a negative frequency dependence for small seeds and a positive frequency dependence for large seeds on being eaten in situ. Moreover, rodents exhibited a negative frequency dependence for large seeds and a positive frequency dependence for small seeds on being eaten after removal and scatter-hoarding. These results reveal the frequency dependence of rodent selection on seed size and provide new insights into animal-mediated seed dispersal and the regeneration of plant populations.

Effects of subtle variation in forest canopy openness on cache pilferage and its implications for forest regeneration

Scatter-hoarding rodents play important roles in plant regeneration and species coexistence in many forest ecosystems. Cache pilferage, the behavior of rodents seeking or relocating seeds cached by other individuals, is ubiquitous during the scatter-hoarding process. The effects of canopy openness on cache pilferage have received considerable attention, most of which have focused on the comparison between full canopy cover and completely open areas, such as forest gaps. However, little attention has been given to whether the subtle variation in forest canopy openness affects cache pilferage, although subtle variation in light environments exists in many forests, especially tropical and subtropical forests, where the overall canopy is large and the forest window is relatively small. Here, we directly tested these questions by simulating 400 artificial caches, each containing one seed from four selected tree species, in a subtropical forest in southwestern China. The overall canopy openness of the forest was relatively small (with a mean value of 11.1%), but subtle spatial variation still existed (ranging from 5.7% to 19.5%). Overall, caches with lower canopy openness were more likely to be pilfered and removed faster, although not all species showed the same pattern. Our study highlights that subtle variation in forest canopy openness, even in a closed primary forest, has significant effects on cache pilferage by rodents, which may influence the following seed germination and forest regeneration processes. Additionally, seedling species composition may further be altered because the canopy effects on cache pilferage are species-specific.

Verified hypotheses on the "nurse" and "burial" effects on introduced Quercus rubra regeneration in a mesic Scots pine forest

A previous study on the encroachment of North American northern red oak Quercus rubra L. into the mesic Scots pine forest (in central Poland) revealed high abundances of seedlings and saplings under shrubs, with lower abundances in open areas or clumps of bilberry Vaccinium myrtillus L. It was unclear whether the regeneration success of Q. rubra is enhanced by the presence of shrubs due to their "nurse effect", and how burying acorns of different sizes in soil or moss affects the survival of oak seeds and seedlings (a "burial effect"). Results of a previous observational study were verified in an experimental study: a pool of 900 large-, medium-, and small-sized acorns was sown under moss cover in open areas and within bilberry clumps and in soil under shrubs in 2018 and monitored for 3 years in natural conditions. The majority of sown acorns were lost, mainly due to acorn pilferage, lack of germination and the death of sprouting acorns. However, acorn and seedling survival depended significantly on acorn size and differed among the microsites studied. Viable seedlings were twice as likely to develop from large- and medium-sized as from small-sized acorns, and they grew mainly from acorns sown under moss cover, confirming a positive "burial effect." Seedling survival was three times higher in bilberry and open areas, than under shrubs; however, seedlings "nursed" by shrubs were less threatened by large ungulates. Only a small part of the pool of sown acorns contributes to the reproductive success of Q. rubra in the mesic Scots pine forest. Microsites characteristic to this type of forest are suitable for northern red oak regeneration; however, bilberry favors acorn survival and germination and early seedling growth, moss cover favors acorn survival and germination, while shrubs protect surviving seedlings from herbivory.

Density-Dependent Seed Predation of Quercus wutaishanica by Rodents in Response to Different Seed States

The predation and/or dispersal of Quercus seeds by rodents play an important role in the creation of the tree species. The present study examined the effects of community habitats on the predation and dispersal of Quercus wutaishanica seeds by rodents. We released seeds with densities set at 2, 4, 8, 16, and 32 seed square meter with litter cover, soil burial, and bare ground in the Liupan Mountains National Nature Reserve in the Ningxia Hui Autonomous Region, northwest China. The results showed that (1) the litter cover and soil burial significantly increased the seed survival probability compared with bare ground treatments, especially the predation in situ (PIS) (p < 0.05). Both the scatter hoarding (SH) and larder hoarding (LH) for litter cover and soil burial were significantly increased compared with bare ground (p < 0.05). (2) The large seeds are preferentially predated after dispersal and their long-distance dispersal (>5 m) was significantly greater than that of small seeds (p < 0.05), while small seeds are more likely to be preyed on in situ or during short-distance dispersal (<3 m). (3) The Q. wutaishanica seed predation by rodents increased at a high density rather than at a low density, indicating a negative density-dependent predation. These findings provide insights into the ecological characteristics of Quercus tree regeneration and shed light on the coexistence between rodents and different-sized seeds.

Seed nutrient content rather than size influences seed dispersal by scatterhoarding rodents in a West African montane forest

Rodents can be important in seed dispersal through their scatterhoarding behaviour, yet, the seed traits that are most influential in seed removal by Afrotropical scatterhoarding rodents remains unclear. Here, we investigated the effect of seed size and nutrient content of four seed species on the scatterhoarding behaviour of rodents in an Afromontane forest, Ngel Nyaki forest, Nigeria. To do this we marked with thread-tags the seeds of Santiria trimera, Beilschmedia mannii, Carapa oreophila and Anthonotha noldeae and observed their fate. We predicted that (1) caching frequency would be higher for larger than smaller seed species; (2) caching frequency would be higher for nutrient-rich than nutrient-poor seeds; (3) larger seeds would be taken across farther distances; and (4) survival of cached seeds would be higher for nutrient-rich seeds. In contrast to studies elsewhere we found no difference in caching probabilities based on seed size, although nutrient-rich (high fat content) seeds had a higher probability of being predated than seeds with lower fat content. Larger and smaller seeds were dispersed over the same distances and nutrient-poor (high fibre content) seeds survived longer in seed caches. Overall, our findings suggest that large, nutrient-rich seed species are less likely to be dispersed by rodents.

Predators and dispersers: Context-dependent outcomes of the interactions between rodents and a megafaunal fruit plant

Many plant species bear fruits that suggest adaptation to seed dispersal by extinct megafauna. Present-day seed dispersal of these megafaunal plants is carried out by rodents, which can act as predators or dispersers; whether this interaction is primarily positive or negative can depend on the context. Here, we parameterized a stochastic model using data from the field and experimental arenas to estimate the effect of rodents on the recruitment of Myrcianthes coquimbensis -an Atacama Desert shrub with megafaunal fruits- and examine whether environmental conditions can alter the sign and strength of these rodent-plant interactions. We show that the outcome of these interactions is context-dependent: in wet conditions seed removal by rodents negatively impacts the recruitment probability of M. coquimbensis; in contrast, in dry conditions, the interaction with rodents increases recruitment success. In all cases, the strength of the effect of rodents on the recruitment success was determined mainly by their role as dispersers, which could be positive or negative. This study demonstrates that by caching seeds, rodents can be effective dispersers of a megafaunal fruit plant, but that the sign and magnitude of their effect on recruitment changes as a function of the environmental context in which the interaction occurs.

Seed removal by scatter-hoarding rodents: the effects of tannin and nutrient concentration

The mutualistic interaction between scatter-hoarding rodents and seed plants have a long co-evolutionary history. Plants are believed to have evolved traits that influence the foraging behavior of rodents, thus increasing the probability of seed removal and caching, which benefits the establishment of seedlings. Tannin and nutrient content in seeds are considered among the most essential factors in this plant-animal interaction. However, most previous studies used different species of plant seeds, rendering it difficult to tease apart the relative effect of each single nutrient on rodent foraging behavior due to confounding combinations of nutrient contents across seed species. Hence, to further explore how tannin and different nutritional traits of seed affect scatter-hoarding rodent foraging preferences, we manipulated tannin, fat, protein and starch content levels, and also seed size levels by using an artificial seed system. Our results showed that both tannin and various nutrients significantly affected rodent foraging preferences, but were also strongly affected by seed size. In general, rodents preferred to remove seeds with less tannin. Fat addition could counteract the negative effect of tannin on seed removal by rodents, while the effect of protein addition was weaker. Starch by itself had no effect, but it interacted with tannin in a complex way. Our findings shed light on the effects of tannin and nutrient content on seed removal by scatter-hoarding rodents. We therefore, believe that these and perhaps other seed traits should interactively influence this important plant-rodent interaction. However, how selection operates on seed traits to counterbalance these competing interests/factors merits further study.