Search

To understand the demographic consequences of seed predation, it is essential to scale-up losses over space and time. We tested how individual tree characteristics, forest attributes, and the local environment affected rates of both pre- and post-dispersal seed predation in Scots pine Pinus sylvestris, in Scotland’s ancient native pinewoods. The sampling strategy was hierarchical, allowing assessment of how the influence of these factors differed with spatial scale. This approach revealed that the fate of a tree’s seeds depended on many factors, at multiple scales. Levels of seed predation were highly variable over space and time. Seed predation by local post-dispersal seed predators (invertebrates and small mammals) varied at the finest spatial scales, whereas mobile avian pre-dispersal seed predators discriminated among seeds at every spatial scale. Tree crown shape affected removal of seed by all seed predators; both crossbills Loxia spp. and generalist avian granivores selected trees with broad crowns. However, trees with narrow crowns were selected by red squirrels Sciurus vulgaris, and the associated tall understory vegetation supported the highest levels of post-dispersal seed predation. Seed chemistry was an effective seed defence that reduced seed loss to all seed predators. In contrast, increases in physical cone defence were ineffective at reducing seed loss and showed a potential trade-off with investment in seed viability. Large seeds carried a survival cost, being preferentially consumed by squirrels and post-dispersal seed predators. All post-dispersal seed predators consumed a higher proportion of seeds at the times and locations where seed fall was greatest.

Why foraging animals sometimes leave patches before consuming all available food items is a key question in behavioural ecology. Abandoning some food would appear to be disadvantageous, yet optimal foraging theory demonstrates that this sometimes optimises energy intake rates. Crossbills Loxia are specialist avian granivores that forage on seeds within the cones of many species of conifer in the northern hemisphere. They often abandon a few seeds within cones they have fed upon. We assessed whether seeds left within Scots Pine Pinus sylvestris cones by foraging crossbills were of lower mass or in particular positions within the cone. Crossbills foraged on long cones which have more seeds per cone and heavier seeds, but avoided the longest cones, which typically have thicker scales that require more energy to open. Cones dropped by foraging crossbills contained over four seeds per cone, representing approximately one fifth of those originally present. Crossbills left mostly small or empty seeds ( 4 mg) were left behind, perhaps mistakenly overlooked during foraging. Such apparent preferential foraging on heavier seeds is probably advantageous, because of the higher energy reward per seed. To directly discriminate between seeds prior to extraction would reduce energy expenditure in foraging. This raises the question of how crossbills could attain this favourable outcome. While cones scales were closed an external cue would be required. After cone scale dehiscence, seeds would be visible to crossbills, allowing them to discriminate visually among seeds and selectively extract heavier seeds, leaving lighter seeds behind within the cone. Dropping cones when few seeds are encountered or as seed mass declines towards the distal scales could be additional components of crossbill foraging strategy.

For plant populations to persist, seedling recruitment is essential, requiring seed germination, seedling survival and growth. Drought and grazing potentially reduce seedling recruitment via increased mortality and reduced growth. We studied these seed–related processes for two species indigenous to the Pamir Mountains of Xinjiang in northwestern China: Saussurea glacialis and Plantago lessingii. Seeds collected from Taxkorgan, Xinjiang, had a viability rate of 15.8% for S. glacialis but 100% for P. lessingii. Of the viable seeds, the highest germination rates were 62.9% for S. glacialis and 45.6% for P. lessingii. In a greenhouse experiment, we imposed a series of stressful conditions, involving a combination of simulated grazing and drought events. These had the most severe impact on younger seedlings. Modelling showed that 89% of S. glacialis mortality was due to early simulated grazing, whereas 80% of P. lessingii mortality was due to early simulated drought. Physiological differences could contribute to their differing resilience. S. glacialis may rely on water storage in leaves to survive drought events, but showed no shifts in biomass allocation that would improve grazing tolerance. P. lessingii appears more reliant on its root system to survive grazing, but the root reserves of younger plants could be insufficient to grow deeper in response to drought. After applying all mortality factors, 17.7 seedlings/parent of P. lessingii survived, while only

1 Despite three decades of intensive research on seed predation, rather little is known about how important a role it plays in plant demography. Here a balance sheet of seed predation is drawn up for Pinus sylvestris, based on mortality caused by its main seed predators. This species is of special interest because trees carry seeds throughout the year and can therefore support specialist as well as generalist seed predators. Also, its regeneration in Scotland is weak and some of its main seed predators (red squirrels, crossbills, wood ants) are themselves threatened. 2 Field data from two P. sylvestris forests were used to estimate rates of predation on seeds due to pre- and post-dispersal predators, including both specialists and generalists. The rates of predation changed systematically over time and were fitted to time-dependent functions. The resulting functions were incorporated into a model, which was solved numerically to determine the fate of a cohort of seeds. 3 On average, approximately 20% of seeds in the cohort were consumed by pre-dispersal seed predators, approximately 25% of seeds by post-dispersal seed predators, and approximately 55% of seeds were still alive at the end of 2 years. This reduces the seed pool for regeneration of P. sylvestris, but is unlikely to be a serious problem for regeneration. 4 Seed survival was most sensitive to the rate at which seeds became invulnerable to further seed predation, and to the rate of seed dispersal from cones on trees. Seed survival was less sensitive to seed predation. However, among seed predators, carabids feeding on loose seeds and birds eating seeds in open cones were the most important. 5 Cones open as they dry out, and the timing of seed dispersal is therefore weather-dependent. Late opening of cones had a deleterious effect on seed survival because it led to an increase in pre-dispersal seed predation. Thus, in a period of climate change, patterns of seed predation would be expected also to change, altering the success of regeneration.