1. The seeds of many plant species produce mucilage on their surfaces that when wetted and dried, firmly adheres seeds to surfaces and substrates. Previous studies have demonstrated that seed anchorage to the ground can reduce seed predation, although only a few species have thus far been tested.2. Here we investigated whether binding to the ground reduces seed removal by harvester ants Pogonomyrmex subdentatus, an important granivore, for 53 species with mucilaginous seeds. We also explored functional traits that associate with seed removal risk to understand the ecological and evolutionary context of this granivory resistance trait.3. Using a field cafeteria choice experiment, we compared harvester ant seed removal of wetted ground-bound seeds to dry unbound control seeds for these 53 species. We developed a simple assay to score dislodgement force. We examined whether this force, seed mass and seed mucilage production explained the interspecific variation in protection that we observed in field seed removal. We integrated these experiments with a broad scale test of correlates of seed attachment using a previously published dataset of attachment potential of mucilaginous seeds for 432 species, examining correlations of attachment potential with 13 plant traits and the climate characteristics of the species' range. 4. Binding to the ground reduced seed removal in 42 of 53 species tested. The benefit increased with seed dislodgement force, which itself increased with mucilage production, but not with seed mass. In the larger dataset, shorter plant life span, higher temperature, more solar radiation, higher humidity, fewer wet days per year and higher seed density correlated positively with the odds of seed attachment. We also found that attachment potential showed a concave down quadratic relationship with latitude, peaking at roughly 30°. No strong evidence that any of the other six predictors correlated with attachment potential was found.5. We demonstrate that protection from granivores is a widespread convergent function of seed mucilage and is associated with mucilage production. We highlight the need for increased mechanistic investigations into this common but poorly studied trait, particularly in relation to functional drivers of the broad patterns we found.

An undercited 1983 paper in Ecology by Patricia Fuller and Mark Hay found that seeds of a mint, when the mucilage was wetted and subsequently dried, became encrusted by the sandy desert substrate and that this coating reduced seed predation by harvester ants and rodents. We tested Fuller and Hay's defensive hypothesis in 53 plant species, finding that a mucilage-bound sand coating reduced predation by harvester ants in almost every species tested. While mucilage has many possible functions and probably did not evolve for the purpose of seed defense, in many dry areas it likely has this ecological function.

Seed mucilage is a common and highly diverse trait shared among thousands of angiosperms. While long recognized that mucilage allows seeds to anchor to substrates (antitelechory), we still lack a mechanistic understanding of this process. We propose a mechanistic model of how mucilage affects substrate anchorage and fluid resistance, ultimately contributing to antitelechory. To test this model, we subjected mucilaginous seeds of 52 species, varying in eight measured seed traits, to a week of continuous water flow at a range of erosion potentials. Supporting our model, mucilage mass increased both dry and wet attachment strength, which explained time to erosion well. A standard deviation increase in log mucilage mass increased seed time to dislodgement by 280 times. Fluid resistance was largely dependent on speed of water flow and the seed's modified drag coefficient, but not the quality of the seed or seed mucilage. Neither mucilage expansion speed nor mucilage decay rate explained dislodgement potential well. Our results suggest that the high substrate anchorage strength is the primary role of mucilage in fostering seed establishment in highly erosive environments. In contrast, other seed and mucilage trait differences among species are of lesser importance to antitelechory.

Infectious disease is an important potential driver of population cycles, but this must occur through delayed density-dependent infection and resulting fitness effects. Delayed density-dependent infection by baculoviruses can be caused by environmental persistence of viral occlusion bodies, which can be influenced by environmental factors. In particular, ultraviolet radiation is potentially important in reducing the environmental persistence of viruses by inactivating viral occlusion bodies. Delayed density-dependent viral infection has rarely been observed empirically at the population level although theory predicts that it is necessary for these pathogens to drive population cycles. Similarly, field studies have not examined the potential effects of ultraviolet radiation on viral infection rates in natural animal populations. We tested if viral infection is delayed density-dependent with the potential to drive cyclic dynamics and if ultraviolet radiation influences viral infection. We censused 18 moth populations across nearly 9 degrees of latitude over two years and quantified the effects of direct and delayed density and ultraviolet radiation on granulovirus infection rate, infection severity, and survival to adulthood. Caterpillars were collected from each population in the field and reared in the laboratory. We found that infection rate, infection severity, and survival to adulthood exhibited delayed density-dependence. Ultraviolet radiation in the previous summer decreased infection severity, and increased survival probability of the virus. Structural equation modelling found that the effect of lagged density on moth survival was mediated through infection rate and infection severity, and was 2.5 fold stronger than the effect of ultraviolet radiation on survival through infection severity.