Plants That Collect Sand Are Less Likely To Get Eaten

Introduction

Many diverse plants entrap substrate, often sand, particles on their stems, petioles, and leaf surfaces with exudates or trichomes (Fig. 1; Appendix S1). Jürgens (1996) termed this adaptation “psammophory” (“sand-carrying”) and described the phenomenon’s occurrence in the flora of in the Namib Desert; Neinhuis et al. (1996) described the phenomenon mechanistically in members of the Amaryllidaceae and Hyacinthaceae. Danin (1996), Lev-Yadun (2007) and Farmer (2014) each independently described the phenomenon in Middle Eastern desert plants with Farmer terming it “sand armor.” In the Americas, many plants in dune or desert communities are psammophorous, perhaps none more so than the various Abronia species (Nyctaginaceae), known colloquially as sand-verbenas. Table 1 lists over 200 species of psammorphorous species from these references and our own observations. These include 88 genera in 34 families of both dicotyledons and monocotyledons across the globe; many more certainly occur in desert, dune, and dry Mediterranean ecosystems worldwide.

Jürgens (1996) presented several explicit hypotheses of the ecological and physiological benefits and costs of this extra-cuticular sand layer, including protection from physical abrasion in sandstorms, reduction of water loss, reduced surface temperature and reduced solar radiation (though potentially limiting photosynthesis). His null hypothesis was that the production of exudates for other purposes (e.g., protection or waste excretion) creates sand entrapment as a side effect. Beal (1878), Jürgens (1996), Neinhuis et al. (1996), Lev-Yadun (2007), and Farmer (2014) all independently hypothesized that the sand layer may protect the plant from herbivores, either by making tissues difficult or costly to consume (resistance) or difficult to find (crypsis).

Many lines of evidence suggest a direct effect of sand on herbivores. Farmer (2014) and Lev-Yadun (2007) justified this hypothesis by comparing sand to the many examples of crystalline structures in plants (phytoliths, raphides, etc.), containing silicon or calcium, that serve as defenses against herbivores. Sand or other substrate material is known to contribute to tooth wear in humans (Molnar et al. 1983), badgers (Da Silva and MacDonald 1989), many ungulates (Jurado et al. 2008) including livestock (Mayland et al. 1977), and rodents (Gomes Rodrigues et al. 2011). This factor may be an important driver in mammalian dental morphology (Janis and Fortelius 1988, Damuth and Janis 2011) and behavior; moles squeeze the soil out of earthworms “like toothpaste from a tube” before consuming them to reduce potential tooth wear (Gorman and Stone 1990). While untested, these examples support the hypothesis that sand coatings may reduce palatability to herbivores and damage to psammophorous plants.

Several artificially parallel cases occur in the applied literature. Nordlander et al. (2009) demonstrated that coating the stems of young conifers in sand (with an adhesive) increased plant survival by reducing pine weevil (Coleoptera: Curculionidae) damage. Similarly, Showler (2002) and Showler (2003) demonstrated that covering cotton plants in kaolin (a clay mineral) dust reduced oviposition by boll weevils (Coleoptera: Curculionidae: Anthonomus grandis grandis), and reduced feeding and oviposition, and caused significant mortality of beet armyworm caterpillars (Lepidoptera: Noctuidae: Spodoptera exigua). It is likely that, in the latter cases, the fine dust interfered with the epicuticular waxes on the insects (Ebeling 1971) as opposed to interfering with feeding or digestion.

Sand coating of plant parts may increase crypsis through the increase in substrate matching, allowing the plant to escape detection by herbivores (i.e., reducing its “apparency” sensu Feeny 1976). Several genera of desert spiders entrap sand grains with small “hairlettes” coming off of their dense body setae; this covering renders the spider almost invisible and presumably reduces predation (Duncan et al. 2007). While Feeny’s apparency hypothesis was proposed to explain interspecific plant variation, Rhoades and Cates (1976) suggested that intraspecific and even within-plant distributions of defenses could be understood by considering the predictability and availability of edible tissue to herbivores. Although these theories have been important in stimulating research in plant–herbivore interactions and have been cited over 1400 times (Endara and Coley 2011), they have rarely been tested explicitly. The dearth of direct tests results from a difficulty in unambiguously measuring “apparency” in most complex, real-world situations (Feeny 1991 cited in Stamp 2003).

We experimentally tested the broad hypothesis that a sand coating provides defense and the two mechanistic hypotheses. We chose two locally available psammophorous plants in northern California, representative of this widespread phenomenon, for experiments. We manipulated the sand coating and measured herbivory on Abronia latifolia (Nyctaginaceae) at the UC-Davis Bodega Marine Reserve (Sonoma County, California, USA) and Navarettia mellita (Polemoniaceae) at the UC-Davis McLaughlin Reserve (Lake County, California, USA). Abronia latifolia is a common dune plant along much of the California coast (Baldwin and Goldman 2012). The short glandular trichomes that coat almost the entire surface of stems, petioles, and the bottoms of leaves, entrap sand while the tops of leaves remain sparsely sandy (Fig. 1). The perennial A. latifolia is grazed by snails (Helix aspersa; personal observation), deer mice (Peromyscus maniculatus: Osborne and Sheppe 1971), possibly deer (as in Crete et al. 2001) and a seasonally common leaf-mining caterpillar (Heliodinidae: Lithariapteryx abroniaeella; personal observation).

Navarretia mellita, as well as several congeners, is densely glandular and often become covered by sandy substrate; it is a ruderal species endemic to California, occurring in small patches along roadsides, in ditches, and eroded areas. These patches are usually small (1–50 individuals) and likely ephemeral. The entire glandular inflorescence, including the spiky bracts, is consumed by a mammalian browser and the damage is consistent with that of black-tailed jackrabbits (personal observation).

Materials and Methods

Results

Abronia field sand reduction

The experimental sand reduction roughly doubled chewing herbivory compared to controls on the same dates (Fig. 3). A model that included treatment as a fixed effect (plus date and pair as random effects) fit the data significantly better than the random intercept-only model (likelihood ratio test, χ² = 16.4, df = 1, P < 0.001). This model predicted 111% (95% CI, 35%–159%; treatment coefficient z = 3.57, df = 176, P < 0.001) more herbivory for the stems in the sand removal treatment. The noted herbivory was primarily small holes both in center and margins of leaves, often with the lower epidermal layer still intact. We could not conclusively assign this damage to any single herbivore; in our experience both arthropod and molluscan herbivores could produce this type of damage.

 

Figure 3.

• Open in figure viewer
• Download Powerpoint slide

Proportion of Abronia latifolia leaves with chewing herbivore damage. For ease of visualization, these results are presented as a proportion, averaged within each treatment for each plant and date; statistics were done on binomial data.

 

Abronia apparency manipulation

Contrary to prediction, both sand treatments (green and brown) suffered less herbivory than the sand-removal treatment (Fig. 4). The model incorporating treatment and the random effects of date and block (of all three treatments) fit significantly better than a model incorporating only the random effects (likelihood ratio test, χ² = 34.3, df = 2, P < 0.001). Tukey post hoc comparisons of the treatments showed no difference between the two sand treatments (P > 0.33). A model incorporating sand presence fit the data better than that of color (AIC 259, 261, respectively) and predicted an increase of 146% (79–240%, z = 5.64, P < 0.001) in the incidence of leaf damage for a sand addition. We also analyzed the damage by mining caterpillars (Lithariapteryx abroniaeella), as they were more common during this spring experiment, though still uncommon: of the total of 225 observations after the first week, we only noted leaf miners in 56 instances. This herbivore was unaffected by our sand treatments, of 25 replicates per treatment, 28% (7) with brown sand, 40% (10) with green sand, and 36% (9) with sand-removal stems had mines, a nonsignificant difference (two-tailed chi-square test, χ² = 0.82, df = 2, P = 0.66).

 

Figure 4.

• Open in figure viewer
• Download Powerpoint slide

Proportion of Abronia latifolia leaves with chewing herbivore damage. For ease of visualization, these results are presented as a proportion, averaged within each treatment for each plant and date; statistics were done on binomial data.