The Role of Wheat Awns in the Seed Dispersal Unit

See allHide authors and affiliations

Science  11 May 2007:
Vol. 316, Issue 5826, pp. 884-886
DOI: 10.1126/science.1140097


The dispersal unit of wild wheat bears two pronounced awns that balance the unit as it falls. We discovered that the awns are also able to propel the seeds on and into the ground. The arrangement of cellulose fibrils causes bending of the awns with changes in humidity. Silicified hairs that cover the awns allow propulsion of the unit only in the direction of the seeds. This suggests that the dead tissue is analogous to a motor. Fueled by the daily humidity cycle, the awns induce the motility required for seed dispersal

Awns and other appendages on the seed dispersal unit of plants aid in dispersing the seed to a germination site (13). Hairs, wings, and hooks influence the route of seeds from the mother plant to a surface by wind or animal dispersal [phase I dispersal (4)]. Hygroscopically active awns propel seeds on the ground through coiling and uncoiling (5, 6) and, thus, mainly affect the movement of the seed after it reaches the resting surface [phase II dispersal (4)]. The nature of the soil affects the ability of the seeds to locate a safe site. Larger seed dispersal units are more easily buried in coarse particle soils, where the lumps are similar in size to the dispersal units. On finer-grained earth, they tend to move along the soil surface (7). Seeds bearing active awns are more abundant in structured soils, meaning soils that contain a stable system of pores and aggregates of different sizes, perhaps because the seeds are easily anchored. Unawned small seeds are prevalent in light sandy soils, where they are trapped, mainly by random movement of earth. Seeds equipped with passive awns are evenly abundant in both environments, with a preference for stable porous soils (8, 9). These observations do not conform to the predictionmade by Chambers et al. (7) that larger seed dispersal units will be trapped preferentially in soils of large particle size, which suggests that passive awns may interact with the soil by a mechanism yet unknown.

To find the specific features that may support this interaction, we studied awns of wild and domesticated tetraploid wheat lines [Triticum turgidum ssp. (10)]. Thin cross sections and oblique sections examined by scanning electron microscopy showed two photosynthetic zones surrounded by cells that provide mechanical support to the awn (Fig. 1). This design is common to passive awns of wheat and other crops (11). A silica layer is deposited on the external surface of the epidermis (11) (Fig. 2A), covering separated papillae and hair cells (fig. S1). This form of tiling creates a surface that is mechanically both hard and tough, to interact with the soil. The hairs are 0.1 to 0.2 mm long and point toward the tip of the awn. More hairs are found on the ridge surface, facing away from the dispersal unit axis (Fig. 1). Back-scattered electron images of the supporting tissue showed no evidence for structural or compositional differences between the cap and the narrower ridge. Nonetheless, scanning acoustic microscopy revealed a huge acoustic impedance contrast in the same sample, related to variations in stiffness (Fig. 2B). The impedance at the ridge was about 1/10th that at the cap. To support this finding, we measured the local effective Young's modulus at the cap and at the ridge using a nano-indentation probe. At the ridge, the reduced Young modulus was 10.0 ± 2.8 GPa (n = 16); however, the cap was much firmer, with a modulus of 20.5 ± 2.6 GPa (n = 7), corresponding to the upper range of spruce wood [7 to 23 GPa (12)]. Wheat stems were studied by a four-point bending test, with reported Young moduli values about half of those we obtained at the ridge [4.76 to 6.58 GPa (13)]. Nano-indentation measures the cell walls specifically and is insensitive to voids in the underlying tissue, so it is likely to yield higher values than macroscopic tests.

Fig. 1.

(Left) Illustration of the wild wheat plant (Triticum turgidum ssp. diccocoides) and two dispersal units (not to scale). Each dispersal unit carries two pronounced awns that balance the dispersal unit as it falls. A scanning electron micrograph in the back scattering mode of a section through a wild wheat awn is shown on the right. The section is taken from the lower third of the awn and is oriented in a way similar to the rectangle drawn over the bottom dispersal unit. The different tissues are indicated by arrows. The cap and ridge are facing in the same direction along the awn.

Fig. 2.

Structural features of the lower part of the wheat awn. (A) Silica deposited at the epidermis, epidermal papillae, and hairs (inset) is detected on a polished sample by a scanning electron microscope in the back-scattered mode. Silica tiles stiffen the epidermis and protect the structure as it interacts with the soil. (B) The differential stiffness between the cap and the ridge is demonstrated by an acoustic impedance map (1 by 1 mm2 field, brightness level is correlated to the relative impedance). Wide angle x-ray scattering patterns of the crystalline cellulose at the cap (C) and at the ridge (D) suggest that the difference in stiffness is created by different organization of the fibrils in the two regions.

To understand the difference in stiffness between the cap and the ridge, we first checked whether lignin is more abundant in the cap. We rejected this hypothesis, as staining by astrablue–safranin indicated an even distribution of lignin in the cross section (fig. S2). Next, we checked whether changes in the cellulose orientation contribute to the observed variation in stiffness (14). It is well known that in the secondary wall of wood cells, cellulose microfibrils are winding helically around the cell. The tilt angle of the fibrils with respect to the cell axis, usually called microfibril angle (MFA), determines the stiffness of the wood cell. Specifically, when the cellulose MFA in different wood types decreases from 50° to 5°, the corresponding stiffness is known to increase from 1 to 14 GPa (12, 15). We used wide-angle x-ray scattering to measure the MFA in different regions of the awns (10). We found that the cellulose fibrils are very well aligned along the long axis of the awn in the cap, with a MFA close to zero (Fig. 2C). At the ridge, the fibrils were found to be randomly oriented (Fig. 2D), similarly to the parenchyma of the wheat stem (16). However, this structure characterizes only the lower part of the awn, close to the seed. At the upper part, the cellulose was found to be aligned both at the ridge and at the cap.

We suggest an active role to this structure in the seed dispersal unit. The two different cellulose arrangements expand differently at ambient humidity conditions, similar to the mechanism of seed explosion in the Acanthaceae capsule (17) and the opening of pine cone scales (18). Water molecules that adsorb to the long crystalline cellulose fibrils will not make them longer but will cause a lateral expansion by swelling of the noncrystalline hemicelluloses between the crystallites. Thus, the awns will generally expand only laterally, except for the lower part of the ridge, where the fibrils are not aligned but are randomly oriented. This part will work like a muscle: expanding in length with humidity, pushing the awns together, and contracting with drying, pulling the awns apart. Such movement was observed long ago in awns of wild wheat (T. turgidum ssp. diccocoides), goat grasses (Aegilops ssp.), and Bromus ssp., but its importance for dispersal was not known (19).

To study the bending of the awns, we scanned the relative humidity of the air between 0.1 and 0.9 at 30°C and followed the change in the distance between the two awns on the same dispersal unit (Fig. 3 and movie S1). The awns bent at a point about 2 cm above the seed, corresponding to the location of the randomly oriented cellulose crystals. Above this position, no bending was observed, in agreement with the finding that the cellulose crystallites were aligned throughout the cross section. In principle, the elongation of the lower ridge relative to the cap may also induce twisting of the structure rather than bending, but we believe that the shape of the cross section with a wide cap, shaped like a half moon (Fig. 2B), may favor bending and creates a joint movement.

Fig. 3.

The principle of the dispersal unit movement. (Top) The average distance (± SD) between the awns of four dispersal units is presented versus relative humidity (r.h.) of the surrounding air. (Bottom) One cycle in the humidity-driven movement of the awns. (I) shows the seeds and part of the awns immersed in soil. The red arrow indicates one of the silica hairs. (II) Because of increased humidity, the awns straighten. The silica hairs lock the awn and prevent an upward movement. As a consequence, the seed has to move downward by a distance d, indicated in (I), to account for the increased length projected onto the vertical. (III) After drying, the awns bend again, which shortens the length projected onto the vertical. Because the silica hairs are locked into place, the seed cannot move back upward, and the awns are drawn further down into the soil (see the silica hair marked by a red arrow). Hence, the net movement of the seed in one cycle corresponds to d.

The movement is reversible; thus, the humidity cycle causes a periodic movement of the awns, which resembles the swimming stroke of frog legs. Most interestingly, there is a humidity cycle in the natural habitat of the wheat in the dry period after the seeds ripen; during the day, the air is dry, but at night, as the temperature goes down, humidity rises. This suggests that the awns may provide the motility required for seed dispersal, as was shown for the hygroscopically active awns (1).

To propel the seed on the soil surface, prominent friction forces must exist. This is most likely supplied by the silicified epidermal hairs that couple the unit to the soil in a ratchet manner. As the awns bend, the hairs slide on the soil particles, allowing movement only in the direction of the seed (Fig. 3). The rough soil lumps, which often contain silicates, are probably not able to burnish away the silicified protrusions even after several cycles of the awns' movement. In this way, the seed is pushed into the soil and, thus, is protected from predators, extreme dryness, and fires. With the first rains, the seed will have a better chance of germinating in a safe site. To show the ability of the seed dispersal unit to propel, we placed it horizontally on a felt cloth, which entangled the epidermal hairs. As we cycled the relative air humidity, the dispersal unit moved in the direction of the seed (movie S2). An identical design is preserved in awns of domesticated wheat, even though their function was lost during the domestication process, as the seeds do not disperse spontaneously. The short evolutionary time since domestication (about 10,000 years) probably allowed the complete loss of awns in several domesticated wheat lines, but not the alteration of the awn structure.

The wheat dispersal unit seems to be optimized to multiply the species in the environment of the Fertile Crescent with long dry summers and short rainy winter seasons. As a self-pollinating annual that grows in dense stands, long-distance dispersal would not improve survival. The local surroundings of the mother plant should be sufficient to support the next generation as well. However, the place where the seed falls may be less ideal than a nearby location. It was shown that active awns are able to propel seeds on the ground for several centimeters (6). We suggest that the paired passive awns of wheat are also able to move the seed along the soil surface, as well as vertically, for burial. The movement, based on a unique arrangement of cellulose fibrils, is fueled by the daily changes in air humidity. The pointed epidermal hairs break the symmetry and allow the unit to move as a ratchet, promoting faster burial. This system increases the chances of the seed to germinate and to reach maturity and may increase the likelihood of a specific stand to proliferate.

The understanding of this seed dispersal mechanism may help in developing new concepts in weed control. The microscopic mechanism found to provide motility to the seed may also serve as a model in biomimetic materials research. Indeed, a hydration-dependent bending movement was recently reported in an artificial system consisting of nano–silicon columns embedded in a hydrogel film (20). From a mechanistic point of view, we have discovered a device for movement that is composed of passive elements. Locomotion is provided by a volume containing nonoriented cellulose crystallites that shortens on drying and pulls the awn like a muscle. The energy source for this active movement is the daily cycle of air humidity.

Supporting Online Material

Materials and Methods

Figs. S1 and S2

Table S1


Movies S1 and S2

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article