Question 1: Can a Perennial Grain Yield As Well As an Annual Grain?
Work at the Land Institute to domesticate perennial grains began in 1978 with an inventory of nearly 300 herbaceous perennial species for their suitability to the environment of central Kansas and their promise as seed crops. A second inventory examined the agronomic potential in 4300 accessions of perennial grass species within the C3 genera Bromus, Festuca, Lolium, Agropyron, and Elymus (Leymus). From these inventories, a handful of perennial species was chosen for exploring the principles of perennial grain agriculture.
Eastern gamagrass (Tripsacum dactyloides [L.] L.) is a large C4 bunchgrass native from the southeastern U.S. and Great Plains southward to Bolivia and Paraguay (Great Plains Flora Association, 1986). A relative of maize (de Wet and Harlan, 1978), eastern gamagrass has long been recognized as a nutritious and highly productive forage. Because of its high-quality seed (27 to 30% protein and 7% fat, Bargman et al., 1989) and large seed size, however, gamagrass shows much promise also as a grain crop for consumption by people, livestock, or both. Ground seed has baking properties similar to those of cornmeal. The major hurdle facing eastern gamagrass as a grain crop is low seed yield (typically around 100 kg/ha, but as high as 250 to 300 kg/ha in some material at the Land Institute, Piper, unpublished data).
Mammoth wildrye (Leymus racemosus (Lam.) Tsveler) is a rhizomatous C3 grass native to Bulgaria, Romania, Turkey, and western parts of the former Soviet Union. Grain of this and closely related species was reportedly eaten by Asian and European people historically, especially in drought years when annual grain crops faltered (Komarov, 1934). As is typical of cool-season grasses, wild rye displays most of its growth in late autumn and early spring. Seed yield in Land Institute trials has been as high 830 kg/ha (Piper, 1993b).
Illinois bundleflower (Desmanthus illinoensis (Michx.) MacM.) is a nitrogen-fixing legume that forms a deep taproot in its first year. It is native to the Great Plains, with a range extending north to Minnesota, east to Florida, and as far west as New Mexico (Great Plains Flora Association, 1986). It grows best during warm weather, flowering from late June onward. Small lenticular seeds are borne within clusters of brown pods beginning in late July. Highest yields have approached 2000 kg/ha (Piper, 1993b; unpublished data). The nutritional quality of the seeds (38% protein, 34% carbohydrate, Piper et al., 1988) suggests great potential as a grain legume.
Wild or Maryland senna (Cassia marilandica) is a legume native to the southeastern region of the Great Plains. Flowering in Kansas takes place from late August to early September, producing racemes of insect-pollinated yellow flowers that become brown-black pods later in the fall. C. marilandica produces thick, deep roots, but does not appear to form symbioses with nitrogen-fixing Rhizobium bacteria. The Land Institute has examined its year-to-year patterns of seed yield to address the biological question of whether a herbaceous perennial can produce a sustained, high yield. It provides a good model for studying the population dynamics of a high-seed-yielding perennial (Piper, 1992).
Maximilian sunflower (Helianthus maximilianii Schrad.) is native to dry to moist open prairies throughout the Great Plains. Its range extends eastward to Maine and North Carolina, and westward to the Rocky Mountains from southern Canada to Texas (Great Plains Flora Association, 1986). Maximum seed yields ranged from 1460 to 1840 kg/ha in Land Institute trial plots in 1996 (J. K. Piper, unpublished data). In addition to its potential value as a food or oil seed crop (seed is 21% oil, Thompson et al., 1981), Maximilian sunflower appears to inhibit weed growth allelopathically, and may therefore be especially important during the establishment phase of a perennial grain field.
Grain sorghum (Sorghum bicolor), a native of the African continent, is grown extensively in the southern Great Plains as a seed crop for animal feed. It is weakly perennial in tropical regions, but is killed by frost at higher latitudes. Johnsongrass (Sorghum halepense), a weedy relative of cultivated sorghum, is in the U.S. a troublesome weed that overwinters by production of rhizomes, fleshy underground stems capable of winter survival. It may be feasible to convert a tetraploid variety of grain sorghum from an annual to a perennial growth habit by combining in hybrids good grain quality with the ability to produce winter-hardy rhizomes (Piper and Kulakow, 1994). The ease of making this transfer will depend on the number of genes controlling the production of rhizomes and whether or not overwintering ability is genetically associated with poor agronomic characteristics.
Under favorable growing conditions, high yields of some perennial grasses can range from 1500 to 2000 kg/ha (Ahring, 1964; Ensign et al., 1983; Mueller-Warrant et al., 1994). Extrapolated seed yields ranging from 1720 to 2090 kg/ha have been recorded for some perennial grasses and legumes in rain-fed, unfertilized plots at the Land Institute (Piper, 1992; 1993b; Piper and Kulakow, 1994). These experimental yields compare favorably with the benchmark yield for Kansas winter wheat of 1800 lb/ac (1960 kg/ha).
Germplasm evaluations at the Land Institute indicate that genetic variability for improvement of such traits as seed yield, loss of seed dormancy, shatter resistance, resistance to viral diseases, and overall vigor is available. Moreover, including in a gamagrass breeding program such favorable types as a gynomonoecious form pro-lifica (a mutant sex form in which all florets are female and thus seed producing) may increase seed yield severalfold (Dewald et al., 1987). Currently, the Land Institute's gamagrass breeding program is focusing on 23 high-yielding, disease-resistant families to create a synthetic variety for further improvement.
Question 2: Can a Perennial Polyculture Overyield?
Competition between individual plants should be stronger in monocultures than in mixtures because conspecific neighbors are most similar morphologically, phe-nologically, and in nutritional requirements. If intraspecific competition is stronger than interspecific competition or if facilitation is occurring, plants should yield relatively better in mixture than in monoculture, resulting in overyielding. Conversely, if the effects of interspecific competition are greater than those of intraspe-cific competition (as when species differ greatly in size, for example), then plants should perform relatively better in monocultures. If the relative effects of different species are neutral, then there is no yield advantage to polyculture.
Overyielding is common in traditional polyculture systems of Latin America, Africa, and Asia (Francis, 1986; Vandermeer, 1989). Whether overyielding can occur in perennial grain systems, and whether it can persist for more than 1 year, is less explored. A few examples from experiments performed at the Land Institute support the model.
In a study of 28 accessions of eastern gamagrass and Illinois bundleflower, a relative yield total (RYT) of 1.19 was obtained (Muto, 1990). This translates into a 19% yield advantage in mixture relative to the monocultures. A three-species study using eastern gamagrass, Illinois bundleflower, and mammoth wildrye produced an RYT of 1.26 (Barker and Piper, 1995). In both of these cases, overyielding occurred in more than 1 year.
Question 3: Can a Perennial Polyculture Provide Its Own Nitrogen Fertility?
The growing environment of perennial systems changes annually. Typical initial conditions include high levels of sunlight and available soil nutrients. With time, concentrations of some available nutrients decline while root mass and shading increase. With canopy closure, species interactions that began as competition for soil nutrients may end as competition primarily for light, with plants exhibiting a trade-off between tolerance of low soil resources and tolerance of shade (Tilman, 1985). Various factors, including root architecture, physiology (C3 vs. C4 photosyn-thetic pathway), and ability to fix atmospheric nitrogen can result in differences among species in net soil resource levels. Understanding the net effects of perennial grains on soil nutrient status is an important consideration for long-term fertility management of perennial stands to be grown with few or no inputs.
The ability of legumes to fix atmospheric nitrogen via symbioses with root nodule-forming bacteria has important implications for agricultural sustainability. It can release a plant from competition with neighbors for soil nitrogen and promote the growth of neighbors if fixed nitrogen subsequently becomes available to them. Nitrogen transfer from legumes to grasses may occur via leakage and excretion from roots (Simpson, 1965), following decay of nodules and roots (Haynes, 1980), and by direct mycorrhizal exchange (van Kessel et al., 1985; Eissenstat, 1990). Perennial grasses may receive from 46 to 80% of their nitrogen directly from companion legumes (Brophy et al., 1987). In agricultural systems, legume nitrogen can prove both energy efficient and cost-effective (Mallarino et al., 1990; Posler et al., 1993).
Ideally, in a sustainable agriculture based on perennial grain mixtures, a significant portion of the available nitrogen should arise from one or more leguminous companion crops. Therefore, an important question is whether or not a perennial polyculture can provide sufficient nitrogen fertility via legume nitrogen fixation to compensate for removal of this nutrient in harvested seed.
Work at the Land Institute has provided some indirect evidence of a benefit of a leguminous grain to companion species. A study monitored over several years showed that Illinois bundleflower grows and yields equally well on high and low nitrogen soil in years with adequate precipitation (Barker and Piper, 1995). This suggests that this legume can compensate for low soil nitrogen without a measurable penalty of lowered growth or seed yield. A piece of evidence is that eastern gama-grass yields better, and its yield does not decline with time, in mixtures with bundleflower relative to gamagrass monoculture (Piper, 1998). Finally, plots containing Illinois bundleflower have higher soil nitrate concentrations 3 to 5 years after establishment than plots containing only grasses (Piper, unpublished data).
Question 4: Can a Perennial Polyculture Manage Weeds, Herbivorous Insects, and Plant Pathogens?
Weeds are a major source of competition with crops for light, water, and nutrients. Because most agricultural weeds are adapted to disturbed habitats, and can oftentimes take up soil water and nutrients faster than crops, they are a chronic problem where soils are repeatedly tilled. Thus, it is not surprising that herbicides account for 69% (by weight) of all pesticide use in the U.S. Nearly 90 million ha, more than half of all U.S. cropland, are treated with herbicide (Pimentel et al., 1991). Perennial grain polycultures are likely to compete well against weeds. Two important aspects of the crop/weed relationship in perennial polyculture are the possible synergistic effects of polycultures on weeds and the direct competitive relationship between perennials and weeds.
In contrast to rotational sequences of monocultures, which can manage weeds over time, intercropping combines the weed-suppressing effects of different crops within a single season. Crop polycultures may intercept more light, water, and nutrients than monocultures. For example, mixtures of C3 and C4 species display active crop growth during a greater proportion of the year, potentially eliminating the need for herbicide application (Evers, 1983). If weed mass in polyculture rows is less than what is predicted by weed mass in the respective monocultures, this may indicate a synergistic effect of polycultures on weeds.
An intercrop advantage may accrue in two ways (Liebman and Dyck, 1993). Greater crop yield may be coupled with lower weed growth in polyculture. This may occur if there is greater resource preemption by crops in the mixture (less available to weeds) or where there is allelopathic suppression by one or more crops. Alternatively, the intercrop may display a yield advantage while failing to suppress weed growth below levels observed within monocultures of the component crops. This may result if the intercrops use resources not exploitable by weeds or where intercrop use of resources is more efficient than in monoculture.
Studies done primarily in traditional cropping systems in the tropics have shown that internal weed management in polycultures is often greater than in monocultures of the respective crops (Liebman, 1988). Similarly, several temperate zone intercropping systems have been shown to improve weed management in important ways. Liebman and Dyck (1993) reviewed 51 published intercropping studies where the main crop was intersown with a "smother" crop. In 47 cases (92.2%), weed biomass was lower in mixture than where the main crop was grown alone. When intercrops were composed of two or more main crops, weed biomass of the intercrop was lower than in all of the component sole crops in 12 cases (50%), intermediate between component crops in 10 cases (41.7%), and higher than all sole crops in only 2 cases (8.3%).
For certain combinations, increasing crop species diversity per se may suppress weeds. In an ongoing experiment at the Land Institute using experimental mixtures of 4, 8, 12, and 16 perennial species (see section on Community Assembly below), percentage cover by annual weeds in the second year was inversely related to the diversity of perennials sown, an effect which was stronger in the third year (see Figure 1B, later).
Second, because of their permanent canopy, deep and extensive root systems, and vigorous regrowth in spring perennial plants should in general compete well against weeds. Several factors affect the relative ability of crop species to suppress weeds, however. Differences in height, canopy thickness, rooting zone, and phenology are likely to influence crop/weed interactions. Studies of prairie restoration typically show an initial period of dominance by weedy annuals, followed by increasing dominance by an array of herbaceous perennials (e.g., Holt et al., 1995).
Effective weed control has occurred in two separate experiments at the Land Institute. In one study, a plot containing rows planted with five densities of Maximilian sunflower and a control (no sunflowers), weed biomass was significantly reduced in the sunflower plots relative to the control (discussed in Piper, 1993b). By the second year, a sunflower density of 3.6 plants/m2 reduced weed biomass between rows to levels only 25 to 50% of the control. In May of the third year, weed biomass in sunflower rows was 44% of weed biomass in the control. Here, effective weed control was maintained across years despite changes in the weed community from predominantly annuals in the first year to perennials by the third year.
In a second experiment that examined weed growth, a triculture comprising wildrye, eastern gamagrass, and Illinois bundleflower at equal densities, species combinations differed in their ability to control weeds (Piper, 1993c). Weed biomass was consistently lowest in rows with eastern gamagrass as a component, despite seasonal and yearly changes in species composition of the weed community. These results point to eastern gamagrass as the primary weed controller among the three species, probably via shading, although unmeasured underground interactions were also likely important throughout the 3-year study.
For many vegetable and grain crops, pest insect density and level of feeding damage on host species are lower within diversified stands than in monoculture (Risch et al., 1983; Andow, 1986; 1991; Coll and Bottrell, 1994). In an extensive review of the literature, Andow (1986) found that 131 of 203 (64.5%) monophagous species were reported to be less abundant in polyculture cropping systems relative to monoculture. This phenomenon, in which host plants associated with other, nonhost plant species suffer less herbivore attack than host plants in monoculture, is known as associational resistance (Tahvanainen and Root, 1972).
Although many of the pest management benefits obtained within diversified annual systems may transfer to perennial polycultures, typically perennial systems are characterized by less soil disturbance and greater year-to-year continuity of the host species. Pests therefore have the opportunity to maintain and even increase population density in host plant patches that are relatively stable and predictable. Unfortunately, the long-term dynamics of insect populations within perennial monocultures vs. polycultures remains relatively unexplored.
The Desmanthus illinoensis/Anomoea flavokansiensis association represents one system for exploring whether perennial grain polycultures can reduce insect density relative to monoculture. The leaf beetle A. flavokansiensis Moldenke (Coleoptera: Chrysomelidae: Clytrinae) specializes locally on Illinois bundleflower. From mid-June to early August adults feed on young bundleflower leaves and inflorescences. At high densities, A. flavokansiensis can reduce seed yield and is thus an important consideration for long-term stands that are to be grown without insecticides.
The potential to manage A. flavokansiensis via intercropping its host species with other, non-host perennial species has been examined (Piper, 1996). Replicated plots, comprising monocultures of Illinois bundleflower, and two- and three-species mixtures of bundleflower with the eastern gamagrass and mammoth wildrye, were established at two sites. Over a 5-year period, the way that the plant species diversity affects A. flavokansiensis density on individual host plants and how beetle density changes with time in perennial stands were examined.
Insects were counted frequently from mid-June to early August. In the first 3 years, beetle density was generally low (<1 per plant), and did not differ among treatments. In the fourth year, however, beetle density peaked at 15 and 25 insects per plant at the two sites, and was highest within bundleflower monoculture for most dates. In year 5, density was again low, but tended to remain higher in monoculture at one site. The results show that beetle density on Illinois bundleflower can be reduced in polyculture and hold promise for the management of this type of insect herbivore within perennial grain polycultures. Differences in beetle density were probably due more to differences in resource concentration, combined with physical barriers provided by the grasses, than an increased presence of natural enemies in polycultures.
It is encouraging that density of this monophagous insect on its perennial host was reduced in some instances in polyculture even 4 and 5 years following establishment. Monitoring of the year-to-year population dynamics of this insect may predict how similar specialist herbivores will react to plant species diversity in perennial grain mixtures.
A potential hazard confronting stands of perennial grains is the establishment and long-term residency of systemic diseases. Perennials may be capable of storing inoculum from year to year, enabling spread of the disease throughout the field in later years. One possibility for the management of such diseases is to grow host plants within polycultures.
Eastern gamagrass is infected by the pathogens sugarcane mosaic virus strain maize dwarf mosaic virus B and maize dwarf mosaic virus. These are aphid-borne, nonpersistently transmitted polyviruses that infect and then overwinter in gamagrass (Seifers et al., 1993). Such diseases have the potential to increase in severity with time. Symptoms of infection vary from a general mosaic to chlorotic and necrotic spots dispersed throughout the leaf. At high levels of severity, these diseases reduce growth and seed yield (Piper et al., 1996).
To examine the potential for polyculture to manage levels of aphid-borne disease in a perennial system, plots comprising monocultures of eastern gamagrass as well as mixtures of gamagrass with two nonhost species, mammoth wildrye and Illinois bundleflower, were established and monitored for 5 years (Piper et al., 1996). For the first 2 years at one of two sites, disease incidence (percent of plants infected) and severity (intensity of symptoms) were generally lower in plots in which bundle-flower was a component. Because there were two biculture treatments in this study, the main factor seemed to be the presence of bundleflower within a plot, not merely lower host plant density relative to monoculture. Illinois bundleflower appears sufficiently different, in terms of architecture and biochemistry, from gamagrass to provide an effective barrier to vector movement. The results suggest that incidence and severity of these viral diseases can be delayed in some mixes of perennial grains, and that the effect can persist for more than 1 year.
This study showed that manipulations of host plant density and diversity may be useful in designing cultural methods for minimizing disease intensity and spread in the early years. Intercropping eastern gamagrass with such nonhost species as Illinois bundleflower appears to represent a disease management strategy complementary to selecting for viral resistance.
Was this article helpful?