Eleusine indica (goose grass)

E. indica is primarily listed as an agricultural and environmental weed (Randall, 2012) and is considered a “serious weed” in at least 42 countries (Holm et al., 1979). This species is described as a “dominant weed” especially in farming systems and annual row-crops where it grows vigorously and produces abundant seedlings (Holm et al., 1979). A single plant may produce more than 50,000 small seeds, which can be easily dispersed by wind and water, attached to animal fur and machinery and as a contaminant in soil (Waterhouse, 1993). E. indica invades disturbed habitats in natural areas and the margins of natural forests and grasslands, marshes, stream banks and coastal areas. It is also a common weed along roads, pavements, and powerline corridors (Queensland Department of Primary Industries and Fisheries, 2011). Currently it is listed as invasive in several countries in Europe, Asia, Central and South America, the Caribbean and on many islands in the Pacific Ocean (see Distribution Table for details). 

The genus Eleusine comprises about 9 species and East Africa is considered its centre of diversification with eight species: E. africana, E. coracana, E. kigeziensis, E. indica, E. floccifolia, E. intermedia, E. multiflora and E. jaegeri occurring in this region. Species within this genus have little morphological differences between them, and include annual and perennial growth forms (Bisht and Mukai, 2002). The diploid (2n = 18) weed, E. indica, is closely related to the tetraploid (2  =36) African crop finger millet (Eleusine coracana) and is assumed to have given rise to it.  

E. indica is a tufted annual grass, prostrate and spreading, or erect to about 40 cm, depending on density of vegetation but not usually rooting at the nodes. The root system is very well developed and strong and the name jongos gras, used in South Africa, implies that it takes a young ox to uproot it. On germination, the first leaf, about 1 cm long, tapers very suddenly to a point and may be pressed quite flat on the soil. Later leaves are flat to V-shaped, up to 8 mm wide, 15 cm long and come to a longer, acute, boat-shaped tip. They are glabrous and usually quite bright, fresh green in colour. The ligule is a very short membraneous rim up to 1 mm long, sparsely fringed with short hairs. The sheaths and stem bases are distinctly flattened. The inflorescence consists of 3-8 racemes, each 5-10 cm long, about 5 mm wide, arranged more-or-less digitately, though one raceme may be inserted about 1 cm below the others. The narrow rachis, about 1 mm wide, has two dense rows of almost glabrous spikelets, each 2.5-3 mm long, 3-5 flowered, the lower and upper glumes about 1.5 and 3 mm long, respectively, and the lemmas very similar in both texture and size to the upper glume. All have a slightly scabrid keel and are acute but not awned. The reddish-brown to black seeds are oblong, about 1 mm long, conspicuously ridged.
 

The geographical origin of E. indica is uncertain but it is considered native to Africa and temperate and tropical Asia (USDA-ARS, 2014). Now it is distributed almost throughout the tropical world and extends significantly into the sub-tropics, especially in North America, Europe and Africa. It occurs up to 2000 m altitude in the tropics.

E. indica has probably been repeatedly introduced in most countries where it is now present, making it very difficult to determine its history of introduction into new habitats. In the USA, this species was introduced around the 1800s. In the West Indies, it was first recorded in 1815 in Cuba, 1867 in Martinique, 1876 in US Virgin islands, and 1885 in Jamaica (US National Herbarium). 

The risk of introduction of E. indica into new habitat is very high. This species is one of the most common agricultural and environmental weeds in tropical and subtropical regions of the world. It has ecophysiological and genetic traits that, coupled with the high number of seeds produced for each individual plant, give it a high score for successful invasion in almost any ecosystem (Holm et al., 1979; Waterhouse, 1994). 

E. indica spreads by seeds. A single plant has the potential to produce more than 50,000 seeds which can be easily dispersed by wind and water, as a contaminant in crop seeds and soils, and attached to animal furs, mud and machinery. Seeds are also eaten by wild animals and by cattle (Waterhouse, 1994). 

E. indica may occur in virtually any annual crop in the tropics and sub-tropics and also in many perennial crops and pastures. It is perhaps most conspicuous in annual row-crops such as cereals, legumes, cotton, tobacco and vegetable crops in which it is able to establish rapidly before there is adequate shading from the crop.

Confusion with other weeds having a digitate inflorescence is possible, e.g. with Digitaria, Dactyloctenium, Cynodon, Chloris or Paspalum spp., but the combination in E. indica of flattened stem, bright green leaves, size and many-flowered character of spikelets, lack of awns, etc. should serve to distinguish it from these species.

E. indica is a typical weedy species of the tropics and sub-tropics, flourishing in cultivated and other disturbed situations on a wide range of soil types, though generally favoured by high fertility.

Many natural enemies have been recorded (see Waterhouse, 1994 for a comprehensive list). The natural enemies listed here are those which are not known to have other hosts. It is not clear that significant damage is caused other than on a local basis but a few organisms have been considered as biocontrol agents (see Control).

Holm et al. (1979) recorded E. indica as a 'serious or principal' weed in 42 countries and it is frequently among the dominant weed species, especially in farming systems which include annual row-crops where lack of shading allows vigorous growth and abundant seeding. Holm et al. (1977) concluded that E. indica was one of the most serious weeds in cotton in 11 named countries, in maize in 10 countries, in upland rice in 8 countries, in sweet potatoes in 4 countries and in sugarcane in 3 countries; it also occurs in a wide range of other crops on a more local basis. These include banana, cassava, pineapple, rape, jute, soyabeans, pawpaw, abaca, cowpea, millet, mango, cacao, sorghum, tobacco, wheat and many vegetable crops. It was later listed among the top seven weed species in a worldwide review of weeds in sugarcane (Cepero and Rodriguez, 1983).

Competitive effects of pure populations of E. indica have rarely been measured, but in groundnut a range of densities of 2 to 32 plants per 10 m of row reduced yield by 2 to 25% and it was estimated that each weed plant per 10 m row reduced yield by 41 kg/ha. The economic threshold for use of sethoxydim to control it was only one plant per 7.4 m of row (McCarty, 1983). In maize, a population of 20 E. indica per maize plant, or 133/m² caused a significant 15% yield reduction. In this crop, E. indica was much more competitive than Euphorbia heterophylla, which required more than double the population to cause equivalent yield loss (Eke and Okereke, 1990). In India, Singh et al. (1996) showed that E. indica was responsible for removal of 20 kg potassium/ha, more than any other weed species present.

Where E. indica has been among the predominant weed species, for example, the second most abundant, representing 30% of the weeds in upland rice, yields have been reduced by 80% in the Philippines (Lourens et al., 1989). In direct-sown rice in Colombia, Fischer et al. (1993) listed E. indica as the dominant in a range of weeds that caused almost total crop loss when not weeded and 20% loss when weeded once at 20 days after sowing. In a further study in Colombia, Fischer and Ramirez (1993) assessed the losses due to weeds emerging 30 days after crop establishment.

Other examples of yield loss where E. indica is a major component of a mixed weed flora include 57% loss of potato and 76% loss of carrot in Brazil (Zagonel et al., 1999a, b). Actual losses due to all weeds in a number of the crops in which E. indica is a major weed (for example, maize and cotton) are of the order of 10-15% and amount to many billions of US dollars (for example, 9 billion in maize, 5 billion in cotton). As a major component of the weed flora in these crops over a substantial proportion of their total area, it is likely that E. indica contributes at least 1 or 2% of this total monetary loss, i.e. many million dollars, to which could be added the time, effort and costs involved in manual weeding. Once E. indica is established it has a notoriously tough root system making it necessary to use a hoe rather than manual uprooting.

While there are no estimates of losses caused, there are reports of stock poisoning where the grass is grazed (Wapshere, 1990a). Furthermore, although the presence of E. indica in crops has occasionally been shown to reduce pest or disease incidence, it may act as an alternate host of important crop pests or diseases, including Pratylenchus zeae on maize (Jordaan and de Waele, 1988); rice ragged stunt (Salamat et al., 1987); rice yellow mottle virus (Okioma et al., 1983); sorghum shoot fly (Granados et al., 1972); and others listed by Holm et al. (1977). There are also reports of increased incidences of Spodoptera frugiperda in maize where the weed is not controlled (van Huis, 1981). Again there is a lack of any quantification of these losses.

Cultural Control

As an annual weed which does not root at the nodes, E. indica is relatively easily removed by hoeing at the early growth stages. Once established, however, the very strong root system makes it difficult to uproot manually. Solarization has been shown to kill seeds of E. indica down to 5 cm (Standifer et al., 1984). Shredded and chopped newspaper has shown potential as a mulching material for control of Echinochloa crus-galli, Chenopodium album, Eleusine indica and Digitaria album in tomato (Monks et al., 1997). However, the effect may vary in different environments and other vegetable crops. E. indica is favoured by zero-tillage techniques but is well suppressed by residues of a rye cover crop (Teasdale et al., 1991). E. indica and the total grass population were higher in fields receiving no tillage in a 5 year study in Honduras. There was a more heterogeneous distribution of species under no tillage suggesting that tillage reduces the diversity of weeds

Chemical Control

E. indica is susceptible to virtually all groups of standard grass-killing herbicides, including arsenicals, substituted ureas (diuron, etc.), uracils (bromacil), triazines (atrazine, etc.), dinitroanilines (trifluralin, etc.), thiolcarbamates (EPTC, etc.), dimethylethers (oxyfluorfen, etc.), graminicides (fluazifop, sethoxydim, etc.), imidazolinones (imazaquin, etc.), propanil, oxadiazon, clomazone, quinclorac, diphenamid, paraquat, glufosinate, glyphosate and flumioxazin. Where broad-spectrum weed control requires the use of herbicide mixtures such as a graminicide with a broad-leaf weed killer, there are risks of antagonism in many combinations, e.g. with 2,4-DB (York et al., 1993). In most cases the antagonism can be avoided by applying the broad-leaf herbicide a day or two later than the graminicide.

Thanks to the wide range of effective compounds, control with herbicide should normally be possible in any broad-leaved or perennial crop and in most cereal crops, with the possible exception of finger millet. This can, however, be compromised by development of resistance to some herbicides.

Herbicide Resistance

Biotypes with resistance to some groups of herbicide have already occurred and are likely to become increasingly important. Resistance to trifluralin was the first to be detected, in the USA, and there is shown to be cross-resistance to most if not all other herbicides in this group (Vaughn et al., 1990). Baird et al. (1996) and Zeng and Baird (1997) have published comprehensive accounts of trifluralin resistance in E. indica. Anthony and Hussey (1999) have studied the molecular basis of the resistance of E. indica to dinitroaniline herbicides.

Resistance to fluazifop-butyl has developed in Malaysia (Leach et al., 1995) with cross resistance to most other graminicides; and to imazapyr in Costa Rica (Valverde et al., 1993) with cross resistance also to some other imidazolinone and sulfonylurea herbicides.

Biotypes of E. indica that are resistant to aryloxyphenoxypropionate and cyclohexanedione herbicides have been reported in parts of Malaysia since 1989, but the weed was easily controlled by glyphosate. However, in late 1997, a fruit grower in Teluk Intan, Perak, Malaysia, reported that glyphosate failed to give adequate control of E. indica in his 4-year-old orchard. Trials confirmed the grower's observation that glyphosate only gave only about 25% control of E. indica ('Teluk Intan' biotype). The 'Teluk Intan' biotype was found to be 8- to 12-fold resistant to glyphosate (Lee LimJung and Ngim, 2000). Dill et al. (2000) reported that there was no significant difference in the uptake and translocation of glyphosate in resistant and sensitive biotypes. Doll (2000) said that the resistance to glyphosate is the result of an altered binding site in the target enzymes, and that resistant E. indica was already infesting 12,500 ha on four oil palm plantations. Glyphosate-resistant populations have also been confirmed in Tennessee (Mueller et al., 2011), and low level resistance has been reported in Rio Grande do Sul, Brazil (Vargas et al., 2013).

Yew NgKwang (2011) reports control of glyphosate-resistant E. indica in Malaysia by application in combination with clethodim.

The repeated use of dinitroaniline herbicides in cotton and soyabean fields in southern USA has resulted in the appearance of resistant biotypes of E. indica. Two biotypes have been characterized: a highly resistant (R) biotype and an intermediate resistant (I) biotype. Each mutation is thought to exert its effect by a different mechanism (Anthony and Hussey, 1999).

Resistance to a number of herbicides has been reported in Malaysia in recent years. Jalaludin et al. (2010) report potentially resistant E. indica biotypes to glufosinate-ammonium. E. indica populations resistant to fluazifop, an acetyl-CoA carboxylase-inhibiting herbicide, have been found in several states in Malaysia (Cha ThyeSan et al., 2014).

An Jing et al. (2014) list paraquat as one of the herbicides to which E. indica has evolved resistance, and investigate the genetic basis of this resistance.

It is very important that herbicide use should take account of the risks of resistant biotypes building up and repeated use of the same, or a closely related, herbicide must be avoided. As so many herbicide groups are active, it should not be difficult to vary the type of compound used, even though this may mean not always using the least expensive product.

Biological Control

Biological control of E. indica has been considered in great detail in Australia, where the related finger millet (E. coracana) does not occur (e.g. Wapshere, 1990a, b; Waterhouse, 1994). For classical biocontrol, potential organisms include the smut fungus Melanopsichium eleusinis, the nematode Heterodera delvii, and certain cecidomyiid gall midges (Contarinia sp.) but all require further study before they could be used. Fungi which might be developed as mycoherbicides include Bipolaris [Cochliobolus] setariae and Pyricularia [Magnaporthe] grisea (Figliola et al., 1988) but no active programme of development of these has yet been reported.

Allelopathy

DIBOA (2,4-dihydroxy-1,4-(2H)-benzoxazine-3-one) and BOA (2-(3H)-benzoxazolinone), the major allelochemicals in rye (Secale cereale), were determined in eight field-grown cultivars. The cultivar Bonel showed the greatest activity on E. indica (Burgos et. al., 1999).