(a) Species description

The following description is based on the literature (Stebler and Schröter 1889; Lubbock 1892; Knuth 1908; Gams 1924; Ball 1968; Gorshkova 1971; Polhill 1981; Kirkbride et al. 2003; Lasseigne 2003) and supplemented with observations of Canadian populations. Measurements are given as the usual range with extremes in parentheses. Many of the characteristics described are illustrated in Figs. 1–3.

Fig. 1.

Goat’s-rue, Galega officinalis. (A) Upper stem with inflorescences and immature fruits; (B) stipules; (C) flower. Scales bars = 1 cm.

Fig. 2.

Goat’s-rue, Galega officinalis. (A) Dense population at a disturbed site along the Ottawa River; (B) flowers. [Colour online.]

Fig. 3.

Goat’s-rue, Galega officinalis. (A) Stipule; (B) seeds; (C) seedling showing narrowly ovate cotyledons and first three true leaves, the later with one, two and three leaflets, respectively; (D) Craspedodromous leaflet venation of G. officinalis; (E) Camptodromous leaflet venation of Astragalus cicer. [Colour online.]

A perennial herbaceous plant from a stout caudex (root crown). Tap root long, fleshy, whitish, with fibrous rootlets; rhizomes absent. Stems hollow, 40–150 (–200) cm tall, more or less erect or sprawling, branched, glabrous to sparsely pubescent, slightly ribbed. Leaves alternate, petiolate, (3–) 8–30 (–40) cm long, more or less glabrous to sparsely hairy, once pinnate; stipules herbaceous, 0.5–1.6 cm long, broadly lanceolate to sagittate with (1–) 2–4 (–6) basal teeth or acute lobes; leaflets (9–) 11–19 (–21), sessile (or rarely a petiolule to 0.5 mm), in opposite pairs (except the terminal one), (7–) 15–50 ×4–17 mm, lanceolate to narrowly ovate, sometimes sparsely pubescent on the margins and (or) veins of the lower surface, apices acute to obtuse, often emarginate, and usually mucronate. Inflorescences on long peduncles, elongate, 8–27 (–30) cm long (including peduncle), axillary, densely to loosely flowered racemes; pedicels filiform, about as long as or shorter than the calyx, with a subtending lanceolate bract 5–7 mm long; calyx of fused sepals with 5 linear subequal teeth about as long as the tube, 4–6 mm long, glabrous or puberulent on teeth; corollas papilionaceous; petals 5 (lower two fused into a keel), (7–) 10–15 mm long, white, bluish, lilac to reddish purple, the banner (standard or vexillum) oblanceolate to obovate, more or less reflexed, the wings (alae) slightly shorter than to about as long as the keel, narrowly obovate, clawed (i.e., with a basal process or auricle), the keel (carina) broad and rounded, not auriculate. Stamens 10, monadelphous (the upper filament fused basally but partly free distally), included in the keel; anthers dimorphic (Endo and Ohashi 1997). Ovaries enclosed in the stamineal sheath. Style filiform, curved upwards, with a small capitate stigma protruding beyond the anthers. Fruits elongate, 2-valved, cylindric pods, (20–) 20–45 (–50) × 2–3 mm, glabrous, striate, erect to spreading at maturity, shallowly torulose (slightly constricted between the seeds), tardily dehiscent along sutures. Seeds (1–) 2–6 (–9), in one series, 2.5–4.5 ×1–2.5 mm, narrowly ellipsoid to somewhat reniform (slightly constricted at the hilum), greyish to yellowish-brown, dull. Seed coat microscopically rugose, with a thin layer of wax (Pandey and Jha 1988). The fruits and seeds of G. officinalis are described in detail and illustrated by Kirkbride et al. (2003).

Cotyledons are obovate-oblong, about 18–27 × 5–8 mm, smooth, entire. The first leaf of a seedling consists of one leaflet, the second leaf has a pair of leaflets, and the third leaf consists of a pair of leaflets and a terminal leaflet (Fig. 3C).

Details of root, stem and leaf anatomy were studied in populations from Turkey by Özbucak et al. (2005) who provide detailed descriptions and illustrations.

In their phylogenetic analysis of some Fabaceae tribes related to Galega, Endo and Ohashi (1997) reviewed various character states reported in the literature for Galega and provided unique observations on pollen grains and embryos of G. officinalis and G. orientalis.

In Canada, Gervais (1979) reported a chromosome count of 2n = 16 from an introduced population at Quebec City. Counts, n = 8 (Kreuter 1930; Senn 1938; Ruíz de Clavijo Jiménez 1990) and 2n = 16 (Tschechow 1930; Polhill 1981; Izmaiłow 1990) have been reported for Eurasian plants.

The entire chloroplast genome of G. officinalis was sequenced and characterized by Du et al. (2021). They reported it to be 125 086 base pairs in length, with a GC content of 34.18% and containing 112 genes.

(b) Distinguishing features

The gross morphology and habitus of Galega officinalis are illustrated in Figs. 1 and 2. The similar species, G. orientalis, which is often cultivated as a forage (Varis 1986; Fairey et al. 2000; Raig et al. 2001), occupies cooler temperate habitats which partly overlap those of G. officinalis in Eurasia (Baležentienė 2011), however, it is not known to occur outside of cultivation in North America. This species differs from G. officinalis in: horizontal rhizomes present (versus rhizomes lacking); reflexed mature pods (versus erect or spreading fruits) which are pubescent (versus glabrous); oval to orbicular stipules (versus deltoid to sagittate); distinctly pubescent calyx (versus glabrous or sparsely pubescent); calyx teeth shorter than the tube (versus about as long as the tube); and, the somewhat larger, more broadly ovate leaflets (30–60 × 10–25, versus 7–50 × 4–17 mm in G. officinalis). The seeds of G. orientalis tend to be more yellowish and lustrous.

The two Galega species can also be readily distinguished by differences in DNA nucleotide sequences at several nuclear and chloroplast loci. Particularly useful genes include the Nod-factor receptor 5 (nfr5) and nodulation receptor kinase (NORK) (Österman et al. 2011; S. Mechanda, (unpublished data); Appendix A).

The genus Astragalus is closely related and usually placed in the tribe Galegeae. With more than 2000 species, it is one of the largest angiosperm genera making morphological generalizations difficult. Species of Galega differ from Astragalus species in the following ways: the primary lateral veins of the leaflets extend to the leaflet margins (craspedodromous venation), while in Astragalus the primary lateral veins join each other and do not reach the leaflet margin (camptodromous venation) (Figs. 3D, 3E); the stamens are more or less monadelphous (versus usually didelphous); and, the pods have prominent oblique veins (versus transverse veins). In a vegetative state G. officinalis can be confused with cicer milk-vetch (Astragalus cicer L.), but the latter species has long rhizomes. In a reproductive state the yellowish flowers and rounded pods easily distinguish A. cicer. Weedy species of Vicia and Lathyrus in Canada are readily distinguished as climbing plants with tendrils present in place of the terminal leaflet. Crown vetch [Securigera varia (L.) Lassen, = Coronilla varia L.] may also be confused with G. officinalis, but the former has smaller flowers that are usually pink in colour in a globose umbel rather than a raceme, and the leaflets are smaller and broadly rounded at the tips without a mucro. Alfalfa (Medicago sativa L.) is also somewhat similar in growth habit, but in this species the pods are coiled, curled or falcate (rather than terete), the seeds are much smaller, and the leaves have only 3 leaflets. In central to western North America, the species might be confused with the native wild licorice (Glycyrrhiza lepidota Pursh), but the latter has solid (not hollow) stems, white to yellowish flowers with a strap-like standard and bur-like seed pods with hooked bristles.

In a survey by Lersten and Horner (2007), leaflets of G. officinalis were found to have calcium oxalate crystals in prisms along the vascular bundles, whereas crystals were lacking in the majority of examined species in the tribe Galegeae, including Astragalus spp. and Oxytropis spp. Peters et al. (2010) suggested that the absence of forisomes (spindle-shaped crystalline P-proteins that regulate phloem transport in the sieve tubes) in G. officinalis and some related species, might be related to the similarly unusual presence of calcium oxalate crystals in those species, both suggesting an unusual mechanism of calcium management in the leaves.

(c) Intraspecific variation

Variation within G. officinalis can be found in the flower colour, which ranges from bluish-purple to white, and in the number and shapes of leaflets in the leaves (see Section 1). Garden cultivars have been bred to intensify colour differences and nurseries offer plants with large white or deep purple racemes. An unusual sport with a simple leaf (not pinnate) has been described from Essex, UK (Mullin 1983).

In Turkey, six naturally occurring populations from the Middle Black Sea region differed significantly in number of flowers, flower nitrogen levels, above-ground biomass, flower biomass, leaf width, leaf length, root biomass, and reproductive effort, suggesting the importance of genotype as well as local environmental conditions on phenotypic plasticity (Özbucak et al. 2005).

Genetic variability was assessed by Wang et al. (2012) in 35 populations of G. officinalis from Europe and Russia using inter-simple sequence repeat (ISSR) and sequence-related amplified polymorphism (SRAP) DNA markers. Considerable variation was detected between the sampled populations. Ten ISSR primers produced a total of 100 bands which were an average of 77% polymorphic (expected heterozygosity = 0.292). Seven SRAP primes produced 88 bands which were an average of 67% polymorphic (expected heterozygosity = 0.257). Similarly, high levels of variability have also been found in morphological characteristics (Wang et al. 2008).

(d) Illustrations

The whole mature plant of Galega officinalis is illustrated in Fig. 1, along with details of a stipule and flower. A dense population at a disturbed site along the Ottawa River (Ontario, Canada) (Fig. 2) shows the competitive nature of the goat’s-rue symbiosis in a disturbed habitat. Morphological details of a stipule, seeds, seedling and leaf venation are illustrated in Fig. 3. Other illustrations of G. officinalis may be found in Stebler and Schröter (1889), Vasey (1893), Step (1896), Gams (1924), Gorshkova (1971), Barneby (1989), Eckel (2004), CFIA (2012) and CABI (2019). An accurate colour plate of morphological characteristics was published in Flora von Deutschland, Österreich und der Schweiz, and there are several copies available at different Internet websites (Thomé 1905).

(a) Detrimental

While Galega officinalis has a long history of economic uses in Europe, it is considered a noxious weed in most areas where it has been introduced. As a weed, it can form dense thickets and monocultures in pastures and meadows, reduce yields of better forage plants, contaminate pedigree seed crops, cause toxic injury to livestock, and can compete with and crowd out native flora (Evans et al. 1997; Wiersema and León 1999; Guitart et al. 2010; Fraiture 2014 ; Oregon Department of Agriculture 2015; CABI 2019). Its weediness and toxicity has been considered a particular problem in the United States, Argentina, Chile and New Zealand (Holm et al. 1991).

In the United States, where it was introduced as a trial forage plant, medicinal plant, and ornamental, the plant has escaped and become weedy, particularly in Utah (Evans 1984) and sporadically in other states (see Section 6). It has been reported as infesting irrigated pastures, roadways, ditch banks, fence lines, wet areas and alfalfa crops (Evans 1984; Patterson 1992, 1993). The species is not only subject to U.S. federal legislation but has also been targeted in state programs for prohibition or eradication (see Section 3c). In Great Britain, G. officinalis is among alien plants listed as persistent, weedy, garden escapes (Salisbury 1961). It is listed as a noxious invasive weed in eastern France with a moderately negative economic impact on local biodiversity and agriculture (EPPO 2008). In Argentina the plant is considered a weed in pastures and vegetable crops where it can be a host to insects that attack crops (Liljeström and Rabinovich 2004). In Chile, where it was introduced as a forage in 1872, it has become a weed of pastures and crops (Oehrens and González 1975; Ellison and Barreto 2004), including vineyards (Longone et al. 2011).

A major concern is the potential for poisoning of livestock due to the toxic properties of alkaloids (Williams 1978, 1980; DiTomaso 1994), including guanidine, galegine (isoamylene guanidine), hydroxygalegin and galuteolin (Barger and White 1923a, 1923b; Pufahl and Schreiber 1961). Galegine is synthesized in seedlings, leaves, flowers and fruit with highest levels occurring in the seed and the levels in plants increasing through the flowering, fruiting and seed maturation stages (Reuter 1962; Oldham 2009; Oldham et al. 2011). However, because of the bitter alkaloids, mature plants are generally unpalatable to livestock (Tingey 1971; Williams 1978, 1980), and intoxication is largely restricted to times of drought or under other conditions which limit the availability of alternative forage, such as when contaminated fodder has been fed to animals (Durieux 1968; Williams 1978; Puyt et al. 1980; Poulet-Wolgust et al. 2012; Fraiture 2014). Poisoning generally occurs when the plant is at flowering or fruiting stages (Parton and Bruere 2002; De Otazúa et al. 2009), with young plants being much less toxic. Similarly, mature seed was more toxic to sheep than semi-mature seed in feeding trials by Keeler et al. (1986).

Observed clinical manifestations after consumption of G. officinalis include dyspnea, anoxia, foaming nasal discharge, vomiting, pulmonary congestion, edema and hydrothorax lesions (Keeler et al. 1988; Lasseigne 2003; Roch et al. 2007). Mortality can occur 24 h or less after ingestion (Durieux 1968; Puyt et al. 1981). Clinical signs were observed at 0.8 g G. officinalis per kg of sheep body weight, and mortality was observed at 10 g G. officinalis per kg of sheep body weight (Keeler et al. 1986, 1988). Mortality was observed in sheep fed approximately 0.7% of body weight of G. officinalis dry matter by De Otazúa et al. (2009), depending on individual plant toxin levels and animal sensitivity. Dried aerial parts fed to rats at a rate of 5 g·kg⁻¹ did not result in any mortalities or any clinical signs of toxicity although there was evidence that significant liver and lung alterations had occurred suggesting these organs are the target of G. officinalis toxicosis (Rasekh et al. 2008). In German tests, alcoholic extracts from seeds and leaves were poisonous to mice, the average lethal dose of galegine sulphate amounting to 77.5 mg·kg⁻¹ body weight (Köhler 1969). Susceptibility of animals to experimental poisoning has shown considerable variation and the reasons for this remain unclear (Keeler et al. 1988).

Confirmed reports of livestock poisoning by G. officinalis are relatively rare. Most cases have been reported from southern and central France, where such reports go back to the late 19th Century (Faliu et al. 1981, 1985). Sheep are the primary victims of poisoning, and mortality rates from 10%–50% have been reported (Faliu et al. 1981; Puyt et al. 1981; Gresham and Booth 1991; Guitart et al. 2010). Severe outbreaks with numerous deaths following consumption of contaminated fodder have been reported in sheep (Durieux 1968; Puyt et al. 1980, 1981; Bézard et al. 2002; Poulet-Wolgust et al. 2012; Fraiture 2014), goats (Lasseigne 2003) and cattle (Roch et al. 2007). In the United Kingdom, a number of sheep died with symptoms of goat’s-rue poisoning when put to graze in a newly-seeded pasture adjacent to an embankment infested with G. officinalis (Gresham and Booth 1991). Poisoning in sheep has also been reported from New Zealand (West 1982). In Canada and the United States, toxicity of the plant has been of little significance to date, because of its limited and scattered distribution (Burrows and Tyrl 2013).

(b) Beneficial

As a folk remedy, G. officinalis has been used for centuries in Europe for a variety of purposes. It has been suggested that it was a plant known to Dioscorides and Pliny the Elder, although it has not been unambiguously identified in their works. Traditional uses include as a poison antidote, mild astringent, vermifuge, plague treatment, anti-convulsive, anti-inflammatory compress, pot or salad herb, egg production stimulant for hens, and galactagogue to increase lactation (Gerard 1597; Culpeper 1650; Loudon 1849; Rasekh et al. 2008). Actual and potential medicinal uses of various secondary metabolites of G. officinalis were reported by Karakaş et al. (2016a, 2016b), Nagalievska et al. (2018) and Atanasov et al. (2019). The common practice of using the plant in livestock feed to increase milk production led to the ancient common name galega (see, for example, Gerard 1597) and to the Linnaean generic name Galega, from the Greek “gala” for milk. Extracts have also been used to treat reproductive disorders and enhance reproduction in goats and sheep (Viegi et al. 2003). As a fodder containing phytoestrogens (see Section 7c), it appears to promote the estrogenic receptors and increase both the length of lactation and milk volume in ewes (González-Andrés et al. 2004; Tabares et al. 2014). A similar effect has also been reported for cows (Hanelt 2001; Witters 2001; González-Andrés et al. 2004) and rabbits (Pałka et al. 2019).

Used in traditional herbal medicine, and now widely available on the Internet as a herbal remedy, there is doubt concerning the safety of extracts of G. officinalis as a galactagogue or hypoglycaemic for human use (Duke et al. 2002; Zuppa et al. 2010; Zecca et al. 2016). Secondary metabolite production has been shown to vary greatly with abiotic stress, indicating that appropriate dosage levels may be difficult to determine in crude plant preparations (Karakaş and Bozat 2020).

Extracts of G. officinalis have been shown to have a significant impact on glucose transport (Neef et al. 1996), although physiological actions are poorly known. In Bulgaria (where G. officinalis is endemic) and Chile (where it is introduced), the aerial parts of the plant have been traditionally used as a hypoglycemic and diuretic (Muñoz et al. 1981; Lemus et al. 1999; Kiselova et al. 2006; Atanasov et al. 2019). A concoction derived from this species had been used in mediaeval Europe to treat symptoms including frequent urination, a symptom of diabetes (Witters 2001). Such practices led to studies early in the 20th Century of potentially useful compounds in the species to treat type 2 diabetes focusing on guanidine and galegine. Subsequently, various new anti-hyperglycaemic compounds were synthesized, including the much less toxic dimethylbiguanide, also known as Metformin or Glucophage (Bailey et al. 1996; Witters 2001; Bailey and Day 2004; Howlett and Bailey 2007). Metformin was introduced into clinical practice in Europe as a treatment for hyperglycemia in the late 1950s (Bailey and Day 2004; Goetz 2007; Goetz and Le Jeune 2008). The drug was introduced to Canada in 1972 and to the United States in 1995 (Bailey et al. 1996; Dowling et al. 2011) and has been prescribed world-wide (Hadden 2005; Mentreddy 2007; Dowling et al. 2011). Metformin has also been studied in the treatment of obesity (Campbell and Howlett 1995; Witters 2001) and polycystic ovary syndrome, as well as showing antiviral and anticancer activity (Dowling et al. 2011). Galegine in G. officinalis has been found to cause weight reduction in mice (Palit et al. 1999; Mooney et al. 2008), but, as with guanidine, subsequent studies have synthesized more effective analogues (Coxon et al. 2009).

In addition to the alkaloids, some of the 48 detected phenolic compounds may contribute to the hypoglycemic activity observed in G. officinalis (Barchuk et al. 2017). Experiments on rats with induced hyperglycemia showed that blood glucose concentration was significantly reduced when treated with saponins, tannins and glycosides extracted from G. officinalis (Luka and Omoniwa 2012).

Galega officinalis has reportedly been used medicinally in hand and foot baths, and to improve skin healing through antibacterial and antifungal activity (Pundarikakshudu et al. 2001; Özbucak et al. 2005; Ertürk 2010; Karakaş et al. 2012). A widely used biguanide compound, chlorhexidine, is a useful germicide and disinfectant (Hadden 2005). Anti-microbiological properties of G. officinalis have been investigated in Turkey, where extracts of leaves and shoots have been found to be effective against bacteria and to a lesser extent against fungi (Özbucak et al. 2005; Karakaş et al. 2012, 2016a, 2016b). Extracts from the plant were found to be inhibitory on both gram-positive and gram-negative bacteria; and, significant tumour inhibition (98% with aqueous extract) was obtained in assays with Agrobacterium tumefaciens-induced potato disk tumors (Karakaş et al. 2012).

An aqueous extract of compounds from G. officinalis was found to have an anti-coagulation effect (Atanasov 1994; Atanasov and Spasov 2000). The fraction inhibiting platelet aggregation contained 19%–23% protein and about 74% polysaccharides, with the high biological activity being attributed to its protein component (Atanasov et al. 2003).

Aqueous extracts (1:4 weight to volume) of G. officinalis were tested on the dagger nematode Xiphinema index in Chile for nematicidal activity (Insunza et al. 2001). Extracts of leaves and flowers killed 100% of the nematodes after 24 h exposure, while root extracts killed 91.7% of the nematodes.

In Europe G. officinalis has been cultivated as forage and green manure (Whyte et al. 1953; Ball 1968; Uphof 1968; Hanelt 2001) as well as for soil amelioration (Našinec and Nĕmcová 1990). Cultivars have been developed that are adapted to acid soils (see Section 5b) and cold climates (Našinec and Nĕmcová 1990). It has been used as a forage crop in Switzerland (Stebler and Schröter 1889), Italy (Peiretti and Gai 2006; Peiretti 2009), Spain (González-Andrés et al. 2004), Finland (Laakso et al. 1990), and China (Xu et al. 2010). As a crop, it is highly productive and might be improved with modern breeding techniques designed to reduce toxicity (Našinec and Nĕmcová 1990). Forage and cutting for fodder are currently best done prior to flowering when alkaloid and phenolic content is low and plants are most palatable and nutritious (Stebler and Schröter 1889; Parton and Bruere 2002; De Otazúa et al. 2009; Oldham et al. 2011; Section 7c).

In Europe and Russia, G. officinalis has been investigated for potential bioremediation of hydrocarbon contaminated soils (Našinec and Nĕmcová 1990; Lyubun and Tychinin 2007).

Pollen of G. officinalis has been a valuable source of food for bee colonies in several areas. It has been grown as bee forage in southern Europe, Germany and Switzerland (Whyte et al. 1953; Hanelt 2001). In Ankara province of north-central Turkey, it was among the top preferred plant species of bumblebees (Apidae: Hymenoptera) (Aytekin et al. 2002). Pollen of G. officinalis has been identified as a component in honey from a few sites in Chile (Montenegro et al. 2004; Montenegro et al. 2010), in one small district (Maniwatu) in New Zealand (Moar 1985), and in Italy (Mercuri and Porrini 1991; Canini et al. 2009).

Various other uses of G. officinalis have been reported. The leaves have been cooked like spinach (Gerard 1597) and plant extracts have been used as a substitute for rennet in the manufacture of curdled milk products (Facciola 1990). An extract claimed to inhibit tyrosinase activity in the production of melanin has been promoted as a skin whitener (Lee et al. 2012) and skin conditioner products containing “galega officinalis extract” (Chemical Abstracts Service registry number 84650-07-7) are available for sale.

In Europe, G. officinalis has been used as a garden ornamental plant for centuries (Gerard 1597; Loudon 1849; Stebler and Schröter 1889; Step 1896; Hedrick 1919; Gams 1924; Hellyer 1955; Salisbury 1961; Polunin 1969) and horticultural trade may have been a significant pathway in its establishment to other countries prior to widespread soil quarantine regulations (see Section 6). Various cultivars have been previously imported for North American gardens (Macoun 1908; Bailey and Bailey 1976), although regulations now prevent such trade.

(c) Legislation

In Canada, G. officinalis is listed as a class 2 primary noxious weed under the Weed Seeds Order (CFIA 2018; Canada Gazette Vol. 150, No. 10 — May 18, 2016) where tolerance levels are set for contamination in traded seed commodities. Its import is also regulated under the Plant Protection Act (CFIA 2018). It is, however, not listed on any provincial weed legislation.

In the United States, it is listed on the U.S. Plant Protection Act and Federal Seed Act (USDA-APHIS 2020). It is also considered a noxious or quarantine weed under State regulations in Alabama, California, Florida, Massachusetts, Minnesota, North Carolina, Nevada, Oregon, Pennsylvania, South Carolina, Vermont and Washington (USDA-NRCS 2020). In 1981 a program was established by the United States Department of Agriculture providing funding for the eradication of goat’s-rue. In Cache County, Utah, the population size was reduced by as much as 95%, but the species has not yet been eliminated (Westbrooks 1993; Evans et al. 1997).

In New Zealand, G. officinalis is a regionally regulated weed. To reduce the occurrence and impact in the Hawke’s Bay region, G. officinalis has been designated as a “total control” pest where land occupiers must destroy all plants before the production of mature seed (Hawke’s Bay Regional Council 2004). In the Auckland region it has been designated a “surveillance pest plant”, where it is banned from sale, propagation, distribution and exhibition (Auckland Council 2020).

(a) Climatic requirements

It is reported that Galega officinalis has a low winter-hardiness (Whyte et al. 1953) which should tend to restrict its spread into and within colder climate zones. In Canada, the species has naturalized in continental climate zones in eastern Canada between 43° and 47° N latitude, and was found at 49° in the Pacific maritime zone of British Columbia. This suggests that it can tolerate an annual minimum temperature of at least –35 °C, which would include most of the important agricultural production areas in southern Canada.

In the United States, the plant has spread widely in Utah in a mid-latitude desert zone (Evans 1984). In greenhouse studies in that state, growth of G. officinalis was favoured by a long photoperiod for flowering (16–18 h), daily temperatures of 26–29 °C and was poorly adapted to large diurnal temperature fluctuations (Patterson 1992, 1993). Growth increased with elevated daytime temperatures from 15 to 29 °C but declined at 36 °C (Patterson 1993). In Pennsylvania it is established (Pennsylvania Department of Agriculture 2011) in a mid-latitude temperate climate. In Oregon and Washington, the plant grows in a Pacific climate zone similar to southern British Columbia.

Using observed growth responses of G. officinalis to temperature and diurnal fluctuations, Patterson (1993) generated a model to predict growth rates at various locations in the United States based on day/night temperature normals from May to September. His model predicts that optimal growth would be achieved in areas with a daytime average of 23 °C or greater and a diurnal change of 7 °C or less, geographically corresponding largely to areas in the southeastern United States. Since this model was formulated from observations of plants reared from surface-sterilized seed established in growth chambers, it is not clear whether effectively nodulating symbionts (see Section 7e) were present or whether their presence would affect growth responses and alter the model predictions.

In Europe, G. officinalis is widespread along river valleys, such as the Rhine, Danube, Thaya and Volga (Gams 1924). Further east, near the Black and Caspian Seas, as in northern Turkey (Özbucak et al. 2005), southern Russia, Ukraine and the Caucasus (Dzyubenko and Dzyubenko 2003), it is found in temperate to mild temperate lowland and sub-montane areas. It is rare in the Mediterranean climate of south-west Turkey (Bennett et al. 1998). It is rare in the United Kingdom outside of central and southeastern England (Biological Records Centre 2014b) and in Switzerland it is naturalized in temperate areas on low-lying lands sheltered from extreme cold (Stebler and Schröter 1889). In Spain, it has been successfully grown in regions with an average annual rainfall 440 mm and average daily maximum and minimum temperatures of 18.8 °C and 6.1 °C (González-Andrés et al. 2004). Through much of Italy it commonly occurs in natural pastures mostly below 1000 m elevation (Peiretti 2009). It is cultivated in areas of northern Italy with high precipitation in April, May and October and little rainfall in summer and winter with mean daily temperatures of 0.5 °C in January and 22 °C in July (Peiretti and Gai 2006). In Argentina, plants are found in areas which are humid and warm with an annual precipitation of 950 mm and maximum daily summer temperature of 22 °C (Ansín 2001). In tropical Ecuador, it is found only at 2500–3500 m asl. (metres above sea level) in the Andes (Jørgensen and Léon-Yanez 1999).

(b) Substratum

In Canada, G. officinalis has been reported from clay soils and rocky landfill in the Ottawa area (Reddoch and Reddoch 2000). Analysis of soil samples from five sites harbouring G. officinalis populations in Ontario, showed a pH of 7.4–7.8 (Bromfield et al. 2019), 140–360 parts per million (ppm) potassium, 120–320 ppm magnesium, 10–37 ppm phosphorus, 1–34 ppm nitrate (NO_3-) and negligible amounts of nitrite (NO_2-); and, the carbon/nitrogen ratio ranged from 7.2–17.0% [Eden Bromfield, (unpublished data)].

In Cache County, Utah, G. officinalis is found in clay loam and loam soil pH 7.3–7.5 with 3.3–5.7% organic matter (Oldham 2009). In New York, it was growing at dumps covered with gritty soil with a high percentage of chert in the form of pebbles (Ahles 1951). In Switzerland, the plant was observed to satisfy its nutritive requirements from the subsoil, so that the quality of the topsoil was deemed of little importance; however, it grew best in deep humus soil where moisture was available to its deep taproots (Stebler and Schröter 1889). Naturalized populations in Spain have been found on solonetzic soils (alfisol typic rhodoxeralf) of pH 7.7, and plants have been cultivated in experimental field plots on regosolic soil (entisol typic xerorthent) of pH 8.0 and solonetzic soil (alfisol typic palexeralf) of pH 8.2 (González-Andrés et al. 2004). In the Black Sea region of Turkey, G. officinalis is a glycophyte usually found in weakly acidic to alkaline soils rich in nitrogen (Özbucak et al. 2005), but around the northern and eastern shores of the Black Sea, the plant is tolerant of saline soils (Dzyubenko and Dzyubenko 2003). Plants thrive in mesic habitats with deep soil in the Czech republic (Kubešová et al. 2010). Along the walls of the upper tidal Thames River in the United Kingdom, plants are commonly found in waste-ground among boulders (Francis and Hoggart 2012), where the underlying soil is of alluvial deposits. In Argentina and central Chile, the plant has spread in rich and humid soils (Whyte et al. 1953).

In Canada, populations of G. officinalis were found at sites with soils above pH 7.0 (Bromfield et al. 2019) which is consistent with reports from the United States (Oldham and Ransom 2009) and Spain (González-Andrés et al. 2004) of plants growing in soils with pH ranges of 7.3–7.5 and 7.7–8.2, respectively. This suggests that the apparent adaptation of G. officinalis to soils above pH 7.0 may act in concert with the high level of specificity between host plant and symbiotic bacterium (see Section 7e) to limit the spread of the plant in new environments (Bromfield et al. 2019).

To assess the adaptation of G. officinalis as a fodder crop to saline soils in northeastern Europe, Egamberdieva et al. (2013) established a greenhouse experiment which involved inoculation of salt-stressed plants either with the symbiotic bacterium Neorhizobium galegae symbiovar officinalis (see Section 7e) alone, or co-inoculation with two growth-promoting bacteria, Pseudomonas extremorientalis and P. trivialis. Increasing salt concentrations decreased the ability of N. galegae alone to colonize the roots, whereas co-inoculation significantly alleviated the effects of salt stress, increasing nodulation and the nitrogen content of the shoots and, to a lesser extent, the roots.

(c) Communities in which the species occurs

In Canada, G. officinalis has naturalized along riverbanks, ditches, channels and other waterways and sometimes along roadsides, abandoned fields or landfill sites, usually not far from waterways (herbarium specimen label data). It can form dense patches (Fig. 2) in communities which are usually open or partly shaded and consist primarily of introduced grasses and herbaceous plants.

In the United States, G. officinalis has also been found mostly near waterways. In Utah, it spread from the forage testing fields into natural seepage areas, ditch-banks, marshes, irrigated pastures, high line canals and drainage systems (Tingey 1971; Evans 1984; Patterson 1992; Welsh et al. 1993). In Philadelphia, it was first detected along roadsides and floodplains in and around the Morris Arboretum (Stokes 1964; Klugh 1998), but has subsequently spread to other parts of Pennsylvania (Pennsylvania Department of Agriculture 2011). In New York, G. officinalis was first seen along river banks in Bronx County, with such plants as Lotus corniculatus L., Coriandrum sativum L., Diplotaxis sp., Artemisia annua L. and Matricaria inodora L. (Ahles 1951).

In its native range, the species occurs in a variety of habitats. In Turkey, it is found on mountain sides, roadsides, edge of lakes, beside fields and streams and in scrub woodlands at low elevations between 5–550 m asl (Davis 1970; Özbucak et al. 2005). In the former USSR (including the southern Russian Federation, southern Ukraine, Georgia, Azerbaijan, Armenia and western Kazakhstan) it occurs on riverbanks and in river valleys, meadows, scrub, beech woods, roadsides and ravines (Gorshkova 1971). It is among pioneer flora in disturbed sites in the western Caucasus, generally present at lower latitudes than G. orientalis (Andronov et al. 2003) and is among native species of closed communities along shoals of western Caucasian rivers at elevations ranging from 70–80 m asl., where occurrence was relatively low, to 800–900 m asl., where occurrence was considerably greater (Akatov and Akatova 2010). In the Italian Alps, G. officinalis is found in grazed natural pastures (Peiretti 2009) at elevations of 260–523 m asl. (Siniscalco et al. 2011). In Tuscany, it occurs primarily in wetlands and is common in pastures, along with hemp-agrimony (Eupatorium cannabinum L.) and field horsetail (Equisetum arvense L.), less frequent along canals and watercourses dominated by galingale (Cyperus longus L.) and rare in deciduous oak-dominated woods and inland swamps dominated by devil’s beggarticks (Bidens frondosa L.) and climbing nightshade (Solanum dulcamara L.) (Giallonardo et al. 2011). In New Zealand it is found mostly along waterways and in wetlands (Webb et al. 1988; Parton and Bruere 2002).

(a) Morphology

The plant’s strong taproot provides access to deep soil moisture under drought conditions and to nutrients in the subsoil when growing in disturbed and nutrient-poor soils. The lax stem bases reduce the impact of mowing for fodder production or control. After cutting, axillary buds at the base of the stem quickly form a new set of stems. Lateral branches originating from buds in the axils of the radical leaves will develop into lateral branches in the first or second year (Stebler and Schröter 1889).

(b) Perennation

A long-lived perennial. Unless killed by frost, the large tap roots (see illustration in King County 2018) over-winter and sprout new stems from the caudex each spring.

(c) Physiology and biochemistry

An early nutritional analysis of G. officinalis as a fodder plant noted that prior to flowering, plants contained 78.7% organic matter, 17.1% nitrogen, 1.4% fat, and 34.4% fibre (Stebler and Schröter 1889). A more recent Italian study simulating forage production have shown a decline in nutritional quality and digestibility through the growth stages of rosette, shooting and budding, although some components were partially restored in the post-cut regrowth stage (Peiretti and Gai 2006; Peiretti 2009). Palmitic, linoleic and α-linolenic represented 82%–85% of the fatty acid content during the mid-season growth stage and in late-season regrowth. These percentages changed little during development; however, α-linolenic proportion was significantly higher at the regrowth stage. Organic matter, neutral detergent fibre, acid detergent fibre, lignin and gross energy increased during maturation, while crude protein, ether extract, ash and organic matter digestibility declined with plant maturation. Amino acid proportions did not vary significantly across growth stages, with only serine showing a significant change with development (highest at shooting and regrowth stages). In overall dry matter and chemical content, dry matter content increased from 111–157 g·kg⁻¹ during first growth stages, and to 180 g·kg⁻¹ at the regrowth stage; total nitrogen decreased from 31.6 g·kg⁻¹ to 20.2 g·kg⁻¹ (dry matter) from vegetative to budding stages, but increased to 33.5 g·kg⁻¹ during the regrowth stage; soluble nitrogen decreased during growth from 5.9 g·kg⁻¹ to 4.8 g·kg⁻¹ and then increased to 5.4 g·kg⁻¹ at regrowth. No significant changes were noted in water soluble carbohydrates (74–82 g·kg⁻¹) or pH (5.6–5.7). The fermentation and quality of ensiled G. officinalis fodder varied with harvest growth stage and post-harvest wilting time.

The free amino acid canavanine is present in seeds of the genus Galega (Turner and Harborne 1967), perhaps functioning as both an anti-herbivore defense and nitrogen source for the embryo.

The most toxic molecules in G. officinalis include the guanidine alkaloids galegine (dimethylallylguanidine), hydroxygalegin and peganine (= vasicine) (Schäfer and Stein 1969). Galegine is found in increasing concentrations from the stem to leaves to the immature pod stage, with no storage in the roots and with some decrease at the ripe seed stage (Reuter 1962; Oldham et al. 2011). The quinazoline alkaloid, vasicine, was once believed to be responsible for the bitter taste avoided by animals, but later analysis suggested that the vasicine content (0.1%–0.35%) was too low to be responsible for reducing the fodder acceptability to livestock (Laakso et al. 1990).

In France, a rare norterpenoid glucoside, dearabinosyl pneumonanthoside, and several phytoestrogens (flavonol triglycosides), were isolated from the vegetative parts of G. officinalis by Champavier et al. (1999, 2000). Flavonoids absorbing UV radiation, isorhamnetin, kaempherol and quercetin, were reported by Kay (1987) in the flowers of G. officinalis. Luka and Omoniwa (2012) identified alkaloids, flavonoids, saponins, tannins, cardiac glycosides, phenols, resins, terpens and steroids from aqueous extracts of whole plants. A number of compounds extracted with methanol were identified by Fukunaga et al. (1987). In a study by Kiselova et al. (2006), aqueous extracts of G. officinalis had a total phenol content of 361.48 (± 1.30) μM (quercetin equivalent), indicating a moderate antioxidant capacity.

In a comparative study of plants grown from seed and nodal explants, Karakaş et al. (2016a) found 25% less total phenolic content (methanol extract) present in explants than in plants grown from seed. Although the total phenolic content was lower, observed levels of apigenin, luteolin and chlorogenic acid were higher in explants (10×, 100× and 2×, respectively). Shymanska et al. (2020) found that phenolic compounds were at the highest concentrations in flower buds and flowers.

Galega officinalis is one of the Fabaceae species with an attenuated extended bundle sheath (EBS) system (paraveinal mesophyll), consisting of digitated cells positioned between the palisade and spongy mesophyll of the leaf. The EBS tissue cells mainly join the spongy mesophyll cells lying between the veins rather than bridging between all of the vein islets, as occurs in species with a continuous EBS system (Franceschi and Giaquinta 1983a, 1983b; Kevekordes et al. 1988). The species is somewhat anomalous in having two layers of bridging cells, but these do not appear to extend throughout the large interveinal regions of the leaf. These specialized cells may act as routes for phloem loading of photosynthesis assimilates (Franceschi and Giaquinta 1983a) or as an EBS recovery system for concentrated amino acid solutes in the transpiration stream (Canny 1988, 1990).

(d) Phenology

The embryo of G. officinalis develops entirely from the apical cell of the zygote, while the basal cell remains part of the suspensor enabling nutrients to reach the developing embryo (Souèges 1949; Packa 2001).

Leaflets of emerging leaves are initially folded along the midrib, and unfurl as they develop (Klugh 1998). Cauline leaves develop alternately on each stem and flowering racemes emerge successively from leaf axils. Flowers and fruits develop acropetally along the racemes. In a laboratory experiment using different temperature regimes (Patterson 1992), the first flower buds appeared at an average 86, 46 and 40 d after seedling emergence at photoperiods of 14, 16 and 18 h, respectively, indicating a large variation in development times depending upon climate and light conditions. Fruits remain attached as they mature and dehisce (Stebler and Schröter 1889). At senescence, which may not occur until first-frost, pods dehisce along both sutures with the valves twisting to release the seeds (Kirkbride et al. 2003). The indeterminant development pattern results in plants bearing flowers and fruits at various stages of development throughout much of the growing season.

Plants in eastern Canadian populations are in flower by late June with fruits developing from early July into September; plant senescence commences in early fall and becomes complete with the first hard frost (herbarium specimen label data). In Pennsylvania, flowering was reported as beginning in June and continuing throughout the growing season (Klugh 1998) and although the upper parts of plants senesced in September, late-season new growth was sometimes observed at plant bases (Stokes 1964). In the Intermountain region of the western United States, the growing season was reported as late April to August (Barneby 1989) and from June to October in Washington State (King County 2018). Two fodder cuttings per season were obtained from plants grown for ensilage in Virginia (Piper 1916).

The flowering season for Europe as a whole was given as June to August by Polunin (1969) and in England, plants were reported to flower from July to August by Gerard (1597). In Switzerland, plants sown in early spring began flowering between the end of June and late July and, after cutting at the early flowering stage, quickly produced new stems from the base and began flowering again in late summer (Stebler and Schröter 1889). In Italy, plants sown in early spring were in full leaf by late June, began to shoot one week later, and began to bud two weeks later (Peiretti and Gai 2006).

Seedling relative growth rate as calculated in Czech trials (Moravcová et al. 2010) was 0.160 g g^–1 day⁻¹. In growth chamber trials (Patterson 1992, 1993), vegetative growth patterns (resource allocation) were significantly affected by variations in day and night temperatures. After 89 d plants attained maximum dry weight at day/night temperatures of 26/22 °C, while plants grown in regimes with greater day/night fluctuation (12–20 °C) or a higher day temperature (34 °C) were smaller. Root biomass, averaged over a period from 14 to 89 d after emergence, was highest at 26/14 °C, and least at 34/22 °C. In a comparison of the root weight ratio with that of alfalfa, G. officinalis tended to partition more biomass into roots than the latter at night temperatures of 18 °C or less, with a tendency for proportional root biomass to increase with decreasing night temperature. The greatest period of root growth was between 28 and 60 d after emergence. Given the response to temperature, photoperiod and moisture, Patterson (1992, 1993) predicted that G. officinalis could potentially flourish in large areas of North America, especially where conditions are suitable for alfalfa production.

(e) Mycorrhizal and bacterial symbioses

Mycorrhiza are fungi that form a symbiotic associations with plant roots and facilitate the uptake of nutrients (Bolan 1991). Mycorrhizal relationships of Galega species have not been well studied, although Palta et al. (2016) reported that roots of G. officinalis were colonized by mycorrhizal fungi and Püschel et al. (2011) reported that the growth of G. orientalis responded positively to soil inoculation with mycorrhizae.

Like most of the Fabaceae, species of Galega form a symbiotic association with soil bacteria (rhizobia) which stimulate the formation of swellings called nodules on the roots of the plant and convert atmospheric nitrogen (N₂) into forms of ammonia (NH₃) that are used by the host plant. In exchange, the bacteria receive energy in the form of fixed carbon from the plant host (Proctor and Moustafa 1962, 1963).

The two most studied species of Galega, G. officinalis and G. orientalis, form a highly specific symbiotic association only with the bacterial species Neorhizobium galegae which consists of two distinct symbiotic varieties (symbiovars): orientalis and officinalis. Both of these symbiovars form nodules on the two plant species but fix nitrogen only on their respective Galega host species (Lindström et al. 1983; Lipsanen and Lindström 1988; Lindström 1989; Kaijalainen and Lindström 1989; Räsänen et al. 1991; Mousavi et al. 2014; Österman et al. 2014; Karasev et al. 2019). The mechanisms controlling the specific interaction between Galega plants and their respective bacterial symbiont are poorly understood (Franche et al. 2009; Österman et al. 2011). Genome sequencing of several strains of N. galegae indicated few differences between the symbiovars officinalis and orientalis, although a single symbiosis related gene was found to be specific for symbiovar orientalis; the function of this gene remains to be determined (Österman et al. 2014, 2015).

In New Zealand, Liu et al. (2012) identified 2 strains of N. galegae isolated from root nodules of G. officinalis. Phylogenetic analyses of four housekeeping (16S rRNA, atpD, glnII, and recA) and two symbiosis gene (nodC and nifH) sequences indicated that 50 bacterial isolates from root-nodules of G. officinalis plants at five Canadian sites were identical to strains of N. galegae symbiovar officinalis originating either from Europe or the Caucasus (Bromfield et al. 2019). Moreover, plant tests with G. officinalis indicated that soils collected from four Canadian sites without a history of agriculture or presence of G. officinalis were deficient in symbiotic bacteria capable of eliciting nodules on this plant. In this connection, it is noteworthy that strains of the species, N. galegae, were not represented in culture collections of rhizobia isolated from diverse sites across Canada (Prévost and Bromfield 2003). Collectively, these data suggest that anthropogenic activities are responsible for the co-introduction of G. officinalis and its specific bacterial symbiont (N. galegae symbiovar officinalis) into Canada from the Old World.

Although G. officinalis can grow and reproduce under cultivation without its effective bacterial symbiont (González-Andrés et al. 2004), establishment and persistence in natural vegetation communities is highly unlikely. Cultivation trials in Estonia showed biomass production of the related G. orientalis was increased by 80% when plants were inoculated with the appropriate bacterial symbiont, but by only 35% when fertilized with supplemental nitrogen in the absence of symbionts (Raig et al. 2001). At sites in Spain, where G. officinalis and its symbiont, N. galegae, are naturalized, 5.9 × 10⁴ infective bacterial cells per gram of soil were found (González-Andrés et al. 2004).

The establishment of G. officinalis populations in Canada outside of cultivation requires the co-introduction of both the plant host and its symbiotic bacterium. This is most likely accomplished through human transport of soil containing both seeds and bacteria (Bromfield et al. 2019).

(a) Floral biology

The flowers, like those of all Fabaceae, are hermaphrodite, however, plants are not able to self-pollinate as enclosed plants do not produce fruits (Fruwirth 1906). Details of floral anatomy and the pollination mechanism are described by Knuth (1908). Plants are pollinated by bees which are attracted to the reddish-yellow pollen grains (Knuth 1908) and nectar is not produced (Knuth 1908; Rodríguez-Riaño et al. 1999). Flowers possess a monadelphous androecium in which the lower part of the filaments are fused to form a tube of ten stamens surrounding the ovary. The stigma protrudes beyond the staminal tube and the anthers dehisce before the flower opens (Knuth 1908).

This species was determined to be a preferred plant species for bumblebees in Turkey (Aytekin et al. 2002) and G. officinalis pollen has been identified in honey from Chile (Montenegro et al. 2004; Montenegro et al. 2010), New Zealand (Moar 1985) and Italy (Mercuri and Porrini 1991; Canini et al. 2009). The plant has been deliberately grown as a bee forage in Switzerland (Hanelt 2001). The plant was listed as being economically important for honey production by Wiersema and León (1999), but is only significant as a pollen source for sustaining colonies and not honey production itself.

(b) Seed production and dispersal

Goat’s-rue is a prolific seed producer. Plants collected in Canada averaged seven racemes per stem, with about 25 flowers per raceme, and 4–6 seeds per pod (approximately 1000 seeds per stem), and an average seed weight of 6.07 mg (S.J. Darbyshire, (unpublished data)). From cultivated plants in Switzerland, Stebler and Schröter (1889) estimated an average seed weight of 7.32 mg. In Pennsylvania, Klugh (1998) reported 20–50 flowers per raceme, producing pods with up to nine seeds. In Utah, Oldham and Ransom (2009) observed seed production ranging from 174–1230 seeds per stem (depending on stem density), while Evans (1984) reported of up to 25 000 seeds per plant averaging six per pod. In the Czech Republic, the average seed weight was reported as 6.56 mg, with 1652 seeds per stem (Moravcová et al. 2010).

Seeds initially fall close to the plant. An experiment to measure speed of seed fall showed a very low capacity for wind dispersal (Moravcová et al. 2010). The spread of the species along waterways or in wetlands (see Section 5c), suggests that water is the main natural dispersal mechanism for seeds (Faliu et al. 1985), individually, in pods, or perhaps occasionally whole plants. In Utah, seeds spread primarily through irrigation systems were subsequently found in nearly all fields downstream of the original infestation (Evans 1984; Keeler et al. 1986). Irrigation channels were also reported to be the main dispersal pathway in Chile (Oehrens and González 1975). This suggests that, although the seeds are not buoyant and have been observed to sink quickly (Klugh 1998; Moravcová et al. 2010), they could be carried in stages by fast-moving water which follows the opening of sluice gates or sudden flooding. Moravcová et al. (2010) reported that some of the seeds in an experimental test remained floating for up to 24 min before sinking. Although seeds do not float for long, they may still approach neutral buoyancy when submerged and be readily transported by moving water.

Seeds of plants not growing near flowing water are likely to remain near the plant, which, together with regrowth from, and expansion of the caudex, may account for the scattered and localized patterns in which many colonies continue to be found in North America.

Soil transport from sites with established G. officinalis populations appears to have contributed to seed dispersal over short to medium distances. In Canada, movement of contaminated soil by excavators and vehicles is believed to have led to the spread of the species in the Ottawa area away from original garden plantings or other locations. In Pennsylvania, dredging equipment at a lake led to dispersal of seed in the sediments, as well as transport of aggregate construction material (Pennsylvania Department of Agriculture 2011). In New York, the species was found where ballast had been dumped (Ahles 1951). In the United Kingdom, plants growing along a roadside in Essex were apparently the source of seed transferred via construction equipment to an embankment which was then converted to pasture for sheep (Gresham and Booth 1991; see Section 3a). Seed can be dispersed by farm equipment or as commodity contaminants (Patterson 1992), but there is little evidence that these are important factors. The plant is not common in arable crops (Evans 1984), with the exception of alfalfa (Patterson 1993) where mowing is usually done before the plants are mature and contamination of alfalfa seed can be prevented by sieving because of the different seed sizes (Tingey 1971). The likelihood of animal dispersal is speculative, however, in a laboratory experiment in the Czech Republic, 51% of seeds scattered over a wild boar pelt remained embedded in the fur following agitation (Moravcová et al. 2010). Mature seeds, which are the most toxic part of the plant, could possibly be eaten by livestock and passed in feces.

The absence of N. galegae in native soils has been implicated in the slow spread of G. officinalis in both Canada (Bromfield et al. 2019) and New Zealand (Proctor 1963). The co-dispersal of G. officinalis seed with the bacterial symbiont would seem most effectively accomplished through soil transport either by water or human actions. Apart from deliberate soil inoculation, the intentional spread of the plant and its symbiont can be achieved by transplanting from gardens where N. galegae is already present or by illegally importing soil containing the bacterium. Inadvertent co-dispersal may occur through the movement of soil or construction aggregate, or on contaminated vehicles from areas with previously established plant-bacteria populations.

(c) Seed banks, seed viability and germination

The continued annual emergence of seedlings following strenuous attempts to eradicate the species indicated the presence of seed banks in infested areas of Utah (Evans et al. 1997), where 15 000–75 000 seeds m⁻² have been observed (Oldham 2009). Seeds germinate rapidly in 1–3 d when scarified and moistened in the laboratory [Patterson 1992; S. Mechanda, (unpublished data)], but germination may take several weeks under natural conditions. Seeds are known to remain viable in the soil for at least 5–10 yr (Evans and Ashcroft 1982). Seed dormancy, as with many Fabaceae, is primarily physical, with seeds germinating rapidly after physical or chemical scarification. Oldham and Ransom (2009) found nearly 100% seed germination after sulfuric acid scarification, but only 8% of un-scarified seeds germinated.

In Utah, greenhouse germination trials used seeds collected from the soil surface as well as seeds collected 26 yr earlier and placed in dry storage (Oldham and Ransom 2009). Seeds were planted at different depths from near-surface to 14 cm. The seeds exhibited dormancy of 80–93%, viability of 91%–100%, and were capable of remaining viable in dry storage for at least 26 y. Dormancy was estimated by comparing germination of the old and new seed with and without sulphuric acid scarification. Of the untreated 26 yr old seed, 61% germinated, while maximum germination (99%) was achieved with a scarification time of 10 min. Only 23% germination occurred in untreated 6 mo seed, but reached 89% germination after 50 min of scarification treatment. These results indicated that dormancy was lower in old seed, although viability was similar in both age classes. Similar observations by Stebler and Schröter (1889) on the germination of “old” and “new” seeds planted for crops in Switzerland found that 12% of new seed and 21% of old seed (age and storage conditions not given) germinated; 88% of the new seed and 21% of the old seed remained hard (i.e., dormant) while the remainder of the old seed (58%) rotted. In Spain, a germination rate of 50% was observed after sandpaper scarified seed was sown (González-Andrés et al. 2004). Other factors have been shown to affect the germination of G. officinalis. In the Czech Republic (Moravcová et al. 2010), germination of freshly harvested seed was tested under three temperature regimes (25/10, 20/5 and 15/5 °C), and dry-stored plus cold-stratified (1–4 °C, 3–5 mo) seed at 25/10 °C (12 h light 12 h dark). Although details of the results were not given, germination under these various conditions ranged from 12%–56%. Testing the effect of temperature on germination in the laboratory, Patterson (1992) observed germination of scarified seeds was 47%–68% one day after planting, with lower germination rates occurring in the 26/14 and 34/14 °C regimes, but after three days, germination was 75%–81%, regardless of temperature conditions. Higher night temperature resulted in high germination rates. In Utah, seedling emergence was inversely related to burial depths; emergence was 93% at 0.5 cm soil depth, 87%–90% at 1–3 cm depth, 56% at 8 cm, and 21% at 10 cm, while no emergence occurred at depths below 12 cm (Oldham and Ransom 2009).

(d) Vegetative reproduction

Roots that survive winter temperatures will send up new shoots in the spring from buds on the caudex (base of the stem). In vitro experiments by Našinec and Nĕmcová (1990) were able to regenerate plants from both callus formation and re-growth from node-bearing stem segments.

(a) Herbivory

(b) Diseases

(c) Higher plant parasites

No information has been located.