Prospects for diversifying temperate pasture systems in south-eastern Australia with serradella (Ornithopus spp.)

The first significant research effort occurred in WA and focused on yellow serradella. It was largely driven by a need for hardseeded annual pasture legumes adapted to deep sandy, acidic soils where subterranean clover, lucerne (Medicago sativa L.) and annual medics (Medicago spp.) along with their specific rhizobia did not persist (Gillespie 1990; Michalk and Revell 1993). The hardseed characteristic of most yellow serradella accessions made them immediately amenable for use as a self-regenerating annual legume in southern Australia. The initial cultivars of yellow serradella such as ‘Pitman’ were found to be too late maturing for the lower-rainfall zones (250–500 mm AAR, Fig. 1) (Gladstones and Devitt 1971; Gladstones 1976), so germplasm collected from the Mediterranean region was evaluated and found to be better adapted to the shorter, milder winters in these lower-rainfall zones. This led to the release of multiple early to mid-season maturing cultivars in the 1980s (Table 3). Adoption of serradella in regions of southern WA and central-western NSW, in which these early–mid-season maturity types were suited, was reported by the early 1990s (Freebairn 1990; Revell 1992). Continued selection for early maturity in yellow serradella led to the release of cultivars such as ‘Charano’ (early) and eventually ‘Yelbini’ (very early) (Tables 2, 3), which enabled their use in regions with as little as 300 mm annual rainfall (Revell et al. 1998a).

A second driving factor for cultivar development in yellow serradellas was improved pod traits. Yellow serradella seed develops in a woody pod and, in some cultivars, a hook on the terminal beak and limited segmentation (i.e. pods that do not readily break into segments) complicated its bulk handling to the point that seed production became expensive and unprofitable (Harrison et al. 2024). Straighter pods were therefore targeted for cultivar releases from the 2000s onward (e.g. ‘Charano’, Table 2) to improve compatibility with bulk-handling equipment. Easily segmented pods (e.g. ‘Serramax’, Loi et al. 2021) that are more amenable to flow through machinery were also selected. However, because segmenting pods are harder to dehull (the process of extracting seed from the pods), cultivar development simultaneously targeted more rapid hardseed breakdown patterns to alleviate the need for dehulling (Revell et al. 1998a, 1998b).

Yellow serradellas typically produce a high level of initial hardseed that is very slow to soften (Revell et al. 1998b, 1999; Newell et al. 2022). This is generally an advantage for long-term persistence and regeneration after cropping phases. However, it can lead to poor regeneration in the year after sowing (Howieson et al. 2021; Hayes et al. 2023). A more rapid hardseed breakdown pattern was a key selection criterion in the development of ‘SerraMax’ (Table 2; Revell et al. 1998b; Loi et al. 2021; Howieson et al. 2021) and is also observed in other cultivars including ‘Pitman’ and ‘Yellotas’ (Newell et al. 2022; Martin et al. 2023). These pod- and seed-trait improvements have helped elevate the utility of yellow serradellas in farming systems in southern Australia.

Renewed interest in French serradella emerged in the mid-1990s, with the distinct advantages that pods could be easily harvested in a single pass by using a conventional cereal harvester and pods broke into segments with minimal processing. Therefore, individual pod segments (containing a single seed) could be sown without further processing (Nutt 1993; Nutt and Paterson 1997; Loi et al. 2005), thereby reducing the cost of seed and subsequent pasture establishment. The release of ‘Cadiz’ (Table 2) provided a mid-season maturity (Table 3) cultivar suitable for a comparable region to which many yellow serradella cultivars had been developed (Nichols et al. 2012). However, as for most French serradella genotypes that had been tested prior to 2002, ‘Cadiz’ was almost entirely softseeded (Bolland 1982; Snowball 1996). To address this issue, breeders developed the first hardseeded cultivars of French serradella, ‘Margurita’ and ‘Erica’, both of which had mid-season maturity (Table 3; Nutt 2008; Harrison et al. 2024). This novel hardseeded trait in French serradellas enabled broader use in ley farming systems and, by 2005, the area sown to French serradella in WA eclipsed yellow serradella by more than two-fold (Nichols et al. 2007). Most recently, an early-maturity hardseeded cultivar, ‘Fran₂o’, was released (Harrison et al. 2024). The latest estimate of land area sown to the older and more widely adopted ‘Margurita’ is over 1 million hectares (Harrison et al. 2024).

There are several adaptations of serradellas that make them a promising legume alternative to subterranean clover in south-eastern Australia, including traits associated with root development and function, stress tolerance, seed morphology and seeding habits. Other traits, such as the specificity of their rhizobial association and herbicide tolerance are challenges to adoption of serradellas in south-eastern Australian pasture systems.

A major benefit of yellow and French serradella is their lower requirement for soil P to achieve maximum yields than that of subterranean clover (Paynter 1992; Paynter and Bolland 1993; Sandral et al. 2019). This lower ‘critical P’ requirement for serradella, than for subterranean clover, is a potentially valuable trait for southern Australia where P deficiency is the largest and most widely encountered soil nutrient constraint after nitrogen supply (Angus et al. 2019). It is estimated that serradella-based permanent pasture systems could achieve near-maximum production at an Olsen P of 10 mg/kg soil, whereas subterranean clover-based pastures would require an Olsen P of 15 mg/kg for equivalent production (Sandral et al. 2019). This indicates potential to reduce both the economic and environmental cost of P fertiliser (Simpson et al. 2014; Simpson and Haling 2021). The superior capacity of serradella to acquire soil P, in comparison to clover species, is due to root morphological traits such as relatively long, thin roots and long root hairs, which provide a large soil interface for P acquisition (Haling et al. 2016; Sandral et al. 2018). The colonisation of roots of serradellas and subterranean clovers by arbuscular mycorrhizal fungi can assist with the acquisition of soil P, but does not change the relative difference in P acquisition efficiency of the two species (Sandral et al. 2019; McLachlan et al. 2020).

French and yellow serradellas are comparatively deeper-rooted species than is subterranean clover, with reports of rooting depths on sandy, acidic soils (unrestricted by moisture deficit, chemical or physical barriers) of 1.7–1.95 m for serradellas compared with 0.6–1.2 m for subterranean clover (Hamblin and Hamblin 1985; Carr et al. 1999; Loi et al. 2005). Expression of this deep rooting trait has been linked to enhanced ability to scavenge deeper soil water reserves later into the season (Hackney et al. 2019, 2021; Monjardino et al. 2022), being particularly valuable on soils with a low water holding capacity, and has also been linked to improved K acquisition (Ozanne et al. 1965; Asher and Ozanne 1966; Hamblin and Hamblin 1985; Sandral et al. 2006). In south-eastern Australia, expression of this trait may be constrained on the widely-occurring duplex soils where high bulk density in the heavy clay subsoil may constrain root penetration to depth (Simpson 2020).

Periodic waterlogging can occur in many parts of the landscape used for pasture systems in south-eastern Australia. In these regions, the risk of waterlogging is likely to be higher than on sandy soils, owing to the combination of higher rainfall and finer soil textures. Laboratory screening of a range of yellow, French and slender serradella genotypes for waterlogging tolerance showed that most of the genotypes could recover from transient waterlogging (<2 weeks) but were unsuited to longer periods of exposure to anerobic conditions (Kidd et al. 2020). Field evaluations in NSW (Sandral et al. 2019; Hackney et al. 2021), South Australia (Ballard et al. 2020) and Victoria (Latta and Moodie 2022) have supported the use of serradella on heavier-textured soils than have been used historically. However, the risk of waterlogging is likely to increase with increasingly finer-textured soils and this may limit adoption. Slender serradella (O. pinnatus) has been observed anecdotally to be more tolerant of waterlogging than is yellow serradella (Freebairn 1990) and has been recommended for use in mixtures in higher-rainfall areas. However, in more detailed investigations, slender serradella was found to have no more direct tolerance of waterlogging than did other serradella species (Kidd et al. 2020). It is assumed, therefore, that its persistence in paddocks that are prone to waterlogging is due more to avoidance, than physiological tolerance of waterlogging, whereby slow growth during winter is followed by a flush of growth in spring when waterlogging has dissipated.

Nitrogen fixation is a critically important attribute of annual pasture legumes. Serradella forms an effective symbiosis with Bradyrhizobium species, which belong to a different genus of rhizobia than those required by other pasture legumes grown in Australia (Rigg et al. 2021; Farquharson et al. 2022). Successful introduction of compatible rhizobia will be essential to the establishment of serradellas in new environments. There have been multiple observations of nodulation failure when sowing serradella into new environments in NSW and TAS (R. Haling and R. Smith, pers. comm.) and, as such, consideration needs to be given to the range of factors that may influence nodulation and symbiotic and saprophytic competence, including paddock history, inoculant type and soil conditions.

The current recommended commercial strain for serradella is inoculant Group S (WSM471), but serradella also nodulates well with the Group G strain for lupins (WU425) (Farquharson et al. 2022). If well-nodulated serradella or lupins have not been previously grown, then Group G or S rhizobia must be introduced when sowing serradella. If paddock history is unknown, it is advisable to refer to soil tests for lupin/serradella (e.g. rNod, a DNA-based soil testing service) that check for rhizobial strains (Farquharson et al. 2022) or inoculate to remove the risk of nodulation failure.

Peat-based inoculants are the oldest and most common form of inoculant used in Australia, typically applied to the surface of the seed as a wet slurry and followed by the addition of lime to help dry the seed and increase rhizobial survival (Farquharson et al. 2022). Higher water volumes are required when inoculating podded seed than with seed alone. Early research with agar cultures of rhizobia found deleterious effects of lime pelleting on serradella nodulation, but no effect when peat-cultured rhizobia were used (Shipton and Parker 1967). Hartley et al. (2004) found that lime pelleting serradella increased rhizobia numbers 24 h after inoculation by 90%, nodulation by 57% and shoot dry matter by 28% compared with unpelleted seeds and is the basis for the general recommendation to lime pellet serradella in all states except WA (Farquharson et al. 2022). Application technologies for inoculants have developed over the past 10–15 years to include dry clay (bentonite) or peat-based granules, freeze-dried powder and liquid formulations (Farquharson et al. 2022). Dry granules and freeze-dried formulations are available for serradella and further evaluation of these forms of rhizobia application for sowing in new environments is warranted.

Acid soil tolerance affects the performance of host plants and the host-specific rhizobia, which enables the capacity of plants to fix nitrogen (Sanford et al. 1994; Howieson and Ballard 2004; Guo et al. 2012). Rhizobia are often more acid sensitive than are the host species (Condon et al. 2021a). The lower end of the optimal pH_Ca range for serradella rhizobia (Bradyrhizobium spp.) is 5.0 compared with 5.5 for subterranean clover (Rhizobium sp.) (Condon et al. 2021a). However, there are few reports (Sanford et al. 1994; Peoples et al. 2012) of the amount of nitrogen fixed by serradella plants to quantify the impact of suboptimal pH (pH_Ca <5) on N fixation. In WA, Sanford et al. (1994) found that 76% of fixed N was derived from atmospheric N₂ (Ndfa), whereas in NSW, estimates were lower at 26–31% Ndfa (Peoples et al. 2012). Further estimates of N fixation across a broader range of environments and among serradella species are recommended to assess the effectiveness of the rhizobial symbiosis in serradella and whether there is value in the development of more acid-tolerant serradella rhizobia.

Dry sowing of serradella seeds in pod is a novel and revolutionary pasture establishment practice that can deliver multiple benefits (Nutt et al. 2021). Benefits include timely sowing operations, enabling producers to shift pasture sowing to summer, before the main winter-crop sowing window. This early sowing also allows for the early germination of serradella with the season’s first rainfalls leading to increased dry matter production and higher seed yields (Nutt et al. 2021). A major challenge with establishing serradella through the sowing of pod (cf. bare seed) is achieving successful nodulation. The rhizobia need to survive desiccation and high temperatures prior to seed softening and autumn rainfall (Bowman et al. 1995). If establishing serradella as part of a pasture and crop rotation, growing a lupin crop within a cropping phase before sowing serradella is one way of establishing a background population of compatible rhizobia. Such a strategy is less feasible in permanent pasture systems. Applying a peat slurry with methyl cellulose adhesive to hulled pod segments of yellow serradella and sowing into dry soil, with germination rains occurring 77 days later resulted in only 22% of plants being nodulated in the first year, which increased to 72% in the following year (Bowman et al. 1995). Bowman et al. (1995) concluded that rhizobia applied to dry sown seed can survive in sufficient numbers to initiate seedling nodulation and to become established in the soil. However, high nodulation in the first year is desired to maximise pasture production, first-year seed set and subsequent pasture production. In a ley pasture system, it is also important to build up the rhizobia population to a level that will ensure survival during the cropping phase. Peat slurry application at double rates is recommended for pulses when dry sowing (Farquharson et al. 2022) and, therefore, double to quadruple inoculation rates may improve nodulation outcomes when sowing serradella in late summer to early autumn.

Research has shown that granular formulations of rhizobia at the standard rate resulted in nodulation in pulses similar to, or greater than, double rates of peat on seed when the dry period exceeded 7 days and the granular product had high rhizobia numbers (Farquharson et al. 2022). Inoculation with dry granules has been used in WA when summer-sowing dormant serradella seeds (hard seeds) into hot, dry soils (Nutt et al. 2020b). Dorji et al. (2023) sowed hardseeded pod segments of French Serradella and bentonite clay granules at four sites in SA in March and April 2023, with opening rains occurring 1–1.5 months later. However, they reported poor nodulation status at all sites without a lupin history, with a mean across sites of 41% plants with nodules. It is possible that the difference between the WA and SA results is the history of lupins being widely grown in WA but not in SA. Supporting this, Dorji et al. (2023) reported high levels of nodulation on their sites with a lupin history and that these sites could be identified using the rNod test. Delaying the sowing of hard seed in a summer sowing system will assist rhizobia survival, but too much delay will reduce the time needed for seed softening, leading to lower plant numbers at the break of season. For pulses, peat granules have provided similar nodulation as peat slurry for lupins (Denton et al. 2009, 2017); so, using this technology when dry sowing serradella deserves further evaluation.

Spraying soil with rhizobia ahead of planting was proposed as an alternative inoculation strategy in lupins and faba bean (Evans et al. 2001). Adapting this strategy for dry sowing systems in serradella (spraying the inoculum after the break of season) may be possible, although rhizobia numbers in proximity to the seed are likely to be lower and nodulation could be compromised.

In permanent pasture systems, the timing of annual regeneration varies between years owing to seasonal conditions. Seed ecology and dormancy traits play a key role in regulating germination, helping optimise seedling regeneration. The main form of seed dormancy in serradella is hardseededness, the primary, physical dormancy that results from the development of an impermeable seed coat at maturity (Taylor 2005). The development of cultivars that vary in both the initial level of hard seed and the rate of hard-seed breakdown (dormancy release) has revolutionised the use of serradella in Australia. Intraspecific variation exists in the expression of hardseededness in yellow (Bolland 1985) and French (Harrison et al. 2024) serradella, with cultivars of yellow serradella typically producing higher levels of hard seed at maturity than those of French serradella. The first French serradella cultivars (released prior to 2005, Table 2) were almost entirely softseeded and this was sometimes problematic because it led to high seedling mortality following ‘false breaks’ to the growing season (e.g. Dear et al. 2002).

Differences in initial levels of hard seed and hard-seed breakdown patterns may have significant implications on the success of establishing serradella pastures in new environments and the prospect for longer-term persistence. Differences in the patterns of hard-seed breakdown have been observed among cultivars (Revell et al. 1998b, Newell et al. 2022; Harrison et al. 2024) and can be influenced by environmental conditions experienced during the period of seed softening (Newell et al. 2022) and by seed burial (Revell et al. 1998b; Taylor 2005). For example, Newell et al. (2022) suggested that in permanent pastures, serradellas that soften slowly over several years are likely to have limited regeneration in the year after pasture establishment, resulting in a high potential for weed invasion. However, where weeds can be managed, this may be an advantageous trait for long-term persistence once a seedbank is established.

A premature end to the growing season with low soil moisture conditions during seed maturation has been shown to substantially reduce seed production and the viability of hard seeds in yellow serradella (Revell et al. 1999) and can result in lower initial hardseededness in French serradella seeds (Nutt 2008). Yellow serradella seeds become hard as they dehydrate and full impermeability is reached at a seed moisture content of about 5–7% (Revell et al. 1999). Rainfall experienced during seed maturation can potentially reduce the level of initial hardseededness (Taylor 2005) if seeds that are normally in a drying phase rehydrate and rupture the seed coat. This may be an issue for temperate environments with long growing seasons in eastern Australia because full hardseededness in both French and yellow serradellas may not be reached, affecting the persistence of seedbanks and levels of seedling regeneration.

Hard-seed breakdown in annual legume species is generally explained by a two-stage conceptual model where the first or preconditioning stage takes place with time at a rate that increases with increasing temperature, followed by the softening stage, which is determined by response to diurnal fluctuations within a particular range (Taylor 2005). Temperature fluctuations for the second stage appear to be quite specific to each legume species and can vary among cultivars (Taylor 2005). Seven cycles of 48°C/15°C successfully softened preconditioned seeds of ‘Serramax’ (formerly known as GEH72-1A) but were less successful for ‘Santorini’ and ‘Charano’ (Taylor and Revell 2002). The softening response to shallow burial of seed in yellow serradella (Revell et al. 1998b, 1999) has been explained by an inhibitory effect of light on the second stage of softening that is overcome by soil cover. The second stage of seed softening of yellow serradella ‘SerraMax’ was inhibited in some seeds by light levels as low as 0.3% of daylight, and in all seeds at a light level between 5% and 25% (Taylor and Revell 1999). Both specific temperature fluctuations and conditions of darkness need to be satisfied before seed softening occurs.

Hard seeds of French serradella (‘Margurita’ and ‘Erica’) have also been shown to soften more rapidly with shallow burial early in summer (Howieson et al. 2021). Nutt (2008) suggested that the two-stage model of seed softening does not fully explain hard-seed breakdown patterns in all French serradella cultivars and offered other factors such as humidity and the exceedance of critical thresholds for seed moisture content as being important.

The influence of seed burial on patterns of hard-seed breakdown in serradella has implications for the use of serradella in permanent pastures. We suggest that further studies of hard-seed breakdown in serradella are required in permanent pasture environments where there is limited opportunity to bury pods mechanically. It should be noted that seed storage is an important consideration in studies of physical dormancy, with reports of faster hard-seed breakdown in seed that has been stored for an extended period than in fresh seed (Harrison et al. 2020). Storing pods for several years prior to sowing may be one way to increase initial emergence rates without the costs of dehulling the seed, but such a practice requires further testing.

Most field-softened seeds of Mediterranean annual pasture legumes fully imbibe within 1 or 2 days of being placed in contact with water (Taylor 2005). However, delayed imbibition is quite marked in some cultivars of yellow serradella (Taylor and Revell 2002; Taylor 2004, 2005). For example, some field-softened seeds of yellow serradella ‘Santorini’ can take up to 6–7 weeks of moist conditions to complete germination (Loi et al. 2021), whereas other cultivars will germinate over 1–5 weeks of moist conditions (Revell et al. 1998b; Taylor and Revell 2002; Loi et al. 2021; Newell et al. 2022). These genotypic differences in the rate of imbibition appear to be a function of differences in seed structures regulating moisture uptake in yellow (Taylor 2004) and French serradella (Nutt 2008) and can be influenced by temperature and state of seed hydration (Taylor 2004). Hydrothermal time models may offer a framework for modelling serradella germination, but need further development (Bradford 2002). A spread of germination has ecological merit, providing a balance between gaining a competitive advantage from an early germination of some seeds (Ross and Harper 1972) while also providing some insurance against seedling mortality in ‘false breaks’ of season (Taylor 2005). In cultivars such as ‘Yellotas’ that display a narrow spread of germination (Martin et al. 2023), long-term persistence may be compromised if there is insufficient residual seed remaining in the seedbank to ensure regeneration in the following year (Newell et al. 2022; Hayes et al. 2023).

Physiological dormancy is a protective mechanism to prevent germination of newly ripened seed during late spring and early summer and is potentially important in temperate environments receiving rainfall during seed formation. Barrett-Lennard and Gladstones (1964) found no evidence of physiological dormancy in seeds of yellow serradella ‘Pitman’, but they did suggest that a water-soluble substance was present in the pod wall and inhibited germination during maturation of the seed. Peck et al. (2023) found that at least one French serradella cultivar (‘Margurita’) has physiological dormancy, but factors influencing its expression are unclear. In subterranean clover, physiological dormancy of newly ripened seed dissipates over time, with higher temperatures increasing the rate of breakdown (Taylor 1970). Artificial techniques to overcome dormancy and enhance germination in subterranean clover include exposure to cool temperatures (1–10°C), atmospheres of CO₂ and ethylene, and flushing seeds with water (Ballard 1961; Esashi and Leopold 1969; Taylor 1970). Genetic variation in the existence and requirements needed to breakdown physiological dormancy in serradella is a potential area of future research.

Adoption of serradellas in new regions will rely on successful sowing strategies suited to the system and environment. To achieve even establishment in the first year of a serradella pasture, seed has been traditionally sown in autumn as bare seed, which has been dehulled and scarified (i.e. fracturing the hard seed coat to allow seed to be permeable to moisture). Post-harvest seed-processing methods have been developed for serradella, which involve mechanical threshing that dehull and scarify the seed (Weeldenburg and Smith 1969).

More recently, techniques for the establishment of serradella have exploited the ease of on-farm seed production and its patterns of hard-seed breakdown. Unprocessed pod segments of suitable cultivars can be sown in late summer–early autumn, soften over autumn and germinate on opening seasonal rainfall (Howieson et al. 2021; Nutt et al. 2021).

Timing is critical in Mediterranean climates. If sowing is delayed to mid-autumn (mid-April), the amount of softened (germinable) seed can be reduced by up to 56% relative to a mid-March sowing date, and by as much as 86% following a mid-May sowing, requiring much higher sowing rates that are unlikely to be economic (Howieson et al. 2021). Late summer–early autumn sowing allows for increased early feed and increased seed set in years with a dry spring (Nutt et al. 2021). The yellow serradella ‘SerraMax’ is particularly suited to sowing in late summer to early autumn (Howieson et al. 2021; Loi et al. 2021). ‘Twin sowing’, where serradella pods are undersown with a cereal or oilseed crop, is a variation of the summer sowing technique. Minimal germination of hardseeded cultivars occurs under the crop in the first year, but during the subsequent summer–autumn, breakdown of hard seeds occurs under the crop stubble to regenerate the pasture on the next break of season rains (Loi et al. 2012). For mixed farming enterprises, summer sowing and twin sowing have allowed farmers to sow pastures at a time of year that does not interfere with crop sowing operations (Loi and Yates 2020). These procedures need to be accompanied by appropriate weed management but offer considerable farm management advantages.

Aerial sowing is a method for introducing improved pasture species into non-arable country and has been used previously to establish pastures on non-arable soils in the rangelands of central NSW (Michalk and Campbell 2002). Michalk and Campbell (2002) reported that a pod:scarified seed mixture of 75:25 was the most suitable ratio for establishing yellow serradella when broadcasting into native grasslands on hardsetting, non-arable, acidic hill country. However, they also reported nodulation failure. Experimentation with the use of clay-based granular inoculation techniques or other rhizobia application methods may warrant investigation for aerial sowing. Hard seeds stored for several years in a laboratory environment (20°C ± 5°C) soften much more readily than do fresh seeds (Harrison et al. 2020). It is possible that storing serradella pods of hardseeded cultivars for 2–3 years may assist their establishment when they are broadcast. Movement of serradella with livestock may be another way of introducing the species into non-arable country. Livestock may move undigested legume seed through landscapes, but Edward et al. (1998) showed that the fraction of undigested seed in serradella was low (10%) relative to other small-seeded legume species. Nevertheless, where high seed loads exist, this still may be a feasible option over time, but the introduction of compatible rhizobia may still be a challenge.

The expansion of serradella adoption into longer season environments will need to consider management of pests and diseases. Serradellas have been reported to have a low incidence of disease infection relative to other annual pasture legumes (Taylor et al. 1979), but have been observed to have foliar diseases such as brown leaf spot (Pleiochaeta setosa), grey mould (Botrytis cinerea), Sclerotinia (Sclerotinia sclerotiorum) and unspecified fungal diseases of the genera Colletotrichum (anthracnose) and Cercospora (NSW DPI 2025), as well as the root disease Rhizoctonia solani (Mackie et al. 1999; You and Lancaster 2008). Diseases caused by viral infections from alfalfa mosaic (AMV) and pea seed-borne mosaic (PSbMV) viruses (Latham and Jones 2001) have also been recorded. The screening of serradella germplasm for disease resistance and the development of disease management packages will be of benefit when considering adaptation to higher-rainfall regions.

Native budworm (Helicoverpa punctigera Wllgr.) is a major pest that causes damage to podded crops and pastures in southern Australia. When the timing of the life cycle of the caterpillar phase of this pest coincides with serradella pod development, there is a risk of seed damage. Early completion of the sensitive seed/pod filling stage before budworm populations build up can minimise seed losses and this was one of the reasons that early maturing cultivars were a primary focus of breeding programs in WA (Revell 1992; Loi et al. 2021). Budworm damage was one of the key problems associated with poor persistence of French serradella ‘Cadiz’ relative to subterranean clover cultivars in southern NSW (Dear et al. 2002). Tolerance to blue green aphid (Acyrthosiphon kondoi Shinji) is high in yellow serradellas but tolerance to red-legged earth mites (RLEM, Halotydeus destructor Tucker) is low–moderate (Revell et al. 1998a). Chemical control options are available to manage RLEM (as well as budworm and aphids) but ongoing work to determine the optimal timing of use around shifting life cycle patterns of the insects may be required, considering changes in temperature and rainfall patterns (Maino et al. 2024).

There are few herbicide options for use on serradella, with only five herbicide active ingredients (flumetsulam, imazamox, imazethapyr, clethodim and 2, 4-DB) being registered specifically for use on serradella in Australia (Australian Pesticides and Veterinary Medicines Authority 2025). In the past, the use of grass-selective herbicides in serradella stands has been shown to help control problem weeds in cropping phases (Revell and Thomas 2004). Research into the tolerance of serradella to broadleaf herbicides by Blake (2002) and Rogers and Lauritsen (2004) showed considerable variation among and within serradella species. Lockley and Wu (2008) demonstrated that French serradella ‘Erica’ recovered well from early applications (3–4-leaf stage) of flumetsulam (2, 4-DB, clethodim, haloxyfop-R, imazamox and imazethapyr), but was severely damaged by MCPA (terbutryn, diflufenican, or bromoxynil). In a separate evaluation in southern NSW, Sandral and Dear (2005) found French serradella ‘Margurita’ to be among the most sensitive of the annual legume species tested to a wide range of herbicides used to control broadleaf weeds. Further research into novel chemistries is needed, and when growing serradella in mixed-species pastures, herbicide selection should be guided by label recommendations specific to each species.

Weed wipers (using glyphosate) have been used effectively in serradella for selective control of some broadleaf weeds such as wild radish (Raphanus raphanistrum L.). The ability to spray-top early-maturing cultivars, such as yellow serradella ‘Yelbini’, with knockdown herbicides when weeds are flowering also provides an opportunity to control herbicide-resistant grasses in cropping rotations. For yellow serradella cultivars such as ‘Santorini’ and ‘Yelbini’, a slow rate of germination (as long as 40–60 days) (Howieson et al. 2021) may also allow knockdown herbicides to be used 10–14 days after opening rains to achieve effective early weed control (Taylor 2005), but the effect on early pasture production owing to the loss of early germinating plants has not been reported. Currently, a lack of registered herbicides is a potential limitation for ensuring effective establishment because of competition from weeds. An increased number of registered herbicides paired with research that investigates the efficacy of a suite of herbicides in improving establishment (and subsequent pasture management) would be valuable.

Grazing management can significantly affect the productivity and persistence of a pasture because of effects on seed production, flowering and interspecific competition within a mixed pasture sward. Because serradellas are aerial-seeded, it is commonly recommended that serradella be rested or at least have reduced grazing pressure during flowering every few years to promote seed production and replenish soil seed reserves (Gladstones and Devitt 1971; Freebairn et al. 1997).

Grazing has been observed to delay the reproductive development of yellow serradella by up to ~2 weeks compared with undefoliated swards (Gladstones and Devitt 1971). There are limited data on the effect of the timing of grazing and grazing intensity on the herbage and seed production of serradella in mixed pasture swards in permanent pasture systems. A study comparing the effect of grazing intensity (light or heavy) and timing (winter, winter–spring or late spring) on seed set of yellow serradella ‘Avila’ and subterranean clover cultivars ‘Karridale’, ‘Clare’ and ‘Trikkala’ found that seed yield was reduced by 30–55% in ‘Avila’ and 30–48% for ‘Clare’ in late spring grazing treatments relative to an ungrazed control (Conlan et al. 1994). In contrast, seed yields of subterranean clover cultivars ‘Karridale’ and ‘Trikkala’ were not different from that of the ungrazed control, which was explained by the capacity of these cultivars to bury their burrs (Conlan et al. 1994). Studies in south-eastern Australia comparing various annual pasture legumes, including French serradella (‘Cadiz’), over a 3-year period concluded that frequent grazing hampered the aerial seeding legume species and provided a competitive advantage to the background subterranean clover populations (Dear et al. 2002).

The nutritive value of yellow and French serradella plant material is comparable to that of subterranean clover (Rossiter et al. 1985; Wilmot et al. 2023), with protein concentrations of green forage in excess of 25% and in vitro dry matter digestibility values of 70–75%. These attributes of feed quality are key drivers of suitability of serradella as an annual pasture legume in temperate Australian environments (Nichols et al. 2012). Sustained late season forage quality in years with extended growing seasons is considered a valuable attribute of serradella pastures (Monjardino et al. 2022). An indeterminate flowering habit is one reason that has been suggested for the late dry-matter accumulation and sustained nutritive value of serradella relative to other annual pasture legumes (Lloveras and Iglesias 1998; Hackney et al. 2021; Monjardino et al. 2022). However, there is inevitably a decline in pasture feed quality as plants mature and leaflets are shed, as observed by declining dry-matter digestibility over time (Iglesias and Lloveras 2000; Norman and Masters 2023). Norman and Masters (2023) reported a more rapid decline in herbage quality of ‘Santorini’ yellow serradella than of ‘Antas’ subclover, but this may simply reflect the earlier maturity of the serradella in that WA environment. Consequently, timing the reproductive phase to ensure a long grazing window with highly digestible forage will remain an important strategy for maximising the value of serradella-based pastures (Thomas et al. 2021).

Yellow and French serradellas are generally regarded as being safe for livestock consumption and may have some relative advantages compared with subterranean clover (Freebairn 1990; Freebairn et al. 1997; Wilmot et al. 2023). Isoflavones are phytoestrogenic compounds that are found in some older subterranean clover cultivars (Wilmot et al. 2023) and have been problematic for sheep reproduction. Recent research has measured low isoflavone concentrations in French serradella ‘Erica’ relative to an oestrogenic clover reference (Wilmot et al. 2023). In a comprehensive screening of a range of yellow and French serradella accessions (n = 129 and 175 respectively) the isoflavone, formononetin, was measured to be 7.4 ± 0.4 and 18.7 ± 0.4 mg/kg dry matter (mean ± s.d.) (Visnevschi-Necrasov et al. 2012), which is below thresholds that have been recorded to cause ewe infertility in subterranean clover pastures (Nichols et al. 2013). These data have provided support for the view that serradellas are safe for the grazing of sheep.

The herbage of yellow serradella ‘Avila’, French serradella ‘Margurita’ and subterranean clover cultivars ‘Coolamon’ and ‘Seaton Park’ was recently compared by Li et al. (2025). The serradella cultivars were shown to have higher concentrations of some plant secondary compounds (PSCs), including tannins and saponins, than did subterranean clover, but similar concentrations of flavonoids and phenols. Plant secondary compounds were once considered anti-nutritional factors but are now widely recognised for their roles in both plant defence and as bioactives for livestock productivity. The higher level of tannins in serradella relative to subterranean clover is believed to reduce the risk of bloat in sheep and cattle (Mueller-Harvey 2006). It remains unclear whether higher concentrations of some PSCs will result in lower emissions of enteric methane in grazed serradella pastures, but early research seems promising (Badgery et al. 2023; Li et al. 2025).

A key principle for cultivar selection is choosing cultivars with a maturity type that is suited to the the length of growing season (e.g. subterranean clover, Dear et al. 2007; Nichols et al. 2013). This ensures that flowering occurs at a time of year when the risk of abiotic stresses (frost, heat, drought) causing damage to flowers and developing seeds is minimised. Flowering as late as possible maximises the length of the vegetative period of growth and, thus, high-quality herbage production for forage (Norman and Masters 2023). However, flowering still needs to occur early enough for seed production to be completed before the growing season ends. In a permanent pasture, a suitable match of flowering date (maturity type) balances herbage production with the need for reliable seed production for pasture regeneration and persistence (Donald 1970).

Most of the development of serradella in Australia has focussed on early-maturing cultivars (Table 3). This has led to rapid adoption of serradella in low-rainfall areas. However, uptake in permanent pastures in colder and higher-rainfall (500–900 mm AAR, Fig. 1) climatic zones has been limited. In part, this can be attributed to a lack of cultivars with suitably adapted flowering traits. For instance, there are few mid- and later-flowering cultivars that are adapted to the colder climate and generally higher-rainfall environments. Suitable maturity is integral for serradellas being a viable alternative to subterranean clover. It ensures competitiveness with well-adapted subterranean clover cultivars and the other botanical components of the mixed-species pastures that are used in these regions (Hayes et al. 2023).

The indeterminate flowering habit of pasture legumes such as serradella can delay senescence and maintain forage quality for longer in the growing season. However flowering is not infinite and many cultivars display a peak period of flowering (Dodd and Orr 1995; Goward et al. 2024a). Ensuring that peak flowering is timed to occur when the risks of abiotic stresses that adversely affect seed production are low assists the reliability of seed production (Aitken 1974). This is crucial for the maintenance of a soil seedbank for subsequent regeneration and for longer-term persistence (Dear et al. 2002).

The late maturing yellow serradella ‘Avila’ has shown promising persistence in permanent pasture systems in the higher-rainfall zones of south-eastern Australia (Drew 1989; Simpson 2020; Hayes et al. 2023). ‘Avila’ is one of three late maturing yellow serradella cultivars released along with ‘Yellotas’ and ‘Pitman’ (Table 3). ‘Avila’ is reported to have greater seedling vigour and production than has ‘Pitman’ (Drew 1989), but is also considered to have a more desirable hard-seed breakdown pattern than do both ‘Pitman’ and ‘Yellotas’ for persistence in permanent pasture systems (Newell et al. 2022). Both ‘Pitman’ and ‘Yellotas’ exhibit a rapid seed softening pattern that can result in the vast majority of hard seeds softening in the initial year after they are set, leaving them at risk of depleting their bank of hard seeds. In contrast, the hard seed of ‘Avila’ softens over several years (Newell et al. 2022). However, high levels of residual hard seeds present an issue when sowing a pasture for the first time because insufficient seeds may soften to germinate in the autumn of the second year. In a permanent pasture, this can hamper establishment where there is a high weed seedbank. For the French serradellas, existing hardseeded cultivars exhibit a gradual seed softening pattern, which is likely to assist in longer-term persistence, but use of these cultivars is limited by a lack of later maturing types with hard seeds (Tables 2, 3). Cultivars of this nature are currently being developed (Haling et al. 2022).

The number of days to first flower (Table 3) is a common metric used to assign maturity type (Nichols et al. 2013) and inform cultivar selection for a given region. The functionality of this metric is reliant on the ranking of cultivars, according to days to first flower, being consistent across environments, as is largely the case for subterranean clover cultivars (Lodge and Harden 2007; Boschma et al. 2019). However, recent studies across four diverse environments (Perth, WA; Tamworth, NSW; Cowra, NSW; and Canberra, Australian Capital Territory (ACT)) suggested that the ranking of certain serradella cultivars can vary among locations (Boschma et al. 2019; Simpson 2020; Simpson et al. 2025). Thus, the current maturity type classification may be insufficient for predicting flowering time of some early–mid-flowering serradella cultivars grown in cool temperate climates.

Strong associations between the requirement for vernalisation and/or photoperiod (‘long-days’) and late-maturity (Goward et al. 2023), and the requirement for vernalisation and flowering date stability (Goward et al. 2024b) have been identified as key factors in the future selection and breeding of serradella cultivars adapted to cool temperate environments (Goward et al. 2023, 2024b). Similar to the development of lupins suitable to low-rainfall zones in WA, it is likely that flowering responsiveness to vernalisation may have been largely bred out of serradella cultivars adapted to low-rainfall environments (Francis and Gladstones 1974; Goward et al. 2023). In contrast, the ‘quantitative long day’ and ‘facultative’ vernalisation responses of late maturing yellow serradella ‘Avila’ and late maturing French serradella ‘Serratas’ have been observed to have similarities with subterranean clover cultivars adapted to longer-season environments (Aitken 1974; Goward et al. 2023).

The rate of flower production (days between flowering on successive nodes) is a key seed yield determinant of annual pasture legumes (Francis and Gladstones 1974), and substantial variation has been observed among serradella cultivars (Goward et al. 2024a). The rate at which seeds mature, defined as the time from first flower to production of a viable (germinable) seed, varies among serradella cultivars and is worthy of consideration, because it influences the estimation of optimal flowering and safe grazing periods (Francis and Gladstones 1974; Thomas et al. 2021). Evaluation of 107 serradella accessions highlighted variation among genotypes in the rate at which seeds mature, with up to 40-day difference between the fastest- and slowest-maturing genotypes when grown under field conditions (Fu et al. 1994). Working with two cultivars of yellow serradella, Revell et al. (1999) found that seed viability and the formation of the impermeable hard seed coat in yellow serradella ‘Avila’ occurred approximately 21–29 days from flowering. Maximum seed weight was achieved about 45 days after flowering in an irrigated environment, but was nearly 10 days earlier (and seeds were smaller) under conditions of moisture stress.

Self-regenerating pastures in south-eastern Australia are subject to a wide window of potential germination dates among years for a given location (Pook et al. 2009; Unkovich 2010; Flohr et al. 2021). The capacity of a given genotype to flower at approximately the same target date irrespective of their germination date (i.e. have stable flowering dates) ensures that plants will utilise the full length of the growing season that the germination date permits. Several studies (Bolland 1982, 1985; Fu et al. 1994; Snowball 1996) have compared the flowering dates of serradella genotypes for one environment and one sowing date, but few (Gladstones and Devitt 1971; Goward et al. 2024b) have closely examined the causes for differences in flowering dates arising from different germination times in modern serradella cultivars. Recent field evaluations (Boschma et al. 2019; Simpson 2020; Haling et al. 2022; Simpson et al. 2025) measured flowering date instability by comparing the start to flowering of cultivars sown in early autumn (March) with those sown later in autumn (May). In these studies, commercially available serradella cultivars were grown alongside subterranean clover cultivars spanning a maturity range of ‘very early’ through to ‘late’. Serradella cultivars were observed to express more flowering date instability than were subterranean clover cultivars of comparable maturity type. That is, for a given location, subterranean clovers most often flowered at a similar date regardless of sowing date. In contrast, serradella cultivars (e.g. ‘Margurita’) often flowered earlier when sown earlier than when sown late (Boschma et al. 2019; Simpson 2020; Haling et al. 2022). Quantification of the vernalisation requirements required for flowering date stability of cultivars of these species for different environments is required, as has been demonstrated for one location (Canberra, ACT) that is representative of a cold NSW Tablelands environment (Goward et al. 2024b). Goward et al. (2024b) showed that cultivars that require less vernalisation to initiate flowering tended to be associated with greater flowering date instability. A modelled sensitivity analysis of vernalisation and photoperiod drivers for target environments could assist in understanding the suitability of existing cultivars for new regions, as well as identifying the combinations of responses to photoperiod and vernalising temperatures that can deliver cultivars of differing maturity types, in combination with stable flowering dates.

Most of the existing cultivars of French serradella commercialised in Australia have either been direct releases of accessions or selections from within the early–mid-maturity cultivars (suggesting a narrow genetic base). Expanding the genetic diversity in serradella can be generated from further exploitation of historical plant collections or through targeted breeding programs. Given the altitudes and latitudes from which French serradella accessions have been collected (Table 1), it is expected that there is scope for selecting more later-maturing cultivars but the development of efficient breeding methods will become increasingly important as more traits (e.g. hardseededness) are desired to be stacked into new cultivars. Nutt et al. (2020a) reported that French serradella is a species with autogamous self-pollination, but with 25% cross-pollination in the field. The percentage of outcrossing in yellow serradella is unknown, but is expected to be lower becauses of the smaller flower size.

Outcrossing could be highly advantageous in breeding programs when combined with traditional methods employed by plant breeders working with inbreeding species. For example, one strategy to stack different traits into a single plant could be to create a diverse population by growing several accessions together, encouraging outcrossing with natural pollinators, followed by selection for desirable traits and selfing of individual plants (to fix the genetic trait) by using speed-breeding methods. Further introgression of traits between yellow and French serradella (such as hardseededness) could be possible through the development of hybrids (as in the example ‘Grasslands Spectra’ (Williams et al. 1987), in a similar way that traits have been transferred between strand (Medicago littoralis Rhode ex Loisel.) and barrel medics (Peck and Damin 2022). Analysis of the breeding system of French serradella has suggested that early flowering may be a dominant trait and that natural selection pressure may result in a genetic shift towards earlier-flowering types (Nutt et al. 2020a; Harrison et al. 2024), although a later-maturing trait could presumably be fixed through recurrent selection.

New speed breeding methodologies allow breeding to be conducted more efficiently than with conventional systems (Pazos-Navarro et al. 2017; Ghosh et al. 2018; Watson et al. 2018; Peck and Damin 2022). Serradella with its mainly autogamous self-pollination (Nutt et al. 2020a), an efficient hybridising method (Veerappan et al. 2014), and low level of physiological seed dormancy (Peck et al. 2023) is expected to be well suited to speed-breeding methods. Currently, only two cultivars have been released in Australia as the result of either inter- or intraspecific crossing programs (‘Grasslands Spectra’ and ‘Fran₂o’, Table 2, Harrison et al. 2024), but recent advances in crossbreeding methods developed in other temperate annual pasture species are expected to make the development of crossbred serradella cultivars more efficient. Harrison et al. (2024) reported 14 successful hybridisations of 97 attempts by using the method of Nutt et al. (2020a), where keel petals are removed and emasculated using fine forceps. Veerappan et al. (2014) reported that in barrel medic, the ‘keel petal incision’ method achieves ~80% success rate, with the flower looking similar to an unopened flower bud before crossing. The method prevents pollen from dislodging from the stigma during subsequent movement of the plant and protects the stigma and pollen from desiccation. As the general structure of a serradella flower is similar to an annual medic flower, it is likely that the ‘keel petal incision method’ can achieve high rates of hybridisation in serradella.

Barrel medic and subterranean clover reference genomes have been sequenced (Rose 2008; Hirakawa et al. 2016; Kaur et al. 2017; Kang et al. 2020; Shirasawa et al. 2023) and as serradella species are of increasing importance in Australia, it is recommended that both yellow and French serradella genomes be sequenced. Molecular markers have been developed for several important traits in annual medics (Oldach et al. 2008; Bogacki et al. 2013; Oldach et al. 2014; Peck et al. 2021) and marker-assisted selection has been used to develop a barrel medic cultivar (Peck et al. 2015). Sequencing a serradella genotype will assist the development of markers for key traits, which in turn will assist breeding and selection, including the use of speed-breeding methods.

Core collections that maximise genetic diversity in a limited number of accessions have been developed for a range of key annual legume species in Australia (Genesys-PGR; Nichols et al. 2013; Smith et al. 2021) but not for serradella. Core collections are a useful starting point when screening a large germplasm collection of a species for the presence of a particular trait. It is suggested that a core collection combining the yellow and French serradella species be developed because of the possibility of interspecific hybridisation and the broad range of genetic diversity that exists across these species in the APG.