Department for Environment, Food and Rural Affairs

S.E.D. = 0.146, 27 D.F.

Such low rates of infection are not detectable by standard seed testing protocols and require a PCR based test to objectively determine presence or absence in a sample of 10,000 seeds. In order to develop such a test it was necessary to determine which genotypes of Colletotrichum were pathogenic on lupins.

The Rothamsted Colletotrichum collection now numbers 75 genotypes, although we can not be certain that there are no duplicates. Of these 22 genotypes were not pathogenic on Lupinus albus and 2 have not been tested.

There is some doubt about the taxonomy of Colletotrichum. We used a benomyl sensitivity test based on growth of isolates on media amended with 5µg/ml benomyl. Literature reports that C. acutatum is insensitive to benomyl and that C. gleosporoides is sensitive (Freeman et al. 1998). In addition we used the primer CaInt2/ITS4 which positively identifies C. acuatatum (Sreenivasaprasad et al. 1996).

Of the 58 isolates tested using both methods; 37 were confirmed as C. acutatum, 3 as C. gloeosporoides and 18 gave unclear results. In pathogenicity tests; 7 of the C. acutatum and all 3 C. gloeosporoides isolates were non pathogenic on white lupin cv. Lucille. Of the 73 isolates tested for pathogenicity 22 were non-pathogenic.

Of the 5 isolates collected from strawberry plants 2 proved pathogenic on white lupin. It had previously been believed that the isolates infecting strawberry were different to those infecting lupins (Mark Sweetingham, Agriculture Western Australia, personal Communication).

To study differences between the Colletrotrichum isolates ribosomal DNA (rDNA) was amplified from the fungi by PCR (primers ITS4/5) and cut by restriction digest enzymes (AluI, RsaI, HaeIII and HhaI). The method is described in Ward and Akrofi (1994).

Of the 75 isolates 52 were treated with all 4 enzymes and 7 were either treated with some but not all or gave indistinguishable band patterns with some enzymes. The data show that 24 isolates had the same banding pattern when treated with these 4 enzymes. Generally these were isolates from the same or similar sources. However this group did include one of the isolates collected from strawberry. The only isolate to show a very different banding pattern was the only one not collected from lupins or strawberries (collected from coconut).

Unfortunately these techniques did not distinguish between pathogenic and non pathogenic isolates on lupin and therefore could not be used as the basis for a molecular screen for seed infection. This would require further work. A company based in Western Australia has developed a method which it offers as a service for seed producers and this has been utilised by the UK lupin industry. The test is objective but not quantitative. Should a quantitative test be developed it may be possible to use seed batches with very low levels of infection providing that the seed is treated with appropriate fungicides. Römer (Südwestsaat GbR, Germany, personal communication) confirmed that iprodione and carbendazim was valuable in preventing crop infection when < 1 infected seed in 10,000 was sown (quantitative testing by growing untreated seeds). In Australia it has been demonstrated that thiabendazole and thiram is effective, a result confirmed by Agrichem (manufacturer) staff in the UK. It has also been demonstrated that storage for 1 year (Greg Thomas, Agriculture Western Australia, personal communication) and dipping seed in hot water reduce the rate of infection (Peter Römer, Südwestsaat GbR, Germany, personal communication). However all of these techniques are only suitable for use with seed showing nil or very low levels of infection.
5.0. Pest control.
Mammals and birds have caused damage to lupin crops from time to time. The problems are similar to those for many other crops. Hares, rabbits and deer are the principal grazers. Rooks and crows can pull young seedlings from the soil around the time of emergence.

Occasionally aphids colonised lupin crops. Several species were observed but the most visible and probably most harmful was Macrosiphon albifrons. This is a large and immobile aphid hence populations on any one plant grow very large and kill that plant. However they spread very slowly creating patches of dead plants by late summer. The effect is much more severe within plots in experiments than on a whole field scale where insecticides are rarely applied.

The time of flowering on the main stem is largely determined by the number of true leaves initiated and the phylochron. In addition there are effects of time to emergence. Julier and Huyghe (1993) described a simple linear model for a selection of autumn sown genotypes in which the thermal time to the opening of the first floret on the main stem was a function of the number of true leaves on the main stem. i.e. the phylochron was constant. Shield et al. (in press) demonstrated that the phylochron was modified by day length in the four determinate cultivars (Lucyanne, Ludet, Lucille and Lunivers), but that this had a very small effect within the narrow range of day-lengths experienced by an autumn sown genotype in Northern Europe. The number of true leaves initiated on the main stem is controlled by the genotypic vernalisation response. See Annexe 2 where empirical models for this and other processes are presented.

In the early years of testing the autumn sown determinate material the theory appeared to be correct. However in the late 1990s reports were made of crops that we not sufficiently dry (seed moisture >25%) for harvesting in early October (e.g. ADAS in SW England). In 1998 and 2000 this was experienced at Rothamsted.

A study was initiated to determine the cause of this unpredictable aspect of crop behaviour. The study is not complete, but our understanding of the processes of maturation of the model white lupin genotypes Lucyanne and Lucille has been improved.

Pod 5 on the main stem, the fifth pod from the base of the inflorescence, was chosen as the representative pod. The time of maximum pod wall dry weight (DW) in the representative pod coincided with maximum plant or crop DW and with observations of leaf canopy degeneration. At this time seed DW was less than 50% of the maximum, the subsequent rapid seed growth is sustained by remobilisation of assimilates as the rate of new carbon fixation slows with leaf degeneration. In addition there is a contribution of pod photosynthesis.

Initially it was thought that the thermal time from flowering to maximum pod wall DW was constant and that the time from maximum pod wall DW to maturity varied. It is now thought that there may be relatively small variation in the time of maximum pod wall DW from season to season and this is under investigation. However, the greatest variation in time to maturity comes after the time of maximum pod wall DW. This variation has been shown to be controlled primarily by the soil moisture deficit and secondarily by the temperature in the 14 days after the maximum pod wall DW is achieved. A soil moisture deficit indicating that the easily available water capacity of the soil has been exhausted and high temperatures ensure early maturation. Dracup et al. (1998) working in Western Australia found that maintaining easily available soil moisture with irrigation interfered with the process of remobilisation of stored assimilates from the pod wall such that it was incomplete at maturity. This may be the mechanism that delays maturation in the UK. In Wales, Ireland and parts of northern and western England, these conditions do not prevail sufficiently frequently to ensure late summer harvesting of dry lupin seed in most years. The data collected relates to the white lupin genotypes Lucyanne and Lucille and appears to be relevant to other white lupin genotypes observed. There were genotypic differences between Lucyanne and Lucille. Lucille matured on average 5 days later than Lucyanne.

Narrow leafed (L. angustifolius) and yellow (L. luteus) genotypes appear to behave differently and are more reliable in cooler damper climates.

However, considerable unexplained variation in yield was recorded with the model determinate genotype (Lucyanne) in a large data set derived from many environments between 1994 and 1996 (Figure 5). The data were filtered to remove individual values with an obvious explanation for anomaly, such as poor weed or disease control. A wide range of factors such as weather and soil variables during all or part of the growing season were used to explain the variance in seed yield in the remaining data. Ultimately the numbers of branches m^-2 was identified as the variable that explained the greatest proportion of the variance in seed yield (Shield et al.2001). Figure 5 shows that there were two distinct relationships between seed yield and the number of branches (or axes, Shield and Milford 1995) carrying yield at harvest; one for high yielding crops and one for more modestly yielding crops.

F
igure 5. The relationship between the number of principal yield bearing axes (main stem and 1° branches) m^-2 and seed yield for 50 crops of cv. Lucyanne grown at sites across the UK between 1994 and 1996.

In 1998-99 the model white lupin was changed as the dwarfing character became available. The model dwarf determinate Lucille showed very similar developmental characters (numbers of leaves and branches) to Lucyanne, but in a more compact plant form. A field experiment in 1999 demonstrated that yield stability in Lucille was affected by the number of branches m^-2 in a very similar way to Lucyanne.

The knowledge gained from this work in conjunction with our understanding of plant development allowed us to refine our sowing date and seed rate recommendations for dwarf determinate genotypes like Lucille. When sowing early (potentially earlier than the beginning of the normal sowing window for a site) the seed rate should be 30 seeds m^-2, this should increase with time until the end of the recommended sowing window, when 40 seeds m^-2 should be sown. An additional factor to be considered when sowing genotypes such as Lucille is the soil type, on clay soils the plants are more compact and a greater population density is recommended than on sandy soils.

These latter two years coincided with a deliberate policy to place the lupin crop on land that had grown lupins in the recent past (1996 and 1998 for the sites used in 2002 and 2003 respectively). Previously all lupin experimentation (except certain pathology experiments) had been placed on new sites. However, close monitoring failed to show any problems associated with cropping history. In 2002 the pressure on the crop from rust was the greatest that had been observed, the crop was given a ‘standard’ two fungicide spray protection. The crop may have responded positively to a third spray. Were that spray to have preserved the 0.5 t ha^-1 that appeared to have been lost in comparison with previous years it would have been economically justified. There was no opportunity to test that theory at the time. In autumn of 2002 on the site of the 2003 crop the soil was not sufficiently re-consolidated following ploughing. It has been observed in the past that over-winter plant losses have been large in poorly re-consolidated soils at Rothamsted. In this case slugs grazing underground and out of the reach of pesticides caused most of the damage. At harvest there were only 6.4 plants m^-2 of Lucille. The severe drought of summer 2003 is not thought to have had a great influence on the seed yield.

One of the great disappointments of the autumn sown lupin crop has been the failure of so many others, whether they are farmers or researchers, to reproduce our yield results. Yield results from our trial partners (ADAS, ARC and Reading University) are presented in Table 5 and reflect this variability. Very acceptable yields have been achieved but must be viewed alongside complete crop failures. As in Table 4 data are selected from the management regime most similar to our recommended agronomy.

– no crop sown, 0.00 crop sown but failed.

9.0. Further improvement of yield.

Lagunes-Espinosa et al. (1999) were able to clearly demonstrate that when a wide range of genotypes exhibiting a range of pod wall proportions (18-32%) were grown the seed yield was negatively correlated with the pod wall proportion. Our unpublished work (Figure 6) with one genotype (cv. Lucille) showed that the environmental conditions that lead to the greatest seed growth also lead to the greatest pod wall DW at harvest, and therefore very little change in pod wall proportion.

Figure 7 shows that the quantity of PAR available had a large influence over the main stem pod (seed plus wall) growth rate in the 6 weeks after flowering. The proportion of the PAR reaching the base of the main stem inflorescence at the beginning of main stem pod growth (flowering) was used as an indicator of the PAR available during the most rapid phase of main stem pod growth (the 6 weeks following flowering). This and Figure 6 indicated that it would not be possible to use agronomy to minimise the pod wall proportion and maximise seed yield.

Assuming that the only role of the pod wall was to act as packaging for the seed was a simplification. The pod wall acts as an intermediate storage organ for assimilates destined to drive seed growth and photosynthesises in its own right. Lagunes-Espinosa et al. (2000) were able to demonstrate that the low pod wall proportion genotypes were able to produce large seeds with high nitrogen (protein) concentrations.

Our work to study pod wall photosynthesis was less clear as large variance was encountered between individual measurements. Later work was carried out in the controlled environment rooms at Rothamsted to lessen the variation due to environmental conditions. This revealed that much variation existed between plants of the same genotype and within pods of similar age on the same plant.

It was demonstrated that lupin pods photosynthesise but that there is only a net fixation of CO₂ at high light intensities (Figure 8). Shading the pods during the light period gave an indication of the extent of the respiration taking place. In this example a light intensity of 1500 µmoles PAR m^-2 s^-1 (bright summer sunshine) was required to maintain the pod at its CO₂ compensation point. This example is typical of the large number of measurements made of field grown plants.

The pods had relatively poor gas exchange characteristics (compared to a leaf) leading to a large build up of CO₂ in the internal gas spaces during the dark period (Figure 9). This CO₂ appeared to be re-fixed during the first few hours of the light period. It was not possible to detect with any confidence differences in pod photosynthesis between genotypes despite differences in pod wall morphology.

10.0. Seed Quality.

However, a large number of leaves per plant also indicates a large number of internodes, and a large mass of stem. Pate et al. (1998) describe stem tissue as an important site for nitrogen accumulation (especially as arginine) and the stem as an important site for nitrogen enrichment of the xylem ensuring that large quantities of nitrogen are supplied to the sinks in the upper parts of the plant. In 1995 in South Devon an experiment was conducted to investigate the affect of a plant growth regulator on Lucyanne and Ludet sown on different dates (Table 6). The growth regulator applied was paclobutrazol, a triazole with fungicidal properties (observed and recoded in a lupin experiment at Rothamsted). Application was on April 19^th, which is very early in the growing season for the fungicidal properties to have an effect on seed quality (see the effects of timing of tebuconazole on seed yield in Summer Diseases above). The proportion of the plant that was still green at harvest was affected by the application of growth regulator. This was an unusual result in that prior experience (in a range of crop types) would suggest that the growth regulator may keep the plants green for longer (especially if there was a fungicide effect). Our results were contrary to this with more of the untreated plants remaining green. The data are representative of one site and one season, but raise the question “is stem an important source of nitrogen to sustain the latter stages of seed growth ?”. Alternatively “are these results explained by later maturity that allowed a greater accumulation of nitrogen in the seed ?”

It was established that the roots system, specifically the cluster roots (previously termed proteoid roots), were highly adapted to exploit a soil environment heterogeneous for lime. This response enabled greater shoot growth and leaf greenness (Kerley 2000a and b)

This work was progressed by comparing L. albus to L. pilosus Murr., which occurs naturally in calcareous soils (Kerley et al. 2000). However, the reasons for the tolerance of L. pilosus to these soils had not been investigated. By comparing L. albus with this tolerant species in both homogeneous and heterogeneous soils, it was anticipated that specific mechanisms of intolerance in L. albus would be elucidated.

The root systems of L. albus and L. pilosus responded to a patch of acid soil within a limed-soil profile through the specific proliferation of cluster roots in the acid soil. This proliferation resulted in a higher rhizospheric citrate concentration on a soil dry weight basis. Although individually the cluster roots were not functionally more active in citric acid exudation, as they were so densely packed it was probable that the zones of citric acid exudation overlapped, ensuring extraction of P and Fe from the whole of the acid soil.

nutritional benefit from the soil. The PAR response curve demonstrated the difference between the plants (Figure 12). There was no response by L. pilosus to the presence of patches of acid soil, in contrast L. albus showed very little ability to respond to the increasing PAR when grown in limed soil, but the response was much greater when acid soil was present. This difference clearly demonstrates the tolerance and intolerance of both species to the limed soil.

Although both L. albus and L. pilosus responded strongly to a heterogeneous soil environment by the proliferation of cluster roots, differences other than PAR response were also apparent between the species. L. pilosus produced 15% of its cluster roots in the limed soil; in contrast fewer were produced by L. albus. Being adapted to calcareous soil, L. pilosus may grow better in the heterogeneous profile and be able to exploit the acid soil more so than L. albus. In addition, the strong secondary and tertiary lateral roots system of L. pilosus may facilitate the exploration of soil. It is possible that the positive response of L. pilosus to an acid soil patch and the greater growth in homogeneous acid soil profiles indicates that although the species tolerates limed soil it might preferentially grow in acid soil.

This work demonstrated important similarities between the two species in their response to limed and heterogeneous soil. L. pilosus however demonstrated a better root development in the limed soil, indicating better nutrient acquisition, which was reflected in the higher photosynthetic capacity of the shoot. The response of the root system as a whole to the stress conditions is considered in the next section

It is difficult in soil-based experiments to determine the changes in root architecture specific to pH, HCO₃^- or Ca. Liquid culture system and two-dimensional root chambers were developed to provide an environment in which these conditions could be evaluated more easily (Figure 13).

The liquid culture system was used to separate pH, HCO₃^- and Ca, effects on the root system of the intolerant cultivar Lucyanne (Kerley and Huyghe 2001). These changes were specific to the type of root, and differed depending on the stress imposed. The study also discriminated between a genotype and cultivar of L. albus and a genotype of L. pilosus, based on quantifiable responses to solution culture pH.
Generalised root measures

Increasing solution culture pH had no effect on root dry weight, weight ratio and Specific Root Length, whereas increasing the Ca concentration affected both the dry weight and weight ratio. In the presence of HCO₃^-, root dry weights and weight ratios were lower than in its absence. The decrease in root weight ratio indicated that the root system was proportionately smaller than the shoot when stressed by HCO₃^- at pH 6.5. Although, the pH responses with HCO₃^- are difficult to interpret, the results indicate that HCO₃^- could be the most important stress of the three at decreasing root exploration.

11.1.1 Tap root architecture

The tap root length was shorter in the presence of higher pH and Ca and HCO₃^- concentrations. The retardation of L. angustifolius L. roots exposed to HCO₃^- has been attributed to decreased respiration, leading to a failure to acidify the apoplast for cell expansion. As the roots were well aerated in this study, a lack of respiration was unlikely. However, both HCO₃^- and pH did decrease root length, most likely through the process of decreased cell elongation. The role of Ca would have been similar; the binding of Ca to cell walls would have prevented elongation.

The tortuousness of the tap root’s pathway increased under pH and Ca stresses from an unimpeded straight line (index 0) to maximum index of 0.36 in 2 mm Ca. Although the physiology of the tortuous response is unclear, it is likely to have been a direct attempt to avoid the unfavourable environment. The decrease in tap root length and increase in tortuousness however, resulted in a root system that would be less efficient for nutrient and water acquisition. Interestingly, HCO₃^- stress did not affect the tortuousness as in the other treatments. Although the reason is unclear, it does demonstrate the presence of different physiological responses to the stresses.

Tap root death was a major response to pH and HCO₃^- stress. In contrast, Ca had no effect on root mortality. This further indicated that a different physiological stress occurred in response to Ca, compared to pH and HCO₃^-, which was related mainly to cell elongation. Death occurred mainly at the tip of the tap root, and resulted in a loss of apical dominance, with the subsequent proliferation of many indeterminate lateral roots.

11.1.2 Lateral root architecture

Stress responses were also apparent in the lateral roots of the cultivar Lucyanne. Fewer lateral roots were produced under all stress conditions, which resulted in a lower density along the tap root. This indicated that the root system under the stress conditions had a much lower capacity to exploit the media than when not stressed.

Lateral roots were defined as either determinate (<3cm) or indeterminate (>3cm); their ratio (determinate : indeterminate) decreased due to alkaline pH, was unchanged by high Ca, and was increased due to HCO₃^- at pH 7.5. At the alkaline pH, the proportionately fewer indeterminate roots accounted for the decreased ratio. The implication of this response is an impaired ability to explore the environment. In the presence of HCO₃^- at pH 7.5, the fewer determinate roots accounted for the increased ratio, and could have implications for the ability of the plant to exploit its immediate surroundings. This difference in response is evidence for these stresses being discrete. Increasing Ca concentration did not change the ratio, however both root types were equally decreased in numbers, indicating that both were similarly affected.

Fewer cluster roots were produced under the Ca and HCO₃^- stress conditions, whereas alkaline pH had no effect on cluster root production. Care must be taken in the interpretation of the solution culture results because of the nutrient solubility, concentration homogeneity, and the regular replacement of the solution. The process of nutrient uptake that relies on the concentration of exudates into the rhizosphere would have been absent under liquid culture conditions. In this respect the use of liquid culture to compare cluster root activity is unreliable. However, it is important to determine that Ca and HCO₃^- do affect cluster root numbers.

By using a drip culture system in which the root system of L. albus can develop observable root architecture, changes in architecture in response to pH, HCO₃^- and Ca stress were quantified. These changes were present in the tap, lateral and cluster roots, and were seen to vary between the stresses imposed. However, as with all liquid culture analyses, care must be taken in the relating these responses to those that may occur in soil systems, due to the complexity involved. This technique was used as the basis of a successful system to screen genotypes (see later section).

11.2 A nutritional basis for tolerance to calcareous or limed soil

The aim of the initial study was to compare a number of nutritional and biomass analyses to evaluate the calcareous or limed soil tolerance of L. albus genotypes, selected from previous trials.

Leaf chlorosis was shown to be an unreliable measure of calcareous soil tolerance in the field (e.g. Kerley et al 2001), whereas a genotype x soil-pH interaction in the expanded leaf number indicated that this analysis might be of use in genotype evaluations. Nutrient concentration differences were apparent between the species and between the L. albus genotypes. This indicated the occurrence of possible tolerance mechanisms including the control of Ca uptake and the partitioning of Fe. Clear differences were apparent between the three species in terms of tolerance to the calcareous soil. Within L. albus important differences were apparent when specific analyses were examined. However, variation between different analyses and at different stages of growth resulted in these differences. Taken as a whole between the L. albus genotypes, they were not of sufficient magnitude to discriminate potentially tolerant from susceptible genotypes. The most interesting differences were apparent in the concentrations of Ca and Fe in the leaves and stems (Table 7).

The only nutrient that increased in concentration when grown in the calcareous soil was Ca. The L. albus genotypes Lucyanne and La 672 showed increased Ca contents in the leaves, suggesting these plants were unable to control Ca uptake into the leaf tissue. In contrast L. pilosus showed no change in leaf Ca concentration between the soil types, indicating that it might be able to control the Ca concentrations in the leaf tissue. Importantly, the response of La 673 was similar, indicating one possible mechanism of tolerance in L. albus.

When the stem tissue was analysed, all the plants increased their stem Ca concentration when grown in the alkaline pH soil. Ca was therefore taken up in greater concentrations by all plants in the alkaline pH soil. However the maintenance of Ca concentrations in the leaf tissue of L. pilosus and La 673, demonstrated a means of nutrient partitioning that may provide a level of tolerance. Additionally, this partitioning may provide a means of comparing genotypes for tolerance.

The potential of Fe and Ca to be nutritional screens led to further research on both nutrients. Fe deficiency is a major cause of lime-induced chlorosis, however despite the appearance of chlorosis, whole-plant total Fe concentrations were not always deficient when L. albus was grown in lime soil. From literature concerning other species, it was possible that differences in the concentration of Fe forms (FeII and FeIII) within leaf tissue might explain the induction of lime-induced chlorosis, even in plants sufficient in Fe. Thus the partitioning of the forms of Fe in the shoot was examined.

Ca is one of the few elements to show an increased shoot concentration when grown in limed or calcareous soil. Within the shoot, the control of Ca is a major tolerance mechanism of calcicole species, and its partitioning might provide another potential screen.
Table 8. The concentration of Ca form (total, insoluble, soluble and the ratio of soluble Ca : total Ca) and Fe form (total, FeIII, FeII and ratio of FeII : total Fe) in leaflet tissue of the L. albus cultivar Lucyanne harvested in April (215 DAS) in neutral or alkaline-pH soil.

Nutrient	Soil pH		P
(tissue dry weight basis)	Neutral	Alkaline
Total Ca (mg g-1)	15.3	36.8	<0.001
Insoluble Ca (mg g-1)	12.3	32.9	<0.001
Soluble Ca (mg g-1)	3.00	4.14	<0.001
Ca ratio (soluble Ca : total Ca)	0.21	0.109	<0.001
Total Fe (μg g-1)	1830	2251	ns
FeIII (μg g-1)	1751	2177	ns
FeII (μg g-1)	79.5	73.5	ns
Fe ratio (FII : total Fe)	0.046	0.037	ns

Plants were grown in neutral pH or limed soil (Kerley et al in press). In April and then in June the shoot tissue was sampled, divided into specific tissue types and analysed for FeII and FeIII, as well as soluble and insoluble Ca fractions. Nutritional and tissue differences were seen between the April and June sampling (Tables 8 and 9).

In April the Ca response was mainly as insoluble Ca and indicated that the excess Ca taken up from the alkaline soil was successfully insolubly sequestered. No differences were apparent in the Fe fractions possibly because the plants were still in an over-wintering form and were less active. The total Fe concentrations in April were considerably higher than values reported for spring sown lupin material harvested in early summer, indicating that continued uptake of Fe occurred over winter by the autumn sown plants. This Fe uptake may explain the observed greenness of the leaves of these plants.

When actively growing in June, the increase in total Ca in the limed soil grown plants was as both the soluble and insoluble forms. The percentage increase was greater in the soluble compared with insoluble form, indicating that although the Ca was being taken up and sequestered, there remained excess Ca in the soluble form. The presence of soluble Ca can affect plant growth e.g. by reducing net assimilation rates through the suppression of stomatal opening by the increase in Ca concentration in guard cells. Although it is unclear whether the concentrations of soluble Ca reported are a cause of a stress response within the plant, the change in soluble and insoluble Ca are important analyses that responded to stress conditions.

In June, the leaflet FeII concentration declined whereas that of FeIII increased. This lower FeII has been observed in calcifuges and is considered to be due to immobilisation of Fe into the FeIII form within tissues and is thus an important response between the calcifuge and calcicole behaviours.
Table 9. The concentration of Ca form (total, insoluble, soluble) and Fe form (total, FeIII, FeII) in leaflet tissue of the L. albus cultivar Lucyanne harvested in June (280 DAS) in neutral or alkaline-pH soil

Nutrient	Tissue	Soil pH		Tissue (s.e.)	Soil pH (s.e.)	TxS (s.e.)
(tissue dry weight basis)		Neutral	Alkaline
Total Ca (mg g-1)	Stem	2.34	3.00	<0.001	<0.001	<0.05
	Leaflet	12.74	14.53	(0.280)	(0.229)	(0.396)
Insoluble Ca (mg g-1)	Stem	2.02	2.66	<0.001	<0.001	ns
	Leaf	11.37	12.69	(0.288)	(0.235)
Soluble Ca (mg g-1)	Stem	0.32	0.34	<0.001	<0.001	 0.001
	Leaflet	1.37	1.85	(0.021)	(0.017)	(0.030)
Total Fe (μg g-1)	Stem	93	125	<0.001	<0.001	<0.05
	Leaflet	358	443	(17.7)	(14.5)	(25.1)
FeIII (μg g-1)	Stem	68	102	<0.001	<0.001	 0.001
	Leaflet	161	321	(18.6)	(15.2)	(26.4)
FeII (μg g-1)	Stem	25	23	<0.001	<0.001	 0.001
	Leaflet	197	122	(6.5)	(5.3)	(9.2)

11.3 New genotype identification: defining the tolerance of ‘Giza1’ and establishing its underpinning physiology.

L. albus

A series of experiments compared a range of Egyptian genotypes with European cultivars and genotypes on neutral pH and limed soils. They were compared using the physiological parameters of photosynthesis, and Fe and Ca form and partitioning established from previous work. The results were used to determine some physiological causes of tolerance (Kerley et al, 2002).

The responses of the Egyptian genotypes to the limed soil differed from either of the intolerant European cultivars Lucyanne or Lublanc. Although leaf greenness is not always consistent within an experiment over time, the absence of chlorosis in any of the Egyptian material was comparable with the tolerant L. pilosus, and indicated that the plants tolerated the stress better than the European material. The leaf light transmission meter (SPAD) analysis discriminated between L. pilosus and the Egyptian genotypes and indicated that the Egyptian genotypes were subjected to a sub-chlorotic level of stress that was greater than L. pilosus. Leaf emergence has been considered a potential means of genotype evaluation, and did effectively discriminate between the Egyptian and intolerant plants. However, this evaluation gives no insight into the tolerance physiology, and was not sufficiently discriminatory in a subsequent UK field trial (see later).

The above screens measured the results of physiological processes; more definitive screens were needed to measure the processes themselves, e.g. Ca and Fe partitioning and photosynthesis. Differences in the leaf soluble Ca in the first experiment confirmed lower concentrations in landraces Egypt 99 and 121 compared with Lucyanne and Lublanc, indicating that some control over soluble Ca was present. Unfortunately this was not apparent in Giza1, indicating that such a process may be minimal in this landrace. However, Giza1 did generally have lower mean values for soluble Ca than Lublanc, with La 675 intermediary; as such it may possess some form of tolerance that requires further study in a truly calcareous soil.

A potential tolerance mechanism was apparent in the Fe response of Giza1 compared with the intolerant La 675 (Table 10). Giza1 showed a smaller partitioning of its Fe as stem Fe³⁺ in limed compared with neutral-pH soil and even increased the fraction of active Fe²⁺ in the leaf tissue, accounting for its lack of chlorosis. This mechanism could facilitate calcicole growth in calcareous soil and might be one explanation why the Egyptian material is cropped in highly calcareous soils. This response also explains the maintenance of P_max values and the absence of a decrease in quantum yield under limed soil conditions, which were in contrast to the non-Egyptian L. albus.

To compare the root growth of Giza1 with a range of European plant types, they were grown in the liquid culture system and their root architectures compared. The root systems showed genotypic differences in response to stress. Some of these could be attributed to differences in vigour or seed size; for example shoot and root dry weight analyses. However, some measurements did show genotype x environment interactions e.g. tap root tortuousness and the indeterminate lateral root analyses. Compared with Lublanc, Giza1 had a lower tap root tortuousness index at pH 7.5 and in 3 mm Ca. Crucially, it also suffered a lower proportion of tap root death in all three stresses. This response, and its longer tap root system, possibly accounted for its higher tortuousness index when grown in HCO₃^- compared with Lublanc. Giza1 also had a higher lateral root number ratio in the HCO₃^- and Ca treatments which was due to a higher number and greater density of indeterminate roots. These responses indicate that Giza1 is better adapted to root growth in these stress conditions than Lublanc, and provides one potential explanation for the improved iron physiology identified above.

Figure 15 summarises the differences found between the current intolerant cultivars such as Lublanc, and the more tolerant Egyptian landraces such as Giza1. However, these landraces are unsuitable for the UK (e.g. see Figure 14) and will require a breeding strategy to incorporate their tolerance mechanisms into a suitable cultivar form for the UK.

Because there is wide variation for tolerance, and tolerance mechanisms within the Egyptian material, there is a need to discriminate more effectively between plant types based on more than a single screening system. Once achieved, this discrimination will facilitate breeding programs designed to develop calcareous soil tolerant agronomic cultivars of L. albus.

Only in the soil chamber experiment did shoot dry weight effectively discriminate the tolerance of Giza1 from an intolerant L. albus cultivar. In the limed-soil field trial the plants may not have been under sufficient stress to show a reduction in shoot dry weight, and in the calcareous-soil trial there was no non-stressed control to determine the potential growth of each plant type. Shoot dry weight may not be a reliable means of evaluation, as although tolerant, in the chamber trial the difference in dry weight of L. pilosus between neutral-pH and limed soil was comparable with Lublanc. Additionally, shoot dry weight is also of little value, as it requires the destruction of the plant.

Analyses of the root system could only be conducted in the soil chamber experiment. No clear response to the limed soil was seen with Lublanc, whereas the more tolerant plants showed a proliferation of cluster roots in the limed soil. In L. pilosus the cluster roots proliferated, possibly at the expense of the shoot growth and lateral roots. In contrast, cluster root proliferation in Giza1 was not concomitant with a reduction in the shoot and lateral root dry weight. These responses corroborate earlier work in which cluster root proliferation of the L. albus cultivar Lucyanne occurred, but with a reduction in shoot and non-cluster root dry weight.
11.5 Fe forms as evaluation screens
In the soil chamber experiment, differences in the partitioning of leaf Fe were apparent between the three plant types (Giza1, Lublanc and L. pilosus) when grown in the limed soil compared with neutral-pH soil. Lublanc maintained its leaf total Fe concentration at levels comparable with those in non-limed soil through an increase of FeIII, whereas there was 25% less FeII. Although in L. pilosus the concentrations of both forms of Fe in limed soil grown plants were less compared with those in neutral-pH soil, the change in FeII was small. In contrast, although Giza1 had a small loss of total Fe concentration, this was due to FeIII, whereas the concentration of the FeII was maintained. This result agrees with recent literature which concludes that immobilization of ‘physiologically less active’ Fe (FeIII) by calcifuges is related to c
hlorosis (Zohlen and Tyler 2000).


Figure 16. Fe form (II v III) expressed plotted for each plant type. L. albus cultivars Lucyanne (◊), Lublanc (□), Primavera 106 (o) and Bardo (∆), genotypes Giza1 (■), La668 (▲), La673 (●) and La675 (♦) and landrace Egy121 (+) and of L. pilosus genotype LD124 (x).
The plant Fe responses in this study indicate that it is the maintenance of high concentrations of FeII in the leaf tissue that is of major importance in the tolerance to a limed or calcareous soil. The concentration of the FeIII is

less important. However it does provide a secondary measure of the Fe physiology, allowing the whole Fe response to be evaluated. The change in the concentration of FeIII between stressed and non-stressed plants may be more important than absolute concentrations. Analysis of Fe form is a potentially valuable screen as it is rapid, non-destructive (to the whole plant), requires a minimal quantity of leaf tissue and can be repeated on a single plant basis during the growing season.

Differences in the partitioning of Fe were apparent in the field trials that support the results from the soil chambers. The plants of Egyptian origin contained the highest concentrations of FeII, whereas the lowest concentrations were present in Lucyanne and Lublanc. In trial 1, the FeIII of the tolerant plants was less compared with plants in the non-limed soil, whereas concentrations were maintained in the intolerant plants. This demonstrates the need to compare FeIII in both stressed and non-stressed plants, and indicates that important differences were present between tolerant and intolerant plants. Such analyses could not be done in trial 2, however most FeIII was found in the intolerant plants and least in Giza1. This indicates that in the intolerant plants, more FeII was converted to FeIII. Fe form appears important in the response of lupins to calcareous soil. The processes that results in their partitioning in the plant is not understood, although literature indicated that the role of bicarbonate ion may be crucial.

Expressing the field trial Fe-form data as a ratio or graphically (Figure 16) showed that Giza1 was at one extreme of the range and Lucyanne and Lublanc were at the other. This created ‘benchmarks of tolerance’ used to compare the other plants. This analysis indicated that Bardo responded similarly to Lucyanne and Lublanc, whereas Primavera 106 represented an improvement over these plants, although it was not as tolerant as Giza1. Some caution must be taken in interpreting physiological effects of the Fe concentrations, as we do not know their critical values or how they vary between the plants. In this absence of knowledge and the comparisons between stressed and non-stressed plants are central in defining the processes occurring.

I
n contrast to L. pilosus, the concentrations of both the soluble and insoluble Ca fractions increased in Lublanc, signifying little control in the uptake of Ca, and that the plant behaved in a calcifuge manner. The Ca concentrations in Giza1 were more comparable with Lublanc than L. pilosus. However, Giza1 differed from Lublanc in the control of the soluble Ca; in Lublanc it increased by 35% in the limed compared with neutral-pH soil of the soil chamber experiment, whereas that of Giza1 increased by only 6% indicating a level of control not present in Lublanc. It is probable that whilst Lublanc showed a general increase in leaf Ca, the increase in insoluble Ca of Giza1 was a means to control the concentrations of the soluble fraction. However the physiology behind the processes require further study.

As with Fe, the Ca form analysis is a repeatable, rapid and non-destructive analysis that is conservative of the whole plant. However, we do not know the critical values for the Ca forms, or whether high concentrations of soluble or insoluble Ca represent stress, tolerance, or sequestration.

Comparing the Ca forms from the field trials as ratios or graphically (Figure 17) did show differences between the plants. As with Fe, Giza1 and Lucyanne were at extremes of the plant types. Interestingly, Egy121 was comparable with Giza1, providing evidence that the Egyptian plants are tolerant of calcareous soil. Primavera 106 had not been evaluated prior to this study, it was the cultivar that was most comparable with the Egyptian material, providing some evidence for greater tolerance.

By not comparing root systems, these protocols can be used in short term (30 day) pot experiments using limed and non-limed to provide and effective, rapid non-destructive screening systems for evaluating plant material either in collections or during breeding projects. It is also the basis up on which a project investigating the molecular basis for tolerance should be reliably and soundly based.

Using these plant types and analyses as evaluation parameters, Egy121 and Primavera 106 were shown to possess levels of tolerance greater than the cultivars Bardo or Lublanc. The magnitude of the difference in Ca concentrations between L. pilosus and L. albus must be considered when using the tolerant species as a control for L. albus. Very different processes may be occurring in L. pilosus that are absent in tolerant L. albus plant

12.0. Knowledge Transfer.

In the autumn of 2000 an extensive web site was introduced that contained all of the agronomic information gained from the work and a recommended management plan for autumn sown lupins. During the first year after the launch the site received an average of 9 hits per day, approximately 67% new comers and 33% returning viewers. Only about 15% of hits came from outside of the UK. By the end of 2003 the number had fallen to 5 hits per day, it is assumed that most people who required the information had acquired it.

From 1999 commercial development of the crop began. Two seed supply companies entered the market with varieties of autumn sown white lupin. Rothamsted Research assisted both with expert advice. Goreham and Bateson marketed seed nationally but were assisted in a development project based in the Welsh Marches, an ideal target area of traditional mixed farms where beef production was struggling to maintain farmer income. The production of all animal feed on farm was seen as directly economically beneficial and ultimately seen as a route to add value through a premium brand. ADAS, Countrywide Farmers and Meadow Valley Livestock formed a strong consortium to push this concept forward. The project undoubtedly had successes, but also suffered some problems largely because the silt soils of the area are not well suited to autumn sown lupins.

Cebeco Seed Innovations also marketed seed of autumn sown varieties. Rothamsted Research assisted them in a national programme of field days based around demonstration crops. The programme began with two large scale days at Rothamsted during 1999. In 2000 this expanded to 5 regional events across England. All were well attended.

All new crop types are introduced with imperfect knowledge and commercial introduction is part of the learning process. Autumn sown white lupins were introduced with one of the best possible knowledge packages yet struggled in commercial agriculture. The principal reason was the complete reliance on the achievement of the correct sowing date for the site. This was compounded by the fact that unusual autumn weather could render the recommended sowing date inappropriate, as happened in 1997-98 and 2000-01. The introduction of dwarf determinate varieties such as Lucille was intended to lessen the reliance on the sowing date; however these varieties were blighted by late delivery of seed from France. This is perhaps understandable as lupin is a low volume crop demanding attention at a seed processing plant at the busiest time of year. It is clear that a number of commercial crops failed because late delivery of seed forced late autumn sowing. The problem can be overcome by over year storage of seed, but this is difficult with new varieties where seed is scarce and demand is high. It also has physical and financial implications for the seed trade.

An additional problem was the lack of suitable approved post emergence broad leafed weed herbicides, and the requirement for 2 summer fungicides. Despite the summer fungicides only being applied at low rates the crop now has an image as requiring too many inputs.

The introduction of a number of diverse spring sown lupin varieties during the period 2000-2003 which were cheap and easy to grow and produced seed yields similar to the winter varieties finally dissuaded farmers from growing winter varieties.

14.0 References.