Seed-borne cucumber mosaic virus infection of subterranean clover in Western Australia

In Western Australia, infection with cucumber mosaic virus (CMV) was widespread in all three subspecies of subterranean clover (Trifolium subterraneum) growing in plots belonging to the Australian National Subterranean Clover Improvement Programme. Seed-borne CMV was detected in seed harvested in 1984–1986 of 18/25 cultivars from two collections of registered cultivars; seed transmission rates ranged up to 8.8%. Seed samples from CMV-inoculated plants of 11 cultivars transmitted the virus to 0.5–8.7% of seedlings. Seed transmission rates greater than 5% were obtained only with cvs Enfield, Green Range and Nangeela. CMV was not detected in seed harvested in 1975–1981 from one of the registered cultivar collections, in 17 commercial seed stocks from 1986 or in a survey of subterranean clover pastures.
Symptoms in subterranean clover naturally infected with CMV included mottle, leaflet downcurling and dwarfing but severity varied with cultivar and selection. CMV isolates from different sources varied in virulence when inoculated to subterranean clover; two (both from subterranean clover) were severe, two moderate and three (including one from subterranean clover) mild. In pot tests, CMV decreased herbage production and root growth (dry wts) of cv. Green Range by 49% and 59% respectively. In spaced-plants growing in plots, CMV decreased herbage production and root growth of cvs Green Range and Northam by 59–630 and seed production of cv. Green Range by 45%. In rows sown with infected seed, aphid spread increased infection levels to 75% in cv. Green Range and 44% in cv. Esperance and losses in herbage production of 42% and 29% respectively were recorded.
CMV isolated from subterranean clover included isolates from both serogroups.     

Subterranean clover decline in permanent pastures in north-eastern Victoria

Experimental sites were established at two locations in north-eastern Victoria to define factors limiting the establishment and growth of Trifolium subterranean L. (subterranean clover). Liming the soil, seed inoculation and fungicide application were used in renovating subterranean clover pasture on two acidic soils (Longwood: brown/grey sandy loam DY 3.14 and Seymour: grey brown light clay DY 3.22, Northcote classification) with mean annual rainfall of 650 mm and 600 mm respectively. Soil acidity, low available soil phosphorus and plant disease were identified as factors limiting clover yield on these soils. Significant yield responses to lime (35–140%) were obtained with subterranean clover at both sites, with corresponding decreases in Al in the 0–10 cm soil horizon. Liming the soil, when combined with seed inoculation, increased the number and effectiveness of root nodules at both sites. Soil P available for plant growth was low at both sites (6.1 and 8.4 μg g⁻¹) resulting in sub-optimal P concentrations in the clover herbage (45 mmol kg⁻¹ at Longwood). Levels of root disease were low but Aphanomyces euteiches and Phytophthora clandestina (causal agents of lateral and tap root rot) were detected frequently on roots. Application of fungicide resulted in higher dry matter yields (p=0.05) at both sites. An assessment of the relative contributions of these limiting factors and the benefits to be obtained from better management would provide a clearer picture of the profitability and sustainability of this farming system.     

Flowering Sequence and the Ordering of Impermeability Breakdown in Seed of Subterranean Clover (Trifolium subterraneum L.)

The pattern of breakdown of seed coat impermeability of seedwithin the burrs of two cultivars of subterranean clover (Trifoliumsubterraneum L.) was closely associated with relative differencesin flowering sequence within the inflorescence. The patternof seed weight variation showed very little relationship withthe pattern of breakdown, while that for seed moisture contentshowed no relationship. Within a burr it was almost invariably the seed which developedfrom the latest flower within the inflorescence which was thefirst to become permeable. An ordering of the timing of breakdownwas evident only within the burr; there was no evidence fora pattern of breakdown of impermeability between the differentburrs along the runner. Trifolium subterraneum L., subterranean clover, seed coat impermeability, seed weight, seed moisture, hardseededness     

Isoflavone Content in Subterranean Clover Germplasm from Sardinia

Subterranean clover (Trifolium subterraneum) is an important pasture legume, and Sardinia is known as a major centre of diversification of this species. As other legumes, this clover produces biologically active flavonoids including the subclass of isoflavones that are natural phytoestrogens with positive health effects. Present sources of isoflavones for medical/nutraceutical treatments are red clover (Trifolium pratense) and soybean (Glycine max). This study assessed the content and composition of flavonoids in 14 subterranean clover genotypes from Sardinia, grown ex‐situ in comparison with two red clover ecotypes, to acquire information on the potential of the species as an alternative source of isoflavones for possible exploitation. Twenty compounds were tentatively identified across the two clovers after HPLC and LC/ESI‐MS analyses, including clovamide, four flavonols, and 15 isoflavones. Most compounds were present as glucosides or glucosyl malonates. Subterranean clover extracts mainly comprised of derivatives of the isoflavones genistein, biochanin A, and formononetin. Compared to red clover, subterranean clover had higher content of total isoflavones and lower concentration of total flavonols. The isoflavone concentration in subterranean clover was higher than literature data for soybean or red clover. The existing genotypic variation warrants the possibility of selecting varieties with high isoflavone concentration for nutraceutical or pharmaceutical purposes.     

Soil-Behaviour of Phytophthora clandestina

Investigations were undertaken to study the nature and behaviour of P. clandestina in soil. The pathogen was recovered only from soil sievings 250–499 μm and 500 μm–0.99 mm, containing small root fragments. In soil the introduced inoculum of the fungus was incapable of saprophytically and competitively colonizing the dead cotyledons of subterranean clover used as bait material. Exposure of the inoculum to increasing numbers of microbes by adding greater proportions of nonsterile fields, oil to the growth medium of the plant had no significant effect on survival rate and fresh shoot weight of subterranean clover. However, microbes present in the field soil reduced the severity of root rot of subterranean clover. P. clandestina was, able to spread between 15–30 mm through pasteurized soil within a period of 20 days.     

Virus diseases of subterranean clover

Subterranean clover (Trifolium subterraneum) is grown as a pasture legume in several temperate regions of the world where the soils are acidic and infertile, and the rainfall is winter dominant and less than 600 mm annually. It is particularly important in southern Australia where more than 16 million ha have been sown with this species as the pasture legume component. Nine viruses have been recorded infecting subterranean clover in the field. These are alfalfa mosaic, bean yellow mosaic (pea mosaic), beet western yellows, clover yellow vein, cucumber mosaic, pea enation mosaic, soybean dwarf (subterranean clover red leaf), subterranean clover mottle and white clover mosaic. In addition there is an important problem referred to as subterranean clover stunt that was assumed to be caused by a virus but whose aetiology is still unknown. The importance of these diseases is reviewed and details on their epidemiologies are outlined together with details on progress towards their control and some comments on matters worthy of attention in the future. Reference is also made to several exotic viruses known to infect subterranean clover experimentally that could possible cause problems if introduced into Australia.     

Screening Subterranean Clover (Trifolium spp.) Germplasm for Resistance to Meloidogyne species

This study was conducted to identify lines of subterranean clover (Trifolium spp.) with resistance to Meloidogyne arenaria (Neal, 1989) Chitwood, 1949, race 1; M. incognita (Kofoid and White, 1919) Chitwood, 1949, race 3; and M. javanica (Treub, 1885) Chitwood, 1949. A collection of 134 subterranean clover lines was evaluated and all had intermediate to high susceptibility. Root galling was negatively correlated with both seed and dry matter yields. Soil fumigation significantly reduced the nematode population in the field. Results indicate there is limited genetic resistance to root-knot nematodes among subterranean clover lines. Alternative sources of variation for this trait should be investigated.     

Investigating and modeling the combined effects of pH and osmotic pressure on seed germination for use in phytoactivity and allelopathic research

Phytoactivity and allelopathic studies are heavily dependent on germination bioassays of water solutions of allelochemical(s), which necessarily imply that pH and osmotic pressure vary among treatments and between treatments and controls and are therefore a confounding factor in the assessment of seed germination responses to allelochemical(s). When the contribution of pH and osmotic pressure to seed germination responses is considered in experimental designs their effects are almost without exceptions examined separately being assumed, without any evidences, that pH and osmotic pressure act independently on seed germination responses. The objectives of this work were to examine experimentally such assumption using wheat, lettuce, and subterranean clover cultivars to evaluate and model the combined effects on germination of pH and osmotic pressure in the range between 3.0–6.0 and 0–100 mOsmol kg^?1, respectively. Empirical equations are fitted, discussed, and the need to consider the simultaneous effects of pH and osmotic pressure firmly established. Finally, the use of the equations fitted and its impact on conclusions is exemplified in a dose-response bioassay of water extracts of Cistus ladanifer on seed germination using subterranean clover as target species where hormesis was found before allelochemical effects were corrected for pH and osmotic pressure values of control and extracts.     

Number of Rhizobia and delayed inoculation influence nodulation of clovers

The relationship between numbers of rhizobia and nodulation response of legumes is of considerable practical importance. Experiments were done under controlled conditions to determine the influence of numbers of Rhizobium leguminosarum biovar. trifolii on nodulation of arrowleaf clover (Trifolium vesiculosum Savi.) and crimson clover (T. incarnatum L.). Numbers of rhizobia in excess of 1000 per seed did not substantially increase earliness of nodulation or total number of nodules formed on the taproot. Nodules, however, were formed nearer the top of the taproot as numbers of rhizobia increased to 100,000 per seed. Delayed inoculation experiments indicated that nodulation sites for these clovers only remained susceptible to infection for less than 1 day. Delaying inoculation for 4 days resulted in only a 1 to 2 day delay in nodulation for arrowleaf and crimson clovers respectively and no delay for subterranean clover (T. subterraneum L.). Apparently, larger seedlings nodulated faster.     

Determining the effectiveness of grazing and trampling by livestock in transmitting white clover mosaic and subterranean clover mottle viruses

Glasshouse and mini-sward experiments were done to determine the relative roles of grazing and trampling by livestock in transmitting white clover mosaic (WC1MV) and subterranean clover mottle (SCMoV) viruses between clover plants in pastures. Wounding due to grazing was simulated by repeatedly cutting plants with serrated scissors (glasshouse) or mowing (mini-swards), while wounding due to trampling was simulated by repeatedly bashing plants with the flat end of a wooden hammer handle (glasshouse) or rolling (mini-swards). In glasshouse experiments, cutting was more effective than bashing in transmitting WC1MV to white clover (Trifolium repens) plants but cutting and bashing transmitted it to subterranean clover (T. subterraneum) plants at similar rates. In an experiment with white clover mini-swards, mowing was more effective than rolling in transmitting WC1MV, and when both were combined, initially spread exceeded that obtained when the spread from mowing and rolling alone was added together. In glasshouse experiments, bashing was more effective than cutting in transmitting SCMoV to subterranean clover plants. In one experiment, neither mowing nor rolling spread SCMoV in mini-swards of subterranean clover. When transmission to subterranean clover cultivars which were ‘susceptible’ or ‘moderately susceptible’ to SCMoV was compared in glasshouse experiments, repeated bashing spread the virus more slowly to the ‘moderately susceptible’ cultivars. When mixed with ruminant saliva, infective sap containing WC1MV or SCMoV was still infective to clover plants after 4 wk storage at room temperature. When infective sap was allowed to dry naturally on a metal surface, SCMoV still infected clover plants when the dried sap was taken up in tap water after 4 but not 14 days, while WC1MV was infective after 24 h but not 4 days. These results suggest that grazing and mowing are more effective than trampling at transmitting WC1MV to white clover plants in pastures, while trampling is more effective at spreading SCMoV to subterranean clover. However, both transmitted WC1MV to subterranean clover at similar rates. Possible reasons for these differences are discussed in relation to differences in clover plant morphology and virus-specific factors.     

Seed-borne cucumber mosaic virus infection of narrow-leafed lupin (Lupinus angustifolius) in Western Australia

In 1986 in Western Australia, cucumber mosaic virus (CMV) infection was widespread in breeders' selections of narrow-leafed lupin (Lupinus angustifolius), and in collections of lupin cvs and wild L. angustifolius lines. When seed of some of these selections and cvs was sown, seed-borne CMV was detected in seedlings. Infection of F₁ progenies was traced to use of infected parent plants. CMV was also widespread in 25 seed crops of the new lupin cv. Wandoo but not in 42 seed crops of the new cv. Danja. When samples of the seed sown in 1986 were tested, CMV was detected in 3 - 34% of seedlings of cv. Wandoo but in none of cv. Danja. Following intensive roguing of symptom-bearing plants in the 1986 seed crop of new lupin cv. Gungurru, the level of seedling infection with CMV in seed samples after harvest was 0·1-0·2%. CMV was detected in 6 - 8%, 0·6-5% and 0 - 18% of seedlings from seed samples of established lupin cvs Chittick, Yandee and Illyarrie respectively. Highest levels of seed transmission were in seed from crops grown in high rainfall areas. When a sample of cv. Wandoo seed was graded for size by sieving, CMV was detected in seedlings grown from seed in all grades, but the smallest grade contained the highest level of infection. When seed was collected from pods at different positions on plants in a CMV-infected crop of cv. Illyarrie, seed from primary pods transmitted the virus to seedlings at a 3% rate, seed from first order lateral pods at 8% while seed from second and third order lateral pods transmitted at 13%. Examination of CMV-infected lupin crops indicated that seed-infected plants competed poorly and tended to be shaded out in dense crops but to survive in sparse crops. In 1987 during drought conditions after seeding, plant mortality was greater with seed-infected seedlings than with healthy seedlings despite wide plant spacing. An isolate of CMV from subterranean clover (Trifolium subterraneum) induced severer symptoms in lupins than four isolates from lupin; only the subterranean clover isolate prevented seed production. In tests at one lupin breeding site, CMV was found in 15 species of weeds and volunteer legumes. Fumaria officinalis, Stachys arvensis and volunteer lupins were most frequently infected.     

Subterranean clover mottle sobemovirus: its host range, resistance in subterranean clover and transmission through seed and by grazing animals

Subterranean clover mottle sobemovirus (SCMV) was transmitted by manual inoculation of sap to 27 cultivars belonging to three sub-species of subterranean clover. The virus readily infected systemically all inoculated plants of five susceptible cultivars of ssp. subterraneum. Ten others showed partial resistance as not all infected plants developed systemic infection; cold winter conditions further delayed or prevented systemic movement in four of them. Two cultivars of spp. brachycalycinum and four of spp. yanninicum failed to develop systemic infection following inoculation and were considered highly resistant. Resistance to SCMV in three of the spp. yanninicum was further confirmed by the failure to establish detectable primary infections in most of the inoculated leaves. Moreover, when the four ssp. yanninicum cultivars were graft-inoculated with SCMV, systemic infection eventually developed in them but the virus concentration was low. SCMV was also transmitted by manual inoculation of sap to a further 23 species of Trifolium, Medicago or Pisum. Three species were non-hosts, five were infected only in inoculated leaves and 18 others developed systemic infection in some or all plants. SCMV reached very high concentrations and was stable in subterranean clover sap. It was transmitted experimentally between subterranean clover plants by brushing infected leaves against healthy ones and in swards was readily transmitted by the trampling and grazing of sheep, but only poorly by mowing. Seed transmission of SCMV to seedlings of five cultivars of subterranean clover was low (0–0.12%). SCMV was not transmitted by Myzus persicae.     

Annual flower strips support pollinators and potentially enhance red clover seed yield

Ecological intensification provides opportunity to increase agricultural productivity while minimizing negative environmental impacts, by supporting ecosystem services such as crop pollination and biological pest control. For this we need to develop targeted management solutions that provide critical resources to service‐providing organisms at the right time and place. We tested whether annual strips of early flowering phacelia Phacelia tanacetifolia support pollinators and natural enemies of seed weevils Protapion spp., by attracting and offering nectar and pollen before the crop flowers. This was expected to increase yield of red clover Trifolium pratense seed. We monitored insect pollinators, pests, natural enemies and seed yields in a total of 50 clover fields along a landscape heterogeneity gradient, over 2 years and across two regions in southern Sweden. About half of the fields were sown with flower strips of 125–2,000 m². The clover fields were pollinated by 60% bumble bees Bombus spp. and 40% honey bees Apis mellifera. The clover seed yield was negatively associated with weevil density, but was unrelated to bee species richness and density. Flower strips enhanced bumble bees species richness in the clover fields, with the strongest influence in heterogeneous landscapes. There were few detectable differences between crop fields with and without flower strips. However, long‐tongued bumble bees were redistributed toward field interiors and during phacelia bloom honey bees toward field edges. Clover seed yield also increased with increasing size of the flower strip. We conclude that annual flower strips of early flower resources can support bumble bee species richness and, if sufficiently large, possibly also increase crop yields. However, clover seed yield was mainly limited by weevil infestation, which was not influenced by the annual flower strips. A future goal should be to design targeted measures for pest control.     

The first genetic maps for subterranean clover (Trifolium subterraneum L.) and comparative genomics with T. pratense L. and Medicago truncatula Gaertn. to identify new molecular markers for breeding

This study reports on the construction of the first genetic maps of subterranean clover (Trifolium subterraneum L.), a diploid, inbreeding annual pasture legume, and alignment of its linkage groups with those of red clover (T. pratense L.) and Medicago truncatula Gaertn. Transferability of red and white clover (T. repens L.) simple sequence repeat (SSR) markers to subterranean clover was observed. A total of 343 SSR loci were mapped into eight subterranean clover linkage groups, with 6?C31 loci per linkage group and 27 loci with similar locations between two distinct F ₂ mapping populations. Phenotypic data obtained for flowering time, content of three isoflavonoids (formononetin, genistein and biochanin A), hardseededness, leaf markings, calyx pigmentation and hairiness of stems were analyzed, together with genotypic data. Genomic intervals influencing each trait were assigned to one to three chromosome regions, accounting for 5.5?C59.8% of the phenotypic variance. Syntenic relationships were observed among subterranean clover, red clover and Medicago truncatula genomes. Comparisons of loci shared between the three species indicated that at least two chromosomal regions have undergone duplications in the subterranean clover genome. Candidate genes for isoflavone content were identified using M. truncatula as a reference genome. Synteny-based segmentation observed in Brassicaceae chromosomes helped to account for the apparent segmental-based relationship between the clover genomes, particularly within the subterranean clover lines. The proposed segmental nature of clover genome could account for the extensive variation observed between the parental genotypes, while not preventing production of fertile intercrosses.     

Determining the effectiveness of resistance to subterranean clover mottle sobemovirus in different genotypes of subterranean clover in the field using the grazing animal as virus vector

The effectiveness of resistance to subterranean clover mottle sobemovirus (SCMoV) previously identified in different genotypes of subterranean clover (Trifolium subterraneum) inoculated with infective sap in the glasshouse, was tested in two field experiments which used the grazing animal as virus vector. Replicated plots each consisting of paired test rows of 20 different genotypes were used. Clover plants infected with SCMoV were transplanted in between the paired test rows and these acted as sources of the virus for spread by grazing sheep. Although used in different years at different sites with different virus isolates, the field exposure methodology employed produced consistent results. The genotypes each behaved similarly in both experiments as regards the relative extents of SCMoV infection that developed, levels ranging from 0–98%. The previously identified resistance in six ‘highly resistant’ and three ‘partially resistant’ cultivars was effective under field conditions. However, the ‘partial resistance’ in three others was overcome, cvs Green Range and Mt Barker developing levels of infection approaching those in ‘susceptible’ cultivars, while an intermediate infection level developed in cv. Karridale. The three cultivars in which partial resistance was not effective all belonged to ssp. subterraneum. In subterranean clover breeding programmes, field screening using the grazing animal as a vector is advisable to determine whether SCMoV resistance found by sap inoculation is still effective under field conditions.     

Effects of pH,Ca and Al on the exudation from clover seedlings of compounds that induce the expression of nodulation genes inRhizobium trifolii

The expression of nodulation genes inR. trifolii is induced by flavone compounds present in clover root exudates. In the present experiments a bioassay with an indicator strain ofR. trifolii, which contained thelacZ gene fromEscherichia coli fused to theR. trifolii nodA gene, was used to measure the level ofnod gene expression inR. trifolii. Compounds that stimulatednodA gene expression were shown to be present in exudates of white clover (Trifolium repens L.) and nine cultivars of subterranan clover (T. subterraneum L.) seedling sgrown at a range of pH between pH 3.0 and pH 8.0. Thenod gene-induction activity of exudates was, however, reduced when seedlings of all clover species were grown at pH>7.0 and at pH<4.0 and pH<5.0 for white clover and subterranean clover respectively. No major differences were apparent in the activity of exudates from seedlings of the various cultivars of subterranean clover.Nod gene-induction activity of exudates was shown to increase markedly with seedling age. The presence of Ca at concentrations up to 10 mM in seedling culture solutions also resulted in marked increases in thenod gene-induction activity of seedling exudates. Increases in activity due to the presence of Ca were most apparent at low pH where between 5 and 10-fold increases were observed for white clover and subterranean clover respectively. Conversely, the presence of Al at concentrations up to 60 M in seedling culture solutions had no effect on thenod gene-induction activity of seedling exudates.The observations that both low pH and Ca concentrations affected thenod gene-induction activity of seedling exudates suggested that the net presence of stimulatory flavones in root exudates was an important contributing factor to the acid-sensitive step in nodule formation.     

Distribution of potassium in the soil profile of a sandplain soil under pasture species

The influence of pasture species and pasture/crop rotations on the fate of K fertilizer in the soil profile of a sandplain soil was investigated. Results for Lupinus cosentinii, subterranean clover and a subterranean clover/wheat rotation are presented. Potassium was applied as KCl at six rates up to 150 kg K ha^-1 for three years; bicarbonate-extractable K was measured at five depths in the profile (0–100 cm) for four years. The net change in available K in the top 100 cm of the profile (kg ha^-1) was calculated. There was a gradual increase in K down the profile under all species with fertilizer application. The increase was largest for L. cosentinii, which also appeared to redistribute K from below 100 cm to the soil surface. The K residual value on this soil type was higher than expected with most of the fertilizer applied over three years being retained in the top 100 cm.     

Accumulation of a sulphur-rich seed albumin from sunflower in the leaves of transgenic subterranean clover (Trifolium subterraneum L.)

A gene encoding a sulphur-rich, sunflower seed albumin (23% cysteine plus methionine) was modified to contain the promoter for the 35S RNA of cauliflower mosaic virus, in order to obtain leaf expression in transgenic plants. In addition, a sequence encoding an endoplasmic reticulum-retention signal was added to the 3 end of the coding region so as to stabilize the protein by diverting it away from the vacuole. The modified gene was introduced into subterranean clover (T. subterraneum L.) and its expression was detected by northern and western blots and by immunogold localization. The albumin was accumulated in the lumen of the endoplasmic reticulum, and, among six independent, transformed lines, it accumulated in the leaves of T₀ transgenic plants at varying levels up to 0.3% of the total extractable protein. The level of accumulation of the sunflower albumin increased with increasing leaf age, and in the older leaves of the most highly expressing plants of the T₁ generation it reached 1.3% of total extractable protein. Expression of the SSA gene was stable in the first and second generation progeny. These results indicate that there is potential for significantly improving the nutritional value of subterranean clover for ruminant animals such as sheep by expressing genes that code for sulphur-rich, rumen-stable proteins in leaves.     

Improving N2 fixation from the plant down: Compatibility of Trifolium subterraneum L. cultivars with soil rhizobia can influence symbiotic performance

Plant genotypes of Trifolium subterraneum L. (subterranean clover) were evaluated for differences in symbiotic N₂ fixation with soil rhizobia, with the long-term aim of using plant selection to overcome sub-optimal N₂ fixation associated with poorly effective soil rhizobia. Symbiotic performance (SP) was assessed for 49 genotypes of subterranean clover with each of four pure Rhizobium strains isolated from soil. Plants were grown in N free media in the greenhouse and their shoot dry weights measured and expressed as a percentage of dry weight with R. leguminosarm bv. trifolii WSM1325, the recommended commercial inoculant. Average SP with two Rhizobium strains (H and J) ranged from completely ineffective to 80% of potential for the subterranean clover genotypes. Two clover cultivars with high (cv. Campeda) and low (cv. Clare) SP values were investigated in more detail. Campeda typically fixed more N₂ than Clare when inoculated with 30 soil extracts (4.2 vs 2.4 mg N₂ fixed/shoot) and with 14 pure strains isolated from those soils (4.2 vs 2.2 mg N₂ fixed/shoot). The poor performance of Clare could be attributed to interruptions at multiple stages of the symbiotic association, from nodule initiation (less nodules), nodule development (small, white nodules), through to reduced nodule function (N₂ fixed/mg nodule) depending on the inoculation treatment. Through the careful use of subterranean clover genotypes by plant breeders it should be possible to make significant gains in the SP of future subterranean clover cultivars.     

Relationship between Colour, Phenolic Content and Impermeability in the Seed Coat of various Trifolium subterraneum L. genotypes

The relationship between seed colour, phenol content of thetesta and water impermeability in dark and light seeded genotypesof Trifolium subterraneum L. was investigated. The developmentof water impermeability and catechol oxidase activity in expandingseeds of two genotypes was monitored. The results show catecholoxidase activity decreases as seed colour changes from greento purple, but the potential to become impermeable with dehydrationis not acquired until later in seed development. Both waterpermeable and impermeable seeds of dark coloured genotypes containsubstantial amounts of phenol in the lumen of the Malpighiancells; light coloured testas do not. It is concluded that the darkening of the testa of Trifoliumsubterraneum is associated with oxidation of phenol by catecholoxidase but that development of impermeability is independentof this process. Trifolium subterraneum L., subterranean clover, seed coat impermeability, catechol oxidase, phenol, testa colour