The pathogenicity of eight Fusarium species causing fusarium head blight (FHB) in barley was studied under controlled conditions. Six barley genotypes varying in resistance to FHB were artificially inoculated with six isolates each of F. acuminatum, F. avenaceum, F. crookwellense, F. culmorum, F. equiseti, F. graminearum, F. poae and F. sporotrichioides 10-14 d after heading. Symptoms of FHB were rated as disease severity using a 0-9 scale, 4, 7, 14, 21 and 28 d after inoculation, and as percentage of infected spikelets (IS) after 21 d. All species tested caused head blight symptoms on the barley genotypes, but only F. crookwellense, F. culmorum and F. graminearum resulted in severe disease development (> 65% IS) and were considered highly pathogenic. Fusarium avenaceum had 48% IS, which was significantly lower than those of the three highly pathogenic species and was moderately pathogenic. The remaining species had < 15% IS and were weakly pathogenic. There were significant differences (P < 0.05) in aggressiveness among isolates within species and in susceptibility among barley genotypes, suggesting that screening for resistance to FHB requires the use of aggressive isolates or a mixture of several isolates. This is the first report showing that F. crookwellense is highly pathogenic and F. avenaceum is moderately pathogenic on barley.
- Fusarium head blight,
- Fusarium spp.,
- Hordeum vulgare,
Le pouvoir pathogène d'espèces de Fusarium responsables de la fusariose de l'épi chez l'orge
Le pouvoir pathogène de huit espèces de Fusarium responsables de la fusariose de l'épi (FÉ) chez l'orge a été étudié en conditions contrôlées. Dix à quatorze j après l'épiaison, six génotypes d'orge avec divers degrés de résistance à la FÉ ont été inoculés artificiellement avec six isolats de chacune des espèces suivantes : F. acuminatum, F. avenaceum, F. crookwellense, F. culmorum, F. equiseti, F. graminearum, F. poae et F. sporotrichioides. Les symptômes de FÉ ont été notés 4, 7, 14, 21 et 28 j après l'inoculation sur une échelle d'intensité de maladie de 0 à 9; ils ont également été notés après 21 j sur le pourcentage d'épillets infectés (PÉI). Toutes les espèces étudiées ont provoqué des symptômes de fusariose de l'épi sur les génotypes d'orge, mais seulement F. crookwellense, F. culmorum et F. graminearum ont causé un développement marqué de la maladie (PÉI > 65 %) et ont été considérés comme fortement pathogènes. Avec un PÉI de 48 %, qui était significativement inférieur à ceux des trois espèces les plus pathogènes, le Fusarium avenaceum a été considéré comme moyennement pathogène. Les autres espèces ont eu un PÉI de moins de 15 % et ont été considérées comme faiblement pathogènes. Des différences significatives (P < 0,05) ont été observées entre les espèces pour l'agressivité parmi les isolats et pour la sensibilité parmi les génotypes d'orge, ce qui suggère que le tri pour la résistance à la FÉ doit faire appel à des isolats agressifs ou à un mélange de plusieurs isolats. C'est la première fois que le F. crookwellense est signalé comme fortement pathogène et le F. avenaceum comme moyennement pathogène sur l'orge.
- Fusariose de l'épi,
- Fusarium spp.,
- Hordeum vulgare,
- pouvoir pathogène
Corps de l’article
Fusarium head blight (FHB) is a destructive disease of barley (Hordeum vulgare L.) in Canada (Legge et al. 2004; Tekauz et al. 2000). This disease reduces grain yield and quality and causes kernel contamination with certain mycotoxins, such as deoxynivalenol and nivalenol (Campbell et al. 2000), which are harmful to livestock and pose a safety concern in human food (Agriculture and Agri-Food Canada 1990; Charmley et al. 1994; Placinta et al. 1999). Contamination by mycotoxins makes barley unacceptable for malting and brewing (Beattie et al. 1998; Salas et al. 1999).
Fusarium acuminatum Ellis & Everhart, F. avenaceum (Corda: Fr.) Sacc., F. culmorum (W.G. Smith) Sacc., F. equiseti (Corda) Sacc., F. graminearum Schwabe, F. poae (Peck) Wollenw. and F. sporotrichioides Sherb. are isolated frequently from FHB-infected kernels (Abramson et al. 1998; Clear et al. 1996; Gordon 1959; McCallum et al. 2000; Sturz and Johnston 1985). Of these species, F. graminearum is considered the most important causal agent of FHB in Canada (Clear et al. 1996; Tekauz et al. 2000). Studies on the pathogenicity of these Fusarium spp. on wheat (Triticum aestivum L.) revealed that only F. culmorum and F. graminearum were highly pathogenic, F. sporotrichioides was intermediate, and the other species were weakly pathogenic (Stack and McMullen 1985; Stack et al. 1997; Wong et al. 1995). Liddell (1985) reported that F. crookwellense Burgess, Nelson & Toussoun is also highly pathogenic on wheat in Australia, causing more severe crown and root rot than F. culmorum and F. graminearum. Fusarium crookwellense was also isolated from scabby wheat kernels and confirmed to cause typical FHB symptoms on wheat and barley in Japan in 1991 (Sugiura et al. 1994). Xue et al. (2004a) further demonstrated that F. crookwellense is highly pathogenic on wheat, causing severe FHB symptoms under controlled conditions. All species, except F. acuminatum and F. equiseti, have been confirmed to cause FHB symptoms on barley (Perkowski et al. 1995; Salas et al. 1999; Sugiura et al. 1994), but little information is available on the comparative pathogenicity of these species. McCallum and Tekauz (2002) reported that barley differs from wheat with regard to the plant growth stage most susceptible to infection and the profile of the pathogenic species causing FHB in nature. When plants were inoculated with a wild-type and trichodiene synthase gene disrupted F. graminearum, Jansen et al. (2005) demonstrated a new pathway of infection in barley, but not in wheat. Given these differences, the pathogenicity of Fusarium spp. on barley may not be the same as on wheat. A better understanding of the comparative pathogenicity of these Fusarium spp. on barley is important in developing FHB-resistant cultivars in Canada and worldwide. In this study, the pathogenicity of eight Fusarium spp. causing FHB on barley was compared. A preliminary report of this research has been published (Xue et al. 2004b).
Materials and Methods
Fungal isolates and inoculum production
Six isolates from each of eight Fusarium spp. used in this study and their origins are listed in Table 1. These 48 isolates were randomly selected from the Canadian Collection of Fungal Cultures at the Eastern Cereal and Oilseed Research Centre (Ottawa, ON), and have designated DAOM (Department of Agriculture, Ottawa, Mycology) numbers at the Canadian National Mycological Herbarium. The 48 isolates were originally isolated from three hosts (barley, oat and wheat) in five provinces (Alberta, Manitoba, Ontario, Quebec and Saskatchewan) during 1965-2001. A single-spore culture of each isolate was established by transferring single germinated conidia to a modified potato dextrose agar (mPDA, 10 g L-1 of agar, 125 g L-1 of white-skinned and unpeeled potato, and 10 g L-1 of dextrose) amended with 20 ppm streptomycin sulfate and incubated at 22-25°C, under mixed UV and fluorescent lighting, on a 12-h light/dark cycle for 14 d. The mPDA medium with reduced sugar prevents possible mutation and vigor loss of the fungi. The cultures were maintained on the mPDA at 4°C and transferred at 3-mo intervals for a maximum of three times.
Inoculum was prepared as previously described (Xue et al. 2004a): 0.5 mL of a concentrated conidial suspension (approx. 107 spores mL-1) from the single-spore culture was spread over the surface of mPDA in 9-cm Petri dishes and incubated as above for 48 h. Each dish then received 10 mL of sterile distilled water containing 0.01% Tween 20 (polyoxyethylene sorbitan monolaurate) and was scraped gently with a sterile microscope slide to dislodge spores. The resulting spore suspension was filtered through two layers of cheesecloth and adjusted to 5 x 104 spores mL-1 for inoculation.
Six barley genotypes with different levels of resistance to FHB were used. Of these, 'AC Metcalfe', 'CDC Guardian' and 'CI 4196' are 2-row type, and 'Chevron', 'Kasota' and 'Myriam' are 6-row type. 'CI 4196' and 'Chevron' are moderately resistant and represent the highest level of resistance available in 2-row and 6-row barleys, respectively (Bai and Shaner 2004; Legge et al. 2004). 'AC Metcalfe' and 'Myriam' are moderately susceptible, and 'CDC Guardian' and 'Kasota' are susceptible (Butler et al. 2003). Seeds were planted in 15-cm pots containing a mixture of loam soil, sand and composted cow manure (1:1:1, volume ratio), and maintained at 23-25°C during the day and 18-20°C at night in a greenhouse. Supplemental light was provided by 300-W metal halide lamps to ensure a 16 h photoperiod and a minimum intensity of 350 μmol•m-2•ms-1. Following emergence, plants were thinned to three per pot and fertilized with a 1% solution of 20-20-20 (N-P-K) once a wk starting 5 wk after planting. Because of genotypic differences in maturity, seeds were planted serially over a 2-wk period, so that all genotypes could be selected from the same growth stage when inoculated.
The six barley genotypes were inoculated with each of the 48 isolates 10-14 d after heading. Prior to inoculation, a maximum of 12 spikes per pot were randomly selected, while the remainder and those from lateral tillers were removed. Plants were sprayed with the spore suspension at 0.2 mL per spike using a DeVilbiss model 15 atomizer (The DeVilbiss Co., Somerset, PA). After the inoculum dried for 30 min, plants were transferred to a polyethylene humidity chamber in a growth chamber for 48 h. The growth chamber was operated at 25°C with a 12-h photoperiod at a light intensity of 250 μmol•m-2•ms-1. The humidity chamber was maintained at or near 100% RH by the continuous operation of two ultrasonic humidifiers. Air temperature and humidity in the chamber were monitored with a portable datalogger (model 21XL micrologger, Campbell Scientific Canada Corp., Edmonton, AB). After incubation, plants were returned to the greenhouse bench. For each isolate and genotype combination, four replicate pots per genotype were used. Pots were arranged in a completely randomized design in both the humidity chamber and the greenhouse after the inoculation. Four pots of 'Kasota' sprayed with sterile distilled water plus the surfactant, and four pots sprayed with a mixture of spore suspensions of three aggressive F. graminearum isolates (Grwon34, Grwon36 and Grwon40) were included as checks. This could help to detect any extraneous airborne inocula in the growth room and ensure suitability of the inoculum and the environment for infection.
Disease assessment and statistical analyses
Symptoms of FHB were rated as disease severity at 4, 7, 14, 21 and 28 d after inoculation and as percentage of infected spikelets (IS) after 21 d, when plants were at the soft dough stage. Disease severity was estimated visually in situ for each inoculated spike on a 0 (no visible FHB symptoms) to 9 (severely diseased, spike dead) scale described by Xue et al. (2004a). Disease severities and percentages of IS from all plants in each pot were averaged and the means per pot of percent IS were used in the statistical analysis. Analysis of variance was conducted as a two fixed factor (species and cultivar) nested design with isolates nested within species. An angular transformation of percent IS was used in the analysis of variance to stabilize variances (Snedecor and Cochran 1980). Treatment means of the detransformed data to the original scale have been presented and were separated by Fisher's least significant difference (LSD) test at a probability level of P ≤ 0.05, based on the analyses of transformed data. Analyses were performed using Proc GLM in SAS/STAT® (SAS Institute Inc., Cary, NC).
The eight Fusarium spp. were different in the rate of FHB symptom development on the six barley genotypes (Fig. 1). Symptoms usually appeared, on susceptible genotypes, 3 d after inoculation with isolates of F. crookwellense, F. culmorum and F. graminearum, and after 7-21 d after inoculation with isolates of the remaining species. The most rapid and severe disease development on these genotypes was observed for F. crookwellense, F. culmorum and F. graminearum, followed by F. avenaceum, which caused minimal infection on 'Chevron', but was as highly pathogenic as the above three species on the other genotypes. Disease progressed slowly and lower severities were generally observed for F. acuminatum, F. equiseti, F. poae and F. sporotrichioides. Disease reached maximum severity 28 d after inoculation, when plants were at or near maturity.
Significant differences (P < 0.05) were observed in percentages of IS among Fusarium spp., among isolates within each species, and among barley genotypes (Table 2). There were also significant Fusarium spp. x genotype interaction and interactions of genotype x isolate for F. acuminatum, F. equiseti, F. graminearum and F. sporotrichioides. On average of the six isolates of the eight Fusarium spp., F. culmorum had the greatest IS (82%), followed by F. graminearum (68%) and F. crookwellense (65%) (Table 3). These three species were considered highly pathogenic. Fusarium avenaceum resulted in 48% IS, which was significantly lower than those of the three highly pathogenic species, and therefore was moderately pathogenic. The remaining species had < 15% IS and were weakly pathogenic.
Of the six barley genotypes, the two moderately resistant genotypes 'CI 4196' and 'Chevron' had relatively low levels of IS when challenged with most of the eight Fusarium spp. (Table 3). 'CI 4196' was significantly better than all other genotypes in reaction to F. crookwellense and F. sporotrichioides while 'Chevron' was better than all others in reaction to F. avenaceum. In addition, the two genotypes differed significantly in their reactions to F. graminearum, where 'CI 4196' was less susceptible. Between the two moderately susceptible genotypes, 'AC Metcalfe' was more susceptible to F. acuminatum and F. avenaceum than 'Myriam'. 'AC Metcalfe' and 'Myriam' were otherwise similar in their responses to the other six Fusarium spp. It is worth mentioning that both 'AC Metcalfe' and 'Myriam' were significantly less susceptible than 'Chevron' to F. graminearum. The two susceptible genotypes 'CDC Guardian' and 'Kasota' had relatively high levels of IS to most of the eight Fusarium spp. However, 'CDC Guardian' was less susceptible to F. acuminatum and F. poae, but more susceptible to F. equiseti than 'Kasota'.
Neither the variation in pathogenicity of the eight Fusarium spp. nor the pathogenicity of F. acuminatum and F. equiseti on barley has been previously investigated. In inoculation tests on wheat, Stack and McMullen (1985) and Wong et al. (1995) reported that only F. culmorum and F. graminearum were recognized to be highly pathogenic, among nine and seven Fusarium spp. examined, respectively. Recent studies by Xue et al. (2004a) further demonstrated that F. crookwellense, which was not included in the previous studies, was among the most pathogenic species, similar to F. culmorum and F. graminearum in causing FHB on wheat. Fusarium avenaceum is weakly pathogenic on wheat (Stack and McMullen 1985; Wong et al. 1995; Xue et al. 2004a), although it has been isolated frequently from both wheat and barley (Abramson et al. 1998; Clear et al. 1996; Gilbert et al. 1995; Martin et al. 1991; Salas et al. 1999; Sturz and Johnston 1985). In this study, F. avenaceum was weakly pathogenic on 'Chevron' but moderately to highly pathogenic on the other genotypes (Fig. 1), suggesting that this species can also be a potentially important FHB causal agent. Fusarium crookwellense was reported to cause crown, foot and root rot on wheat in Australia (Liddell 1985) and FHB symptoms of wheat and barley in Japan (Sugiura et al. 1994). This research demonstrated that F. crookwellense is highly pathogenic and can potentially be an important causal agent of FHB on barley in areas where the pathogen is present.
Of the four highly and moderately pathogenic species, the percentage of IS on the six barley genotypes varied between 75-93% for isolates of F. culmorum, 52-81% for F. graminearum, 50-76% for F. crookwellense, and 34-56% for F. avenaceum (Table 1). The differences in aggressiveness among isolates within each of these species were significant (Table 2). The presence of different levels of aggressiveness among isolates has practical implications that must be considered when screening barley for FHB resistance. It is important that isolates with a known level of aggressiveness be used in variety evaluation trials, so that valid comparisons can be made between genotypes.
A genotype x isolate interaction was observed for F. acuminatum, F. culmorum, F. equiseti, F. graminearum and F. sporotrichioides (Table 2). However, the effect of genotype x isolate interactions contributed only to < 1% of the total variance, which was too low to differentiate any possible races among the isolates tested. These results are in agreement with Takeda et al. (1995) who tested 12 isolates of F. graminearum on 10 varieties each of wheat and barley in Japan and found that the genotype x isolate interaction, although significant, was very small in comparison to the variation of the host resistance and pathogenicity. Similarly, Bai and Shaner (1996) and Xue et al. (2004a) found no strong evidence for the existence of pathogenic variation in F. graminearum on wheat in the United States and in Canada, respectively.
Although the differences in reaction to the eight Fusarium spp. were generally similar to field observations of these genotypes in FHB resistance, there were significant genotype x species interactions observed in the present study (Table 3). 'Chevron', for instance, is commonly known as a moderately resistant genotype (Bai and Shaner 2004; Butler et al. 2003; Legge et al. 2004), but here it was significantly more susceptible than 'CI 4196', 'AC Metcalfe' and 'Myriam' in reaction to F. graminearum, the predominating species causing FHB in Canada. Similarly, 'AC Metcalfe' was less susceptible than 'CDC Guardian' and 'Kasota' in a field evaluation (Butler et al. 2003), but in the present study it was significantly more susceptible than all other genotypes to F. acuminatum. To our knowledge, the significant barley genotype x Fusarium species interaction in FHB etiology has not been previously reported. The genotype x species interaction was more apparent on the two most resistant genotypes, 'CI 4196' and 'Chevron' (Table 3). The results indicate that these two barley genotypes may each possess different genes for resistance to the respective Fusarium species. Further research is needed to confirm the presence and heritability of these resistance genes to the different Fusarium spp. and their usefulness in future cultivar development.
We thank F. Sabo, Y. Chen, M. Kuc and T. Potter for their technical assistance, and J. McCarthy for her assistance in preparing the graphics.
- Abramson, D., R.M. Clear, E. Usleber, R. Gessler, T.W. Nowicki, and E. Martlbauer. 1998. Fusarium species and 8-ketotrichothecene mycotoxins in Manitoba barley. Cereal Chem. 75 : 137-141.
- Agriculture and Agri-Food Canada. 1990. Mycotoxins in livestock feeds. Communications Branch, Agriculture and Agri-Food Canada, Ottawa, ON, Trade Memorandum T-3-116.
- Bai, G.-H., and G. Shaner. 1996. Variation in Fusarium graminearum and cultivar resistance to wheat scab. Plant Dis. 80 : 975-979.
- Bai, G.-H., and G. Shaner. 2004. Management and resistance in wheat and barley to fusarium head blight. Annu. Rev. Phytopathol. 42 : 135-161.
- Beattie, S., P.B. Schwarz. R. Horsley, J. Barr, and H.H. Casper. 1998. The effect of grain storage conditions on the viability of Fusarium and deoxynivalenol production in infested malting barley. J. Food Prot. 61 : 103-106.
- Butler, G., K.M. Ho, M.J. Morrison, A.G. Xue, W.G. Legge, S. Rioux, R.A. Martin, Q. Shen, J. Yang, J.R. Tucker, M.E. Savard, J.L. Gale, M. Kuc, and C. Danjou. 2003. Collaborative study of barley fusarium head blight nurseries. Page 131 in Proceedings of the 3rd Canadian Workshop on Fusarium Head Blight. Winnipeg, MB, Canada, December 2003.
- Campbell, H., T.M. Choo, B. Vigier, and L. Underhill. 2000. Mycotoxins in barley and oat samples from Eastern Canada. Can. J. Plant Sci. 80 : 977-980.
- Charmley, L.L., A. Rosenberg, and H.L. Trenholm. 1994. Factors responsible for economic losses due to Fusarium mycotoxin contamination of grains, foods and feedstuffs. Page 471-486 in J.D Miller and H.L. Trenholm (eds.), Mycotoxins in grain compounds other than aflatoxin. Eagan Press, St. Paul, MN, USA.
- Clear, R.M., S.K. Patrick, R.G. Platford, and M. Desjardins. 1996. Occurrence and distribution of Fusarium species in barley and oat seed from Manitoba in 1993 and 1994. Can. J. Plant Pathol. 18 : 409-414.
- Gilbert, J., A. Tekauz, R. Kaethler, E. Mueller, and U. Kromer. 1995. Occurrence of fusarium head blight in Manitoba in 1994. Can. Plant Dis. Surv. 75 : 124-125.
- Gordon, W.L. 1959. The occurrence of Fusarium species in Canada. VI. Taxonomy and geographic distribution of Fusarium species on plants, insects, and fungi. Can. J. Bot. 37 : 257-290.
- Jansen, C., D.V. Wettstein, W. Schäafer, K.-H. Kogel, A. Felk, and F.J. Maier. 2005. Infection patterns in barley and wheat spikes inoculated with wild-type and trichodiene synthase gene disrupted Fusarium graminearum. Proc. Natl. Acad. Sci. USA 102 : 16892-16897.
- Legge, W.G., M.C. Therrien, J.R. Tucker, M. Banik, A. Tekauz, D. Sommers, M.E. Savard, B.G. Rossnagel, E. Lefol, D. Voth, T. Zatorski, B.L. Harvey, and G. Scoles. 2004. Progress in breeding for resistance to fusarium head blight in barley. Can. J. Plant Pathol. 26 : 436-442.
- Liddell, C.M. 1985. The comparative pathogenicity of Fusarium graminearum Group 1, Fusarium culmorum and Fusarium crookwellense as crown, foot and root rot pathogens of wheat. Australas. Plant Pathol. 14 (2) : 29-32.
- Martin, R.A., J.A. MacLeod, and C. Caldwell. 1991. Influence of production inputs on incidence of infection by Fusarium species on cereal seed. Plant Dis. 75 : 784-788.
- McCallum, B.D., and A. Tekauz. 2002. Influence of inoculation method and growth stage on fusarium head blight in barley. Can. J. Plant Pathol. 24 : 77-80.
- McCallum, B., A. Tekauz, J. Gilbert, J. Gold, M. Idris, E. Mueller, R. Kaethler, M. Stulzer, and U. Kromer. 2000. Fusarium head blight of barley in Manitoba in 1999. Can. Plant Dis. Surv. 80 : 36.
- Perkowski, J., I. Kiecana, and J. Chelkowski. 1995. Susceptibility of barley cultivars and lines to Fusarium infection and mycotoxin accumulation in kernels. J. Phytopathol. 143 : 547-551.
- Placinta, C.M., J.P.F. D'Mello, and A.M.C. Macdonald. 1999. A review of worldwide contamination of cereal grains and animal feed with Fusarium mycotoxins. Anim. Feed Sci. Technol. 78 : 21-37.
- Salas, B., B.J. Steffenson, H.H. Casper, B. Tacke, L.K. Prom, T.G.Jr. Fetch, and P.B. Schwarz. 1999.Fusarium species pathogenic to barley and their associated mycotoxins. Plant Dis. 83 : 667-674.
- Snedecor, G.W., and W.G. Cochran. 1980. Statistical Methods. 8th edition. The Iowa State University Press, Ames, IA, USA. 503 pp.
- Stack, R.W., and M.P. McMullen. 1985. Head blighting potential of Fusarium species associated with spring wheat heads. Can. J. Plant Pathol. 7 : 79-82.
- Stack, R.W., R.C. Frogberg, and H. Casper. 1997. Reaction of spring wheats incorporating Sumai-3 derived resistance to inoculation with seven Fusarium species. Cereal Res. Commun. 25 : 667-671.
- Sturz, A.V., and H.W. Johnson. 1985. Characterization of Fusarium colonization of spring barley and wheat produced on stubble or fallow soil. Can. J. Plant Pathol. 7 : 720-726.
- Sugiura, Y., H. Saito, T. Tanaka, M. Ichinoe, and Y. Ueno. 1994.Fusarium crookwellense, a newly isolated fungus from wheat in Japan: Its mycotoxin production and pathogenicity to wheat and barley. Mycoscience 35 : 77-82.
- Takeda, K, J. Wu, and H. Heta. 1995. Analysis of host-pathogen interaction in fusarium head blight of wheat and barley. Breed. Sci. 45 : 349-356.
- Tekauz, A., B.D. McCallum, and J. Gilbert. 2000. Fusarium head blight of barley in western Canada: a review. Can. J. Plant Pathol. 22 : 9-16.
- Wong, L.S.L., D. Abramson, A. Tekauz, D. Leisle, and R.I.H. McKenzie. 1995. Pathogenicity and mycotoxin production of Fusarium species causing head blight in wheat cultivars varying in resistance. Can. J. Plant Sci. 75 : 261-267.
- Xue, A.G., K.C. Armstrong, H.D. Voldeng, G. Fedak, and C. Babcock. 2004a. Comparative aggressiveness of isolates of Fusarium species causing head blight on wheat in Canada. Can. J. Plant Pathol. 26 : 81-88.
- Xue, A.G., K.M. Ho, C. Babcock, Y. Chen, F. Sabo, and M. Kuc. 2004b. Pathogenicity of Fusarium species causing head blight in barley. Can. J. Plant Pathol. 26 : 429 (abstract).