Effects of transgenic corn and Cry1Ab protein on the nematode, Caenorhabditis elegans

The effects of the insecticidal Cry1Ab protein from Bacillus thuringiensis (Bt) on the nematode, Caenorhabditis elegans, were studied with soil from experimental fields cultivated with transgenic Bt corn (MON810) and with trypsinized Cry1Ab protein expressed in Escherichia coli. The content of Cry1Ab protein was above the detection limit of an ELISA test in only half of the soil samples obtained from transgenic plots, ranging from 0.19 to 1.31 ng g⁻¹ dry weight. In a laboratory bioassay, C. elegans was exposed to rhizosphere and bulk soil from fields with isogenic or transgenic corn or to solutions of Cry1Ab protein (0, 24, 41, 63, 118, and 200 mg l⁻¹) over a period of 96 h, with growth and reproduction serving as the test parameters. Nematode reproduction and growth were significantly reduced in rhizosphere and bulk soil of Bt corn compared with soil from isogenic corn and were significantly correlated with concentrations of the Cry1Ab protein in the soil samples. Moreover, the toxicity of pure Cry1Ab protein to the reproduction and growth of C. elegans was concentration-dependent. As significant inhibition occurred at relatively high concentrations of the Cry1Ab protein (41 mg l⁻¹), the effects of the soil samples from Bt corn could not be assigned directly to the toxicity of the Cry1Ab protein. The results demonstrate that bioassays with the nematode, C. elegans, provide a promising tool for monitoring the potential effects of Bt toxins in aqueous medium and soils.     

Using Transgenic Caenorhabditis elegans in Soil Toxicity Testing

Soil bioassays are important tools for evaluating toxicological effects within the terrestrial environment. The American Society for Testing and Materials E2172-01 Standard Guide outlines a method for conducting laboratory soil toxicity tests using the nematode Caenorhabditis elegans. This method is an efficient tool for extracting C. elegans from soil samples and can be carried out after a 24-h exposure period using relatively small amounts of soil. Drawbacks of this method include problems with (1) recovery of nematodes from soils containing a high percentage of organic matter, and (2) distinguishing indigenous nematode species from nematodes added for the laboratory test. Due in part to these issues, C. elegans has not been extensively accepted for use in soil testing. To address these concerns and improve upon the American Society for Testing and Materials method, this project focused on using transgenic strains of C. elegans carrying a GFP-expressing element. Lethality and behavior tests revealed that the transgenic nematodes respond similarly to the wild-type N2 strain, indicating that they can be used in the same manner in soil testing. The GFP marker is easily identifiable not only within soils containing a large amount of organic matter, but also in field-collected soils containing indigenous nematodes. These results support the use of transgenic GFP C. elegans in soil bioassays as a tool to further the reliability of laboratory toxicity tests.     

A soil toxicity test using the nematode Caenorhabditis elegans and an effective method of recovery

A new method for recovering nematodes from soils in an efficient, reproducible, and non-destructive manner has been developed. It was used to conduct short-term soil toxicity tests using the soil-dwelling nematode Caenorhabditis elegans and several different soil types spiked with copper chloride. The recovery method, which involves centrifugation through a colloidal silica suspension, allows the nematodes to be extracted from the soil matrix so that lethality can be assessed. The nematodes are unharmed by the recovery procedure, and both live and dead individuals are recovered with high efficiency (well over 80%), allowing reproducible concentration-response curves to be made after a 24-h exposure. The LC₅₀s for copper were increased about tenfold by the presence of soil, and different soils had significantly different effects on toxicity. Toxicity of copper ion was also influenced by the concentration of sodium chloride and potassium chloride in the test solution, and the presence of bacteria increased the toxicity of copper ion in some soils. The LC₅₀s in soil were close to the LC₅₀ for the 2-week earthworm soil toxicity test, suggesting that a 24-h nematode toxicity test may be comparable to the 2-week earthworm test in terms of sensitivity.     

Toxicity assessment of sediments from three European river basins using a sediment contact test battery

The toxicity of four polluted sediments and their corresponding reference sediments from three European river basins were investigated using a battery of six sediment contact tests representing three different trophic levels. The tests included were chronic tests with the oligochaete Lumbriculus variegatus, the nematode Caenorhabditis elegans and the mudsnail Potamopyrgus antipodarum, a sub-chronic test with the midge Chironomus riparius, an early life stage test with the zebra fish Danio rerio, and an acute test with the luminescent bacterium Vibrio fischeri. The endpoints, namely survival, growth, reproduction, embryo development and light inhibition, differed between tests. The measured effects were compared to sediment contamination translated into toxic units (TU) on the basis of acute toxicity to Daphnia magna and Pimephales promelas, and multi-substance Potentially Affected Fractions of species (msPAF) as an estimate for expected community effects. The test battery could clearly detect toxicity of the polluted sediments with test-specific responses to the different sediments. The msPAF and TU-based toxicity estimations confirmed the results of the biotests by predicting a higher toxic risk for the polluted sediments compared to the corresponding reference sediments, but partly having a different emphasis from the biotests. The results demonstrate differences in the sensitivities of species and emphasize the need for data on multiple species, when estimating the effects of sediment pollution on the benthic community.     

Toxic Effects of Acetochlor on Mortality, Reproduction and Growth of Caenorhabditis elegans and Pristionchus pacificus

The effects of acetochlor on the mortality, growth and reproduction of two nematode species were assessed. The LC₅₀ values for Caenorhabditis elegans and Pristionchus pacificus were 1,296 and 210.7 mg/L at 24 h, and 540.0 and 126.4 mg/L at 48 h exposure, respectively. In three succession generations, reproductive capacity was more sensitive in P. pacificus than in C. elegans. Moreover, the sublethal test endpoint of final length was more sensitive with P. pacificus. This study suggested that acetochlor had no long-term effects on C. elegans at lower concentrations. The higher concentrations of acetochlor (from 40 to 160 mg/L) revealed sublethal toxicity to the two tested species, with P. pacificus being more sensitive than C. elegans.     

Unexpected reduction in reproduction of Collembola exposed to an arsenic-contaminated soil

To assess the problems posed by ecological testing of the toxicity of polluted soils, we applied the Collembola reproduction test to the soil from a mining site polluted by metals. Our aim was to quantify the chronic toxicity of contaminated soil using the Folsomia candida reproduction test. Briefly, the polluted soil, which contains in particular large quantities of arsenic, is mixed in various proportions with a reference soil, and the number of juveniles is counted five weeks later. The results were quite unexpected in that a stronger effect was observed on reproduction with dilute contaminated soil, 0.78 to 3.1%, than with much higher concentrations. The pure contaminated soil is not highly toxic. Arsenic (As) dosage in pore water shows that a change in its bioavailability occurs that may explain the surprising toxicity results. As the main contaminant of the soil was arsenic, we showed that the lowest-observed-effect concentration and the median effective concentration of this metal for the reproduction of F. candida in International Organization for Standardization (Geneva, Switzerland) soil experimentally contaminated with arsenic were low (2.22 and 2.19 microg/g, respectively). Likewise, we showed that pH modifies the bioavailability and the toxicity of As, but we do not think that this parameter explains the test results. These results show that this bioassay should be used cautiously to test toxicity of polluted soils.     

Evaluation of the Leaching Potential of Anthranilamide Insecticides Through the Soil

The soil nematode Caenorhabditis elegans was used in 24-h acute exposures to arsenic (As), copper (Cu) and glyphosate (GPS) and to mixtures of As/Cu and As/GPS to investigate the effects of mixture exposures in the worms. A synergistic type of interaction was observed for acute toxicity with the As/Cu and As/GPS mixtures. Sublethal 24-h exposures of 1/1000, 1/100 and 1/10 of the LC50 concentrations for As, Cu and GPS individually and for As/Cu and As/GPS mixtures were conducted to observe responses in locomotory behavior (head thrashing), reproduction, and heat shock protein expression. Head thrash frequency and reproduction exhibited concentration dependent decreases in both individual and combined exposures to the tested chemical stressors, and showed synergistic interactions even at micromolar concentrations. Furthermore, the HSP70 protein level was significantly increased following exposure to individual and combined chemical stressors in wild-type C. elegans. Our findings establish for the first time the effects of exposure to As/GPS and As/Cu mixtures in C. elegans.     

Long-term responses of snails exposed to cadmium-contaminated soils in a partial life-cycle experiment

Juvenile snails were exposed during their growth period to Cd-contaminated field and artificial soils and then transferred to uncontaminated soil to assess the sequels of previous exposure on adult reproduction. Growth modelling highlighted growth inhibitions of 5% and 10% after 70 and 84 days of exposure to 20 and 100 mg Cd kg(-1) in artificial soils, respectively. Growth disruption was accompanied by a decrease in the clutch number and a 4-week delay in the egg-laying cycle. Although it was also contaminated at 20 mg Cd kg(-1), the contaminated field soil did not lead to detectable effects in snails, suggesting a lower Cd bioavailability confirmed by the bioaccumulation analysis. We demonstrated that the 28-day growth test, as advised by the ISO-guideline, may not be sufficient to assess sublethal toxic effects of realistically contaminated soils. For this purpose, a life cycle experimental set-up is proposed, allowing a thorough assessment of toxicity during successive life stages.     

The Enchytraeid Reproduction Test (ERT): A Potentially Quick and Affordable Tool for the Assessment of Metal Contaminated Soils in Emerging Economies

The enchytraeid reproduction test (ERT) was used to assess the ecotoxicity of selected mine tailings and agricultural soils from South Africa. The mine tailings had higher cumulative metal concentrations than agricultural soils. The most contaminated mine tailings significantly reduced the survival of the oligochaete Enchytraeus doerjesi whose reproduction was suppressed in all mine waste substrates. Because it reliably singled out the most contaminated substrate and was found easy to perform, we suggest that the ERT could be a quick and affordable tool for assigning intervention values for soil remediation in emerging economies such as South Africa.     

A semi-fluid gellan gum medium improves nematode toxicity testing

This study examined an alternative test medium for nematodes that use gellan gum as the gelling agent instead of agar. The semi-fluid consistency of the gel-like component nematode growth gellan gum (CNGG) supports three-dimensional distribution of the nematodes and food bacteria, but still allows free movement of the former. Moreover, flexible preparation of the medium and easy recovery of the test organisms are possible. Here, the effects of the nematicides ivermectin (pharmaceutical) and aldicarb (pesticide) and of the metal cadmium on the growth and reproduction of the free-living nematodes Caenorhabditis elegans and Panagrolaimus cf. thienemanni were studied in CNGG media. Results were compared to those obtained with the standard liquid test media in order to evaluate the applicability of CNGG for nematode toxicity testing. The sensitivity of P. cf. thienemanni to all three substances was found to be higher than that of C. elegans, but both nematodes showed the highest sensitivity to ivermectin exposure. This raises concerns about the risk posed by the pharmaceutical to non-target nematodes. In contrast to ivermectin bioassays carried out in CNGG medium, those conducted in liquid medium resulted in wide-ranging variability between and within replicates. Thus, CNGG seems to be particularly valuable for testing hydrophobic substances with a high sorption affinity as it favors their sorption to food bacteria and minimizes contact with the surfaces of the test vessels. However, the medium was less suitable for deriving toxicity thresholds for cadmium and may likewise not be an appropriate choice for testing other metals. The medium introduced herein was shown to be appropriate for sublethal nematode toxicity testing and likely provides a convenient environment for testing other nematode species. Besides improved testing of hydrophobic substances, CNGG also offers advantages for long-term studies, such as full life-cycle experiments, in which fresh medium is regularly needed. Moreover it may be beneficial for testing other poorly soluble or insoluble substances, such as nanoparticles.     

Interlaboratory comparison of a standardized toxicity test using the nematode Caenorhabditis elegans (ISO 10872)

A ring test was carried out within the standardization process of ISO 10872 to evaluate the precision of the toxicity test for the nematode Caenorhabditis elegans. Eight different laboratories tested aqueous solutions of the reference substance benzylcetyldimethylammonium chloride as well as native sediments and soils for toxic effects on the growth and reproduction of C. elegans. Validity criteria were met in all laboratories. Average median- and low-effect concentrations were determined to be 15.1 mg L(-1) (EC50) and 8.7 mg L(-1) (EC10) for growth and 7.5 mg L(-1) (EC50) and 3.8 mg L(-1) (EC10) for reproduction of C. elegans, with ECx values showing a high degree of reproducibility (CV(R) : <21% and <11% for EC10 and EC50, respectively) and repeatability (CV(r) : <20% and <7% for EC10 and EC50, respectively). The toxic effects of the sediments and soils revealed by the different laboratories were well related to each samples' degree of chemical contamination. Moreover, the effects showed an acceptable reproducibility (CV(R) : 5-33% and 0-28% for growth and reproduction, respectively) and repeatability (CV(r) : 3-13% and 0-12% for growth and reproduction, respectively). The present study confirms that the toxicity test with C. elegans according to ISO 10872 is a reliable and precise tool to assess the toxicity of aqueous media, freshwater sediments, and soils.     

Effect of Single and Paired Metal Inputs in Soil on a Stress-Inducible Transgenic Nematode

A toxicity test using a transgenic strain of the free-living soil nematode Caenorhabditis elegans carrying a stress-inducible β-galactosidase reporter has been adapted for use in soil biomonitoring. High concentrations (250 μg · g⁻¹) of cadmium are required to induce the stress response in worms exposed to Lufa 2.2 soil. Even at relatively high concentrations, the response to copper and zinc additions alone is minimal, yet combinations of cadmium and copper in the test soil induce a larger response than with cadmium alone at the same concentration. In contrast, the addition of both zinc and cadmium induces a lower response than cadmium additions alone. Analysis of the interstitial water suggests that there is preferential occupation by copper of sorption sites in the soil, allowing more cadmium to remain in solution. Conversely, cadmium and zinc would appear to interact similarly with the soil constituents, resulting in an increase of both metals in solution with increased additions to the soil. Aquatic tests mimic the results of the soil test, so it is not increased cadmium availability alone that causes an increased stress response when both cadmium and copper are present. The presence of other metals could reduce the amount of cadmium available, which may be one factor in the zinc moderation of the stress response to cadmium. Intracellular mechanisms may also contribute to the copper enhancement of the stress response to cadmium. Received: 30 November 1998/Accepted: 12 May 1999     

Influence of aging on copper bioavailability in soils

Because of long-term chemical processes, metal bioavailability in field soils decreases with time. Metal toxicity may, therefore, be overestimated if toxicity data with freshly spiked soils are used to derive soil quality criteria, a current practice. In the present study, effects of the long-term processes, called aging, on copper partitioning and ecotoxicity are investigated. Twenty-five field soils contaminated by copper runoff from bronze statues and 25 uncontaminated control soils sampled at 5-m distance from these statues were collected in Flanders (Belgium). The soils were selected so that parameters affecting copper bioavailability (pH, cation-exchange capacity, organic matter content, etc.) varied considerably. To assess the effect of aging on copper toxicity, control soils were spiked at total copper concentrations comparable to those of historically contaminated soils. Pore-water copper concentrations and 0.01 M CaCl2-extracted copper concentrations were significantly higher in freshly spiked soils compared to contaminated field soils. However, this could be a pH effect, because pH decreased after spiking. Acute toxicity to Enchytraeus albidus (14 d) as well as chronic toxicity to Folsomia candida (28-d reproduction) and Trifolium pratense (14-d growth) indicated a dose-response relationship between copper toxicity and pore-water copper concentration or the CaCl2-extracted copper fraction.     

Acanthopanax sessiliflorus stem confers increased resistance to environmental stresses and lifespan extension in Caenorhabditis elegans

BACKGROUND/OBJECTIVES

Acanthopanax sessiliflorus is a native Korean plant and used as traditional medicine or an ingredient in many Korean foods. The free radical theory of aging suggests that cellular oxidative stress caused by free radicals is the main cause of aging. Free radicals can be removed by cellular anti-oxidants.

MATERIALS/METHODS

Here, we examined the anti-oxidant activity of Acanthopanax sessiliflorus extract both in vitro and in vivo. Survival of nematode C. elegans under stress conditions was also compared between control and Acanthopanax sessiliflorus extract-treated groups. Then, anti-aging effect of Acanthopanax sessiliflorus extract was monitored in C. elegans.

RESULTS

Stem extract significantly reduced oxidative DNA damage in lymphocyte, which was not observed by leaves or root extract. Survival of C. elegans under oxidative-stress conditions was significantly enhanced by Acanthopanax sessiliflorus stem extract. In addition, Acanthopanax sessiliflorus stem increased resistance to other environmental stresses, including heat shock and ultraviolet irradiation. Treatment with Acanthopanax sessiliflorus stem extract significantly extended both mean and maximum lifespan in C. elegans. However, fertility was not affected by Acanthopanax sessiliflorus stem.

CONCLUSION

Different parts of Acanthopanax sessiliflorus have different bioactivities and stem extract have strong anti-oxidant activity in both rat lymphocytes and C. elegans, and conferred a longevity phenotype without reduced reproduction in C. elegans, which provides conclusive evidence to support the free radical theory of aging.     

Discrepancy of the microbial response to elevated copper between freshly spiked and long-term contaminated soils

A systematic comparison of Cu toxicity thresholds was made between freshly spiked soils and soils in which elevated Cu concentrations have been present for various times. Three uncontaminated soils were spiked and experimentally leached or incubated outdoors for up to 18 months. Additionally, five field-contaminated soils with a 6- to 80-year-old Cu contamination were sampled, and corresponding uncontaminated soils were spiked to identical total concentrations. All soil samples were subjected to three microbial assays (nitrification potential, glucose-induced respiration, and maize residue C-mineralization). Experimental leaching or soil incubation after spiking reduced Cu toxicity (1.3- or 2.3-fold increase of dose, respectively, to inhibit process by 50% [ED50]). No significant effects of soil type, aging time (6, 12, or 18 months), or bioassay on the factor change of ED50 were found. Significant reductions of microbial activity in field-contaminated soils were only identified in 2 of the 15 series (three assays in five soils), whereas freshly spiking the corresponding control soils significantly affected these processes in 12 series. Soil solution Cu concentrations significantly decreased on leaching at corresponding total soil Cu, and smaller decreases were found during additional aging. Soil solution Cu concentrations largely explain changes in Cu toxicity on leaching and aging, although additional variation may be related to changes in the sensitivity of microbial populations. It is concluded that total Cu toxicity thresholds are lower in freshly spiked soils compared to soils in which Cu salts have equilibrated and leaching has removed excess soluble salts. The large variability of soil microbial processes creates a large uncertainty about the magnitude of the factor by which aging mitigates Cu toxicity.     

Cerium Oxide Nanoparticles are More Toxic than Equimolar Bulk Cerium Oxide in Caenorhabditis elegans

Engineered cerium oxide nanoparticles (CeO₂ NPs) are widely used in biomedical and engineering manufacturing industries. Previous research has shown the ability of CeO₂ NPs to act as a redox catalyst, suggesting potential to both induce and alleviate oxidative stress in organisms. In this study, Caenorhabditis elegans and zebrafish (Danio rerio) were dosed with commercially available CeO₂ NPs. Non-nano cerium oxide powder (CeO₂) was used as a positive control for cerium toxicity. CeO₂ NPs suspended in standard United States Environmental Protection Agency reconstituted moderately hard water, used to culture the C. elegans, quickly formed large polydisperse aggregates. Dosing solutions were renewed daily for 3 days. Exposure of wild-type nematodes resulted in dose-dependent growth inhibition detected for all 3 days (p < 0.0001). Non-nano CeO₂ also caused significant growth inhibition (p < 0.0001), but the scale of inhibition was less at equivalent mass exposures compared with CeO₂ NP exposure. Some metal and oxidative stress-sensitive mutant nematode strains showed mildly altered growth relative to the wild-type when dosed with 5 mg/L CeO₂ NPs on days 2 and 3, thus providing weak evidence for a role for oxidative stress or metal sensitivity in CeO₂ NP toxicity. Zebrafish microinjected with CeO₂ NPs or CeO₂ did not exhibit increased gross developmental defects compared with controls. Hyperspectral imaging showed that CeO₂ NPs were ingested but not detectable inside the cells of C. elegans. Growth inhibition observed in C. elegans may be explained at least in part by a non-specific inhibition of feeding caused by CeO₂ NPs aggregating around bacterial food and/or inside the gut tract.     

Effect of leaching and aging on the bioavailability of lead to the springtail Folsomia candida

Because it is unclear if leaching can account for differences in metal bioavailability observed between metal-spiked soils and historically contaminated field soils, we simultaneously assessed Pb toxicity to the springtail Folsomia candida in three transects of Pb-contaminated soils and in leached and unleached soils spiked at similar total Pb concentrations. Total Pb concentrations of 3,877 mg/kg dry weight and higher always caused significant effects on F. candida reproduction in the spiked soils. In the transects, only the soil with the highest Pb concentration of 14,436 mg/kg dry weight significantly affected reproduction. When expressed as pore-water concentrations, reproduction was never significantly affected at Pb concentrations of 0.539 mg/L, whereas reproduction was always significantly affected at Pb concentrations of 0.678 mg/L and higher, independent of the soil treatment. These results indicate that pore-water Pb concentrations can explain, at least in part, the observed differences in the toxicity data expressed as total Pb concentrations. Leaching after the spiking procedure only caused small differences in Pb toxicity and, therefore, cannot account for toxicity differences between laboratory-spiked soils and historically contaminated field soils.     

The influence of soil properties on the toxicity of molybdenum to three species of soil invertebrates

Mo toxicity to earthworms (Eisenia andrei), Collembola (Folsomia candida) and enchytraeids (Enchytraeus crypticus) was determined in 10 European soils and a standard artificial soil, freshly spiked with Na₂MoO₄, after 28 days exposure. Mo affected survival only in three low pH sandy soils; in all other soils LC50 was >3200 mg Mo/kg dry soil. EC50 values for the reproduction toxicity of Mo were 129-2378 mg/kg for earthworms, 72->3396 mg/kg for Collembola, and 301->2820 mg/kg for enchytraeids. Variation in toxicity among soils could not be explained by differences in available (pore water, water and 0.01 M CaCl₂ extractable) Mo concentrations. Clay content best predicted the EC50 for Mo toxicity to earthworms, while toxicity of Mo for enchytraeids was best described by soil pH. For Collembola no relationships could be derived due to the absence of toxicity in most soils. Soil properties had a strong but species-specific effect on Mo toxicity to soil invertebrates.     

Toxicity to Eisenia andrei and Folsomia candida of a metal mixture applied to soil directly or via an organic matrix

Regulatory limits for chemicals and ecological risk assessment are usually based on the effects of single compounds, not taking into account mixture effects. The ecotoxicity of metal-contaminated sludge may, however, not only be due to its metal content. Both the sludge matrix and the presence of other toxicants may mitigate or promote metal toxicity. To test this assumption, the toxicity of soils recently amended with an industrial sludge predominantly contaminated with chromium, copper, nickel, and zinc and soils freshly spiked with the same mixture of metals was evaluated through earthworm (Eisenia andrei) and collembolan (Folsomia candida) reproduction tests. The sludge was less toxic than the spiked metal mixture for E. andrei but more toxic for F. candida. Results obtained for the earthworms suggest a decrease in metal bioavailability promoted by the high organic matter content of the sludge. The higher toxicity of the sludge for F. candida was probably due to the additive toxic effect of other pollutants.     

Tolerance of the Nematode Caenorhabditis elegans to pH,Salinity, and Hardness in Aquatic Media

The toxicity of many chemicals depends on the physical conditions of the test environment, and any change or adjustment made to the tests can alter the results. Therefore it is important to establish the sensitivity of the test organism over a range of test conditions to determine when it is necessary to make adjustment and to what extent. In this study, we established the tolerance range of the nematode Caenorhabditis elegans for pH, salinity and hardness using 24- (without food source) and 96-h (with food source) aquatic toxicity tests. The tests were performed in two media: K-medium and moderately hard reconstituted water (MHRW). C. elegans has high tolerance under these test conditions. In K-medium worms survived a pH range of 3.1 to 11.9 for 24 h and 3.2 to 11.8 for 96 h without significant (p > 0.05) lethality. In MHRW the pH range was 3.4 to 11.9 for 24 h and 3.4 to 11.7 for 96 h. Salinity tolerance tests were approximated with NaCl and KCl individually. Up to 15.46 g/L NaCl and 11.51 g/L KCl were tolerated by C. elegans in K-medium without significant lethality (p > 0.05). In MHRW higher salt concentrations were tolerated; about 20.5 g/L NaCl and 18.85 g/L KCl did not show any adverse effect compared to control. Hardness tolerance was tested by adding NaHCO₃. The nematode could tolerate 0.236 to 0.246 g/L of NaHCO₃. The high tolerance of C. elegans to these test conditions (pH, salinity, and hardness) allows more versatility than other organisms commonly used in aquatic toxicity tests. It also allows the monitoring of effluents and receiving waters from freshwater or estuarine sources without dilution or adjustment. Received: 14 February 1996/Revised: 20 June 1996