The use of cluster analysis for plant grouping by their tolerance to soil contamination with hydrocarbons at the germination stage.

Clustering was employed for the analysis of obtained experimental data set (42 plants in total) on seed germination in leached chernozem contaminated with kerosene. Among investigated plants were 31 cultivated plants from 11 families (27 species and 20 varieties) and 11 wild plant species from 7 families, 23 annual and 19 perennial/biannual plant species, 11 monocotyledonous and 31 dicotyledonous plants. Two-dimensional (two-parameter) clustering approach, allowing the estimation of tolerance of germinating seeds using a pair of independent parameters (С75%, V7%) was found to be most effective. These parameters characterized the ability of seeds to both withstand high concentrations of contaminants without the significant reduction of the germination, and maintain high germination rate within certain contaminant concentrations. The performed clustering revealed a number of plant features, which define the relation of a particular plant to a particular tolerance cluster; it has also demonstrated the possibility of generalizing the kerosene results for n-tridecane, which is one of the typical kerosene components. In contrast to the "manual" plant ranking based on the assessment of germination at discrete concentrations of the contaminant, the proposed clustering approach allowed a generalized characterization of the seed tolerance/sensitivity to hydrocarbon contaminants.

Tolerance of cultivated and wild plants of different taxonomy to soil contamination by kerosene.

In laboratory experiments on leached chernozem contaminated by kerosene (1-15 wt.%), germination of 50 plants from 21 families (cultivated and wild, annual and perennial, mono- and dicotyledonous) as affected by kerosene type and concentration and plant features was determined. Tested plants formed three groups: more tolerant, less tolerant, and intolerant, in which relative germination was more than 70%, 30-70% and less than 30%, respectively. As parameters of soil phytotoxicity, effective kerosene concentrations (EC) causing germination depression of 10%, 25% and 50% were determined. EC values depended on the plant species and varied in a wide range of kerosene concentrations: 0.02-7.3% (EC(10)), 0.05-8.1% (EC(25)), and 0.2-12.7% (EC(50)). The reported data on germination in soils contaminated by oil and petrochemicals were generalized. The comparison showed that at very high contamination levels (10 and 15%) kerosene was 1.3-1.6 times more phytotoxic than diesel fuel and 1.3-1.4 times more toxic than crude oil, and at low (1 and 2%) and medium (3 and 5%) levels the toxicity of these contaminants was close differing by a factor of 1.1-1.2. Tolerance of plants to soil contamination had a species-specific nature and, on the average, decreased in the following range of families: Fabaceae (germination decrease of 10-60% as compared to an uncontaminated control)>Brassicaceae (5-70%)>Asteraceae (25-95%)>Poaceae (10-100%). The monocotyledonous species tested were characterized as medium- and low-stable to contamination, whereas representatives of dicotyledonous plants were met in all groups of tolerance. Tested wild plants, contrary to reference data on oil toxicity, were more sensitive to kerosene than cultivated. No correlation was observed between degree of plant tolerance to kerosene and mass of seeds. The evidence indicates factors as structure and properties of testa, structure of germ, type of storage compounds, and type of seed germination (underground or aboveground) are more important.