Silver nitrate

Toxicity to aquatic algae and cyanobacteria

Description of key information

Data for three species were used to derive the freshwater PNEC by statistical extrapolation. These were an EC10 of 0.41 µg Ag/L for Pseudokirchneriella subcapitata, an EC10 of 0.54 µg Ag/L for Chlamydomonas reinhardtii (Hiriart-Baer et al. 2006) and an EC10 of 0.41 µg Ag/L for Nostoc muscorum (geometric mean from Rai and Raizada (1985), Rai and Raizada (1987) and Rai et al. (1990)). The lowest reliable chronic value is an EC10 of 0.16 µg Ag/L for the blue-green algae Nostoc muscorum (Rai et al. 1990). This has been selected as the chronic ERV for classification purposes. Additional chronic toxicity data of similar sensitivity are also available for the Brown Trout Salmo trutta (217 d EC10 of 0.19 µg Ag/L) and Oncorhynchus mykiss (196 d EC10 of 0.17 µg Ag/L) obtained from Davies et al. (1998). A NOEC of 1.2 µg Ag/L for Champia parvula (Steele et al. 1983) was used as supporting data for the derivation of the marine PNEC.

Additional information

Key long-term data for three algal species are available. Data for Pseudokirchneriella subcapitata and Chlamydomonas reinhardtii are reported by Hiriart-Baer et al. (2006). Exposures to silver were conducted in the presence and absence of thiosulphate (an inorganic ligand) in “turbidostat” culture vessels, which maintain populations of algae at optimum conditions for growth and allow the measurement of the rate of population increase. Only the exposures conducted in the absence of thiosulphate were considered sufficiently reliable for PNEC derivation. Three replicate turbidostats per algal species were used to determine the growth rate of algae after successive increases in the concentration of silver in the turbidostat nutrient solution influent. The population growth rate was allowed to stabilise between successive increases (usually for between 18-24 hours). Results were reported on the basis of measured dissolved silver concentrations in the turbidostat effluent. EC10 values for population growth rate were calculated as 0.41 and 0.54 µg/L dissolved silver (measured) for C. reinhardtii and P. subcapitata, respectively.

Data are also available for Nostoc muscorm, a species of blue-green algae (cyanobacteria), from a series of three papers published between 1985 and 1990 by Rai et al. (Rai and Raizada (1985), Rai and Raizada (1987), Rai et al. (1990)). In all of the three studies algae were cultured under static conditions for 15 days in 150 ml Erlenmeyer flasks at 25 ± 1 °C containing 50 ml of buffered (pH 7.5) modified CHU₁₀ nutrient media. Light intensity was 2,500 lux with a 14:10 light:dark photoperiod. Optical density at 663 nm was measured every three days in a spectrophotometer as a surrogate measurement for growth (although only 15-day data are reported). Silver exposure was based on nominal concentrations. EC10s for algal yield obtained from four independent tests ranged from 0.16 to 0.66 μg/L dissolved silver. The geometric mean of the EC10s was calculated as 0.41 µg/L dissolved silver, which is consistent with the data for other algal species.

The lowest long-term result for marine algae is a 14-day reproductive NOEC of 1.2 µg/L total silver for Champia parvula (Steele and Thursby 1983). Day 0 analytical values were used to derive the toxicity values as the lowest three concentrations were below the reported limit of detection (1.0 µg/L). The Day 0 data for undetectable concentrations were based on the dilution of the volume of standard added to the flasks. EDTA was present at 100 µg/L in the exposure medium and it is likely that some of the silver would have complexed with this and potentially been unavailable to the test organisms. The loss of silver over the 7-day period in the concentrations was approximately 50% (most likely as a result of sorption to algal cells and the test vessel). Similar results are reported by Thursby and Steele (1986) using the same methodology, but without the addition of EDTA. However, no analytical confirmation of the test concentrations was carried out and there is limited reporting of the method and results. These data, together, are considered sufficiently robust to be used as supporting data for determination of the AF when deriving the PNEC, but not directly for PNEC derivation.