The biological effects of phytotoxic compounds will be dependent upon uptake, the amounts reaching the site of toxic action, and the toxicity at cell level. When the materials are applied as sprays to growing plants absorption through the roots, retention by and penetration into the shoot, transport and localized accumulation are factors which may determine differences in response either between compounds or between species. The precise assessment of relative toxicity must therefore involve studies of the effects on whole plants and at cell level. For such assessments it is essential to determine in the first place the change in the biological effects brought about by a wide range of dosage. The percentage inhibition of germination bears a sigmoid relationship to the amount of toxicant in the external medium, while the same relationship holds between percentage mortality and the concentration of the spray solution. Accurate comparisons of relative toxicity can only be obtained when the variation in reaponse is measured at several dosages and the data treated by the methods of probit analysis. Using such techniques in germination or spraying experiments it has been demonstrated that the relative toxicities of chlorinated phenoxycetic acids, alkyl phenylcarbamates, dinitro-alkylphenols, pentachlorophenol, thioacetic acid, and formamide are greatly dependent on the species. In fact, the order of toxicity may be reversed from one species to another, while between compounds the results of germination tests may or may not be comparable with those obtained by spray applications. The physical characteristics of the spray solution will in part determine the degrees of retention and penetration. For compounds with a low solubility in water the addition of a hygroscopic substance may increase the percentage kill. According to the species, spray solutions of a low surface activity may be more or less toxic than those with a high activity, while the relative effects of oil emulsions and aqueous sprays vary between species. For compounds which are freely translocated, the methods of growth analysis are of value for assessing the toxic effects, especially of non-lethal dosages. Since the effects on the growth of the component parts of the plant may be widely divergent, conclusions based on a single criterion are likely to be erroneous. Where movement of the compound is restricted, such as with hydrocarbons, an assessment of toxicity can be obtained by measuring the degree of localized damage following on the application to the leaves of individual droplets of varying size. Lemna minor has the twin advantages that the experimental conditions can be controlled and that in some respects its response to phytotoxic compounds is akin to that of unicellular organisms. Since with some compounds at any one dose the depression in the growth rate remains constant with time (e.g. nitrophenols), while for others the depressant effect is cumulative (e.g. dichloro-phenoxyacetic acid, cupric salts), the nature of the growth response must first be established before comparisons between compounds can be made. For studies of relative toxicity at cell level two methods have ken employed. The external concentrations in the agar medium required to halve the growth rate of Trichoderma viride have been determined, or the dosages required to bring about a standard effect on the respiration of yeast have been estimated. It is concluded that only by using a range of species and a number of techniques can relative toxicity be established with precision.