As part of a general programme for the development of new techniques of selective weed control, numerous experiments have been carried out during 1943–7 on the eradication of annual weeds from spring-sown oats and barley. These multifactorial experiments consisted essentially of simultaneous comparisons of the relative merits of sulphuric acid, copper salts, dinitro-o-cresol and the two growth-regulating substances 2:4-dichloro-phenoxyacetic acid and 2-methyl-4-chloro-phenoxyacetic acid. All the compounds were applied as aqueous solutions or suspensions at a spray rate of 100 gal./acre.

Within groups of compounds, cupric chloride is more toxic than cupric nitrate, and this in turn is more toxic than cupric sulphate to such weeds as Raphanus raphanistrwn. Similarly, dinitro-o-cresol or its ammonium salt is more effective than the sodium salt for the control of R. raphanistrum and Galeopsis tetrahit. There is some evidence that sodium methyl-chloro-phenoxyacetate is slightly superior to sodium dichloro-phenoxyacetate for the control of Galeopsis tetrahit and Sinapis arvensis.

Between groups of compounds, as long as the weeds are in the young seedling stage there is little to choose between sulphuric acid, cupric chloride, ammonium dinitro-o-cresylate and the growth-regulating substances for the control of Sinapis arvensis, Raphanus raphanistrum, Thlaspi arvense and Polygonum convolvulus when the concentration ranges for the four groups are respectively 9–18, 1–4, 0·4–0·8 and 0·1–0·4%. Within these limits of concentration P. persicaria is best controlled with sulphuric acid and Atriplex patula by dinitro-o-cresol.

In general, there is a decrease in the toxic effects with increasing age. There are however complex interrelationships between weed species, compound and stage of development. Whereas for cupric chloride its effectiveness against Sinapis arvensis is little changed between the cotyledon and flowering stage there is with flowering a marked falling off in the efficiency of sodium methyl-chloro-phenoxyacetate. In the intermediate phase of development the growth-regulating substances are most toxic to Raphanus raphanistrum.

The results for crop yield demonstrate that sulphuric acid and cupric chloride are less selective than dinitro-o-cresol or the substituted phenoxyacetic acids, since for equivalent degrees of weed control smaller increases in grain production have often been recorded. These injurious effects are maximal at the higher concentrations and when spraying is delayed until after tillering is completed. There is a similar age trend for dinitro-o-cresol and the two growth-regulating substances, and in the case of the latter some indication of a concentration effect, i.e. the compounds are most selective at concentrations of 0·2% or less.

For the twenty experiments in which the yields were recorded the most appropriate herbicidal treatments have resulted in gains in yield over the controls ranging from 0·4 to 82·2% with a mean of 23·8.