The research looked specifically at the eating habits of snails in grasslands - where it is known that some plants survive a mollusc attack better than others. "We quantified seedling selection by snails (Cornu aspersum) for 14 common, European grassland species," writes Dr Mick Hanley, Assoc Prof in Plant-Animal Interactions. "Seedling acceptability was subsequently compared with species-specific expression of constitutive secondary defence metabolites (CSDMs), and VOCs (volatile organic compounds)."
In other words, researchers compared the snails' selection of plants based on their taste (CSDMs) and on their smell (VOCs).

The results showed there was no relationship between the snails’ choices and the presence of CSDMs, but there was strong correlation with VOCs. It seems that certain VOCs, such as green leaf volatiles (3-hexen-1-ol acetate), were particularly offensive to the molluscs; whereas the taste of specific monoterpenes (β-ocimene) were identified as likely attractants.

Futher research is needed to identify exactly what volatile compounds repel snails. These could be used as naturally occuring herbicide controls.

Garden Organic members with long memories will remember the research led by Lawrence Hills on Perfume Against Pests in 1960s, as well as the Members' Experiment in 1971 of using garlic oil as a natural resistance to slugs and snails.

For Dr Hanley's full report and results, see here in the Annals of Botany.