The final paragraph answers the question: > Why choose a nerve agent over a hail of bullets, as used in the assassination of the Iranian nuclear scientist Mohs

The final paragraph answers the question: > Why choose a nerve agent over a hail of bullets, as used in the assassination of the Iranian nuclear scientist Mohsen Fakhrizadeh, probably by Israel, in November? The logic behind using nerve agents in the Skripal and Navalny cases was that the target could be poisoned through contact with their own belongings, a method that allowed the would-be killers to operate at a safe distance. The death would be formally deniable but the result plain enough to send a message. In both cases the plot ran into the classic problem for assassins: the difficulty of actually killing the victim. It’s surprisingly common for people to survive assassination attempts, particularly when single gunshots or explosives are used. The recent record of attempted nerve agent assassinations leads to the suspicion that they, too, are a poor choice of weapon. Contrary to popular belief, they aren’t always fatal, since they aren’t quickly absorbed through the skin (death is quicker if they’re inhaled, or enter through the eyes). They’re expensive to produce and their effects can be treated with atropine. In this respect they aren’t a dramatic improvement over traditional poisons: hydrogen cyanide, pentobarbital, potassium chloride – or fentanyl derivatives of the kind used in the attempted assassination of the former Hamas leader Khaled Meshal in 1997 – do just as good a job. A mundane explanation for their recent use is that the FSB unit in charge of assassinations happens to be the division that oversaw the old Novichok labs. They used what they had.