causal connection. This is sufficient to justify requiring the employer to assume responsibility for causing or materially contributing to the onset of the mesothelioma when, in the present state of medical knowledge, no more exact causal connection is ever capable of being established. Given the present state of medical science, this outcome may cast responsibility on a defendant whose exposure of a claimant to the risk of contracting the disease had in fact no causative effect. But the unattractiveness of casting the net of responsibility as widely as this is far outweighed by the unattractiveness of the alternative outcome. He suggested that policy considerations would be key to further application of the test: Policy questions will loom large when a court has to decide whether the difficulties of proof confronting the plaintiff justify taking this exceptional course. It is impossible to be more specific. Lord Hoffman elaborated on the concepts underpinning rules of legal causation (at para. 56): The concepts of fairness, justice and reason underlie the rules which state the causal requirements of liability for a particular form of conduct (or noncausal limits on that liability) just as much as they underlie the rules which determine that conduct to be tortious. And the two are inextricably linked together: the purpose of the causal requirement rules is to produce a just result by delimiting the scope of liability in a way which relates to the reasons why liability for the conduct in question exists in the first place. In Barker v Corus (UK) plc [2006] UKHL 20 the UK's House of Lords addressed two questions undecided by the decision in Fairchild: • First, what are the limits of the exception? In Fairchild the causal agent (asbestos dust) was the same in every case, the claimants had all been exposed in the course of employment, all the exposures which might have caused the disease involved breaches of duty by employers or occupiers and although it was likely that only one breach of duty had been causative, science could not establish which one it was. Must all these factors be present? • Secondly, what is the extent of liability? Is any defendant who is liable under the exception deemed to have caused the disease? On orthodox principles, all defendants who have actually caused the damage are jointly and severally liable. Or is the damage caused by a defendant in a Fairchild case the creation of a risk that the claimant will contract the disease? In that case, each defendant will be liable only for his aliquot contribution to the total risk of the claimant contracting the disease - a risk which is known to have materialised. Lord Hoffman concluded that it did not matter whether all the sources of risk were 42