It is perhaps relevant to remark that I work for a nuclear operator, my principal responsibilities being to negotiate regulatory authorisations for discharge and disposal of radioactive wastes and to supervise my employer’s compliance arrangements. I also have specialist knowledge of waste sentencing processes and procedures for potentially contaminated materials, and a fair understanding of the public perception and acceptance difficulties to do with very low level or exempt waste disposals. Given this background, my remarks are related solely to that part of the draft Recommendations concerned with exclusion of radiation sources. In principle, it is entirely sensible to establish minimum levels of radioactivity for which it could ever be appropriate to undertake regulation. That said, my focus is then on the levels recommended, and the language used when discussing the topic. Recommended Exclusion Levels (Tables S2 and 10): Sections 8.1 and 8.2 seem calculated to reinforce the common misconceptions that: (a) natural radioactivity is somehow inherently less hazardous than artificial radioactivity; and (b) artificial radioactivity is so hazardous that it should be regulated down to levels generally beyond the range of measurement techniques that are reasonably practicable for bulk wastes. I do not believe ICRP would intend either of these, but clarification of the intent of these two sub-sections would help to prevent false conclusions being drawn and used as a basis for inappropriate levels of regulatory control. Considering the exclusion levels suggested in Tables S2 and 10 it might be thought that ICRP considers artificial alpha-emitters to be 100 times more dangerous than natural alpha-emitters. In fact the discrimination can be still greater because instances of secular equilibrium for natural decay series could imply a total excluded activity (e.g. for uranium-238) of 14 Bq/g (i.e. 8 Bq/g alpha plus 6 Bq/g beta) – more than 3 orders of magnitude, rather than ‘just’ two. Some discrimination is justifiable, because the concentration of artificial activity should be amenable to control, but it would be better to make it clear that whatever discrimination is applied is a consequence solely of the control aspect, not the risk per becquerel. I find it difficult to accept that ICRP should be recommending exclusion levels on the basis of quoting ‘lowest common denominators’ identified in documents from other sources without, apparently, critically evaluating their validity and questioning whether such low levels are objectively justifiable. Finally on this point, the considerable variability of natural background suggests that discrimination for control beyond a factor of 10 would be highly unlikely to offer a net benefit and would therefore be pointless. It would also greatly hamper the progress of cost-effective decommissioning and site restoration, and would potentially add very substantially to the difficulty and cost of waste disposal. The Language Used: Section 2.3 defines the purpose and nature of exclusion and exemption. The text is clear but is nevertheless set out in a manner that risks selective quotation. Paragraph (25), for example, divides neatly into halves: if I were, for example, a member of an anti-nuclear single interest group, I would certainly quote only the first half and cite it (wrongly, perhaps, but how many would realise that?) as ‘ICRP endorsement for the need to regulate radioactivity all the way down to the levels given in Table 10’. The first half of Paragraph (25) is: 'Sources and exposures that are not excluded are within the scope of the system of protection. These sources and exposures should be subject to appropriate authorization by the relevant regulatory agency.' A consequence, as noted above, could be that site restoration and de-licensing could become even more difficult and costly to achieve and to verify, with potentially large additional volumes of waste requiring their disposal to be regulated because of the ‘radioactive’ labelling.