General The first drafts of the new ICRP recommendations put forward a drastic change in comparison with the 1990 recommendations in ICRP Publication 60. It was therefore then obvious that an entirely new presentation was needed. The drafter, the ICRP Chairman, suggested that satisfactory protection could be achieved on the basis of individual dose restrictions alone and that further protection efforts were not essential. His assumption was that “if the risk of harm to the health of the most exposed individual is trivial, then the total risk is trivial – irrespective of how many people are exposed”. This assumption was not accepted by most of the radiation protection community and was not maintained in the latest drafts. The consultation draft presents, and gives great weight to source related dose constraints but also, as in Publication 60, recommends that optimisation of protection “is an important component of a successful radiological protection programme”. So, like in Publication 60, the two basic elements, individual dose restriction and optimisation of protection are both still there. The difference is now in verbal presentation. In Publication 60, the main emphasis was on optimisation of protection, since this procedure determined the eventual level of protection. Dose constraints were introduced as a boundary condition, to make sure that the optimisation result was not applied at the cost of unreasonable high risks to the most exposed individuals. From the ethical point of view, the Commission here dealt with two ethical principles: utilitarian and deontological ethics. This was clear and easy to explain. The present presentation begins with the dose limitation offered by the source-related constraints and says (in paragraph S5) that this will “provide a level of protection for the most exposed individuals”. It then, somewhat half-heartedly (in paragraph S6) says that “the constraints are complemented by the requirement to optimise the level of protection achieved”. But it is not true that the constraints will provide the level of protection, they will only make sure that no individual runs an unreasonable risk. The level of protection is still provided by the optimisation requirements. So, there is no difference from Publication 60 except in the emphasis on words and order of presentation. And this difference lingers from the first drafts, where a drastic difference was intended. To give the restraints an apparent weight different from that provided by Publication 60 will only confuse the reader who will see no reason for the difference in presentation. The Aim of the Recommendations In order to make sure that no change in the basic philosophy is intended, it is suggested that the first line of paragraph S3 is expanded to read “. . . concluded that its recommendations should continue to be based . . .” The Principles of Protection Paragraph S5 should be amended. A “level of protection” cannot be a “restriction”. Furthermore, the “most fundamental” (or rather actual) level of protection is provided by the optimisation procedure. The dose constraints do not provide a level of protection (nor can they be regarded as such), they make sure that a certain level of protection is not exceeded. In paragraph S5 reference is made to “a class of exposure”. However, this concept is not introduced and explained until in paragraph 129. In Publication 60 the term was “type”, not “class”. Paragraph S6 should also be amended. Here the wagon is before the horse. The optimisation procedure is the essential one, because it determines the actual level of protection. The constraints only guarantee a minimum level of protection. Paragraph S7 and Table S1. There is an inconsistency between the text and the table. The text says that the constraints would “probably” not be lower than 1/10 of the values in the table. However, the last value in the table is 0.01 mSv/a as a minimum for “any constraints”. But who would consider 0.01 mSv/a as a constraint where the maximum is 100 or 20 mSv/a? The table has been misunderstood by several people who have only looked at it briefly, as presenting recommended values rather than maximum values. It would perhaps be more in line with the text to give recommended intervals, e.g. 1—100, 0.2—20 and 0.01—1. Paragraph S8. Surely, it should not be said that “The level of protection . . . is the dose limit”. Also, it is the minimum level of protection that is guaranteed by the dose limit. Optimisation of Protection No comments here, except that optimisation if treated half-heartedly. The relevant comments are on section 7. Exclusion of radiation sources No comments here. The development of effective dose The change of name from equivalent dose to radiation weighted dose is probably an improvement but is likely to meet objections from people who already feel that the Commission changes too often and plays with too many terms. The equivalent dose is mentioned here but not in the main text (paragraph 51) when the radiation weighted dose is introduced; this may cause some confusion. One way of overcoming the problem of finding a new name for the unit (so that the name sievert is used only with the effective dose) is to use the name gray for both the radiation weighted and the unweighted dose. This would be no more confusing than having the unit J kg-1 for both. A third name in addition to gray and sievert may meet difficulties in being accepted. Protection of non-human species Many reader may feel that the Commission is overdoing this item in order to live up to (ignorant?) expectations. The problem is merely to find justification for the statement in paragraph 16 of Publication 60. Many would think that this can be made more easily and with less complication. The key issue is whether protection against stochastic effects is relevant in protection of non-human species. It may be possible to exclude stochastic effects. This would reduce the problem drastically. One would only have to study threshold doses and arrange protection so that these are not exceeded for any species or any large population. Paragraph S19 says that there is a need for international standards of discharges into the environment. The way in which this is said gives the impression that discharge limits depend on the risks for non-human species. However, in most cases such limits are set on the basis of source-related constraints and optimisation of protection of human. The main text (paragraphs 1—251) The comments already made will require corresponding amendments in the main text. Additional comments are made here only for some paragraphs. Paragraph 8: The first feature is not “Recommending dose constraints” since that recommendation was already made in Publication 60. The text should read “Recommending values for the dose constraints…” Paragraph 8: In accordance with the previous general comments, it is really not “Complementing the constraints and limits with the requirement for optimisation…” but rather “Complementing the optimisation . . . with the boundary conditions presented by the constraints . . .” Paragraph 51: See previous comments- Paragraph 117: Severe mental retardation and IQ losses are presented as two different effects. However, in Publication 60 it was shown (para. B171) that an IQ shift of 30 units per sievert would result in the expected mental retardation. The text here that “The review . . . now supports a true dose-threshold . . .” is too simplistic: a better description was made in Publication 60 (B171). Section 7.2 This section reads well, but the described policy does not hold for a thorough analysis. The main objection to the usefulness of the collective dose is said to be that it may aggregate information excessively. Let us examine the validity of that statement. Most of the recommendations relates to protection against controllable sources, where doses are below the recommended constraints. However, there are also accidental situations where some individuals receive high doses. In such situations the statement is true – a single value of a collective dose would not only aggregate the situation excessively, it would be meaningless. However, the main purpose of the recommendations is to provide a policy for normal situations where doses are below constraints. With a few exceptions (mainly radon in dwellings) radiation individual doses to the public are then so low that the total biological detriment rather than individual risk should be the main concern. With the LNT assumption accepted by the Commission, the total (global) collective dose is a measure of that detriment irrespective of the distribution of individual doses. In this case there is no “excessive aggregation” unless some further assumptions are introduced. Such assumptions could be: (a) The LNT assumption is not valid for very low doses; (b) Deaths in distant countries carry less weight than deaths “at home” (discrimination). To reject the total collective dose as a measure of the total detriment, therefore, implies that assumptions (a) and/or (b) have been made. However, the Commission has not made any of these assumptions. It may be appropriate to say that the collective dose is not the only valid input to the optimization procedure. There may be other harm than the biological detriment, e.g. in public relations. If the collective dose is assessed without discrimination, it may still be justified, for such reasons, to make extra efforts to further reduce the highest individual doses. But this should involve extra total expenses that may not be in proportion to the number of individuals involved and should therefore not be achieved by giving extra weight to part of the collective dose. It therefore seems that the total (global) collective dose, without any weighting of the individual dose contributions, is a necessary input to the optimization assessment but not necessarily the only one. It does not aggregate any information needed for the assessment of the total biological harm. However, additional information on the distribution of individual doses may be needed for other purposes. Section 9: This section deals with both diagnostic and therapeutic exposures. This is confusing to the reader because the situations and the requirements are so different. It would be better to divide this section into a diagnostic and a therapeutic sub-section.