The report discusses the topic authoritatively and is a useful exploration of the approaches one can take to assessing doses to the average member of the critical group, now called the representative individual. The comments below, by paragraph, are on the main text and Appendix B. Some of them will also apply to corresponding text in the Abstract and Summary. Technical points (7) It is inaccurate to suggest that it is only now with sophisticated computer and software tools that uncertainties can be propagated through calculations and that only point estimates of dose could be made in the past. Uncertainties had to be propagated; comparison of a point value with a limit is meaningless unless the uncertainty in the value has been estimated. (18) It is difficult to see why an emphasis on source-related dose constraints now makes estimating doses to individuals more important. That argument seems a red herring; demonstrating compliance has always entailed, inter alia, estimating doses to individuals. A more appropriate purpose would be the idea of introducing more structure into the determination and application of the critical group in the light of experience and advances in assessment tools; notably software. (19) “As a result, the Commission believes there should be a small probability that some individual may exceed the dose constraint.” A more reasonable belief than “should be a small probability . . .” would be that a dose constraint could be considered complied with even if there is a small probability that some individual exceeds it. (24) The reader here might well expect a discussion of the issues or difficulties that the Commission sees with the current definition and application of critical group, as a preamble to the following sections. The following sections could then point to where the more elaborate approach described in them, involving a representative individual, help resolve those difficulties. Sections 2.1, 2.2 and 2.3 are succinct and helpful description of the assessment process. (45) It would be more accurate, in discussing “dose constraints, weighting factors, and dose coefficients”, to state that they are defined as point values, rather then, as in the text, that they are assumed to be point values. (46) “Uncertainties associated with estimation of dose may be taken into account either deterministically by using single values for parameters chosen to take account of uncertainty, . . .” At this stage in the text it is a mystery how a single value can take into account uncertainty. There needs to be some sense given here that values are chosen to yield doses on the high side – i.e., conservatively, though that term strictly would need some explanation. This is all discussed later; perhaps merely writing here “as discussed below” would mollify the reader. (46) “. . . the goal should be to perform a realistic evaluation of dose.” This needs some explanation too. Section 2.4 The discussion here of the types of uncertainty with which one has to contend is helpful. Implicit in the discussion here and in later sections (for example, in (B14)) is that it is important when analyzing the results of an assessment to know where in the various pathways and variables the main contributors to the overall uncertainty arise. It is a point worth making explicitly somewhere in the main text. (51) Rather than stating “With advances in technology, decision makers have available powerful tools that allow the variability of data needed to estimate dose to be taken into account.” it might be fairer to write: “The propagation of uncertainties in data through assessments of dose is more readily accomplished with the computer technology that is available now, compared with when the critical group concept was first developed, or even two decades ago.” Also, “decision makers” are usually not the ones using the tools; the rewording takes care of this. (63) “. . . the 95th percentile of consumption rates for many staple foods tend to exceed the mean value of the distribution by approximately a factor of 3 . . Therefore, with deterministic methods, using the 95th percentile of behaviour is considered to represent a cautious, but acceptable, assumption for defining a reasonable and sustainable intake rate.” Why this is cautious and acceptable is not clear. Is it reflecting the point made later (in para (69)) to the effect that “ . . the group should be regarded as homogenous if the distribution of individual doses lies substantially within a total range of a factor of 10 (i.e., a factor of about 3 on either side of the mean).”? If so, para (69) should be earlier. Also, “defining a reasonable and sustainable intake” need only be “defining a reasonable intake” since para (68) reads “Sustainability and homogeneity are aspects of reasonableness”. These points also occur in (S12). (88) “. . . the Commission recommends that the representative individual be identified such that the probability is less than about 5% that a person drawn at random from the hypothetical population will receive an annual effective dose exceeding the dose constraint.” This seems to an inaccurate definition of the representative individual. Perhaps the intention is “The representative individual is identified such that the dose to that individual is greater that 95% of the population. Then, if the dose to the representative individual is less than the dose constraint, less than 5% of the members of the population receive doses greater than the constraint.” (The wording in B43 is the same.) Also “ . . .drawn at random from the hypothetical population. . .” seems to conflict with para (B9) “The hypothetical individual cannot reasonably be interpreted as a random individual chosen from the affected population.” (89) “Determining the existence of a homogeneous group in a probabilistic assessment can be problematic.” This is certainly the case and this point and the discussion following (and in para (B45)) on the difficulties that can arise in identifying the homogeneous group that is the basis for defining the representative individual might make readers feel they are back where they started with the critical group of previous documents. It would be worthwhile after Section 4 having a review of how the approach presented here provides the assessor and the decision-maker with an improved way of estimating and using estimates of individual doses. Section 4.3 The discussion of stakeholder involvement seems to go beyond the scope of the report. The comment already made in para (89) is sufficient for this document: “Close attention must be paid to suggestions from members of the public. . .” Section B3. This is an useful discussion on handling distributions. I’m surprised more of paras (B10) through (B14) is not the main text. (B10) (3) “Distributions of this kind, combined with risk estimates, could be useful in estimating the number of health effects arising from the collective dose.” Doesn’t the Commission discourage such estimation of health effects? (B16) “The reason for the suggested omission . . .” Which suggested omission? Wording points (10) There are many occurrences of “situation” being used after a noun. Avoiding such constructions improves readability. Examples: “exposure situations” should be “exposures”; “emergency situations” should be “emergencies”. Previous occurrences were in paras (3) and (8). There are many more in subsequent paras. (36) “During emergency planning activities, prospective assessments may be made modelling potential source terms and the populations around a particular source . . .” The ambiguity of whether it is during planning in an emergency or during planning for emergencies would be avoided by writing (if the latter meaning is the intent) “In planning for emergencies, prospective . . .” (B15) “The crudest level of exposure scenarios assumes that . . .” would be better as “In the simplest exposure scenarios, the assumptions made are . . .”