The Commission is commended for the consultation process that it implemented for its 2005 recommendations. Such a process can only improve the acceptance of the recommendations and improve the science behind them. It is my view that there appears to be a widespread misconception concerning the recommendations of the ICRP and the Commission may wish to dispel this misconception more forcefully. It is clear that the ICRP 2005 recommendations do not attempt to present a summary of the best possible scientific information regarding the effects of radiation on humans. Instead, it presents a practical system of protection based on the best scientific evidence. The Commission appears to take into account the practicality of its recommendations, the inherent uncertainty of the data, the combined evidence of physical measurements and epidemiological data. Specific suggestions: (S5) and (189 – 193) The Commission must emphasize that the risks from radiation should never be assessed in isolation. There are many instances of decisions based solely on radiation protection that make absolutely no sense when the overall context is considered. As an example, it is absurd to expose workers to a chemical carcinogen in order to avoid any exposure to small radioactive sources. (51 and 52) The radiation weighted dose should not have units of J/kg. This definition of the sievert does not respect the principle of conservation of energy. It is almost impossible to teach radiation protection units to intelligent students when the units lack basic internal consistency. The definition of the sievert should be distinct from the gray. The conversion factor from gray to sievert should be the radiation weighting factor with units of Sv/Gy. (67) The Commission must do a better job of explaining its rationale for the choice of radiation weighting factors. It is clear that the Commission thinks that there are models that better describe the actual risk in a given set of circumstances. However, the Commission chose to use simpler models because they are more practical and usable. This also gives the Commission the flexibility to adjust the weighting factors on the basis of experimental and epidemiological data. This is reasonable and commendable, but it begs for more explanations, particularly when it states that the weighting factors are ‘not for the retrospective assessment of individual risks of stochastic effects from radiation exposure or for use in epidemiological evaluations’. (73) If the Commission decides to deviate from a direct correlation between qE and wR, it should support the deviation with experimental data. It is NOT the deviation from the status quo that needs justification. It is the deviation from scientific evidence. (89) (93) The Commission bases its internal dosimetry dose conversion factors on computational models of the body that takes into account the mass and size of the various organs. This is reasonable and it works quite well. In the case of the dose to the thyroid from radioactive iodine, the situation is rather special. The size of the organ does not seem to have a very large impact on the dose conversion factors (HP 84(3)-334). The dietary intake of stable iodine has a much more important impact on the coefficients and yet it is not considered in the current evaluations. Dietary variations of more than a factor 10 are not uncommon between countries. In Canada, the dietary intake of iodine is estimated at 1000 microgram/day while in Ukraine, it was less than 75 microgram/day before the Chernobyl accident. The Commission should mention its importance in the 2005 recommendations and the new dose conversion factors should take into account this factor. This alone, will have a large impact on emergency plans for reactor accidents. (164) The dose constraints of 20 and 100 mSv pear year during an emergency change considerably the whole philosophy of protective actions after an accident. Up to now, the protective actions for reducing stochastic effects were justified when the dose averted exceeded a dose intervention level. The Commission must explain how it thinks this new principle translates in practice. (B17) ICRP’s desire to limit its recommendations for the protection of the non-human biota is understandable: it is a complex and immature area of our science that still rests on great uncertainties. In its effort to avoid some of the more controversial issues, the ICRP proposes a set of reference animals that would facilitate communication between scientists. It also discusses dose or derived consideration levels that remain vague in application. However, in doing so, not only does the ICRP NOT avoid controversy, but it adds complexity to an already overwhelmingly complex discussion. Its position is a refinement on details that will not resolve the issue without a proper framework. And it is this overall framework for the assessment of impacts on non-human biota that is needed. Thus it is suggested that the ICRP has two basic options: 1) stay away from the debate completely and refrain from making recommendations in this area until the subject is more mature or 2) get make recommendations that include a detailed, practical framework. I would not presume to offer a solution, but by way of an example, we can take the framework for the protection of humans. For the protection of humans, there is a reference man that forms the basis of the dosimetric models. Right from the start, the Commission could have selected a series of reference men, women, children of different races, but it did not, until the models were quite mature. There is an end point, which is the detriment associated with the stochastic effects. We have radiation weighting factors that are relevant to the chosen endpoints. We have dose limits. If the same approach was transferred to non-human biota, the Commission would need to propose all these things. Unavoidably, the assessment of the risk to non-human biota will require dosimetric models. However, considering the level of uncertainties and the sparseness of the data, we should strive to simplify the dosimetric models. The current proposal appears to be based on the FASSET-ERICA program, which is rather complex. Instead of having 20-odd reference biota, it would be a lot better to limit the recommendations to the smallest initial set of generic biota, such as: • generic aquatic flora for seawater, • generic aquatic fauna for seawater, • generic aquatic flora for freshwater, • generic aquatic fauna for freshwater, • generic terrestrial flora and, • generic terrestrial fauna. The generic biota should be sufficiently representative to guarantee that it is unlikely that a real biota would receive a dose higher than the generic biota. Also, a specific end point (such a no effects observed on the population, or no effect on the reproductive ability of the individual) should be suggested. RBE factors and annual absorbed dose limits corresponding to the chosen endpoint should also be specified for each generic biota. This would harmonize the framework for the protection of the non-human species with the framework for the protection of humans.