As suggested by the title, the authors explore the possibility of, and conclude that, the HONO + Criegee reaction acting is the dominant sink for Criegee and is the missing non-photolytic source of OH. Their actual results do not properly justify these strong statements of importance. Their own results show that the first part is greatly overstated, while the second part remains speculative given the fairly limited nature of their simulations. There are also significant shortcomings in the methodologies that could result in significant errors. Improved discussion of many aspects is needed.

The following shortcomings in the rate analysis should be addressed:

• The description of their rate calculations seems to imply that they evaluate the overall rate constant as a product of a separately calculated rate for the formation of RC1 and a master equation calculation for the branching in the thermal dissociation of RC1 (between forward reaction, kuni, and back dissociation to reactants). This product of terms would be appropriate if RC1 was being formed in the high pressure limit. But it most certainly is not. In this case, the master equation should instead be used to directly obtain the rate constant for proceeding from the reactants to the bimolecular products.
• There is also some possibility that the PCi complexes are collisionally stabilized since they are fairly deep wells on their PESs. Such stabilization would be important as it would reduce the rate of forming OH. Thus, properly formulated master equations should include the PCi complexes and some rate for their decomposition.
• The authors refer to a KTOOLS code for estimating the formation rate. The authors should also briefly describe the physical assumption behind the calculation in KTOOLS. This point is significant because their formation rates appear to be about an order of magnitude less than what would be expected.
• The wells and TSs (and perhaps the reactants) appear to have hindered rotational modes, some of which might have multiple distinct minima. At the very least, the authors should describe how they treated those torsional motions, and whether or not they searched for multiple torsional minima to ensure they had found the global minimum conformational states.

There authors discussion of their electronic structure calculations needs various clarifications:

• It appears that the TS energy reported here for CH₂OO + HONO is about 8 kcal/mol below what was reported in an earlier report from the same group. This is rather odd since the electronic structure methods are very similar. Should I presume that the uncertainty in the energy is truly that large. Some comment on this discrepancy is needed, and ideally the authors would provide some indication of the expected uncertainty in their energies.
• It is well known that CH₂OO has significant multireference character that often disappears in the TSs for its reaction. This commonly results in about a 1 kcal/mol raising of the barrier heights relative to CCSD(T)/CBS estimates. Some comment on this shortcoming in their estimates would be helpful.
• The authors claim that 0.04 angstrom geometry errors clearly suggest that M062X geometries are accurate. For this statement to be true, the authors should provide some estimate of how large an error could arise from such bond length errors. In principle, that is straightforward from some consideration of typical force constants. Simply stating that the geometry errors are small is not helpful. Similarly, the authors claim that 250 cm^-1 frequency errors imply that M062X is appropriate for frequency calculations. From my experience, those sorts of frequency errors are extraordinarily large, and would make me wonder if I had done something wrong. My expectation is that they could yield order of magnitude sorts of errors in the predicted rates. Some more appropriate discussion of the meaning of those shortcomings is needed.

The discussion of the meaning of their calculated rates also needs improvement:

• The focus on just the bimolecular rate constants in their discussion of the effective rate constants is misleading. For the (CH₃)₂COO case, the unimolecular decomposition rate near room temperature is about 400 s^-1, which swamps their effective bimolecular rates.
• With that in mind, their suggestion that the CI + HONO reactions are the major sink for the CI requires some indication as to how rapidly the unimolecular decay rates decrease with temperature.
• It would be helpful to have some estimate of the expected uncertainty in their rate predictions.
• The actual HONO concentration used in their keff calculations should be explicitly stated.
• The model simulations are limited enough in scope that I consider them to be highly speculative at best.