Comments on 16/08/2004 draft. ============================= Title: SNe -> supernovae Introduction ============ You should mention Garbiele's work. See http://panisse.lbl.gov/collab/papers/garavini_evolution/index.html Section 3 ========= 1st paragraph There is LaTex error above the number 1.2. 2nd paragraph The term "plausibly Ia" is a bit vague. Unfortunately, it may be hard to be more precise since the data comes from different groups and the classifications adopted by the different groups are largely qualitative. Section 4 ========= 1st paragraph. In general I am against sentences that point out deficiencies in the analysis of other groups unless it is really necessary for the paper. I prefer that you state what you have done and to leave it for others to judge the merits of your analysis. I would like to see the last sentence deleted in this paragraph. In general, and more so for high-z SNe, the rest frame B- and V-band measurements will be taken at different times. How do you handle this? Figure 2 ======== Ideogram - what is that. Section 5 ========= I guess that you only use SNe with at least three points within the desired interval to determine beta and to derive an estimate of the goodness of fit. If you are willing to give up the goodness of fit, there are probably a number of other SNe with two well measured points within the desired interval that could be used. I guess that 2000fr is one of them. Is it worth measuring beta for those with accurate photometry? This would give you a few more high z SNe to check the universality of beta. Figures 4 and 6. =============== Instead of stacking the histogram, I suggest that you overlay them. The left and right hand axes can be for labeling the different vertical scales. Table 5 ======= Some of these cuts have units, e.g. days. Section 6.1 =========== Are you planning to do a plot which is similar to figure 13 in Rob's HST paper? I think that it is worth showing that the value of alpha derived for B_BV0.6 is half of that for B_max. I think that you should compare the 0.12 magnitudes of dispersion with that derived by other methods, as you do in your methods document. Section 6.2 =========== 1st paragraph after the analysis cuts. allowed -> applied. I found the discussion on probcut difficult to read and follow. Perhaps you can find a better way of expressing the 4th sentence in this paragraph. In table 5, I would have expected a number greater than 1. This probably means that I have not fully understood the cut "... eliminate 62 of these." It is not clear to which sample (the 119 or 131) the relative pronoun is referring. In the 3rd paragraph of section 2, it may be worth including the word baseline to describe the sample of 131 SNe that is mentioned there. Section 8 ========= Can you double check the derived slopes "S". From figure 11, S=1 looks better than S~0.7. Did you include the errors in the fit? It might be necessary for you to include error bars in figure 11. How much does Omega_1 change if you include the restriction that only SNe with at least two points in the CMAGIC region can be used in the fit. Earlier, you had expressed some concern about this cut. From the analysis I've seen so far, I do not share this concern, and I think that the current cut (i.e. at least one point in the CMAGIC region) is adequate. Note that your jackknife test showed that the results are sensitive to the removal of SN2002kd. Section 9.1 =========== In Rob's paper, one of the largest sources of systematic error was the assumed U-B colour for the supernova. Here, you have shown that you are much senstive to this error. I think that you should explicitly mentioning this. Here are a couple of other points to consider ============================================= The light curve fits are done differently for different SNe. In some cases s_V=s_B, in other cases, these values are left free. How is the cosmology affected if we restrict the fits so that s_B=s_V for all light curve fits? I think that you are planning to do this (p. 8 of your methods document).