Abstract: the two consecutive adjectives in "the low-luminosity dwarf galaxy population" are a bit redundant. I realize they do not mean the same but perhaps for the abstract just low-luminosity would suffice. --Fixed Intro: 2nd paragraph: Since when you talk about the origin of the UV and Halpha light you mention OB stars perhaps it would be better than when you talk about the origin of the nIR emission instead of "old stellar populations" you say "old giant stars" instead, for consistency. --Do you know of a reference that quantifies the fractions of NIR light from giant stars and average/small stars? The Sample: 1st paragraph: When you mention the ANGST survey you only talk about it augmenting the database with HST data. However, there are also UV observations of the LVL that were taken as part of a GALEX GI proposal led by the ANGST team. The PI was E.Skillman (GI Cycle 3, program 061). --Done Section 3.2: Last paragraph: I would say: "the correction for the starburst galaxy NGC3034 is 83%" --Done Section 4: 2nd paragraph: This is weird. So, the 2MASS images do not come with the airmass term of the flux calibration already corrected? Just checking. --Yes. Here is an excerpt of the relevant email: =============================================================== From: Tom Jarrett Subject: Re: 2mass airmass Date: Fri, 28 Jul 2006 11:45:25 -0700 Yes, the 2MASS photometry is calibrated to an airmass of 1.0. =============================================================== Section 4.1: 1st paragraph: I would say "from the Extended Source Catalog are 0.5-2 mag lower than what one would expect from the extrapolation of the IRAC 3.6 and 4.5 um data assuming a ...???..." Is this what you meant right? --Yes. I say "similar extrapolations" instead of the additional wording you give above, since it would be redundant/repetitive with the previous sentence. I hope it's clear. Section 4.3: This is where most of my questions arise. Some of them philosophical, so please ignore them if you think they make no sense. First, I think we should say what "undetected" means. Is when the flux is below the level of the sky and our attempt of computing the aperture fluxes give a NaN, or is it when we just do not see (by eye and without smoothing) the source, or it is when the flux (even if we measure it) it is below the upper limit given below? I would think that the latter option is what it makes most sense, but it is not clear for the text that is what it was meant. --Excellent point. I have added to the text that the measured fluxes are indeed smaller than the 3sigma estimates for our upper limit cases. In this same section. I just wonder it would be adequate to provide a reference for this expression for the upper limits. I believe this expression is valid for an extended source filling up the aperture. For a point source the upper limit would be lower, right? Perhaps we should mention something like that in the text. --I have modified the text here. Finally, still regarding the upper limits. I find the way the limits for 2MASS were computed should be explained in more detail. In particular how the conversion from a PS limit to our XS limits were obtained. Perhaps it is trivial but I think it is worth a couple of sentences. --Good suggestion. I've simply added the relevant equation. Your question prompted me to find an error in my math, too -- thanks! Section 5.1: 1st paragraph: "the the" -> "the" --Thanks Section 5.3: Footnote: You should add the latest GALEX GI Cycle 4 awarded to 11HUGS (PI: J. Lee, GI4 - 095). --Done In this same section, when you add a random (gaussian) error to the MIPS data to do your fits, these are uncorrelated errors, right? It might be worth adding that word. You do not expect the errors in the three bands to be correlated, e.g. they do not include calibration errors (which might or not be correlated), right? --Stansberry et al. (2007) explain that the 160um calibration is tied to the 24 and 70um calibrations. I've done Monte Carlo simulations where I take this aspect into account, and I've done simulations just blindly using Gaussian deviates of the total ef24, ef70, and ef160. These details don't make much difference in the fits (i.e., the statistical portion of the uncertainty is significant), and I'd rather not get into so many details when I ultimately don't analyze the fits themselves. Last paragraph: I assume that for figure 10 you will only use one panel in the printed version and all in the electronic one (as a figureset). If that is correct perhaps it would be good that that single panel in the printed version would include the examples you give in this paragraph for each of the SED types and the leave them the way the are now for the electronic edition that has all objects. --That's a good idea. I haven't decided yet which way to go with this figure, but I'll keep your suggestion in mind. Section 5.4: 1st paragraph: I agree there is a significant difference in the dispersion between the f8/f24 compared to the f70/f160 ratio but the numbers given are rounded up for the former and down for the latter. Perhaps we could relax a bit the sentence. --Done (I added "nearly" to the 8/24 description). In this same paragraph: "the the" -> "the". --Thanks In this same paragraph perhaps we could emphasize that we do not have strong AGNs just because these are rare in a volume-limited sample like ours. --I've tweaked the wording. 2nd paragraph: The comments regarding the monochromatic-to-bolometric ratios seem to imply (at least in some parts in the text) that the total IR is better indicator of the SFR than the monochromatic estimates. However, it is not that clear since, e.g., the 24um luminosity has been found by several authors to better trace/estimate the current SFR than the FIR and certainly better than the TIR. --I understand your concern. I re-read this (sub)section, and I think the way it is worded is appropriate: though I do mention that people use TIR and FIR for SFR, the remainder of the (sub)section focuses simply on how closely monochromatic fluxes trace TIR. It's left to the reader to infer any implications. In other words, I don't say "fnu(X) is better for SFR since sigma(fnu(X)/TIR) is smallest." 3rd paragraph: When you say that one possibility to explain the fact that the f_8dust/f_TIR is very large is that late-type spirals and irregulars are less abundant in heavy metals I get a bit confused. In principle the f_8dust/f_TIR to first order is not so dependent on the absolute abundance of heavy metals as both the dust grains responsible for the FIR emission and PAHs are made of metals, the former a bit more based on Oxygen and the latter on Carbon, but they both need them. As you perhaps remember, Juan Carlos Muņoz-Mateos suggests in his paper that this could be due to the earlier release of Oxygen by SNII (mainly) compared with the delayed release of Carbon, mainly by AGBs. Of course, the issue of the harder radiation fields is the another possibility that you already comment. --Figure 21 of Draine et al. (2007) shows a clear difference in q_PAH for Z<8.1 and Z>8.1. As Bruce said in his email to Juan Carlos, it's unclear why this is the case, but it is an empirical fact. What I don't understand is why Juan Carlos did not see the same trend for the radial decompositions (and why are the metallicities so much higher than in Bruce's plot?) -- did the low-Z dwarfs not make it into Juan Carlos' Figure 12? 4th paragraph: I would add "and a dispersion of 0.15 dex for a given 70/160 ratio" since I believe that is what it is, right? --No. It's the entire dispersion for all 70/160, not the dispersion for a given 70/160. Section 5.5: I would not say that the IR-to-UV ratio is a coarse measure of the dust extinction given that (I believe) is the best one we have. It is true that the relation to the reddening of the UV spectra is not so good but the attenuation it is (in my opinion). Perhaps you meant something like this "The IR-to-UV ratio is a measure of the dust extinction in the ultraviolet and thus can be related (coarsely though) to the amount of reddening in the ultraviolet spectra". --Karl said the same thing. I've updated the text accordingly. One of the issues here is the qualification for 'star-forming' galaxies that I added to the sentence -- some of the LVL targets are not star-forming, so we need to be a bit careful with the wording. 2nd paragraph: The latter option to explain the offset of the LVL galaxies compared to the "normal galaxies" in the IRX-beta diagram, namely "or have lower IR-to-UV ratios for a given ultraviolet color (less dusty?)" does not sound completely right: (Maybe I did not understand it correctly). If it would be only that then the UV slope should be also affected and be quite blue, so those objects would also move to the left in the diagram (except if their stellar populations are intrinsically redder, but in that case we would be in the former situation). --If there is reddening due to ageing, then we can still talk about low IR/UV without necessarily also requiring blue colors. I've tried to re-work the wording to make it more effective. 4th paragraph: Carlos-Muņoz, J. -> "Muņoz-Mateos, J.C." Below, "is a measure of the HII line emission" -> "is a measure of the line emission in HII regions" --Thanks 5th: This paragraph suggests that the star formation history effects on the IRX-beta are mostly related to effects associated to different amounts of dust attenuating the young and old stellar populations differently. However there is also the effect of the change in the intrinsic beta value of the stellar population, which is roughly FUV-NUV=0 for continuous SF or for a relatively young burst but that can get much redder (without involving any dust) for an old burst or for an episodic star formation history. --I agree, but I guess my wording wasn't clear. --> I said "...effects that young and old stellar populations ... have on the intrinsic ... luminosities..." which is my confusing way of saying that ageing effects can make them red. I've reworded things to be hopefully more explicit. For example, I've replaced 'intrinsic' with 'age-reddened'. Below, in the same paragraph, when you say "Assuming these theoretical spectra" I think we should add "and a continuous star formation". Just after this, you say "the star formation timescales range from ..." This sounds a bit confusing since these are actually ages (right?) and timescales here might get confused with the timescales (tau values) of exponentially decaying star formation histories. --Excellent suggestion. Done References: "Carlos-Muņoz, J." -> "Muņoz-Mateos, J.C.," I believe the reference on "Lee, H., et al. 2009, in prep.:" is missing. --Thanks Table 1: I would put the third column as "v_{$\odot$}" instead of cz. The z values from which you computed cz are heliocentric, right? --I prefer 'cz'. I've added 'heliocentric' to the table note. Table 2: Caption: I would say "IRAC imaging taken from the Spitzer archive" and "MIPS imaging taken from the Spitzer archive". BTW, are these images processed with the same pipeline versions as the LVL own data. It would good to mention it. --Done. I mention that we do the same data processing (and pipeline) in Section 3.3 Figure 2: There is one object in the bottom-left panel that belongs to 11HUGS apparently but has a |b|<30 dg. There should not be any object there, right? The same for the bottom-right panel were there are objects supposedly part of the ANGST that are beyond 3.5 Mpc. --Bottom left: Good catch -- my binning was such that an 11HUG galaxy just above 30 degrees fit in a bin that extended just above 30 degrees. Bottom right: I've updated the sample description to point out that ANGST includes an extension to the M81 group and the Sculptor filament. (ANGST objects go out as high as 4.2 Mpc). Figure 4: Maybe it would be nice to show an horizontal line that would show the flux of a 10sigma point source, just as a reference. --Interesting idea. That value is so low, though (0.0008) that I'm not sure it'd be useful. Figure 7: It might worth noting (maybe you did already and I missed it) that the highly deviant objects are either objects that could not be well isolated from a companion in the IRAS images (NGC5194 and NGC5195) or that are so large that likely suffer for aperture effects (SMC & LMC). Figure 10: Probably when refering to the GALEX data is better to use Lee et al. 2009 since Janice is the one writing the paper with the UV photometry and the comparison with Halpha. --OK, I'll leave it out of the caption and leave the text to refer to Lee et al. Last panel of this figure: The label "z" is on the bottom central part of the plot instead of in the bottom-left one as for all other panels. --Thanks Figure 11: Do the models used here (Dale & Helou and Dale et al.) include the stellar contribution to 8um? It might be worth saying whether or not that is the case. --Yes, they're dust-only at 8um. I've added that to the caption. I think it would be nice to add a couple of figures showing the dependence of these properties on the luminosity, either the B-band or (even easier/better) the 3.6 um luminosity, given that the LVL explores an unprecedented range in luminosities down to the least luminous galaxies known. --I've added a version of the color plot as a function of M_B, which shows that much of the 8_dust/TIR scatter comes from the faint dwarfs (Fabian asked for the same plot). Figure 12: It would be nice to express the right side of the y-axis in units of A_FUV using the expression by Buat et al. (2005). We should mention what attenuation law was used to relate the A_V arrow to the FUV-NUV color excess and the TIR/FUV (via A_FUV, I suppose) ratio. When the data from Calzetti et al. (1995) and Kong et al. (2004) is used I think it is fair to mention that these TIR/FUV ratios were not obtained through matched apertures since FUV comes from the IUE aperture and the TIR from the IRAS total fluxes. It would be also fair to acknowledge that the difference between the total and the aperture-based IR fluxes is probably not very large given that some of our points -those of the most actively star forming ones- fall close to the starburst relationship. --I had the same idea after reading Juan Carlos' draft. I've added this axis and explained the reddening vector in the caption.