Figure 5: I would like to see this same figure for only HDFN objects with secure identifiations in the literature for which we are positive the counterparts are the right one. It worries me that maybe the solid histogram shows lower values than the dotted one beacuse of random associations with faint foreground/background objects. Also in the caption say which are the 2 calibration uncertatinties, or refer the reader to a section explaining it.

- Observations and Data Reduction

- 3.1

I would not even mention the existance of the very early data we took with SO1 and SO2, and just say: "Data were obtain during 17 observing runs between October 2007 and February 2010". Also I think the correct reference for COSMOS is an Scoville paper, not Capak's (don't have internet as I write so can;t check it). 

In this section you should also say what callibrations we take and how often. Arcs and twilight flats at the beggining and end of the night, and standard star.

The flux limit should state if it corresponds to flux at the telescope, or at the top of the atmosphere, or outside the galaxy (i.e corrected for galactic extinction). Also at what airmass. I think it is more meaningful to show the median flux limit for the whole survey (computed from the error frames). Maybe we should show both, and that is a good oportunity to talk about the median atmospheric extinction that hits us.

- 3.2

I would capitalize VACCINE, and call it a "pipeline" more than a set of scripts an routines. I would rephrase: "Data reduction was carried out using the custom pipeline VACCINE. The pipeline consists of a series of scripts and FORTRAN routines developed at UT Austin by JA and GB."


--- skipping sections that I have not read yet


- Section 6: Sample Properties

I think your paper should definitively show the following plots:

flux versus redshift with overploted flux limit (I would recomend with different symbols for different types of objects and color coded by field)

observed redhsift distribution overploted over predicted redshift distribution from the Gronwall or Ouchi LFs. I will look at this in a lot of detail in the luminosity function paper, but I think your paper will show them and maybe say that we will discuss it in further detail in an accompaning paper.

numbercounts for OII and Ly-Alpha emmiters (i.e line flux distributions)

R-band distribution of both [OII] and LAEs

EW distribution of [OII] and LAEs accompanied by a discussion of our classification criteria.

(for the last 3 plots I would overplot the distributions for literature confirmed HDFN objects alone which hopefully will be similar those of the whole sample)

Also, although originally I was thinkg on doing the LBG color-color diagrams for our LAEs in my paper, I think it would be better to do that here. I want to focus my paper on the luminosity functions, the uv-slopes, and the escape fraction, so the overlap with LBGs won't fit that smoothly, and shoulod definitively go in your paper.

- Table 2: Add column with total area surveyed on each field. For simplicity I would merge this table with table 4. And state in more detail the number of each type of galaxies we detect. I would show per field the number of [OII], [OIII], Hbeta, CIV, and Ly-Alpha emitters taht we detect, putting in parenthesis by each number the number of objects in each category that show X-ray emission. 

- Table 3: I would replace all the "None" and the "-99" in this table with dashes. Also I think the name of the objects should say the field in which they were detected. I proposo H#, C#, M#, X#, for HDFN, COSMOS, MUNICS, and XMM respectively. Also, both fluxes and EW should show errors. The redshifts should show only significant figures, and we should estimate the accuracy to which we can measure them, a conservative estimate should include the wavelength solution rms and the uncertainty with which we recover the centroid of S/N=5 sources. Also, for comfort and to avoid paper waisting I would comment the bulk of this table in future drafts that you circulate leaving only a couple of rows.