Dear Editor, We have submitted a revised version of the following paper: MS#ApJ83620 Title: The WiggleZ Dark Energy Survey: Galaxy Evolution at 0.250.5 bins are sampled somewhat deeper, and the z<0.5 samples, somewhat shallower. The new results and conclusions are essentially the same as those in the first version we submitted. That is, the error in the k-correction has a very minor effect in our work (mostly small changes in the derived M* and f_red, but with very little impact in the conclusions). Tables 3 and 4, and Figures 11, 12, 13, 14, and 15 have been updated to reflect the changes in the numbers. In the text, bold-face fonts are used where the numbers are changed from the original version. Sections 5.2.2 (The Evolution of the GLF) and 5.3.1 (The Redshift Dependence of f_red) have been rewritten more extensively, and bold-face fonts are not used. We note that the changes in these two key sections are in the details, and essentially there is no changes in the conclusions. We summarize the changes below. For Section 5.2.2, the changes in M* measurements produce changes in the Q values (primarily producing a smaller Q for the blue galaxies). The most significant modification is that the Q value for the `Red' sample is now different from that of the `Blue sample' only at the 1.3 sigma level (instead of 2.8). Also, in comparing to the Faber et al. results, the difference is now primarily for the red galaxies only (at the 2 sigma level, compared to previously at 1.5 sigma), with the blue galaxies from both studies having consistent Q values. The discussion in Section 5.2.2 reflects the newly measured Q values and their significance. For Section 5.3.1, the conclusions are essentially identical, but with somewhat lower statistical significance. Furthermore, because we are able to sample deeper into the luminosity function for the z=0.7 bin, we now do our analysis of the redshift dependence of f_red using a magnitude limit of M*+1.4 instead of M*+1.0. This effectively changes the actual f_red values in the discussion by a small amount. The f_red at z=0.7 still differs from the lower redshift values at the 3.5sigma (instead of 4). We also modified our discussion of the comparison with the Iovino et al result using their deeper samples (to M*+1.5). The central results in the discussion in the section are identical to those from the earlier version. ___________________________________________ Replies to the Referee's comments: -------------------------------- > This paper is proposing to use the Wiggle-z survey to probe galaxy > evolution > together with the RCS2 dataset. Although I was skeptical at first, I > believe that > the authors have developed an interesting methodology, that could be > used for other > "biased" (in terms of their target selection) spectroscopic survey. I > thus recommend > publication after some improvements. > > My main concern is that everything in the paper assumes that "the marker > galaxies > and their neighbors are located at the same redshift". This is most > likely correct, > when looking at figure 5. However I think one would be even more > convinced if any > spectroscopic confirmation of this fact could be shown. Either by > looking at the > brightest pair (separated by less than 250kpc) within Wiggle-z data, or > using other > deeper spectroscopic surveys (assuming some overlap exists between > Wiggle-z and > other surveys, or by mimicking Wiggle-z selection in VVDS or COSMOS > surveys for > example). > It would be nice to be able to put some confidence level on how often this > assumption is correct, and whether we can safely do the proposed work, > despite their > exist some pairs that are pure chance alignment. > It would be also important to understand whether all marker galaxies > have a neighbor > and if not, what is the fraction of marker galaxies with one or more > neighbor. > Indeed within a distance of 250 kpc, it is likely that the marker and > the neighbor > galaxies are gravitationally bounded, and thus possibly sharing the same > dark matter > halo. So the evolution traced by this work would then be relevant to > gravitationally bounded system and may not represent the full diversity > of galaxies, > for example those galaxies with no neighbor. > We find that we can actually use the WiggleZ spectroscopic sample itself to confirm the statistical background subtraction method, due to the very large sample. We find that within the counting apertures of the markers, the ratio of wiggleZ galaxies associated with a marker (based on having delta z <0.002(1+z) from the marker) to the total number of wiggleZ galaxies within the aperture is statistically identical to the ratio of excess galaxies to total galaxies in the counting aperture using the photometric sample. We added a paragraph to Section 5.1 in the discussion of the robustness of the technique. Below we quote the key points in the paragraph: "... We perform this test by comparing the ratio of the net excess galaxy counts to the total counts within the $r_p=0.25$Mpc aperture in both the photometric and WiggleZ redshift samples..... The most significant selection effect that produces a significant difference in the redshift distributions between a $r'$-band selected sample and the WiggleZ sample is the low-redshift- rejection (LRR) applied to the selection of WiggleZ targets (see \S2.3). Thus, we limit our comparison to using data from WiggleZ markers in the three high-redshift bins of our sample ($0.45 My second main comment is the sudden drop of the f_red fraction at > z>0.7, there is > no good explanation it seems to me to such a drop, and I am worrying > whether this is > not due to some selection issues (for example the fact we may > considering only > system with at least 2 galaxies that are gravitationally bounded). > Although this is > probably difficult to disentangle with just the wiggle-z data, and one > would > probably need to compare the same strategy used here using a mass > selected or > magnitude selected redshift survey. > We have tested, as much as the data allow, whether the drop of f_red at z~0.7 is due to some selection difference in the z~0.6 and z~0.7 sample by choosing subsamples from these redshift bins that are common in absolute NUV and absolute r magnitude, and found that the same drop is seen when these are controlled (see the last 2 paragraphs of section 5.2.1). If there are other selection issues (such as bounded galaxies), they would be similar in the 2 redshift bins, and hence it would be unlikely the cause of the drop.. > I will know give my detailed comments. > > * section 2 > What is the typical spectral resolution of the blue and red arm. The resolution for the blue & red arms are 3.5 and 5.3 Angstroms. We have added this information in Section 2.1.2. "Using the 580V and 385R gratings for the blue and red arms with the 670nm dichroic, the spectra have a wavelength range from 4700\AA to 9500\AA, with a dispersion of $\sim$1.1\1AA/pix in the blue arm and $\sim$1.6\AA/pix in the red arm, providing spectral resolutions of $\sim3.5$\AA and $\sim5.3$\AA, respectively." > > I would suggest to show the g-r,r-z color-color diagram and the wiggle-z > color-box > used in the RCS2 fields. I would suggest to add the N(z) of wiggle-z > galaxies in > figure 1 (corresponding to the black points shown in figure 2). We are not sure what the referee is referring to in this comment. Figure 1 is the N(z) distribution of the WiggleZ galaxies in the RCS2 sample, which corresponds to the black points in Figure 2. We do not quite know what the first part of the comments refers to. > > The purple line is not visible in figure 2 - use a blue line instead? We have changed it into a blue color. > > * section 3 > > " we identify all galaxies within a comoving radius of rp Mpc": I > understand this is > angular distance on the sky converted to comoving distance at the > redshift of the > wiggle-z galaxy, correct? The way it is written we could think that it > is a 3D > comoving distance. We add 'projected' to make it clearer wherever applicable: Section 3.1: "...within a projected comoving radius of r_p Mpc to a WiggleZ galaxy." Section 4: "All the CCM cubes are made using galaxies within a projected comoving radius r_p=0.25Mpc from the markers." > > What is the magnitude difference distribution between the wiggle-z > galaxy and the > neighbor galaxy? What is the magnitude distribution of the overall > neighbors? How > does it compares to the full galaxy population? Is it representative? Generally speaking, the neighbors have a magnitude distribution similar to galaxies in the full RCS2 catalogs, ranging between ~18 to 24 (the cutoff) mag at low-z and ~21-24 at high-z; i.e., they have apparent magnitudes equivalent to an absolute magnitude range of about -23 to the absolute magnitude limit of each redshift bin (see Section 4.5). The WiggleZ galaxies, however, are 20 < r < 22.5 at all redshifts due to the survey selection, and hence have an effectively different (apparent) magnitude distribution from the neighbors. > *section 4 > > Calculation of the EW: I am a bit worried that you are using the noisy > continuum > spectra to characterized the OII EQW. Why not using the broad band > magnitude? The broad band magnitude allows us to work out the scale to obtain proper/reasonable flux (and other parameters such as SFR). This scale is cancelled out in computing the EW, i.e., the flux in the continuum and in the window where the line is. Also, the WiggleZ spectra are not flux calibrated, since the survey's main interest focuses on obtaining redshift. This makes using the broad band magnitude to derive the EW not practical. > > Figure 6: can you add the selection box of the wiggle-z galaxy in this > diagram? Or > add a new panel where you discuss the selection box. You say that the dusty > star-forming galaxies is insignificant for both markers, can you > indicate where they > would be expected in the color-color diagram? Figure 6 is updated with dashed boxes to indicate the regions for dusty galaxies for the three low-z bins. The WiggleZ LRR rejecting criteria are now overplotted as well. > The color scale is not fully used, it spans 0 to 50, but only the 0 to > ~30 is really > used in the display. What are those units? We have revised the scale of the plots, and also the plotting pixel size for better visual presentation. The unit is in net counts per pixel, where a pixel is 0.0125X0.0125 in colors. The original pixel size was 0.05 mag. To produce a smoother intensity contour diagram, we subdivide the larger pixels into 4X4, and then apply a 10X10 (small pixel) smooth kernel. A description of the smoothing is given in section 4.3.1: "For better visual presentation, the pixels ($\Delta(g'-r') \times \Delta(r'-z')$) in the color-color intensity plot are subdivided by a factor of 4 into smaller pixels in units of 0.0125 mag, and then smoothed by a kernel of 10$\times$10 pixels. The intensity scale is in units of counts per small pixels after normalizing the net counts to $1\times10^4$ in each redshift bin." Full range of the scale is used in the revised version, and the scale information is now included in the caption. A similar process is also performed for the color-magnitude diagrams (Figures 11 and 13). > > figure 8: the NUV magnitude range indicated are missing one significant > digit (for > example, in some panels we can read: -19 > I think the sentence saying that the properties of the neighbors are not > significantly dependent on the characteristics of the Wiggle-z galaxies, > is a bit > too strong. It is not by just looking at figure 7 and 8 we can conclude > this. It > would be important to quantify this better. It does also go somewhat > against one of > the conclusion stating that galaxy environments is important for galaxy > evolution. We believe we have addressed this point in the original version of the paper. We used the K-S test on collapsed g-r and r-z distributions of the different panels (within the same redshift bin) to test whether the color distributions from the different bins in Figure 7 are different. In Section4.3.2, we stated: "Kolmogorov-Smirnov tests of pairs of both $g'-r'$ and $r'-z'$ distributions find no significant difference between the different EW([OII]$\lambda$3727\AA) bins within the same redshift bin. The smallest significant level in all the pair-wise comparisons is 0.22. This implies that the properties of the neighbors do not strongly depend on the properties of the WiggleZ galaxies." and for Figure 8: "As in the case for EW([OII]$\lambda$3727\AA), we observe, and confirm with Kolmogorov-Smirnov tests, that within a fixed redshift bin the color distributions of the neighbors around markers with different $NUV$ fluxes are statistically identical." However, when data from all redshift bins are summed, we do find a 2sigma difference in f_red between neighbors of galaxies with strongest and weakest OII, we added at the end of Section 4.6 that: "Thus, there is some evidence that the neighbors of markers with the largest EW([OII]$\lambda$3727\AA~may have a lower average \fred, indicating regions around galaxies with large specific star-formation rates may have a larger fraction of star-formation galaxies." We do not see why the lack of strong correlation between the properties of the WiggleZ marker and the properties of the neighbors would "go against" the conclusion that environment is important for galaxy evolution. We believe this is simply a reflection that there is not a large range of environments for the WiggleZ galaxies. This is not unreasonable, since they are all chosen to be star forming. > Figure 14: the grey lines are barely visible. > The gray dotted lines are now replaced by black ones. > section 5: > > What is the redshift difference distribution between the marker and the > wiggle-z > galaxies within the rp=0.25 Mpc radius? Can we draw some conclusion > regarding the > assumption saying that neighbor galaxies have the same redshift as the > markers? > Please see our reply to the first comment. > Figure 15: the point at z~0.8 is not discussed in the figure caption > (but is in the > text). The purple and gray lines are barely visible (increase the width > of the > lines, and chose color that are easier to differentiate). > The Figure is updated with better line colors. We also enlarged the figure to make it more visible, since it is the main science figure of the paper. A caption statement for z~0.8 point has also been added.