Sept 15, 2002

Dear Dr. Leibert and Anonymous Referee,

We have resubmitted our manuscript entitled "The Complex Chemical 
Abundances and Evolution of the Sagittarius Dwarf Spheroidal Galaxy" 
by Tammy A. Smecker-Hane and Andrew McWilliam (MS #56093), to the
APJ web page. 

We have responded to all the referees comments, as summarized below.

Most importantly, we have included all the data (stellar photometry,
atmosphere parameters, line lists, equivalent widths, abundances, ages,
errors, etc.) relevant to the derivation of the chemical abundances 
discussed in the paper (Al, Na, Si, Ca, Ti, Fe, Y, Ba, La, Eu). 
In addition, we thoroughly describe the derivation of the stellar
parameters (temperatures, gravities, masses, and ages) and the 
chemical abundances.  This more than doubles the size of the present 
paper.  We had planned to submit a larger paper with the complete set 
of data to the ApJ Supplements, but because the referee objected to the 
data not being published simultaneously with the analysis (with merit!)
we have greatly expanded our present paper and will present the 
data on the remaining elements (O, Mg, Sc, V, Cr, Mn, Co, Ni, Cu, 
Zr, Mo, Ba, Ce, Pr, Nd) in a subsequent ApJ paper.
 
We have added a much more thorough discussion of the abundance work
of Bonafacio et al. (2000; 2 Sgr field stars) and Brown, Wallerstein
& Gonzalez (1999; 5 stars in M54, a Sgr dSph globular cluster)
in section 7.5. discussing how the data and analyses differ, 
and how the abundances derived are very similar. In addition, 
we have added points showing the Bonafacio et al. data on our figures.

We also have added Section 7.7 to the paper, which compares
the Sgr dSph abundances we find to those for other dSph stars
(Shetrone, Cote & Sargent 2001) and Galactic stars (Fulbright
2000, 2002) and discusses the constraints we can place on the
evolution of the proto-galactic fragments that merged to
form the Milky Way.

In addition, the referee wrote ...

> Since the observations were 
> taken (1996-1998), considerable progress has been made on the 
> atomic data for La and Eu by Lawler and collaborators.  Has 
> the present analysis used these new atomic data? 

In the original paper, we did not. But we have now used them.
For La, we use 5 lines in the abundance analysis. 
Lawler et al. (2001a) give new log gf values for only 2 of these 
lines. Neither log gf value changed appreciately (0 and 0.02 dex). 
For Eu, we use only 1 line. For this line, the new value of 
log gf = +0.12 (Lawler, et al. 2001b) is significantly different
than the old, log gf = +0.204, which we initially adopted for 
our abundance analysis. This change increased all our derived 
[Eu/Fe] ratios by 0.084 dex. 

As an aside, the referee wanted to know how bad the situation was 
with respect to the Ba lines being too strong for a reliable 
abundance analysis. We will reserve discussion of Ba for our future 
paper, but to satisfy his/her curiosity, the EWs of the Ba II lines 
at 5853.69, 6141.72, 6496.90, 6141.72 and 6496.90 A range 
from 150 to 400 mA.  They're very strong! 

> Nonetheless, the present observational data set does go 
> well beyond that of Bonafacio et al., and thus is of very
> high value to the community.  The Bonifacio paper presents
> only two stars of roughly equal metallicity.  The present
> paper reports the results for 11 stars of young age and 
> relatively high metallicity and 3 stars at intermediate and 
> old age and lower metallicity.   Thus the properties of the
> young population are very well constrained.  The weakness
> of the data set is the relative lack of older, metal-poor stars 
> (although Figure 1 shows that there appear to be plenty of them,
> so hopefully future observations will give a better view of
> the old and intermediate age stars too).  Herein lies a 
> another problem though - the dataset provides the basis
> for an excellent study of the young population in Sgr
> dSph, but one of the major emphases of the paper is on
> the evolution of Sgr dSph - which, due to the sampling,
> is very poorly constrained.

We deliberately chose our stars to evenly sample the width 
in color of the RGB, which is sizable, with the hope of probing 
the full range in [Fe/H]. What we wound up finding were predominantly
metal rich stars. There's nothing we can do about that. 
Note that since 1 of the 2 Sarajedini & Layden fields we used 
contain the M54 globular cluster, these would not be the data 
set one would choose to use to further probe the metal-poor
population because the giant branch of M54 contaminates 
half the sample. But other data sets, e.g., Bellazzini, et al.
do exist from which one could draw another sample. 
We heartily agree that more work on this galaxy is warranted!

> The lack of abundance for intermediate and old stars is
> made up for in some respects by including the clusters studied
> by Layden & Sarajedini (2000).  However, an age-metallicity
> relationship for Sgr was already derived by LS00 using both
> the clusters and the mean field components (their figure 18).  
> A discussion of what is different here from that presentation 
> is needed.

The difference is that we've derived [Fe/H] and abundance ratios
from spectroscopy, Layden & Sarajedini derived it photometrically
from the shape of the red giant branch. Thus we feel our discussion of 
the age-[Fe/H] and age-[alpha/H] relationships derived from the
giants are warranted. We have add to the text in section 7.1 
to stress this point.

> In fact, the two simple models are not that different
> (the effective yield from LS00 is about 50% higher than
> that derived here).  The present paper has two stars which,
> at first glance, are not well fit by the simple age-metallicity
> relationship, but when the uncertainties in the ages of 
> these two star are considered, it is clear that there is
> really no conflict at all.  Due to the lack of intermediate
> age stars in the present sample and the large uncertainty
> in ages, it appears that the study of LS00 provides better
> constraints on the metallicity evolution of Sgr.

Our age estimates errors are large for old stars (> 3 Gyr)
because the red giant branch isn't as sensitive to age as 
the main-sequence turnoff. But they're not that bad for the younger
ages. Also our metallicities are much better determined than 
LS00's photometric metallicities. So, again, we feel this 
discussion is warranted.

> While these are the main two points, some additional points
> need attention.  In Section 5.1, 2nd para, some comments are
> made about the relative numbers of metal rich and metal poor
> stars, contesting the comment that 80-90% of the SGR stars are
> metal-poor.  Note that the RGB has not been equitably sampled
> spectroscopically, so comments on the relative sizes of the 
> metal poor and metal rich populations should be offered with
> care.

Its sampled as equally as you can get with the current 
photometric data base (Figure 1) and the size of our sample. 
There isn't a substantial bias.

> In section 5.2, the comparison of the alpha/Fe vs. Fe/H
> to that of the Galactic disk seems a bit obscure.  For systems 
> where the bulk of the star formation has occurred more than 1 Gyr
> ago, you expect to find roughly the solar alpha/Fe at the 
> present Fe/H.  At lower Fe/H. you expected alpha/Fe to increase 
> to the SNII yield.  That is what is seen here.  By attributing
> this to "slower star formation" and discussing e-folding
> timescales, one is thinking in terms of implicit quasi-continuous
> star formation histories, which are not necessary for the
> interpretation of this result.

We agree. We noted in the paper that it could be an episodic 
SFR or one which operates on a slower timescale than the solar 
neighborhood.  We've beefed up this discussion substantially 
to get this point across. 

> Since a good deal of the alpha elements are produced in
> the 20 solar mass range (assuming standard IMF), it is not
> clear how strong a constraint can be placed on the universality
> of the upper IMF by the observed alpha/Fe ratios.  Calculations
> (or a reference to some) should be provided to support this
> "crucial point".

The referee was absolutely right. We had in my mind the
old yields of Woosley & Weaver (1992) for solar metallicity for
which the yields were a strong function of the IMF slope. 
We then calculated the yields from newer, low metallicity (Z = 10^-2
and 10^-4 times solar), type II SNe models by Woosley & Weaver (1995) 
assuming different IMF slopes, and discussed the 
constraint on the IMF in section 7.2.

> The proposed explanation for the fascinating case of the Al and 
> Na deficiencies, which is given as "a paucity of nucleosynthesis
> products from massive stars," is not really clear or convincing.
> Do the authors mean to imply preferential mass loss of SNII
> products, true variances in the yields, or what?

We mean that these metal-rich stars have been enriched by a 
fewer number of type II SNe, and more type Ia SNe, than stars
in the solar neighborhood at the same [Fe/H]. Again, we conclude
that the star formation rate has either been episodic (meaning
the dSph had periods with little or no star formation), or the 
star formation rate had a smaller e-folding timescale than that 
of the Galactic disk in the solar neighborhood, which 
Scalo (1986) suggests has had an (average SFR)/(current SFR) = 
2 to 3.  In short, for the metal-rich Sgr stars the 
average SFR/SFR at the time of their birth was higher than
in the solar neighborhood stars at the same metallicity.
We have added text to sections 7.3, 7.4 and 7.6 clears 
up this point. 

We've also shown that the deficits in the Na, Al, and alpha
element ratios are all qualitatively and quantitatively 
consistent with 60 to 70% of the Fe in the metal-rich Sgr population 
being synthesized by type Ia SNe and the remainder from type II SNe.
We've added a discussion of how this relates to the chemically 
peculiar stars that Fulbright (2000, 2002) sees in the Galaxy
that may be remnants of proto-galactic fragments that
formed stars on long enought timescales that type Ia SNe
have enriched them. In addition, we discuss the abundances seen
in the Sgr dSph stars, Fullbright's stars, and stars done in
other dSph galaxies by Shetrone, Cote \& Sargent (2001).

We thank the referee for his/her careful reading of the manuscript
and thoughtful comments, which undoubtedly have made this a 
better paper.

With regards,

Tammy Smecker-Hane & Andy McWilliam