Dear Editor and Referees,

We would like to thank the journal and the reviewers for their time
and effort to improve our draft. We have interspersed our responses to
your comments starting with a "==>" and ending with a "<==". If anything is not
clear or if you have any questions please feel free to contact us.

=====================================================================

Re: LK12305
    Search for supersymmetry with gauge-mediated breaking in diphoton
    events with missing transverse energy at CDF II
    by T. Aaltonen, J. Adelman, B. \'{A}lvarezGonz\'{a}lez, et al.

Dear Dr. Lee,

The above manuscript has been reviewed by our referee(s).  We ask you
to consider the enclosed comments from the report(s).

While we cannot make a definite commitment, the probable course of
action if you choose to resubmit is indicated below.

(x) Acceptance, if the editors can judge that all or most of the
    criticism has been met.

( ) Return to the previous referee(s) for review if available.

( ) Submittal to new referee(s) for review.

With any resubmittal, please include a summary of changes made                 
and a brief response to all recommendations and criticisms.

Yours sincerely,

Robert Garisto
Associate Editor
Physical Review Letters
Email: prl@ridge.aps.org
Fax: 631-591-4141
http://prl.aps.org/

Physics - spotlighting exceptional research: http://physics.aps.org/

P.S. Another referee was consulted but we now assume that no
     report will be received.  If a useful report is received,
     we will contact you.

==> Good to know. Thank you. 
<==
----------------------------------------------------------------------
Report of Referee A -- LK12305/Aaltonen
----------------------------------------------------------------------


In this paper, the results of a search for gauge-mediated
supersymmetry with the CDF experiment are reported. Without doubt
supersymmetry is one of the most discussed extensions of the standard
model, and in particular models with gauge-mediated supersymmetry
breaking are in the focus of theoretical and experimental particle
physics. The channel investigated here, consisting of two photons plus
missing transverse energy, can be considered one of the benchmark
channels for new physics at colliders.

Similar results have been reported before both by the CDF and D0
collaborations.

However, several substantial improvements are included here. The used
datasets are about an order of magnitude larger than previous CDF
results, and more than a factor of two compared to the most recent D0
result. This is the main reason why this new result can improve the
mass limit on the lightest neutralino from 125 GeV (D0) to 149 GeV.
Some of the improvement is due to new experimental techniques like
photon timing which are however also described elsewhere. Presenting
the results also as a function of the neutralino lifetime is
additional information useful to constrain new physics. All in all,
the improvements are certainly large enough to make this an important
result for the larger community, and I fully support publication in
PRL.

==> Thank you. 
<==

The quality of the paper is already very high, and I have only minor
suggestions for improvement. The analysis is scientifically sound, and
most details (exceptions below) are described in sufficient detail.
The text is accessible to the general reader.

==> We greatly appreciate the extremely thorough reading and thoughtful 
comments of the refree. 
<==

The method to predict the missing ET distribution of the standard
model background (called "METMODEL" by the authors) is hard to
understand. It remains unclear to me in how far there remains a
dependence on the detector simulation with this method. 

==> We agree that this sophistcated modeling of the MET is difficut to
describe in the short confines of a PRL length manuscrip. It is indeed
unfortunate that the PRD with more details on the method was not submitted 
simultaneously. However, that paper has now been submitted and will be
referenced with its full arXiv citation, which is arXiv:0910.5170. 
In addition we have worked to make the text clearer on these issues.
We note, as described in the arXiv, that the METMODEL is not designed to 
predict just the missing ET distribution. Rather, it estimates the significance of
the measured MET according to the measured variations of the measured jets and
underlying event from the event itself. Prameterization of jet energy resolution 
(JER) functions is based on dijet and Z-jet MC samples. Possible differences in 
JER between the two are studied by comparing dijet and Z-jet balance in
data and MC. We observe a good agreement between balance results from data and
MC. Nonetheless, we still assign a small systematic uncertainty (~5%) on
modeling of JER in MC. The entire performance of the METMODEL is checked on
Z->ee events in both data and MC. We also found a good agreement in these two
tests. We have also repeated the entire analysis on the control sample of
loose diphoton events in data. This sample is made of events where both
photons pass relaxed ID requirements, but at least one of the photons fails
tight ID requirements used in our reported analysis. The diphoton control sample is 
dominated by events with at least one fake photon. It contains only ~5% or real 
diphoton events. Overall, all performed tests indicate that there are no indications 
of unaccounted data-MC differences in the performance of the METMODEL. To help make 
this more explicit we have added a line on page 2, line 56:
"Uncertainties are dominated by the statistics of the MC sample"
-->
"Uncertainties are dominated by the statistics of the MC sample, but also
include the small differences between the measured response of the METMODEL to
MC simulation events and real data" 
<==

Also the selection criteria used for the samples in this method remain unclear,
as well as the determination of systematic uncertainties. All of this
can probably be improved by small modifications to the text. One
underlying problem here is that in many places a reference is made to
[15], "Search for Anomalous Production...", a paper that is however
not yet submitted for publication!

==> The systematic uncertainties are from the one sigma variation of the parameters 
of jet resolution functions and MET resolution due to smearing by unclustered
energy (underlying event and multiple interactions). Due to the limit of paper 
lenth further details of the description is not possible. However, we are pleased 
that the PRD has been submitted as noted above. This provide the full details, 
and, along with the above text addition, should help.
<==

Backgrounds from cosmic rays and beam halo:

- Cosmic ray induced events are selected by an out-of-time
requirement. Then, it is stated that "Similarly, we create a beam
halo-enhanced sample." Isn't beam halo typically in-time with the
beams? If so, the selection of the beam-halo sample cannot be similar
in that respect, and the text should be corrected.

==> This text was not clear enough so we have fixed as follows:
"Similarly, we create a beam halo-enhanced sample."
-->
"Similarly, we create a beam halo-enhanced sample from events that were
filtered out from our signal sample by the beam halo rejection requirements." 
<==


- It is stated that the ratio of out-of-time to in-time (signal
region) events is used to estimate the non-collision background. Is
this really what has been done? Isn't one rather extrapolating from
out-of-time to the in-time region?

==> The reviewer is correct that the two methods are essentially the same.
Since the extra detail (especially for such a small background) was confusing
we have elected to remove it. We have simplified the sentence more as follows:
"We estimate the non-collision background events in the signal region using
the ratio of events outside the timing requirements to events inside the
signal region and using the measured efficiencies of the non-collision
rejection requirements [15]"
-->
"We estimate the non-collision background events in the signal region using
extrapolation techniques and the measured efficiencies of the non-collision
rejection requirements [15]"
<==

The authors write that they simulate "events from all SUSY processes
considered". It would help the reader if at least the most important
ones are listed here.

==> This was explicitly mentioned on page 1, in the introduction, line 30 that 
"At the Tevatron sparticle production is predicted to result primarily into gaugino pairs...".
<==

The authors write that "the high-ET diphotons with large HT from SM
background are mostly back-to-back with fake MET; the delta phi
(gamma1, gamma2) requirements... reduces... these backgrounds." The
delta-phi requirement is delta phi (gamma1, gamma2) < pi - 0.35, i.e.
against back-to-back *photons*. It seems to me that if the diphotons
are back-to-back against the fake MET, they are typically *not*
back-to-back with respect to each other. Somehow the statements as
they are written do not appear to be logical. I can imagine the
authors wanted to say that *one* of the photons is back-to-back to the
fake MET?

==> The refree is correct that the sentence is ambiguous, and could be
potentially be interpreted as meaning as saying that the photons
were opposite to the MET. We fixed it to describe what we did more explicitly as follows:
"Also, the high-ET diphotons with large HT from SM background are mostly
back-to-back with fake MET."
-->
"The SM backgrounds with fake MET and large HT also have photons which are
mostly back-to-back." 
<==

Ref. [10] are the LEP results on GMSB searches. Could the authors
please make sure that they are consistent with the final LEP results?
On first sight, it seems that they are different from what is given on
the LEP joint SUSY working group at http://lepsusy.cern.ch. References
to journal publications by the collaborations should be preferred over
conference proceedings.

==> The referee is correct that these needed to be updated. We have updated:
"A. Heister et al, (ALEPH Collaboration), Eur. Phys. J. C 25, 339 (2002)"
-->
"R. Barate et al, (ALEPH Collaboration), Eur, Phys. J. C 28, 1 (2003)"
We have left other three remain as they were (OPAL, L3, DELPHI). Two conference 
proceedings (OPAL, L3) are preferred over out-dated published papers 
by their authors since they are most up-to-date and contain more data. 
<==

Page 9: "3.1-m long drift chamber" -> "3.1 m long drift chamber"

==> After reviewing a couple of published PRL papers we decided not to put the
information on exact size of the detectors. So all 3-m, 5-m, and 3.1-m are
removed.
<==

Ref. [23]: remove upper page number

==> Fixed.
<==

----------------------------------------------------------------------
Report of Referee B -- LK12305/Aaltonen
----------------------------------------------------------------------


The manuscript "Search for Supersymmetry with Gauge-Mediated Breaking
in Diphoton Events with Missing Transverse Energy at CDF II" by the
CDF Collaboration presents the results of an updated search for
supersymmetry with gauge-mediated breaking in the diphoton plus
missing transverse energy final state using 2.6 fb-1 of p pbar data
taken with the CDF II detector at the Fermilab Tevatron collider.
Models with gauge-mediated supersymmetry breaking (GMSB) solve the
naturalness problem and provide a low-mass dark matter candidate that
is consistent with cosmological observations. In such models the
lightest neutralino decays to a stable gravitino and a photon leading
to missing transverse energy due to the unobserved gravitino.

The measurement appears to have been carried out in a careful manner
using the already well-tested CDF detector and analysis algorithms.
The analysis benefits from a new photon timing system and a new model
for the missing transverse energy resolution. The search is also
extended to neutralino lifetimes up to 2 ns.

Standard Model backgrounds include production of gamma-gamma,
gamma-jet, and jet-jet final states, where jets can be misidentified
as photons. The missing transverse energy can be real (neutrinos) or
fake, due to mismeasurement in the calorimeter or fluctuations.

Cross section limits are set and the prediction of the number of
events is obtained from a GMSB model. No events in the data pass the
final event selection. The neutralino mass lower limit is determined
to be 149 GeV for a lifetime << 1 ns, as compared with an expected
limit of 141 GeV.

This manuscript is recommended for publication in Physical Review
Letters after the following comments are addressed or corrections
made.

==> Thank you.
<==

The authors should address the sensitivity of this search as compared
with the most recent D0 result (Ref. 9), which had an expected lower
limit on the lightest neutralino mass of 135 GeV with 1.1 fb-1 of
integrated luminosity, as compared with the expected limit presented
here of 141 GeV with more than twice as much data.

==> The text reads:"The m_chi reach, based on the predicted and observed
number of events for tau<<1 ns, is 141 GeV and 149 GeV, respectively. These
limts significantly extend the search sensitivity beyond the results of D0[9]",
which explicitly points out that the expected sensitivity is better than D0,
and gives the reference for more information. Additional information about the
D0 result, when no more is given about other experiments, seems unwarrented
and we have left the text as is. For the edification of the refree we calculated 
a quick sensitivity with 1.1 fb-1, by scaling down all background prediction 
linearly with fixing their uncertainty ratio. With this simple assumption we got 
our neutralino lower mass limit as 125 GeV, which looks worse than 135 GeV from D0. 
The fact that D0's expected limit is still better than CDF's can be explained
by the exceptional efficiency and rejection power of their calorimeter. For example 
their signal acceptance is about 20%, which is 2.5 times lager than 8% in our 
analysis. We note that the W(tau,nu)+photon (where tau->fake photon), Z(mumu)+dipho 
and Z(tautau)+dipho events are considerable sources of background. However, we 
couldn't find any indication of these sources in the D0 paper if these sources
were considered or taken care of in their analysis. 
<==

p. 8, line 4: "compelling rationale" should be "a compelling
rationale"

==> Fixed.
<==

p. 13, second paragraph, line 2: "is well modeled" should be "are well
modeled" ("data" is plural)

==> Fixed.
<==