1. As you suggested, we carefully proofread the paper and corrected some
   English mistakes.



2. We completely agree with you that it is important to describe 1kt
   detector calibrations. We would like to add a paragraph to the
   detector description on the 3rd page.



3. Yes, we think that the list of systematic uncertainties in Table 2(A)
   is important.  For the DIS and coherent pion production cross section,
   the difference between models is adopted as the systematic
   uncertainty. For the CC component subtraction error, the uncertainty
   on CC/NC cross section ratio is estimated based on TABLE I in
   Phys.Rev.Lett.41(1978)728. For pion nuclear effects in O16, Figure 4
   in Nucl.Phys.Proc.Suppl.112(2002)171 was used. We would like to make
   a modification in Table 2(A). The expression in the bottom row
   "Pion production outside ..." was incorrect and has been fixed to read
   "Pion interaction outside ...". This systematic uncertainty was
   estimated by comparing our calculation with data in
   Nucl.Phys.A350(1980)301 and Nucl.Phys.B76(1974)15. Though effects of
   the other items are small, uncertainties on MA and total cross section
   were estimated by referring Nucl.Phys.Proc.Suppl.112(2002)77 and
   Nucl.Phys.Proc.Suppl.112(2002)171, respectively. We have added this
   information on the 4th page in the last paragraph.



4. Our explanation was not sufficient. Though the reconstructed shift
   from the true vertex is rather large (about 20cm) for single pi0
   events, the size of the correction is small (2%). The vertex
   distribution of neutrino events within the detector, and thus pi^0
   events, is very flat. Therefore, the number of events brought into the
   fiducial volume from outside is almost same as the number of events
   leaving the fiducial volume due to this reconstruction bias. Since the
   above two numbers are almost same, the fiducial volume correction is
   only 2%. Since the reconstructed vertex distribution of observed pi0
   events is in good agreement with that of the Monte Carlo sample, we
   apply this 2% correction to the data and then estimate the systematic
   error in the fiducial volume correction by moving the boundaries of
   the fiducial volume along the beam direction by +-30cm, which is the
   vertex resolution of single pi0 events. We would like to modify the
   paragraph about the fiducial volume correction on the 5th page.



5. As the referee notes, the pi0 sample is selected from 2-ring events
   while a large part of the numuCC sample consists of single-ring
   events. As described in text, the vertex position of single-ring and
   multi-ring events are determined using different algorithms. In
   addition, pi0 events are kinematically selected using the pi0
   invariant mass cut. Therefore, the vertex determination of pi0 events
   has almost no correlation to that of numuCC events.  For the same
   reason (i.e. the kinematical selection from the pi0 mass cut), other
   systematic errors related to event reconstruction have no correlation
   between the NC1pi0 sample and the numuCC sample. In the last page, we
   add a sentence which explains how to combine the systematic errors. 



6. The systematic uncertainty on ring counting is estimated from
   differences in the likelihood distributions of ring candidates for
   data and Monte Carlo. The systematic uncertainty on PID is estimated
   from differences in the PID parameter distributions for data and Monte
   Carlo. The shapes of these distributions are affected by the amount of
   scattered light and other water parameters in calibration. We account
   for the effects to estimate the systematic errors.

   The energy scale uncertainty in the detector is not 1% but +2-3% as
   described in text. The uncertainty of energy scale affects the
   efficiency of pi^0 mass cut, where the effect is less than 1%.



7. As you pointed out, the pi0 momentum distribution for data and Monte
   Carlo simulation is not consistent around 300MeV/c and 600MeV/c if
   only statistical error bars are drawn. However, if we consider the
   systematic uncertainties on our neutrino interaction models including
   pion nuclear effects in 16O, these uncertainties change the shape of
   the distribution as shown by the size of outer boxes in Figure 4.

   In Figure 4, we normalize the MC events with the number of events in
   data. Therefore, the error boxes in the MC events represent only the
   uncertainty on the shape. So, the uncertainty which drives all the MC
   points higher or lower is not taken account in Figure 4. The largest
   error on the shape comes from the uncertainty of pion inelastic
   scattering in 16O. However this uncertainty affects little (less than
   1%) on the cross section measurement.

   The primary goal of this paper is to show the first result of high
   statistics NC1pi0 cross section measurement in water. We expect that
   it is the next stage to study the uncertainties of nuclear effect and
   tune the neutrino MC simulation to match the data. In order to clarify
   this point, we would like to modify the paragraph where we explain
   FIG4, the caption of FIG.4, and abstract.