Mail to:
nishikaw, gajewski, wins%"hsobel@uci.edu", oyamay, kenzo, kanzaki, 
wins%"suzuki@phys01.phys.kobe-u.ac.jp", icrkm1::suzuki

Gentlemen,
 This is a short note on what we have discussed. 
Regards, 
Sakuda
-------------------------------------------------
Informal meeting on Front Detector (FD) design
April 5, 18:00 , 1995 (after E362 meeting)
Present: H. Sobel, W. Gajewski, M. Sakuda, K. Nishikawa
April 6, 14:00 , 1995 
Present: H. Sobel, W. Gajewski, M. Sakuda 

(1) Nishikawa: Can we use Liq.Scintillator (Carbon) as target instead of
water target? Then, the design should be easy.
Sobel: We will get an imediate criticism that the cross-section for
water (oxygen) is different from that for Carbon, even if the theory 
may predict a small difference.

(2)Gajewski: The neutrino flux per 1mrad (seen from the target point) is 
different at L=300m, 500m and 250km(SK). Relative corrections are 
2.2, 1.5 and 1 (normalized), respectively. This means that the beam
content is complicated and thus the extrapolation from FD to SK may not
be justified. We must use pi0 events (NC) as normalization.
(Ohyama showed at today's meeting that pi0 events will be about 25 
at 10**20 POT. 20% statistical error at SK.) 
Nishikawa: The neutrino energy spectra at L=300m, 500m and SK are nearly
the same within 5% as shown in the proposal. The corrections can be 
understood if we consider the fact that the neutrino source positions
(=pi-decays) are about 100m downstream from the target. Sobel/Gejewski agreed.

(3)Gajewski: It is very hard to design FD which can measure both QE events
with small background (say <20%).
Sobel: Is it necessary to measure QE events at FD? Why not measure just
muons in an inclusive way with water-tank detector
and compare the rate CC/NC both at FD and SK?
Sakuda: With water-tank as FD, CC should have >15% error on N(target)
 (=fiducial volume) due to poor VTX resolution. NC may have the similar error. 
On the other hand, SK has large volume and the fiducial volume is determined 
with 5% accuracy or better. So, the ratio CC/NC behaves quite differently
at FD and SK. Further, NC at SK has larger statistical error than 20% 
untill 10**20 POT. For the first two years, we must use the extrapolation 
of the beam from FD to SK. For that purpose, we must demonstrate that 
we understand the neutrino energy spectrum very well at least by measuring 
QE events (with E-mu and cos-th) at FD. All past neutrino oscillation 
experiments did it. 
 I showed at the meeting that the momentum spectrum of QE muons reflects
the oscillation pattern of parent neutrino and that the sensitivity is
very good. At SK we can select QE events with small background. So, I think
 it consistent to measure the momentum spectrum of muons from QE 
reactions at FD (with water as target). It is the simplest and
 best-tested reaction. This feature does not hold for inelastic reactions. 
  I will estimate the background (inelastic events) to QE events with
 a simple design of FD(fine-grained type) using Monte Carlo.

Gajewski: The design that I showed at the meeting is the best at present.
Sakuda: At L=300m, the weight of FD can be 4 times less than that at
L=500m. I will seek a possibility of L=300m.

(4) Sobel proposed to set the chambers between beam dump and FD.
They are set in the direction of the beam center as well as 
in 2-3 mrad off axis. They will monitor the muon rate of the neutrino 
interactions at the wall. They will show the uniformity of the beam
with respect to the angle (<3mrad). It will give another confidence
on the validity of the beam simulation. 

(5) Gajewski checked the momentum spectrum of muons from inelastic
channels and compared it with that presented by Sakuda. Differences
between two Monte Carlo programs exists but they are not very big.
 
(memos by Sakuda, I am sorry if the memo is not very accurate.)