One of the most interesting open questions in particle physics is the possibility that neutrinos have non-vanishing masses and, therefore, that oscillations among the different families may occur. Massive neutrinos are natural candidates for the hot component of the dark matter of the universe [1]. Neutrino oscillation has been long considered a likely explanation of the solar neutrino deficit [2, 3, 4, 5] and is also a plausible hypothesis to explain the atmospheric neutrino problem [6].
The interpretation of the present experimental data in terms of neutrino oscillation
is not a straightforward issue.
Matter-enhanced oscillation [7]
can naturally explain the solar neutrino deficit
with parameters 
- 
and
- 
[8].
The atmospheric neutrino deficit requires parameters
and
.
In addition, the recent claim from the LSND
experiment [9] suggests the existence of
oscillation with 
.
It is hard to accommodate the above parameters
with three neutrino flavours [10, 11].
One can resort to radical explanations such as adding a fourth (sterile) neutrino
[12] or disregarding some of the experimental evidence.
Several scenarios have been proposed, differing in their initial assumptions and predicting
rather different mass patterns [11, 13, 14, 15, 16, 17].
In addition, recent Cold-Hot Dark Matter
(CHDM) models [18] prefer neutrino masses in the
range of a few eV.
The scenario with two light neutrinos and one heavy 
in the mass
range of 
is plausible.
The CHORUS [20] and NOMAD [21] experiments,
presently taking data at the SPS were inspired
[19, 15] by the argument that a massive is considered a
good candidate for hot dark matter in the universe.
Relating this mass of the to the mass of the 
by the see-saw mechanism
the solar neutrino deficit would then also be explained.
In the near future, important new results are expected to shed some light on
the experimental status of the various hints for neutrino oscillation.
The SNO experiment
[22],
and in particular their measurement of the neutral current rate,
can help to decide whether neutrino oscillation is the origin of the solar neutrino deficit.
The Borexino experiment
[23] can address the energy-dependence of the deficit, thereby constraining
oscillation-based explanations.
Forthcoming results from the Superkamiokande experiment and, in particular,
more accurate results on the zenith angle dependence of the 
ratio,
can clarify the atmospheric neutrino situation.
The CHOOZ experiment
[24] will soon measure the disappearance of
thus testing one of the possible origins of the atmospheric
neutrino anomaly.
More data from LSND will improve the understanding of their measurements and
an independent check of the effect will also come from the recently
upgraded KARMEN experiment [25].
CHORUS and NOMAD have recently reported no
evidence for oscillation at the level of the present best limit [26]
using only a small fraction of their available data.
They will reach
a sensitivity of 
for 
,
corresponding, within the above framework, to a with a mass
of about 10 eV.
Their sensitivity to a lighter , in the range of 
is 
.
This sensitivity can be greatly improved by a next-generation short base line
experiment (SBLE).
If oscillation occurs at the present best limit,
CHORUS and NOMAD would each find 60 events.
Even if a few 
candidates would be unambiguously detected, this would have
strong implications to our understanding of the neutrino oscillation scenario.
This would definitely call for a new high sensitivity exploration.
At present, a next generation of accelerator-based experiments is being
planned to search for neutrino oscillation.
At Fermilab, two new experiments,
MINOS [27] and COSMOS [28] are approved.
MINOS is a long base line experiment (LBLE), which will use a
beam from Fermilab to the Soudan Mine 730 km away, to test the hypothesis of neutrino
oscillation with 
.
COSMOS is a short base line experiment, similar
in design to CHORUS, but expecting a total data sample one order of
magnitude larger.
Several LBLE's [29, 30, 31] have been proposed
using a beam from CERN to the Gran Sasso Laboratory in Italy, 732 km away.
A combination of a LBLE and an intermediate base line experiment has
also been proposed recently [32],
which would use the present SPS beam and a
detector of 400 tons located in the Jura mountain, 17 km from CERN,
to explore neutrino oscillation in the region of .
We believe that a next-generation 
short base line experiment
(SBLE), in the CERN-SPS neutrino beam can be designed based on a hybrid detector.
One can combine an emulsion target weighing a few tons and a high-resolution
electronic tracker resulting in an experiment with 20 times
more sensitivity than CHORUS and NOMAD and twice the sensitivity of COSMOS.
Such an experiment will have a high discovery potential as it will be sensitive to
at full mixing, while simultaneously
being able to explore a domain of higher
down to mixing angles of the order of 
.
One hundred unambiguous events would be observed in this experiment if
five events were detected in CHORUS and NOMAD.
Some analyses of present data suggest that oscillation may be found at a
and with small effective mixing angle (i.e. 
).
To illustrate this point we briefly describe in an appendix
several recent studies which attempt to explain all presently available results.
In this scenario this new SBLE will detect about 20 candidates.
In conclusion, neutrino oscillation is an appealing field of study, with a number of experiments either taking data or planned. The situation is rapidly evolving. New indications as to which regions are more relevant to be explored are expected to be refined in the near future. Candidate events from CHORUS and NOMAD will contribute to clarify the neutrino oscillation scenario, naturally calling for a follow-up experiment.
This document is organized as follows. In section 2 we briefly describe the activities at CERN which have lead to the design of this experiment. In section 3 we describe the improved SPS neutrino beam to be used by this experiment. In section 4 we present the basic design of the experiment and the principles of the search. The different detectors of the apparatus are described in section 5. Improvements still needing feasibility studies are presented in section 6. In section 7 we show the expected event selection efficiency, background estimates and the achievable results. These results are shown in terms of the expected exclusion plots to illustrate the reach of the experiment in a model-independent way. The conclusions are given in section 8.