External events are produced by neutrino interaction in the rock around
the detector so the effective target is larger. The detector sensitivity
depends on the surface and not on the active mass. Two kinds of events
can be detected: a) Stopping muons: low energy muons produced from neutrino
interaction in the rock and stopping inside the detector. In the interpretation
of the data there are problems connected with the low energy neutrino cross-section
and with a possible background due to large angle pions produced from the
cosmic ray muons. b) Through-going muons: upward-going muons produced from
neutrino interactions in the rock below the apparatus and crossing the
entire apparatus. If the oscillation hypothesis is true and with the value
suggested by internal events, it should be possible to see a clear signature
in the angular distribution of the up-going muons. In fact vertical neutrinos
(having a path length of the order of 13000 Km) should be attenuated (a
factor 2 for maximum mixing) , while horizontal neutrinos (having a shorter
path length) should not be attenuated. Data on up-going through-going muons
come from Kamiokande, Superkamiokande (April 1997), IMB, Baksan and MACRO.
The Baksan detector, located in the Baksan underground laboratory in Russia
at a minimum depth of 850 hg/cm2, consists of 3150 liquid scintillator
counters. MACRO consists of liquid scintillators and streamer-tube chambers.
It is located in the Gran Sasso underground laboratory. The neutrino induced
upward-going muons must be separated from a huge background due to down-going
muons. The selection is based on the directionality of the Cerenkov light
or on the time of flight measurement (Baksan, MACRO). Table
summarises the available statistics. The angular distribution of the data
is at the moment not conclusive for the presence of oscillations.
Table: Upward-going muons data. The prediction is based
on the Bartol neutrino flux. The theoretical uncertainty (17% ) is not
included in the errors.