424
M.
C.
Nucci
4
Singularity manifold equations and recip-
rocal aut
o-B
acklund transformat ions
In 1983 the concept of the singularity manifold equation
was
intro-
duced [44].
While it plays
an
important rBle in establishing if
a
partial differential equation possesses the Painlevd property, the sin-
gularity manifold equation did not seem
to
have other applications.
In
an
appendix
to
one of his papers
[38],
Weiss noticed that the sin-
gularity manifold equation of the Korteweg-de Vries equation could
be transformed into the Harry Dym equation via an inverse trans-
formation plus
a
change of dependent variable. In [36], Rogers and
Wong derived
a
reciprocal auto-Backlund transformation [19] for the
Harry Dym equation. All these results were found
to
be true for the
Korteweg-de Vries hierarchy [34]: the Harry Dym hierarchy was ob-
tained by applying
a
reciprocal Backlund transformation [19] to the
equations satisfied by the potentials of the singularity manifold func-
tions. In 2+
1
dimensions, substituting the Kadomtsev-Petviashvili
equation for the Korteweg-de Vries equation, Rogers [32] obtained
the (2+1)-dimensional Harry Dym equation and its auto-Backlund
transformation using the same approach. He noticed that the recip-
rocal
auto-Blklund transformation
was
induced by the invariance
of the singularity manifold equation under the Mobius group.
In
[33],
analogous results were obtained for both the Caudrey-Dodd-
Gibbon and Kaup-Kuperschmidt hierarchies and two different hier-
archies based on Kawamoto-type equations were derived. In [29], it
was shown that these results are not exclusive
to
few cases but they
are generalizable
so
that any singularity manifold equation generates
a
novel equation and its auto-Backlund transformation. Let us
see
how this works. In
1+1
dimensions,
a
singularity manifold equation
in
4(x,t)
is made of combinations of terms such
as:
4t
4X’
-
and their derivatives, which are invariant under the Mobius group: