
We consider the partial average i.e., the Lagrange average with respect to
{\it just one} of the two mean anomalies, of the Newtonian part of the
perturbing function in the threebody problem Hamiltonian. We prove that such
a partial average exhibits a nontrivial first integral. We show that this
integral is fully responsible of certain cancellations in the averaged
Newtonian potential, including a property noticed by Harrington in the 60s. We
also highlight its joint r\^ole (together with certain symmetries) in the
appearance of the so called "Herman resonance". Finally, we discuss an
application and an open problem.

In this paper we discuss about the possibility of {\it coexistence} of stable
and unstable quasiperiodic {\sc kam} tori in a region of phase space of the
threebody problem. The {argument of proof} goes along {{\sc kam} theory and,
especially,} the production of two non smoothly related systems of canonical
coordinates in the same region of the phase space, the possibility of which is
foreseen, for `properlydegenerate' systems, by a theorem of Nekhorossev and
Mi{\v{s}}{\v{c}}enko and Fomenko. The two coordinate systems are alternative to
the classical reduction of the nodes by Jacobi, described, e.g.,
in~[V.I.~Arnold, Small denominators and problems of stability of motion in
classical and celestial mechanics, 18, 85 (1963); p. 141].

Following the techniques of [4], we formulate a Normal Form Lemma suited to
close to be integrable Hamiltonian systems where not all the coordinates are
action angles. The Lemma turns to be useful in the theory of KAM tori of
SunEarthAsteroids systems.

We propose a new analysis of the twocentre problem particularly suited to
be used as a basis to study the dynamics of SunEarthAsteroid systems. Our
method, based on a tricky choice of initial coordinates, allows us to evaluate
the risk of collisions between the Asteroid and Earth. Moreover, it seems to be
fitted to prove the existence of chains of transition tori in the planar
SunEarthAsteroid systems.

We prove the existence of an almost full measure set of $(3n2)$dimensional
quasi periodic motions in the planetary problem with $(1+n)$ masses, with
eccentricities arbitrarily close to the LeviCivita limiting value and
relatively high inclinations. This extends previous results, where smallness of
eccentricities and inclinations was assumed. The question had been previously
considered by V.I.Arnold [2, Ch III, \S 1, n. 6, p. 128] in the 60s, for the
particular case of the planar threebody problem, where, due to the limited
number of degrees of freedom, it was enough to use the invariance of the system
by the SO(3) group.
The proof exploits nice parity properties of a new set of coordinates for the
planetary problem, which reduces completely the number of degrees of freedom
for the system (in particular, its degeneracy due to rotations) and, moreover,
is well fitted to its reflection invariance. It allows the explicit
construction of an associated close to be integrable system, replacing Birkhoff
normal form, common tool of previous literature.

In 1963, V. I. Arnold stated his celebrated Thorem on the Stability of
Planetary Motions. The general proof of his wonderful statement (that he
provided completely only for the particular case of three bodies constrained in
a plane) turned out to be more difficult than expected and was next completed
by J. Laskar, P. Robutel, M. Herman, J. F\'ejoz, L. Chierchia and the author.
We refer the reader to the technical papers \cite{arnold63, laskarR95, rob95,
maligeRL02, herman09, fej04, pinzarith09, ChierchiaPi11b} for detailed
information; to \cite{fejoz13, chierchia13, chierchiaPi14}, or the introduction
of \cite{pinzari13} for reviews.
The complete understanding of Arnold's Theorem relied on an analytic part and
a geometric one, both highly non trivial. Of such two aspects, the analytic
part was basically settled out since \cite{arnold63} (notwithstanding
refinements next given in \cite{fejoz04, chierchiaPi10}). The geometrical
aspects were instead mostly unexplored after his 1963's paper and have been
only recently clarified \cite{pinzarith09, chierchiaPi11b}. In fact, switching
from the threebody case to the manybody one needed to develop new
constructions, because of a dramatic degeneracy due to its invariance by
rotations, which, if not suitably treated, prevents the application of Arnold's
1963's strategy.
The purpose of this note is to provide a historical survey of this latter
part. We shall describe previous classical approaches going back to Delaunay,
Poincar\'e, Jacobi and point out more recent progresses, based on the papers
\cite{pinzarith09, chierchiaPi11a, chierchiaPi11b, chierchiaPi11c, pinzari13,
pinzari14}.
In the final part of the paper we discuss a set of coordinates recently found
by the author which reduces completely its integrals, puts the unperturbed part
in Keplerian form, preserves the symmetry by reflections and is regular for
zero inclinations.

We improve a result in [L. Chierchia and G. Pinzari, Invent. Math. 2011] by
proving the existence of a positive measure set of $(3n2)$dimensional
quasiperiodic motions in the spacial, planetary $(1+n)$body problem away
from coplanar, circular motions. We also prove that such quasiperiodic
motions reach with continuity corresponding $(2n1)$dimensional ones of the
planar problem, once the mutual inclinations go to zero (this is related to a
speculation in [V. I. Arnold. Russ. Math. Surv. 1963]).
The main tool is a full reduction of the SO(3)symmetry, which, in
particular, retains symmetry by reflections and highlights a quasiintegrable
structure, with a small remainder, independently of eccentricities and
inclinations.

The discovery in [G. Pinzari. PhD thesis. Univ. Roma Tre. 2009], [L.
Chierchia and G. Pinzari, Invent. Math. 2011] of the Birkhoff normal form for
the planetary manybody problem opened new insights and hopes for the
comprehension of the dynamics of this problem. Remarkably, it allowed to give a
{\sl direct} proof of the celebrated Arnold's Theorem [V. I. Arnold. Uspehi
Math. Nauk. 1963] on the stability of planetary motions. In this paper, using a
"ad hoc" set of symplectic variables, we develop an asymptotic formula for this
normal form that may turn to be useful in applications. As an example, we
provide two very simple applications to the threebody problem: we prove a
conjecture by [V. I. Arnold. cit] on the "Kolmogorov set"of this problem and,
using Nehoro{\v{s}}ev Theory [Nehoro{\v{s}}ev. Uspehi Math. Nauk. 1977], we
prove, in the planar case, stability of all planetary actions over
exponentiallylong times, provided meanmotion resonances are excluded. We
also briefly discuss perspectives and problems for full generalization of the
results in the paper.

I defended my PhD Thesis in Rome, Universit\`a Roma Tre, on April, 23, 2009,
under the direction of Professor Luigi Chierchia. The judging committee was
composed by Professors M. Berti, A. Celletti, C. Falcolini, J. F\'ejoz.
Professors M. Berti and J. F\'ejoz refereed my thesis. The main result of my
thesis is the first direct proof (the first general proof was given in [J.
F\'ejoz, ETDS, 2004]) of a famous statement by V. I. Arnold (1963), usually
referred to as "Arnold's Planetary Theorem". My proof of Arnold's Planetary
Theorem relies on the rediscovery, during the year 2008, of a symplectic set of
actionangle variables (described in \S 4 of my thesis) which perform
explicitly the reduction of rotation invariance of the system. Indeed, even
though in a different form, they had been previously considered by [F. Boigey,
Cel. Mech. Dyn. Astr., 1982] and [A. Deprit, Cel. Mech. Dyn. Astr., 1983]. The
version I found in 2008 corresponds to the "planetary" form of BoigeyDeprit
variables, since it includes the elliptic elements of the instantaneous
ellipses of the planets around the sun and for this reason is especially fitted
to this problem .
I then regularized "my" planetary variables to include coplanar and
cocircular motions. This regularization leads to a set of mixed actionangle
and rectangular variables analogous to Poincar\'e' variables but better fitted
to rotation invariance of the system, since they exhibit a cyclic couple of
conjugated variables. I finally applied my regularized variables to the
problem, checked nontrivial torsion and obtained the proof of the theorem.
I wish to thank J. F\'ejoz for mentioning my contribution to the proof of
Arnold's Theorem, and especially my rediscovery of Deprit's reduction, in his
paper [J. Fejoz, DCDSA, 2013].