Definition 2.35.
The second fundamental form of a distribution
$H$
on a Riemannian manifold
$(M,g)$
is the tensor
$\mathcal{S}$
(which takes any vector
$X\in {T}_{x}M$
at a point
$x\in M$
to a linear map
${\mathcal{S}}_{X}:{T}_{x}M\to {T}_{x}M$
) given by
$$g({\mathcal{S}}_{M}N,X)=\frac{1}{2}g({\nabla}_{M}^{g}N+{\nabla}_{N}^{g}M,X),g({\mathcal{S}}_{M}N,L)=0,g({\mathcal{S}}_{M}X,Y)=0,g({\mathcal{S}}_{M}X,N)=g({\mathcal{S}}_{M}N,X),{\mathcal{S}}_{X}=0,$$
where
$X,Y\in \mathcal{X}(\mathcal{\mathcal{F}})$
and
$L,M,N\in {C}^{\infty}(M,{F}^{\perp})$
,
${\nabla}^{g}$
denote the LeviCivita connection determined by
$g$
.
Definition 2.36.
A map
$f:M\to B$
of Riemannian manifolds
$M$
and
$B$
is called a Riemannian submersion, if the tangent map
$d{f}_{m}:{T}_{m}M\to {T}_{f\left(m\right)}B$
at any point
$m\in M$
is surjective and induces an isometric map between the normal space
${T}_{m}M/{T}_{m}{f}^{1}\left(f\right(m\left)\right)$
to the level set
${f}^{1}\left(f\right(m\left)\right)$
of
$f$
at
$m$
and the tangent space
${T}_{f\left(m\right)}B$
to
$B$
at
$f\left(m\right)$
.
Theorem 2.37.
A foliation
$(M,\mathcal{\mathcal{F}})$
is Riemannian if and only if there is a Riemannian metric
$g$
on
$M$
, satisfying any of the following equivalent conditions:

(1)
The distribution
$H={F}^{\perp}$
is totally geodesic.

(2)
The second fundamental form of
$H$
vanishes.

(3)
The induced metric
${g}_{\tau}$
on the normal bundle
$\tau $
is holonomy invariant:
${\stackrel{\circ}{\nabla}}_{X}{g}_{\tau}(M,N)=0$
for any
$X\in \mathcal{X}(\mathcal{\mathcal{F}})$
and for any
$M,N\in {C}^{\infty}(M,\tau )$
, where, by definition,
$${\stackrel{\circ}{\nabla}}_{X}{g}_{\tau}(M,N)=X\left[{g}_{\tau}\right(M,N\left)\right]{g}_{\tau}({\stackrel{\circ}{\nabla}}_{X}M,N){g}_{\tau}(M,{\stackrel{\circ}{\nabla}}_{X}N).$$

(4)
For any vector fields
$M$
and
$N$
, which are defined on an open set, are orthogonal to the leaves and are infinitesimal transformations of the foliation, and for any
$X\in \mathcal{X}(\mathcal{\mathcal{F}})$
, we have
$X\left[g\right(M,N\left)\right]=0.$

(5)
In any foliated chart
$\phi :U\to {I}^{p}\times {I}^{q}$
with the local coordinates
$(x,y)$
, the restriction
${g}_{H}$
of
$g$
to
$H$
is written in the form
$${g}_{H}={\sum}_{\alpha ,\beta =1}^{q}{g}_{\alpha \beta}\left(y\right){\theta}^{\alpha}{\theta}^{\beta},$$
where
${\theta}^{\alpha}\in {H}^{*}$
is the 1form, corresponding to the form
$d{y}^{\alpha}$
under the isomorphism
${H}^{*}\stackrel{\sim =}{\to}{T}^{*}{\mathbb{R}}^{q}$
, and
${g}_{\alpha \beta}\left(y\right)$
depend only on the transverse variables
$y\in {\mathbb{R}}^{q}$
.

(6)
$(M,\mathcal{\mathcal{F}})$
locally has the structure of a Riemannian submersion, i. e., for any foliated chart
$\phi :U\to {I}^{p}\times {I}^{q}$
, there exists a Riemannian metric on
${I}^{q}$
such that the corresponding distinguished map
$p{r}_{nq}\circ \phi :U\to {I}^{q}$
is a Riemannian submersion.

(7)
The adapted connection
$\nabla $
on the normal bundle
$\tau $
given by ( 2.2 ) is a (torsionfree) Riemannian connection: for any
$Y\in \mathcal{X}\left(M\right)$
and
$M,N\in {C}^{\infty}(M,\tau )$
$$Y\left[{g}_{\tau}\right(M,N\left)\right]={g}_{\tau}({\nabla}_{Y}M,N)+{g}_{\tau}(M,{\nabla}_{Y}N).$$

(8)
The holonomy group of the adapted connection
$\nabla $
on the normal bundle
$\tau $
given by ( 2.2 ) at any point preserves the metric.
Definition 2.38.
Any Riemannian metric on
$M$
, satisfying the equivalent conditions of Theorem 2.37 , is called bundlelike.
One can prove that a (torsionfree) Riemannian connection on the normal bundle
$\tau $
to a Riemannian foliation
$\mathcal{\mathcal{F}}$
is unique. It is uniquely determined by the transverse metric
${g}_{\tau}$
and is called the transverse LeviCivita connection for
$\mathcal{\mathcal{F}}$
. Thus, the transverse LeviCivita connection is an adapted connection.
Moreover, the transverse LeviCivita connection turns out to be a holonomy invariant and, therefore, a basic connection. In particular, this shows the existence of a basic connection for any Riemannian foliation.
The existence of a bundlelike metric on a foliated manifold imposes strong restrictions on the geometry of the foliation. There are structure theorems for Riemannian foliations obtained by Molino (cf. [
136,
135]
). Using these structure theorems, many questions, concerning to Riemannian foliations, can be reduced to the case of Lie foliations, that is, of foliations, whose transverse structure is modeled by a finitedimensional Lie group (cf. Example 2.41 ).
Example 2.39.
Any foliation defined by a submersion
$\pi :M\to B$
is Riemannian.
Example 2.40.
The orbits of a locally free isometric action of a Lie group on a Riemannian manifold define a Riemannian foliation.
On the other hand, flows, whose orbits form a Riemannian foliation, are called Riemannian flows. There are examples of Riemannian flows, which are not isometric. Concerning to Riemannian flows, see, for instance, [
32]
, and also [
137,AppendixA]
.
Example 2.41.
Let
$M$
be a smooth manifold,
$\mathfrak{g}$
a real finitedimensional Lie algebra and
$\omega $
an 1form on
$M$
with values in
$\mathfrak{g}$
, satisfying the conditions:

(1)
the map
${\omega}_{x}:{T}_{x}M\to \mathfrak{g}$
is surjective for any
$x\in M$
;

(2)
$d\omega +\frac{1}{2}[\omega ,\omega ]=0$
.
The distribution
${F}_{x}=ker{\omega}_{x}$
is integrable and, therefore, defines a codimension
$q=dim\mathfrak{g}$
foliation on
$M$
. Such a foliation is called a Lie
$\mathfrak{g}$
foliation. The class of Lie foliations was introduced in [
69]
.
Any Lie foliation is Riemannian. In the case
$\mathfrak{g}=\mathbb{R}$
, a Lie foliation is precisely a codimension one foliation given by a nonvanishing closed 1form. Actually, it can be easily seen that a codimension one foliation is Riemannian if and only if it is given by a nonvanishing closed 1form.
Example 2.42.
A foliation
$\mathcal{\mathcal{F}}$
on a manifold
$M=\stackrel{~}{B}{\times}_{\Gamma}F$
, obtained by the suspension from a manifold
$B$
and a homomorphism
$\phi :\Gamma \to Diff\left(F\right)$
of the fundamental group
$\Gamma ={\pi}_{1}\left(B\right)$
is Riemannian if and only if for any
$\gamma \in \Gamma $
the diffeomorphism
$\phi \left(\gamma \right)$
preserves a Riemannian metric on
$F$
.
Let
$\mathcal{\mathcal{F}}$
be a transversely oriented Riemannian foliation and
$g$
a bundlelike Riemannian metric. The induced metric on the normal bundle
$\tau $
yields the transverse volume form
${v}_{\tau}\in {C}^{\infty}(M,{\wedge}^{q}{\tau}^{*})={C}^{\infty}(M,{\wedge}^{q}{N}^{*}\mathcal{\mathcal{F}})$
, which is holonomy invariant and defines, therefore, a holonomy invariant transverse measure on
$\mathcal{\mathcal{F}}$
.
2.6 Symplectic foliations
Definition 2.43.
A foliation
$(M,\mathcal{\mathcal{F}})$
is called symplectic, if it has a transverse symplectic structure. In other words, a foliation
$(M,\mathcal{\mathcal{F}})$
is called symplectic, if there is a cover
$\left\{{U}_{i}\right\}$
of
$M$
by foliated coordinate charts,
${\phi}_{i}:{U}_{i}\to {I}^{p}\times {I}^{q}$
, and symplectic forms
${\omega}_{i}$
, defined on the local bases
${I}^{q}$
of
$\mathcal{\mathcal{F}}$
such that, for any coordinate transformation
$${\phi}_{ij}(x,y)=\left({\alpha}_{ij}\right(x,y),{\gamma}_{ij}(y\left)\right),(x,y)\in {\phi}_{j}({U}_{i}\cap {U}_{j}),$$
the map
${\gamma}_{ij}$
preserves the symplectic form,
${\gamma}_{ij}^{*}{\omega}_{j}={\omega}_{i}$
.
Definition 2.44.
A presymplectic manifold is a manifold equipped with a closed
$2$
form of constant rank.
A transverse symplectic structure on
$(M,\mathcal{\mathcal{F}})$
uniquely determines a presymplectic structure
$\omega $
on
$M$
such that
$T\mathcal{\mathcal{F}}$
coincides with the kernel of
$\omega $
. On the other hand, if
$(M,\omega )$
is a presymplectic manifold, then the kernel of
$\omega $
determines an integrable distribution on
$M$
and
$\omega $
induces a transverse symplectic structure on the corresponding foliation
$(M,\mathcal{\mathcal{F}})$
(cf. for instance, [16, 20] , in [16] symplectic foliations are called Hamiltonian). If
$(M,\mathcal{\mathcal{F}})$
is a symplectic foliation and
$\omega $
is the corresponding presymplectic structure on
$M$
, then the
$q$
form
${\wedge}^{q}\omega $
defines the holonomy invariant transverse density
$\left{\wedge}^{q}\omega \right\in \left{\wedge}^{q}{\tau}^{*}\right$
, and, therefore, a holonomy invariant transverse measure, which can be naturally called the transverse Liouville measure.
Example 2.45.
A foliation
$\mathcal{\mathcal{F}}$
on a manifold
$M=\stackrel{~}{B}{\times}_{\Gamma}F$
, obtained by the suspension from a manifold
$B$
and a homomorphism
$\phi :\Gamma \to Diff\left(F\right)$
of the fundamental group
$\Gamma ={\pi}_{1}\left(B\right)$
is symplectic if and only if for any
$\gamma \in \Gamma $
the diffeomorphism
$\phi \left(\gamma \right)$
preserves a symplectic structure on
$F$
.
Example 2.46.
Recall (cf., for instance, [
120]
) that a submanifold
$\Sigma $
of a symplectic manifold
$X$
is called coisotropic, if, for any
$\sigma \in \Sigma $
, the skeworthogonal complement
$({T}_{\sigma}\Sigma {)}^{\perp}$
of
${T}_{\sigma}\Sigma $
is contained in
${T}_{\sigma}\Sigma $
. If
$\Sigma $
is a coisotropic submanifold, then the distribution
$({T}_{\sigma}\Sigma {)}^{\perp}$
is integrable, and the corresponding foliation
${\mathcal{\mathcal{F}}}_{\Sigma}$
is called the characteristic foliation of the coisotropic submanifold
$\Sigma $
. It is wellknown that there is a canonical symplectic structure on
${T}_{\sigma}\Sigma /({T}_{\sigma}\Sigma {)}^{\perp}$
, therefore, the foliation
${\mathcal{\mathcal{F}}}_{\Sigma}$
is symplectic. Moreover, if
${\mathcal{\mathcal{F}}}_{\Sigma}$
is simple, then the set
$\Gamma \subset {T}^{*}M\times {T}^{*}M$
, which consists of all
$(\nu ,{\nu}^{\prime})\in {T}^{*}M\times {T}^{*}M$
such that
$\nu $
and
${\nu}^{\prime}$
lie on the same leaf of
${\mathcal{\mathcal{F}}}_{\Sigma}$
is a canonical relation.
As shown in [
20]
(see also [
80]
), any presymplectic manifold can be obtained, using this construction, that is, as a coisotropic submanifold of a symplectic manifold. A particular example of the construction described above is the following one. For a foliated manifold
$(M,\mathcal{\mathcal{F}})$
, consider
${T}^{*}M$
as a symplectic manifold with the standard symplectic structure.
Then
${N}^{*}\mathcal{\mathcal{F}}$
is a coisotropic submanifold in
${T}^{*}M$
. The corresponding characteristic foliation
${\mathcal{\mathcal{F}}}_{N}$
is the natural lift of
$\mathcal{\mathcal{F}}$
to the conormal bundle and is called a horizontal (or linearized) foliation. Thus, the linearized foliation
${\mathcal{\mathcal{F}}}_{N}$
is symplectic. The coordinate chart
${\phi}_{n}:{N}^{*}\mathcal{\mathcal{F}}\to {I}^{p}\times {I}^{q}\times {\mathbb{R}}^{q}$
determined by a foliated coordinate chart
$\phi $
on
$M$
(cf. Section 2.1 ) is a foliated chart for
${\mathcal{\mathcal{F}}}_{N}$
with plaques given by the level sets
$y=const,\eta =const$
. Informally speaking, the leaf space
${N}^{*}\mathcal{\mathcal{F}}/{\mathcal{\mathcal{F}}}_{N}$
of
${\mathcal{\mathcal{F}}}_{N}$
can be considered as the cotangent bundle to the leaf space
$M/\mathcal{\mathcal{F}}$
of
$\mathcal{\mathcal{F}}$
.
2.7 Differential operators
Let
$(M,\mathcal{\mathcal{F}})$
be a compact foliated manifold and
$E$
a smooth vector bundle on
$M$
(unless otherwise is stated, we will assume that vector bundles under consideration are smooth and complex). We start with some general definitions, concerning to differential operators on
$M$
.
Definition 2.47.
A linear differential operator
$A$
of order
$\mu $
, acting in
${C}^{\infty}(M,E)$
, is called a tangential differential operator, if, in any foliated chart
$\phi :U\subset M\to {I}^{p}\times {I}^{q}$
and any trivialization of
$E$
over it,
$A$
is of the form
$$\begin{array}{c}A={\sum}_{\left\alpha \right\le \mu}{a}_{\alpha}(x,y){D}_{x}^{\alpha},(x,y)\in {I}^{p}\times {I}^{q},\end{array}$$ 
(2.4)

where
${a}_{\alpha}$
are matrixvalued function on
${I}^{p}\times {I}^{q}$
,
${D}_{x}=\frac{1}{i}\frac{\partial}{\partial x}$
.
Definition 2.48.
For a tangential differential operator
$A$
given by ( 2.4 ) in some foliated chart
$\phi :U\subset M\to {I}^{p}\times {I}^{q}$
and a trivialization of
$E$
over it, define its tangential (complete) symbol
$$\sigma (x,y,\xi )={\sum}_{\left\alpha \right\le \mu}{a}_{\alpha}(x,y){\xi}^{\alpha},(x,y)\in {I}^{p}\times {I}^{q},\xi \in {\mathbb{R}}^{p},$$
and its tangential principal symbol
$${\sigma}_{\mu}(x,y,\xi )={\sum}_{\left\alpha \right=\mu}{a}_{\alpha}(x,y){\xi}^{\alpha},(x,y)\in {I}^{p}\times {I}^{q},\xi \in {\mathbb{R}}^{p}.$$
The tangential principal symbol is invariantly defined as a section of the bundle
$\mathcal{\mathcal{L}}\left({\pi}_{F}^{*}E\right)$
on
${T}^{*}\mathcal{\mathcal{F}}$
(where
${\pi}_{F}:{T}^{*}\mathcal{\mathcal{F}}\to M$
is the natural projection).
Definition 2.49.
A tangential differential operator
$A$
is called tangentially elliptic, if its tangential principal symbol
${\sigma}_{\mu}$
is invertible for
$\xi \ne 0$
.
Let
${D}^{m}(M,E)$
denote the set of all differential operators of order
$m$
and
${D}^{\mu}(\mathcal{\mathcal{F}},E)$
denote the set of all tangential differential operators of order
$\mu $
, acting in
${C}^{\infty}(M,E)$
.
Introduce classes
${D}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
, which are linearly generated by arbitrary compositions of tangential differential operators of order
$\mu $
and differential operators of order
$m$
on
$M$
. In other words, an operator
$A\in {D}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
is of the form
$A={\sum}_{\alpha}{B}_{\alpha}{C}_{\alpha}$
, where
${B}_{\alpha}\in {D}^{m}(M,E)$
,
${C}_{\alpha}\in {D}^{\mu}(\mathcal{\mathcal{F}},E)$
. If
${A}_{1}\in {D}^{{m}_{1},{\mu}_{1}}(M,\mathcal{\mathcal{F}},E)$
,
${A}_{2}\in {D}^{{m}_{2},{\mu}_{2}}(M,\mathcal{\mathcal{F}},E)$
, then
${A}_{1}\circ {A}_{2}\in {D}^{{m}_{1}+{m}_{2},{\mu}_{1}+{\mu}_{2}}(M,\mathcal{\mathcal{F}},E)$
and, if
$A\in {D}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
, then
${A}^{*}\in {D}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
. The classes
${D}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
can be extended to classes
${\Psi}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
, which contain, for instance, parametrices for elliptic operators of class
${D}^{m,\mu}(M,\mathcal{\mathcal{F}},E)$
(cf. [
113]
).
We will use the standard classes of pseudodifferential operators
${\Psi}^{k}(M,E)$
(for the theory of pseudodifferential operators see, for instance, [
101,
175,
179,
167]
). Recall that a pseudodifferential operator on
$M$
is a linear operator
$P:{C}^{\infty}\left(M\right)\to {\mathcal{D}}^{\prime}\left(M\right)$
, which can be represented in a coordinate patch
$X\subset {\mathbb{R}}^{n}$
as
$$Pu\left(x\right)=\int {e}^{(xy)\xi}p(x,\xi )u\left(y\right)dyd\xi ,x\in X,$$
where
$u\in {C}_{c}^{\infty}\left(X\right)$
,
$p(x,\xi )\in {S}^{m}(X\times {\mathbb{R}}^{n})$
is the complete symbol of
$P$
.
Usually, we will assume that the complete symbol
$p$
can be represented as an asymptotic sum
$p\sim {p}_{m}+{p}_{m1}+\dots $
, where
${p}_{k}$
is homogeneous of degree
$k$
in
$\xi $
for
$\left\xi \right>1$
. The principal symbol
${p}_{m}$
of
$P$
is welldefined as a section of the bundle
$\mathcal{\mathcal{L}}\left({\pi}^{*}E\right)$
on
${\stackrel{~}{T}}^{*}M={T}^{*}M\backslash \left\{0\right\}$
, where
$\pi :{T}^{*}M\to M$
is the natural projection.
Definition 2.50.
The transversal principal symbol
${\sigma}_{P}$
of an operator
$P\in {\Psi}^{m}(M,E)$
is the restriction of its principal symbol
${p}_{m}$
to
${\stackrel{~}{N}}^{*}\mathcal{\mathcal{F}}={N}^{*}\mathcal{\mathcal{F}}\backslash \left\{0\right\}$
.
Definition 2.51.
A operator
$P\in {\Psi}^{m}(M,E)$
is called transversally elliptic, if its transversal principal symbol
${\sigma}_{P}\left(\nu \right)$
is invertible for any
$\nu \in {\stackrel{~}{N}}^{*}\mathcal{\mathcal{F}}$
.
Now suppose that a closed foliated manifold
$(M,\mathcal{\mathcal{F}})$
is equipped with a Riemannian metric
${g}_{M}$
. Let
$H$
be the orthogonal complement of
$F=T\mathcal{\mathcal{F}}$
.
Thus, there is a decomposition of
$TM$
into the direct sum
$TM=F\oplus H$
and the corresponding bigrading of the exterior power bundle
${\Lambda}^{*}{T}^{*}M$
:
$${\Lambda}^{k}{T}^{*}M={\oplus}_{i+j=k}{\Lambda}^{i,j}{T}^{*}M,{\Lambda}^{i,j}{T}^{*}M={\Lambda}^{i}{H}^{*}\otimes {\Lambda}^{j}{F}^{*}.$$
There is (cf., for instance, [
15,Proposition10.1]
,[
177]
) the corresponding decomposition of the de Rham differential
$d$
into the sum of bigraded components of the form
$$\begin{array}{c}d={d}_{F}+{d}_{H}+\theta .\end{array}$$ 
(2.5)

Here

(1)
${d}_{F}={d}_{0,1}:{C}^{\infty}(M,{\Lambda}^{i,j}{T}^{*}M)\to {C}^{\infty}(M,{\Lambda}^{i,j+1}{T}^{*}M)$
is the tangential de Rham differential, which is a first order tangentially elliptic operator, independent of the choice of
$g$
;

(2)
${d}_{H}={d}_{1,0}:{C}^{\infty}(M,{\Lambda}^{i,j}{T}^{*}M)\to {C}^{\infty}(M,{\Lambda}^{i+1,j}{T}^{*}M)$
is the transversal de Rham differential, which is a first order transversally elliptic operator;

(3)
$\theta ={d}_{2,1}:{C}^{\infty}(M,{\Lambda}^{i,j}{T}^{*}M)\to {C}^{\infty}(M,{\Lambda}^{i+2,j1}{T}^{*}M)$
is a zero order differential operator, which is the contraction operator by the 2form
$\theta $
on
$M$
with values in
$F$
,
$\theta \in {C}^{\infty}(M,F\otimes {\Lambda}^{2}{\tau}^{*})$
, given by
$$\theta (X,Y)={p}_{F}\left(\right[X,Y\left]\right),X,Y\in {C}^{\infty}(M,H),$$
where
${P}_{F}:TM\to F$
is the natural projection. In particular,
$\theta $
vanishes if and only if
$H$
is integrable.
There is a similar decomposition for the adjoint:
$$\begin{array}{c}\delta ={\delta}_{F}+{\delta}_{H}+{\theta}^{*},\end{array}$$ 
(2.6)

where
${\delta}_{F}$
,
${\delta}_{H}$
and
${\theta}^{*}$
are the adjoints of
${d}_{F}$
,
${d}_{H}$
and
$\theta $
in the Hilbert space
${L}^{2}(M,\Lambda {T}^{*}M)$
accordingly.
The Laplace operator
${\Delta}_{g}=d\delta +\delta d$
of the metric
${g}_{M}$
can be written in the form
$$\begin{array}{c}{\Delta}_{g}={\Delta}_{F}+{\Delta}_{H}+{\Delta}_{1,2}+{K}_{1}+{K}_{2}+{K}_{3},\end{array}$$ 
(2.7)

where

$\bullet $
${\Delta}_{F}={d}_{F}{\delta}_{F}+{\delta}_{F}{d}_{F}\in {D}^{0,2}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
is the tangential Laplacian.

$\bullet $
${\Delta}_{H}={d}_{H}{\delta}_{H}+{\delta}_{H}{d}_{H}\in {D}^{2,0}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
is the transversal Laplacian.

$\bullet $
${\Delta}_{1,2}=\theta {\theta}^{*}+{\theta}^{*}\theta \in {D}^{0,0}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
.

$\bullet $
${K}_{1}={d}_{F}{\delta}_{H}+{\delta}_{H}{d}_{F}+{\delta}_{F}{d}_{H}+{d}_{H}{\delta}_{F}\in {D}^{1,0}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
.

$\bullet $
${K}_{2}={d}_{F}{\theta}^{*}+{\theta}^{*}{d}_{F}+{\delta}_{F}\theta +\theta {\delta}_{F}\in {D}^{0,0}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M).$

$\bullet $
${K}_{3}={d}_{H}{\theta}^{*}+{\theta}^{*}{d}_{H}+{\delta}_{H}\theta +\theta {\delta}_{H}\in {D}^{1,0}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
.
We also introduce the first order differential operator
${D}_{H}={d}_{H}+{d}_{H}^{*}$
in
${C}^{\infty}(M,\wedge {H}^{*})$
, which is called the transverse signature operator.
A basic property of geometric operators on manifolds equipped with Riemannian foliation is that, if
$\mathcal{\mathcal{F}}$
is a Riemannian foliation and
${g}_{M}$
is a bundlelike metric, then the operators
${d}_{F}{\delta}_{H}+{\delta}_{H}{d}_{F}$
and
${\delta}_{F}{d}_{H}+{d}_{H}{\delta}_{F}$
belong to
${D}^{0,1}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
. In particular,
${K}_{1}\in {D}^{0,1}(M,\mathcal{\mathcal{F}},\Lambda {T}^{*}M)$
.
3 Operator algebras of foliations
In this Section, we will describe the noncommutative algebras associated with the leaf space of a foliation. First, we will define an algebra, consisting of very nice functions, on which all basic operators of analysis are defined, then, depending on a problem in question, we will complete this algebra and obtain an analogue of the algebra of measurable, continuous or smooth functions. The role of a “nice” algebra is played by the algebra
${C}_{c}^{\infty}\left(G\right)$
of smooth compactly supported functions on the holonomy groupoid
$G$
of the foliation. Therefore, we start with the notion of holonomy groupoid of a foliation.
3.1 Holonomy groupoid
A foliation
$\mathcal{\mathcal{F}}$
defines an equivalence relation
$\mathcal{\mathcal{R}}\subset M\times M$
on
$M$
:
$(x,y)\in \mathcal{\mathcal{R}}$
if and only if
$x$
and
$y$
lie on the same leaf of the foliation
$\mathcal{\mathcal{F}}$
. Generally,
$\mathcal{\mathcal{R}}$
is not a smooth manifold, but one can resolve its singularity, constructing a smooth manifold
$G$
, called the holonomy groupoid or the graph of the foliation, which “almost every where” coincides with
$\mathcal{\mathcal{R}}$
and which can be used in many cases as a substitution for
$\mathcal{\mathcal{R}}$
. The idea of the holonomy groupoid appeared in the papers of Ehresmann, Reeb and Thom (cf. [
67,
176]
) and was completely realized by Winkelnkemper [
184]
. First of all, we give the general definition of a groupoid (see [
156,
137,
31,
126,
146]
for groupoids and related subjects).
Definition 3.1.
We say that a set
$G$
has the structure of a groupoid with the set of units
${G}^{\left(0\right)}$
, if there are defined maps

$\bullet $
$\Delta :{G}^{\left(0\right)}\to G$
(the diagonal map or the unit map);

$\bullet $
an involution
$i:G\to G$
called the inversion and written as
$i\left(\gamma \right)={\gamma}^{1}$
;

$\bullet $
a range map
$r:G\to {G}^{\left(0\right)}$
and a source map
$s:G\to {G}^{\left(0\right)}$
;

$\bullet $
an associative multiplication
$m:(\gamma ,{\gamma}^{\prime})\to \gamma {\gamma}^{\prime}$
defined on the set
$${G}^{\left(2\right)}=\left\{\right(\gamma ,{\gamma}^{\prime})\in G\times G:r({\gamma}^{\prime})=s(\gamma \left)\right\},$$
satisfying the conditions

$\bullet $
$r(\Delta (x\left)\right)=s(\Delta (x\left)\right)=x$
and
$\gamma \Delta \left(s\right(\gamma \left)\right)=\gamma $
,
$\Delta \left(r\right(\gamma \left)\right)\gamma =\gamma $
;

$\bullet $
$r\left({\gamma}^{1}\right)=s\left(\gamma \right)$
and
$\gamma {\gamma}^{1}=\Delta \left(r\right(\gamma \left)\right)$
.
Alternatively, one can define a groupoid as a small category, where each morphism is an isomorphism.
It is convenient to think of an element
$\gamma \in G$
as an arrow
$\gamma :x\to y$
, going from
$x=s\left(\gamma \right)$
to
$y=r\left(\gamma \right)$
.
We will use the standard notation (for
$x,y\in {G}^{\left(0\right)}$
):

$\bullet $
${G}^{x}=\{\gamma \in G:r(\gamma )=x\}={r}^{1}\left(x\right)$
,

$\bullet $
${G}_{x}=\{\gamma \in G:s(\gamma )=x\}={s}^{1}\left(x\right)$
,

$\bullet $
${G}_{y}^{x}=\{\gamma \in G:s(\gamma )=x,r(\gamma )=y\}$
.
Definition 3.2.
A groupoid
$G$
is called smooth (or a Lie groupoid), if
${G}^{\left(0\right)}$
,
$G$
and
${G}^{\left(2\right)}$
are smooth manifolds,
$r$
,
$s$
,
$i$
and
$m$
are smooth maps,
$r$
and
$s$
are submersions, and
$\Delta $
is an immersion.
Example 3.3.
Lie groups. A Lie group
$H$
defines a smooth groupoid as follows:
$G=H$
,
${G}^{\left(0\right)}$
consists of a single point, the maps
$i$
and
$m$
are given by the group operations in
$H$
.
Example 3.4.
Trivial groupoid. Let
$X$
be an arbitrary set. Put
$G=X$
,
${G}^{\left(0\right)}=X$
, the maps
$s$
and
$r$
are the identity maps (that is, in other words, each element
$x\in {G}^{\left(0\right)}=X$
is identified with an unique element
$\gamma :x\to x$
).
Example 3.5.
Equivalence relations. Any equivalence relation
$R\subset X\times X$
defines a groupoid, if we put
${G}^{\left(0\right)}=X$
,
$G=R$
, the maps
$s:R\to X$
and
$r:R\to X$
are given by
$s(x,y)=y$
,
$r(x,y)=x$
. Thus, pairs
$({x}_{1},{y}_{1})$
and
$({x}_{2},{y}_{2})$
can be multiplied if and only if
${y}_{1}={x}_{2}$
, and in this case
$({x}_{1},{y}_{1})({x}_{2},{y}_{2})=({x}_{1},{y}_{2})$
.
In the particular case of
$R=X\times X$
, we get a so called principal or pair groupoid.
Example 3.6.
Group actions. Let a Lie group
$H$
act smoothly from the left on a smooth manifold
$X$
. The crossed product groupoid
$X\mathbb{\setminus}H$
is defined as follows:
${G}^{\left(0\right)}=X$
,
$G=X\times H$
. The maps
$s:X\times H\to X$
and
$r:X\times H\to X$
have the form
$s(x,h)={h}^{1}x$
,
$r(x,h)=x$
. Thus, pairs
$({x}_{1},{h}_{1})$
and
$({x}_{2},{h}_{2})$
can be multiplied if and only if
${x}_{2}={h}_{1}^{1}{x}_{1}$
, and in this case
$({x}_{1},{h}_{1})({x}_{2},{h}_{2})=({x}_{1},{h}_{1}{h}_{2})$
.
Example 3.7.
Fundamental groupoid (cf. for instance, [
169]
). Let
$X$
be a topological space,
$G=\Pi \left(X\right)$
the set of homotopy classes of paths in
$X$
with all possible endpoints. More precisely, if
$\gamma :[0,1]\to X$
is a path from
$x=\gamma \left(0\right)$
to
$y=\gamma \left(1\right)$
, then we denote by
$\left[\gamma \right]$
the homotopy class of
$\gamma $
with fixed
$x$
and
$y$
. Define the groupoid
$\Pi \left(X\right)$
as the set of triples
$(x,[\gamma ],y)$
, where
$x,y\in X$
,
$\gamma $
is a path with the initial point
$x=\gamma \left(0\right)$
and the final point
$y=\gamma \left(1\right)$
, where the multiplication is given by the product of paths. The groupoid
$\Pi \left(X\right)$
is called the fundamental groupoid of
$X$
.
Example 3.8.
The Haefliger groupoid
${\Gamma}_{n}$
[
23,
89,
91]
. Let
$M$
be a smooth manifold. A groupoid
${\Gamma}_{M}$
consists of the germs of local diffeomorphisms of
$M$
at arbitrary points of
$M$
.
$({\Gamma}_{M}{)}^{\left(0\right)}=M$
. If
$\gamma \in {\Gamma}_{M}$
is the germ at
$x\in M$
of a diffeomorphism
$f$
from some neighborhood
$U$
of
$x$
on an open set
$f\left(U\right)$
, then
$s\left(\gamma \right)=x$
,
$r\left(\gamma \right)=f\left(x\right)$
.
The multiplication in
${\Gamma}_{M}$
is given by the composition of maps. If
$M={\mathbb{R}}^{n}$
, then the groupoid
${\Gamma}_{M}$
is denoted by
${\Gamma}_{n}$
.
The holonomy groupoid
$G=G(M,\mathcal{\mathcal{F}})$
of a foliated manifold
$(M,\mathcal{\mathcal{F}})$
is defined in the following way. Let
${\sim}_{h}$
be an equivalence relation on the set of continuous leafwise paths
$\gamma :[0,1]\to M$
, setting
${\gamma}_{1}{\sim}_{h}{\gamma}_{2}$
, if
${\gamma}_{1}$
and
${\gamma}_{2}$
have the same initial and final points and the same holonomy maps:
${h}_{{\gamma}_{1}}={h}_{{\gamma}_{2}}$
. The holonomy groupoid
$G$
is the set of
${\sim}_{h}$
equivalence classes of leafwise paths. The set of units
${G}^{\left(0\right)}$
is a manifold
$M$
. The multiplication in
$G$
is given by the product of paths. The corresponding range and source maps
$s,r:G\to M$
are given by
$s\left(\gamma \right)=\gamma \left(0\right)$
and
$r\left(\gamma \right)=\gamma \left(1\right)$
. Finally, the diagonal map
$\Delta :M\to G$
takes any
$x\in M$
to the element in
$G$
given by the constant path
$\gamma \left(t\right)=x,t\in [0,1]$
. To simplify the notation, we will identify
$x\in M$
with
$\Delta \left(x\right)\in G$
.
For any
$x\in M$
the map
$s$
maps
${G}^{x}$
on the leaf
${L}_{x}$
through
$x$
. The group
${G}_{x}^{x}$
coincides with the holonomy group of
${L}_{x}$
. The map
$s:{G}^{x}\to {L}_{x}$
is the covering map associated with the group
${G}_{x}^{x}$
, called the holonomy covering. One can also introduce the holonomy groupoid
$G\left(L\right)$
of a leaf
$L$
of
$\mathcal{\mathcal{F}}$
as the set of
${\sim}_{h}$
equivalence classes of piecewise smooth paths in
$L$
.
The holonomy groupoid
$G$
has the structure of a smooth (in general, nonHausdorff and nonparacompact) manifold of dimension
$2p+q$
. Recall the construction of an atlas on
$G$
[
34]
.
Let
$\phi :U\to {I}^{p}\times {I}^{q},{\phi}^{\prime}:{U}^{\prime}\to {I}^{p}\times {I}^{q}$
be two foliated charts,
$\pi =p{r}_{nq}\circ \phi :U\to {\mathbb{R}}^{q}$
,
${\pi}^{\prime}=p{r}_{nq}\circ {\phi}^{\prime}:{U}^{\prime}\to {\mathbb{R}}^{q}$
the corresponding distinguished maps. The foliated charts
$\phi $
,
${\phi}^{\prime}$
are called compatible, if, for any
$m\in U$
and
${m}^{\prime}\in {U}^{\prime}$
with
$\pi \left(m\right)={\pi}^{\prime}\left({m}^{\prime}\right)$
, there is a leafwise path
$\gamma $
from
$m$
to
${m}^{\prime}$
such that the corresponding holonomy map
${h}_{\gamma}$
takes the germ
${\pi}_{m}$
of
$\pi $
at
$m$
to the germ
${\pi}_{{m}^{\prime}}^{\prime}$
of
${\pi}^{\prime}$
at
${m}^{\prime}$
.
For any pair of compatible foliated charts
$\phi $
and
${\phi}^{\prime}$
denote by
$W(\phi ,{\phi}^{\prime})$
the subset in
$G$
, consisting of all
$\gamma \in G$
from
$s\left(\gamma \right)=m={\phi}^{1}(x,y)\in U$
to
$r\left(\gamma \right)={m}^{\prime}={{\phi}^{\prime}}^{1}({x}^{\prime},y)\in {U}^{\prime}$
such that the corresponding holonomy map
${h}_{\gamma}$
takes the germ
${\pi}_{m}$
of the map
$\pi =p{r}_{nq}\circ \phi $
at
$m$
to the germ
${\pi}_{{m}^{\prime}}^{\prime}$
of the map
${\pi}^{\prime}=p{r}_{nq}\circ {\phi}^{\prime}$
at
${m}^{\prime}$
. There is a coordinate map
$$\begin{array}{c}\Gamma :W(\phi ,{\phi}^{\prime})\to {I}^{p}\times {I}^{p}\times {I}^{q},\end{array}$$ 
(3.1)

which takes each element
$\gamma \in W(\phi ,{\phi}^{\prime})$
such that
$s\left(\gamma \right)=m={\mathbb{\varkappa}}^{1}(x,y)$
,
$r\left(\gamma \right)={m}^{\prime}={{\phi}^{\prime}}^{1}({x}^{\prime},y)$
and
${h}_{\gamma}{\pi}_{m}={\pi}_{{m}^{\prime}}^{\prime}$
to the triple
$(x,{x}^{\prime},y)\in {I}^{p}\times {I}^{p}\times {I}^{q}$
.
As shown in [
34]
, the coordinate neighborhoods
$W(\phi ,{\phi}^{\prime})$
form an atlas of a
$(2p+q)$
dimensional manifold (in general, nonHausdorff and nonparacompact) on
$G$
. Moreover, the groupoid
$G$
is a smooth groupoid.
NonHausdorffness of the holonomy groupoid is related with the phenomenon of onesided holonomy. The simplest example of a foliation with the nonHausdorff holonomy groupoid is given by the trajectories of a nonsingular vector field on the plane, having a onesided limit cycle. As shown in [
184]
, the holonomy groupoid is Hausdorff if and only if the holonomy maps
${H}_{{T}_{0}{T}_{1}}\left({\gamma}_{1}\right)$
and
${H}_{{T}_{0}{T}_{1}}\left({\gamma}_{2}\right)$
along any leafwise paths
${\gamma}_{1}$
and
${\gamma}_{2}$
with the initial point
$x$
and the final point
$y$
, given by smooth transversals
${T}_{0}$
and
${T}_{1}$
, passing through
$x$
and
$y$
accordingly, coincide, if they coincide on some open subset
$U\subset {T}_{0}$
such that
$x\in \overline{U}$
. In particular, the holonomy groupoid is Hausdorff, if the holonomy is trivial, or real analytic. Moreover, the holonomy groupoid of a Riemannian foliation is Hausdorff. In the sequel, we will always assume that
$G$
is a Hausdorff manifold.
Example 3.9.
If a simple foliation
$\mathcal{\mathcal{F}}$
is defined by a submersion
$\pi :M\to B$
, then its holonomy groupoid
$G$
consists of all
$(x,y)\in M\times M$
such that
$\pi \left(x\right)=\pi \left(y\right)$
, and, moreover,
${G}^{\left(0\right)}=M$
, the maps
$s:G\to M$
and
$r:G\to M$
are given by
$s(x,y)=y$
,
$r(x,y)=x$
.
Example 3.10.
If a foliation
$\mathcal{\mathcal{F}}$
is given by the orbits of a free smooth action of a connected Lie group
$H$
on a manifold
$M$
, then its holonomy groupoid coincides with the crossed product groupoid
$M\mathbb{\setminus}H$
.
Besides the holonomy groupoid, there are another groupoids, which can be associated with the foliation. First of all, it is the groupoid given by the equivalence relation on
$M$
, setting points
$x$
and
$y$
to be equivalent, if they lie on the same leaf of the foliation (the coarse groupoid). As noted above, this groupoid is not smooth. One can also consider the fundamental groupoid of the foliation
$\Pi (M,\mathcal{\mathcal{F}})$
, which also consists of equivalence classes of leafwise paths, where two leafwise paths are called equivalent, if they are homotopic in the class of leafwise paths with fixed endpoints. The fundamental groupoid of the foliation
$\Pi (M,\mathcal{\mathcal{F}})$
is a smooth groupoid (cf., for instance, [149] ).
There is a foliation
$\mathcal{G}$
of dimension
$2p$
on the holonomy groupoid
$G$
. In any coordinate chart
$W(\phi ,{\phi}^{\prime})$
given by a pair of compatible foliated charts
$\phi $
and
${\phi}^{\prime}$
, the leaves of
$\mathcal{G}$
are given by equations of the form
$y=const$
. The leaf of
$\mathcal{G}$
through
$\gamma \in G$
consists of all
${\gamma}^{\prime}\in G$
such that
$r\left(\gamma \right)$
and
$r\left({\gamma}^{\prime}\right)$
lie on the same leaf of
$\mathcal{\mathcal{F}}$
and coincides with the holonomy groupoid of this leaf.
The holonomy group of a leaf of
$\mathcal{G}$
coincides with the holonomy group of the corresponding leaf of
$\mathcal{\mathcal{F}}$
. (The last statement corrects an erroneous one made in [
184]
. This fact was noted, for instance, by Molino in his review of [
184]
in Mathematical Reviews (see MR 85j:57043).) The differential of the map
$(r,s):G\to M\times M$
maps isomorphically the tangent bundle
$T\mathcal{G}$
to
$\mathcal{G}$
to the bundle
$F\u22a0F$
on
$M\times M$
, therefore, there is a canonical isomorphism
$T\mathcal{G}\sim ={r}^{*}F\oplus {s}^{*}F$
.
A distribution
$H$
on
$M$
transverse to
$\mathcal{\mathcal{F}}$
determines a distribution
$HG$
on
$G$
transverse to
$\mathcal{G}$
. For any
$X\in {H}_{y}$
, there is a unique vector
$\widehat{X}\in {T}_{\gamma}G$
such that
$ds\left(\widehat{X}\right)=d{h}_{\gamma}^{1}\left(X\right)$
and
$dr\left(\widehat{X}\right)=X$
, where
$d{h}_{\gamma}:{H}_{x}\to {H}_{y}$
is the linear holonomy map associated with
$\gamma $
. The space
${H}_{\gamma}G$
consists of all vectors of the form
$\widehat{X}\in {T}_{\gamma}G$
for different
$X\in {H}_{y}$
. In any coordinate chart
$W(\phi ,{\phi}^{\prime})$
on
$G$
, the tangent space
${T}_{\gamma}\mathcal{G}$
to
$\mathcal{G}$
at some
$\gamma $
with the coordinates
$(x,{x}^{\prime},y)$
consists of vectors of the form
$X\frac{\partial}{\partial x}+{X}^{\prime}\frac{\partial}{\partial {x}^{\prime}}$
and the distribution
${H}_{\gamma}G$
consists of vectors
$X\frac{\partial}{\partial x}+{X}^{\prime}\frac{\partial}{\partial {x}^{\prime}}+Y\frac{\partial}{\partial y}$
such that
$X\frac{\partial}{\partial x}+Y\frac{\partial}{\partial y}\in {H}_{(x,y)}$
and
${X}^{\prime}\frac{\partial}{\partial {x}^{\prime}}+Y\frac{\partial}{\partial y}\in {H}_{({x}^{\prime},y)}$
.
Let