Similarly as before, we can define the spectral flow
${sf}_{\ell}\left\{{A}_{s}\right\}$
. It counts the number of spectral lines of
${A}_{s}$
coming from the negative side of
$\ell $
to the nonnegative side of
$\ell $
. For the details, see [
33]
.
Examples 2.2.
a) In the above selfadjoint case,
$\ell =\sqrt{1}(\epsilon ,\epsilon )$
(
$\epsilon >0$
) with coorientation from left to right. Then a selfadjoint operator
$A$
is admissible with respect to
$\ell $
if and only if
$A$
is Fredholm. b) Another important case is that
$\ell =(1\epsilon ,1+\epsilon )$
(
$\epsilon \in (0,1)$
) with coorientation from downward to upward, and all
${A}_{s}$
unitary. A unitary operator
$A$
is admissible with respect to
$\ell $
if and only if
$AI$
is Fredholm.
The spectral flow has the following properties (cf. [
23]
and Lemma 2.6 and Proposition 2.2 in [
33]
).
Proposition 2.3.
Let
$\ell \in \mathcal{A}(\mathbb{C})$
be admissible and let
$\left\{{A}_{s}\right\}$
,
$0\le s\le 1$
be a curve in
${\mathcal{A}}_{\ell}\left(X\right)$
. Then the spectral flow
${sf}_{\ell}\left\{{A}_{s}\right\}$
is well defined, and the following holds:

(i)
Catenation. Assume
$t\in [0,1]$
. Then we have
$$\begin{array}{c}{sf}_{\ell}\{{A}_{s};0\le s\le t\}+{sf}_{\ell}\{{A}_{s};t\le s\le 1\}={sf}_{\ell}\{{A}_{s};0\le s\le 1\}.\end{array}$$ 
(2.3)


(ii)
Homotopy invariance. Let
$A(s,t)$
,
$(s,t)\in [0,1]\times [0,1]$
be a continuous family in
${\mathcal{A}}_{\ell}\left(X\right)$
. Then we have
$$\begin{array}{c}{sf}_{\ell}\left\{A\right(s,t);(s,t)\in \partial ([0,1]\times [0,1]\left)\right\}=0.\end{array}$$ 
(2.4)


(iii)
Endpoint dependence for Riesz continuity. Let
${\mathcal{\mathcal{B}}}^{sa}\left(X\right)$
, respectively
${\mathcal{C}}^{sa}\left(X\right)$
denote the spaces of bounded, respectively closed selfadjoint operators in
$X$
. Let
$$\begin{array}{ccccc}R& :& {\mathcal{C}}^{sa}& \to & {\mathcal{\mathcal{B}}}^{sa}\left(X\right)\end{array}$$  
$$\begin{array}{ccccc}& & A& \mapsto & A({A}^{2}+I{)}^{\frac{1}{2}}\end{array}$$  
denote the Riesz transformation. Let
${A}_{s}\in {\mathcal{C}}^{sa}\left(X\right)$
for
$s\in [0,1]$
. Assume that
$\left\{R\right({A}_{s}\left)\right\}$
,
$0\le s\le 1$
is a continuous family.
If
${m}^{}\left({A}_{0}\right)<+\infty $
, then
${m}^{}\left({A}_{1}\right)<+\infty $
and we have
$$\begin{array}{c}sf\left\{{A}_{s}\right\}={m}^{}\left({A}_{0}\right){m}^{}\left({A}_{1}\right).\end{array}$$ 
(2.5)


(iv)
Product. Let
$\left\{{P}_{s}\right\}$
be a curve of projections on
$X$
such that
${P}_{s}{A}_{s}\subset {A}_{s}{P}_{s}$
for all
$s\in [0,1]$
. Set
${Q}_{s}=I{P}_{s}$
. Then we have
${P}_{s}{A}_{s}{P}_{s}\in {\mathcal{A}}_{\ell}(im{P}_{s})\subset \mathcal{C}(im{P}_{s})$
,
${Q}_{s}{A}_{s}{Q}_{s}\in {\mathcal{A}}_{\ell}(im{Q}_{s})\subset \mathcal{C}(im{Q}_{s})$
, and
$$\begin{array}{c}{sf}_{\ell}\left\{{A}_{s}\right\}={sf}_{\ell}\left\{{P}_{s}{A}_{s}{P}_{s}\right\}+{sf}_{\ell}\left\{{Q}_{s}{A}_{s}{Q}_{s}\right\}.\end{array}$$ 
(2.6)


(v)
Bound. For
$A\in {\mathcal{A}}_{\ell}\left(X\right)$
, there exists a neighborhood
$\mathcal{N}$
of
$A$
in
$\mathcal{C}\left(X\right)$
such that
$\mathcal{N}\subset {\mathcal{A}}_{\ell}\left(X\right)$
, and for curves
$\left\{{A}_{s}\right\}$
in
$\mathcal{N}$
with endpoints
${A}_{0}=:A$
and
${A}_{1}=:B$
, the relative Morse index
${I}_{\ell}(A,B):={sf}_{\ell}\{{A}_{s},0;\le s\le 1\}$
is well defined and satisfies
$$\begin{array}{c}0\le {I}_{\ell}(A,B)\le {\nu}_{h,\ell}\left(A\right){\nu}_{h,\ell}\left(B\right).\end{array}$$ 
(2.7)


(vi)
Reverse orientation. Let
$\hat{\ell}$
denote the curve
$\ell $
with opposite coorientation. Then we have
$$\begin{array}{c}{sf}_{\ell}\left\{{A}_{s}\right\}+{sf}_{\hat{\ell}}\left\{{A}_{s}\right\}={\nu}_{h,\ell}\left({A}_{1}\right){\nu}_{h,\ell}\left({A}_{0}\right).\end{array}$$ 
(2.8)


(vii)
Zero. Suppose that
${\nu}_{h,\ell}\left({A}_{s}\right)$
is constant for
$s\in [0,1]$
. Then
${sf}_{\ell}\left\{{A}_{s}\right\}=0$
.

(viii)
Invariance. Let
$\{{T}_{s}{\}}_{s\in [0,1]}$
be a curve of bounded invertible operators. Then we have
$$\begin{array}{c}{sf}_{\ell}\left\{{T}_{s}^{1}{A}_{s}{T}_{s}\right\}={sf}_{\ell}\left\{{A}_{s}\right\}.\end{array}$$ 
(2.9)

Now we give a method of calculating the spectral flow of differentiable curves, inspired among others by J.J. Duistermaat [
14]
and J. Robbin and D. Salamon [
27]
.
Definition 2.4.
Let
$\ell \in \mathcal{A}(\mathbb{C})$
be admissible and
$\{{A}_{s}{\}}_{s\in [0,1]}$
be a curve in
${\mathcal{A}}_{\ell}\left(X\right)$
.

(i)
A crossing for
${A}_{s}$
is a number
$t\in [0,1]$
such that
${\nu}_{h,\ell}\left({A}_{t}\right)\ne 0$
.

(ii)
Set
${P}_{s}={P}_{\ell}^{0}{A}_{s}$
. A crossing
$t$
is called regular if
$dom\left({A}_{s}\right)=D$
fixed for
$s$
near
$t$
,
${A}_{s}x$
is differentiable at
$s=t$
for all
$x\in D$
, and
${P}_{t}\dot{{A}_{t}}{P}_{t}$
is hyperbolic, i.e.
${\nu}_{h,\ell}\left({P}_{t}\dot{{A}_{t}}{P}_{t}\right)=0$
, where
$\dot{{A}_{s}}$
is the unbounded operator with domain
$D$
defined by
$$\dot{{A}_{s}}x=\frac{d}{ds}{A}_{s}x$$
for all
$x\in D$
.

(iii)
A crossing
$t$
is called simple if it is regular and
${\nu}_{h,\ell}\left({A}_{t}\right)=1$
.
Proposition 2.5 (cf. Theorem 4.1 of [33] ).
Let
$X$
be a Banach space and
$\ell =\sqrt{1}(\epsilon ,\epsilon )$
(
$\epsilon >0$
) with coorientation from left to right. Let
${A}_{s}$
,
$\epsilon \le s\le \epsilon $
(
$\epsilon >0$
), be a curve in
${\mathcal{A}}_{\ell}\left(X\right)$
. Suppose that
$0$
is a regular crossing of
${A}_{s}$
. Set
$P={P}_{\ell}^{0}\left({A}_{0}\right)$
,
$A={A}_{0}$
and
$B={\dot{A}}_{s}{}_{s=0}$
. Assume that
$$\begin{array}{c}P(ABBA)P=0.\end{array}$$ 
(2.10)

Then there is a
$\delta \in (0,\epsilon )$
such that
${\nu}_{h,\ell}\left({A}_{s}\right)=0$
for all
$s\in [\delta ,0)\cup (0,\delta ]$
and
$$\begin{array}{ccc}& & {sf}_{\ell}\{{A}_{s};0\le s\le \delta \}={m}^{}\left(PBP\right),\end{array}$$ 
(2.11)

$$\begin{array}{ccc}& & {sf}_{\ell}\{{A}_{s};\delta \le s\le 0\}={m}^{+}\left(PBP\right).\end{array}$$ 
(2.12)

Here we denote by
${m}^{+}\left(PBP\right)$
$($
${m}^{}\left(PBP\right)$
$)$
the total algebraic multiplicity of eigenvalues of
$PBP$
with positive
$($
negative
$)$
imaginary part respectively.
2.2 Symplectic functional analysis and Maslov index
A main feature of symplectic analysis is the study of the Maslov index. It is an intersection index between a path of Lagrangian subspaces with the Maslov cycle, or, more generally, with another path of Lagrangian subspaces. The Maslov index assigns an integer to each continuous path of Fredholm pairs of Lagrangian subspaces of a fixed Hilbert space with continuously varying symplectic structures.
Firstly we define symplectic Hilbert spaces and Lagrangian subspaces.
Definition 2.6.
Let
$H$
be a complex vector space. A mapping
$$\omega :H\times H\u27f6\mathbb{C}$$
is called a (weak) symplectic form on
$H$
, if it is sesquilinear, skewhermitian, and nondegenerate, i.e., (i)
$\omega (x,y)$
is linear in
$x$
and conjugate linear in
$y$
; (ii)
$\omega (y,x)=\overline{\omega (y,x)}$
; (iii)
${H}^{\omega}:=\{x\in H\omega (x,y)=0\text{for all}y\in H\text{}\}=\{0\}$
.
Then we call
$(H,\omega )$
a complex symplectic vector space.
Definition 2.7.
Let
$(H,\omega )$
be a complex symplectic vector space.
(a) The
annihilator of a subspace
$\lambda $
of
$H$
is defined by
${\lambda}^{\omega}:=\{y\in H\omega (x,y)=0\text{for all}x\in \lambda \text{}\}.$
(b) A subspace
$\lambda $
is called isotropic, coisotropic, or Lagrangian if
$$\lambda \subset {\lambda}^{\omega},\lambda \supset {\lambda}^{\omega},\lambda ={\lambda}^{\omega},$$
respectively.
(c) The
Lagrangian Grassmannian
$\mathcal{\mathcal{L}}(H,\omega )$
consists of all Lagrangian subspaces of
$(H,\omega )$
.
Definition 2.8.
Let
$H$
be a complex Hilbert space. A mapping
$\omega :H\times H\to \mathbb{C}$
is called a (strong) symplectic form on
$H$
, if
$\omega (x,y)=\langle Jx,y{\rangle}_{H}$
for some bounded invertible skewadjoint operator
$J$
.
$(H,\omega )$
is called a (strong) symplectic Hilbert space.
Before giving a rigorous definition of the Maslov index, we fix the terminology and give a simple lemma.
We recall:
Definition 2.9.
(a) The space of (algebraic) Fredholm pairs of linear subspaces of a vector space
$H$
is defined by
$$\begin{array}{c}{\mathcal{\mathcal{F}}}_{alg}^{2}\left(H\right):=\left\{\right(\lambda ,\mu \left)\rightdim\left(\lambda \cap \mu \right)<+\infty \text{and}dim\left(H/(\lambda +\mu )\right)<+\infty \}\end{array}$$ 
(2.13)

with
$$\begin{array}{c}\text{index}(\lambda ,\mu ):=dim(\lambda \cap \mu )dim(H/(\lambda +\mu \left)\right).\end{array}$$ 
(2.14)

(b) In a Banach space
$H$
, the space of (topological) Fredholm pairs is defined by
(2.15)
ℱ
2
(
H
)
:
=
{
(
λ
,
μ
)
∈
ℱ
alg
2
(
H
)

λ
,
μ
,
and
λ
+
μ
⊂
H
closed
}
.
We need the following wellknown lemma (see, e.g., [
11,Lemma1.7]
).
Lemma 2.10.
Let
$(H,\omega )$
be a (strong) symplectic Hilbert space. Then

(i)
there is a
$1$

$1$
correspondence between the space
$${\mathcal{U}}^{J}=\{U\in \mathcal{\mathcal{B}}\left({H}^{+},{H}^{}\right){U}^{*}J{}_{{H}^{}}U=J{}_{{H}^{+}}\}$$
and
$\mathcal{\mathcal{L}}(H,\omega )$
under the mapping
$U\to L:=\mathfrak{G}\left(U\right)$
$($
= graph of
$U$
$)$
, where
${H}^{\pm}={H}^{\mp}\left(\sqrt{1}J\right)$
in the sense of the decomposition 2.1 ;

(ii)
if
$U,V\in {\mathcal{U}}^{J}$
and
$\lambda :=\mathfrak{G}\left(U\right)$
,
$\mu :=\mathfrak{G}\left(V\right)$
, then
$(\lambda ,\mu )$
is a Fredholm pair if and only if
$UV$
, or, equivalently,
$U{V}^{1}I$
is Fredholm. Moreover, we have a natural isomorphism
$$\begin{array}{c}ker(U{V}^{1}I)\simeq \lambda \cap \mu .\end{array}$$ 
(2.16)

Definition 2.11.
Let
$(H,\langle \cdot ,\cdot {\rangle}_{s})$
,
$s\in [0,1]$
be a continuous family of Hilbert spaces, and
${\omega}_{s}(x,y)=\langle {J}_{s}x,y{\rangle}_{s}$
be a continuous family of symplectic forms on
$H$
, i.e.,
$\left\{{A}_{s,0}\right\}$
and
$\left\{{J}_{s}\right\}$
are two continuous families of bounded invertible operators, where
${A}_{s,0}$
is defined by
$$\langle x,y{\rangle}_{s}=\langle {A}_{s,0}x,y{\rangle}_{0}\text{for all}x,y\in H.$$
Let
$\left\{\right({\lambda}_{s},{\mu}_{s}\left)\right\}$
be a continuous family of Fredholm pairs of Lagrangian subspaces of
$(H,\langle \cdot ,\cdot {\rangle}_{s},{\omega}_{s})$
. Then there is a continuous splitting
$$\begin{array}{c}H={H}_{s}^{}\left(\sqrt{1}{J}_{s}\right)\oplus {H}_{s}^{+}\left(\sqrt{1}{J}_{s}\right)\end{array}$$ 
(2.17)

associated to the selfadjoint operator
$\sqrt{1}{J}_{s}\in \mathcal{\mathcal{B}}(H,\langle \cdot ,\cdot {\rangle}_{s})$
for each
$s\in [0,1]$
. By Lemma 2.10 ,
${\lambda}_{s}={\mathfrak{G}}_{s}\left({U}_{s}\right)$
and
${\mu}_{s}={\mathfrak{G}}_{s}\left({V}_{s}\right)$
with
${U}_{s}$
,
${V}_{s}\in {\mathcal{U}}^{{J}_{s}}$
, where
${\mathfrak{G}}_{s}$
denotes the graph associated to the splitting ( 2.17 ). We define the Maslov index
$Mas\{{\lambda}_{s},{\mu}_{s}\}$
by
$$\begin{array}{c}Mas\{{\lambda}_{s},{\mu}_{s}\}={sf}_{\ell}\left\{{U}_{s}{V}_{s}^{{}^{1}}\right\},\end{array}$$ 
(2.18)

where
$\ell :=(1\epsilon ,1+\epsilon )$
with,
$\epsilon \in (0,1)$
and with upward coorientation.
Remark 2.12.
For finitedimensional
$H$
, constant
${\mu}_{s}={\mu}_{0}$
, and a loop
$\left\{{\lambda}_{s}\right\}$
, i.e., for
${\lambda}_{0}={\lambda}_{1}$
, we notice that
$Mas\{{\lambda}_{s},{\mu}_{s}\}$
is the winding number of the closed curve
$\{det({U}_{s}^{1}{V}_{0}){\}}_{s\in [0,1]}$
. This is the original definition of the Maslov index as explained in Arnol'd, [
2]
.
Lemma 2.13.
The Maslov index is independent of the choice of the complete inner product of
$H$
.

Proof.
Let
$\langle \cdot ,\cdot {\rangle}_{s,k}$
,
$s\in [0,1]$
with
$k=0,1$
be two continuous families of complete inner products of
$H$
. We define
$$\langle \cdot ,\cdot {\rangle}_{s,t}=(1t)\langle \cdot ,\cdot {\rangle}_{s,0}+t\langle \cdot ,\cdot {\rangle}_{s,1}$$
for each
$(s,t)\in [0,1]\times [0,1]$
. Let
$({\lambda}_{s},{\mu}_{s})$
be a continuous family of Fredholm pairs of Lagrangian subspaces of
$(H,{\omega}_{s})$
. For each inner product
$\langle \cdot ,\cdot {\rangle}_{s,t}$
, we denote by
${U}_{s,t}$
and
${V}_{s,t}$
the associated generated ”unitary” operators of
${\lambda}_{s}$
and
${\mu}_{s}$
respectively.
We also denote by
${Mas}_{t}$
the Maslov index defined with
$\langle \cdot ,\cdot {\rangle}_{s,t}$
for each
$t\in [0,1]$
.
By Proposition
2.3 we have
$$\begin{array}{ccc}{Mas}_{0}\{{\lambda}_{s},{\mu}_{s}\}& & {Mas}_{1}\{{\lambda}_{s},{\mu}_{s}\}\end{array}$$  
$$\begin{array}{ccc}& =& {sf}_{\ell}\left\{{U}_{s,0}{V}_{s,0}^{{}^{1}}\right\}+{sf}_{\ell}\left\{{U}_{s,1}{V}_{s,1}^{{}^{1}}\right\}\end{array}$$  
$$\begin{array}{ccc}& =& {sf}_{\ell}\{{U}_{s,t}{V}_{s,t}^{{}^{1}};(s,t)\in \partial \left([0,1]\times [0,1]\right)\}\end{array}$$  
$$\begin{array}{ccc}& =& 0.\end{array}$$  
□
Now we give a method of using the crossing form to calculate Maslov indices (cf. [14] , [27] , [5,Theorem2.1] ; for a full proof of the following Proposition see [32,Corollary3.1] ).
Let
$\lambda =\{{\lambda}_{s}{\}}_{s\in [0,1]}$
be a
${C}^{1}$
curve of Lagrangian subspaces of
$H$
. Let
$W$
be a fixed Lagrangian complement of
${\lambda}_{t}$
. For
$v\in {\lambda}_{t}$
and
$st$
small, define
$w\left(s\right)\in W$
by
$v+w\left(s\right)\in {\lambda}_{s}$
. The form
$$\begin{array}{c}Q(\lambda ,t):=Q(\lambda ,W,t)(u,v)=\frac{d}{ds}{}_{s=t}\omega (u,w(s\left)\right),\forall u,v\in {\lambda}_{t}\end{array}$$ 
(2.19)

is independent of the choice of
$W$
. Let
$\left\{\right({\lambda}_{s},{\mu}_{s}\left)\right\}$
,
$0\le s\le 1$
be a curve of Fredholm pairs of Lagrangian subspaces of
$H$
. For
$t\in [0,1]$
, the crossing form
$\Gamma (\lambda ,\mu ,t)$
is a quadratic form on
${\lambda}_{t}\cap {\mu}_{t}$
defined by
$$\begin{array}{c}\Gamma (\lambda ,\mu ,t)(u,v)=Q(\lambda ,t)(u,v)Q(\mu ,t)(u,v),\forall u,v\in {\lambda}_{t}\cap {\mu}_{t}.\end{array}$$ 
(2.20)

A crossing is a time
$t\in [0,1]$
such that
${\lambda}_{t}\cap {\mu}_{t}\ne \left\{0\right\}$
. A crossing is called regular if
$\Gamma (\lambda ,\mu ,t)$
is nondegenerate. It is called simple if it is regular and
${\lambda}_{t}\cap {\mu}_{t}$
is onedimensional.
Proposition 2.14.
Let
$(H,\omega )$
be a symplectic Hilbert space and
$\left\{\right({\lambda}_{s},{\mu}_{s}\left)\right\}$
,
$0\le s\le 1$
be a
${C}^{1}$
curve of Fredholm pairs of Lagrangian subspaces of
$H$
with only regular crossings. Then we have
$$\begin{array}{c}Mas\{\lambda ,\mu \}={m}^{+}(\Gamma (\lambda ,\mu ,0\left)\right){m}^{}(\Gamma (\lambda ,\mu ,1\left)\right)+{\sum}_{0<t<1}sign(\Gamma (\lambda ,\mu ,t\left)\right).\end{array}$$ 
(2.21)

3 Symplectic analysis of symmetric operators
3.1 Local stability of weak inner UCP
Let
$X$
be a complex Hilbert space and
$A\in \mathcal{C}\left(X\right)$
a linear, closed, densely defined operator in
$X$
. We assume that
$A$
is symmetric, i.e.,
${A}^{*}\supset A$
where
${A}^{*}$
denotes the adjoint operator. We denote the domains of
$A$
by
${D}_{m}$
(the minimal domain) and of
${A}^{*}$
by
${D}_{max}$
(the maximal domain).
Definition 3.1.
Let
$X$
be a Hilbert space and
$A\in \mathcal{C}\left(X\right)$
with
$domA={D}_{m}$
and
${A}^{*}\supset A$
. We shall say that the operator
$A$
satisfies the weak inner Unique Continuation Property (UCP) if
$kerA=\left\{0\right\}$
.
It is well known that weak UCP and weak inner UCP can be established for a large class of Dirac type operators, see the first author with Wojciechowski [10,Chapter8] , and the first author with M. Marcolli and B.L. Wang [9] . However, it is not valid for all linear elliptic differential operators of first order as shown by one of the Pliś counterexamples [26] . Moreover, one has various quite elementary examples of linear and nonlinear perturbations which invalidate weak inner UCP for Dirac operators. Two such examples are listed in [9] . In the same paper, however, it was shown that weak UCP is preserved under certain `small' perturbations of Dirac type operators. Here we show an elementary result, namely the local stability of weak inner UCP.
Lemma 3.2.
Let
$X$
be a Hilbert space. Let
${A}_{s}\in \mathcal{C}\left(X\right)$
,
$0\le s\le 1$
be a family of symmetric operators with
$dom{A}_{s}={D}_{m}$
and
$dom{A}_{s}^{*}={D}_{max}$
independent of
$s$
. Assume that
$\{{A}_{s}^{*}:{D}_{max}\to X\}$
is a continuous curve of bounded operators, where the norm on
${D}_{max}$
is the graph norm induced by
${A}_{0}^{*}$
. If
${A}_{0}$
satisfies weak inner UCP and there exists a selfadjoint Fredholm extension
${A}_{0}^{*}{}_{D}$
of
${A}_{0}$
, then for all
$s\ll 1$
the operators
${A}_{s}^{*}$
are surjective and the operators
${A}_{s}$
satisfy weak inner UCP.

Proof.
By our assumptions,
$im{A}_{0}^{*}{}_{D}$
is closed and is of finite codimension. Since
$im{A}_{0}^{*}{}_{D}\subset im{A}_{0}^{*}\subset X$
, the full range
$im{A}_{0}^{*}$
is closed. Since
${A}_{0}$
satisfies weak inner UCP,
$im{A}_{0}^{*}=X$
. Then
${A}_{0}^{*}$
is semiFredholm. By Theorem IV.5.17 of Kato [17] we have
$im{A}_{s}^{*}=X$
for
$s\ll 1$
. Since
${A}_{s}$
are symmetric,
${A}_{s}$
satisfy weak inner UCP for
$s\ll 1$
. □
3.2 Continuity of the family
$\left\{{A}_{s,{D}_{s}}\right\}$
Let
$X$
be a complex Hilbert space, and
$M,N\subset X$
be two closed linear subspaces.
Let
${P}_{M},{P}_{N}$
be the orthogonal projections onto
$M$
,
$N$
respectively. Then the distance
$d(M,N)$
is defined by
$d(M,N)=\parallel {P}_{M}{P}_{N}\parallel $
and called the gap between
$M$
and
$N$
. For any two closed operators
$A,B$
on
$X$
, we define
$d(A,B)$
as the distance between their graphs.
Let
$A\in \mathcal{C}\left(X\right)$
be a linear, closed, densely defined operator in
$X$
. By Footnote 1 (page 198), Theorems IV.1.1 and IV.2.14 in [
17]
, it is easy to verify the following
Lemma 3.3.
Let
$B\in \mathcal{\mathcal{B}}(dom(A),X)$
be a bounded operator, where the norm on
$dom\left(A\right)$
is the graph norm
${\mathcal{G}}_{A}$
induced by
$A$
. Let
$d:=\parallel BA{\parallel}_{{\mathcal{G}}_{A}}<\frac{1}{2}$
. Then we have

(i)
$B\in \mathcal{C}\left(X\right)$
, and it holds that
$(12d)\langle x,x{\rangle}_{{\mathcal{G}}_{A}}\le \langle x,x{\rangle}_{{\mathcal{G}}_{B}}\le (1+d{)}^{2}\langle x,x{\rangle}_{{\mathcal{G}}_{A}}\text{for}x\in D\text{}.$

(ii)
$d(B,A)\le \frac{\sqrt{2}d}{(1d{)}^{1}}$
.
Lemma 3.4.
Let
$X$
be a Hilbert space, and
$Y$
be a closed linear subspace of
$H$
. Then there exists a bijection between the space of closed linear subspaces of
$X$
containing
$Y$
and that of closed linear subspaces of
$X/Y$
which preserves the metric.

Proof.
We view
$X/Y$
as
${Y}^{\perp}$
. Let
$M,N\subset {Y}^{\perp}$
be two closed subspaces and
${P}_{M},{P}_{N}$
be the orthogonal projections onto
$M$
,
$N$
respectively. Then we have
$$d(M+Y,N+Y)=\parallel {P}_{M+Y}{P}_{N+Y}\parallel =\parallel {P}_{M}{P}_{N}\parallel =d(M,N).$$
□
From the definition of the gap norm and by some computations we have
Lemma 3.5.
Let
${D}_{m}\subset {D}_{max}\subset X$
be three Hilbert spaces such that
${D}_{m}$
is a closed subspace of
${D}_{max}$
and a dense subspace of
$X$
. Let
${\left\{{A}_{s}\in \mathcal{C}\left(X\right)\right\}}_{s\in [0,1]}$
be a family of densely defined symmetric operators with domain
${D}_{m}$
, and
${\left\{{D}_{s}\right\}}_{s\in [0,1]}$
be a family of closed subspaces of
${D}_{max}$
containing
${D}_{m}$
. We assume that
$dom\left({A}_{s}^{*}\right)={D}_{max}$
, each graph norm
${\mathcal{G}}_{s}$
of
${D}_{max}$
induced by
${A}_{s}^{*}$
is equivalent to the original norm
$\mathcal{G}$
of
${D}_{max}$
, and
$\left\{{A}_{s}^{*}\in \mathcal{\mathcal{B}}({D}_{max},X)\right\}$
,
$\left\{{D}_{s}/{D}_{m}\subset {D}_{max}/{D}_{m}\right\}$
are two continuous families. Then
${\left\{{A}_{s,{D}_{s}}\in \mathcal{C}\left(X\right)\right\}}_{s\in [0,1]}$
is a continuous family of closed operators.
3.3 Continuity of natural Cauchy data spaces
In this subsection we generalize the proof of the continuity of Cauchy data spaces given in [
5,Section3.3]
. We need the following
Proposition 3.6 (Proposition 3.5 of [5] ).
Let
$X$
be a Hilbert space, and
$A\in \mathcal{C}\left(X\right)$
be a symmetric operator. Set
${D}_{m}=dom\left(A\right)$
and
${D}_{max}=dom\left({A}^{*}\right)$
.
If
$A$
admits a selfadjoint Fredholm extension with domain
$D$
, then the quotient space
$D/{D}_{m}$
and the natural Cauchy data space
$(ker{A}^{*}+{D}_{m})/{D}_{m}$
form a Fredholm pair of Lagrangian subspaces of the
$($
strong
$)$
symplectic Hilbert space
${D}_{max}/{D}_{m}$
$($
introduced above in Subsection 1.3
$$
, Item
$\left(ii\right))$
.
Remark 3.7.
At present (March 2005), it is not known whether all linear formally selfadjoint elliptic differential operators of first order over a compact smooth Riemannian manifold with smooth boundary admit a selfadjoint Fredholm extension. Recently in [
8]
, however, that crucial property has been established under the additional assumption of selfadjoint principal symbol of the “tangential operator” at the boundary.
Now we can prove
Proposition 3.8.
Let
$X$
be a Hilbert space, and
${D}_{m}\subset {D}_{max}$
be two dense linear subspaces of
$X$
. Let
$\{{A}_{s}:{D}_{m}\to X{\}}_{s\in [0,1]}$
be a family of closed symmetric densely defined operators in
$X$
. We assume that

(i)
each
${A}_{s}$
admits a selfadjoint Fredholm extension with domain
${D}_{s}$
;

(ii)
$dom\left({A}_{s}^{*}\right)={D}_{max}$
independent of
$s$
and that all graph norms
${\mathcal{G}}_{s}$
of
${D}_{max}$
induced by
${A}_{s}^{*}$
are mutually equivalent;

(iii)
each
${A}_{s}$
satisfies weak inner UCP relative to
${D}_{m}$
; and

(iv)
$\{{A}_{s}^{*}:{D}_{max}\to X\}$
forms a continuous family of bounded operators, where the norm on
${D}_{max}$
is the graph norm
$\mathcal{G}$
induced by
${A}_{0}$
.
Then the natural Cauchy data spaces
$\left({D}_{m}+ker{A}_{s}^{*}\right)/{D}_{m}$
are continuously varying in
${D}_{max}/{D}_{m}$
.

Proof.
We denote the projection of
${D}_{max}$
onto
${D}_{max}/{D}_{m}$
by
$\gamma $
. Note that
$ker{A}_{s}^{*}$
is closed in
${D}_{max}$
.
To prove the continuity, we need only to consider the local situation at
$s=0$
.
First we show that
$\{ker{A}_{s}^{*}{\}}_{s\in [0,1]}$
is a continuous family of subspaces of
${D}_{max}$
; then we show that
$\gamma (ker{A}_{s}^{*})$
is a continuous family in
${D}_{max}/{D}_{m}$
.
We consider the bounded operator
$$\begin{array}{cccc}{F}_{s}:& {D}_{max}& \u27f6& X\oplus ker{A}_{0}^{*}\end{array}$$  
$$\begin{array}{cccc}& x& \mapsto & \left({A}_{s}^{*}\left(x\right),{P}_{0}x\right)\end{array}$$  
,
where
${P}_{0}:{D}_{max}\to ker{A}_{0}^{*}$
denotes the orthogonal projection of the Hilbert space
${D}_{max}$
onto the closed subspace
$ker{A}_{0}^{*}$
. By definition, the family
$\left\{{F}_{s}\right\}$
is a continuous family of bounded operators.
Clearly,
${F}_{0}$
is injective. Since
$im{A}_{0}^{*}{}_{{D}_{0}}\subset im{A}_{0}^{*}\subset X$
and
${A}_{0}^{*}{}_{{D}_{0}}$
is Fredholm,
$im{A}_{0}^{*}$
is closed. From weak inner UCP we get
$im{A}_{0}^{*}=X$
. So the operator
${F}_{0}$
is also surjective. This proves that
${F}_{0}$
is invertible with bounded inverse. Then all operators
${F}_{s}$
are invertible for small
$s\ge 0$
, since
${F}_{s}$
is a continuous family of operators. Note that
$${F}_{s}(ker{A}_{s}^{*})\subset \left\{0\right\}\oplus ker{A}_{0}^{*},\left({F}_{s}{)}^{1}\right(\left\{0\right\}\oplus ker{A}_{0}^{*})\subset ker{A}_{s}^{*}.$$
Since
${F}_{s}$
are invertible for small
$s\ge 0$
, we have
$$\begin{array}{c}{F}_{s}(ker{A}_{s}^{*})=\left\{0\right\}\oplus ker{A}_{0}^{*}.\end{array}$$ 
(3.1)

We define
$${\phi}_{s}:={F}_{s}^{1}\circ {F}_{0}:{D}_{max}\sim ={D}_{max}\text{and}{\phi}_{s}^{1}={F}_{0}^{1}\circ {F}_{s}:{D}_{max}\sim ={D}_{max}$$
for
$s$
small. Since
${F}_{s}$
are invertible for small
$s\ge 0$
, from 3.1 we obtain that
$$\begin{array}{c}{\phi}_{s}(ker{A}_{0}^{*})=ker{A}_{s}^{*}.\end{array}$$ 
(3.2)

From 3.2 we get that
$$\{{P}_{s}:={\phi}_{s}{P}_{0}{\phi}_{s}^{1}:{D}_{max}\u27f6ker{A}_{s}^{*}\}$$
is a continuous family of projections onto the solution spaces
$ker{A}_{s}^{*}$
. The projections are not necessarily orthogonal, but can be orthogonalized and remain continuous in
$s$
like in [
10,Lemma12.8]
. This proves the continuity of the family
$\{ker{A}_{s}^{*}\}$
in
${D}_{max}$
.
Now we must show that
$\left\{\gamma \right(ker{A}_{s}^{*}\left)\right\}$
is a continuous family in the quotient space
${D}_{max}/{D}_{m}$
. This is not proved by the formula
$\gamma (ker{A}_{s}^{*})=\gamma \left({\phi}_{s}\right(ker{A}_{0}^{*}\left)\right)$
alone.
We must modify the endomorphism
${\phi}_{s}$
of
${D}_{max}$
in such a way that it keeps the subspace
${D}_{m}$
invariant.
By Proposition (
3.6 ), the Cauchy data space
$\gamma ({D}_{m}+ker{A}_{0}^{*})$
is closed in
${D}_{max}/{D}_{m}$
.
So
${D}_{m}+ker{A}_{0}^{*}$
is closed in
${D}_{max}$
. We define a continuous family of mappings by
$$\begin{array}{cccccc}{\psi}_{s}:{D}_{max}=& {D}_{m}+ker{A}_{0}^{*}& +& ({D}_{m}+ker{A}_{0}^{*}{)}^{\perp}& \u27f6& {D}_{max}\end{array}$$  
$$\begin{array}{cccccc}& x+y& +& z& \mapsto & x+{\phi}_{s}\left(y\right)+z\end{array}$$  
with
${\psi}_{0}=id$
. Hence all
${\psi}_{s}$
are invertible for
$s\ll 1$
, and
${\psi}_{s}\left({D}_{m}\right)={D}_{m}$
for such small
$s$
. Hence we obtain a continuous family of mappings
$\{{\stackrel{~}{\psi}}_{s}:{D}_{max}/{D}_{m}\to {D}_{max}/{D}_{m}\}$
with
${\stackrel{~}{\psi}}_{s}\left(\gamma \right(ker{A}_{0}^{*}\left)\right)=\gamma (ker{A}_{s}^{*})$
. From that we obtain a continuous family of projections as above. □
Remark 3.9.
From the preceding arguments it also follows that the Cauchy data spaces form a differentiable family, if
$\left\{{A}_{s}^{*}\right\}$
is a differentiable family.
3.4 Proof of the spectral flow formula
We begin with a simple case.
Lemma 3.10.
Let
$X$
be a Hilbert space, and
$A\in \mathcal{C}\left(X\right)$
be a symmetric operator with
$dom\left(A\right)={D}_{m}$
and
$dom\left({A}^{*}\right)={D}_{max}$
. Let
${A}_{D}:={A}^{*}{}_{D}$
be a selfadjoint Fredholm extension of
$A$
. We assume that
$A$
satisfies weak inner UCP. Then there exists an
$\epsilon >0$
such that
${A}_{D}+aI$
is Fredholm and satisfies weak inner UCP for each
$a\in [0,\epsilon ]$
. Let
$\gamma :{D}_{max}\to {D}_{max}/{D}_{m}$
be the natural projection. Then we have
$$sf\{{A}_{D}+aI;a\in [0,\epsilon \left]\right\}=Mas\left\{\gamma \right(D),\gamma (ker({A}^{*}+aI));a\in [0,\epsilon \left]\right\}.$$

Proof.
By the definition of the spectral flow we have
$$\begin{array}{c}sf\{{A}_{D}+aI;a\in [0,\epsilon \left]\right\}={\sum}_{a\in (0,\epsilon ]}dimker({A}_{D}+aI).\end{array}$$ 
(3.3)

Let
$\omega $
be the Green form on
${D}_{max}$
induced by
${A}^{*}$
. Let
$W\in \mathcal{\mathcal{L}}({D}_{max}/{D}_{m})$
be a Lagrangian complement of
$\gamma (ker({A}^{*}+{a}_{0}I\left)\right)$
. By Proposition 3.8 ,
$\gamma (ker({A}^{*}+aI\left)\right)$
and
$ker({A}^{*}+aI)$
are two differentiable families. For each
$y\left({a}_{0}\right)\in ker({A}_{D}+{a}_{0}I)$
, there exists a continuous family
$w\left(a\right)\in W+{D}_{m}$
,
$a{a}_{0}$
small, such that
$w\left({a}_{0}\right)=0$
and
$y\left(a\right):=y\left({a}_{0}\right)+w\left(a\right)\in ker({A}^{*}+aI)$
. Since
${A}^{*}\left(y\right(a\left)\right)=ay\left(a\right)$
and the family
$\left\{y\right(a\left)\right\}$
is continuous in
${D}_{max}$
, the family
$\left\{y\right(a\left)\right\}$
is also continuous in
$X$
. For all
$x\left({a}_{0}\right)\in ker({A}_{D}+{a}_{0}I)$
, we have
$$\begin{array}{ccc}& & \omega \left(\gamma \right(x\left({a}_{0}\right)),\gamma (w\left(a\right))\end{array}$$  
$$\begin{array}{ccc}& =& \langle {A}^{*}\left(x\right({a}_{0}\left)\right),y\left(a\right)y\left({a}_{0}\right)\rangle \langle x\left({a}_{0}\right),{A}^{*}\left(w\right(a\left)\right)\rangle \end{array}$$  
$$\begin{array}{ccc}& =& \langle {a}_{0}x\left({a}_{0}\right),y\left(a\right)y\left({a}_{0}\right)\rangle \langle x\left({a}_{0}\right),{A}^{*}\left(y\right(a\left)\right){A}^{*}\left(y\right({a}_{0}\left)\right)\rangle \end{array}$$  
$$\begin{array}{ccc}& =& \langle {a}_{0}x\left({a}_{0}\right),y\left(a\right)y\left({a}_{0}\right)\rangle \langle {x}_{(}{a}_{0}),ay(a)+{a}_{0}y({a}_{0})\rangle \end{array}$$  
$$\begin{array}{ccc}& =& (a{a}_{0})\langle x\left({a}_{0}\right),y\left(a\right)\rangle \end{array}$$  
Let the crossing forms
$Q$
and
$\Gamma $
be defined by ( 2.19 ) and ( 2.20 ) respectively. Then we have
$Q\left(\gamma \right(ker({A}^{*}+aI)),{a}_{0})\left(\gamma \right(x\left({a}_{0}\right)),\gamma (y\left({a}_{0}\right))=\langle x({a}_{0}),y({a}_{0})\rangle $
and
$$\Gamma \left(\gamma \right(D),\gamma (ker({A}^{*}+aI)),{a}_{0})\left(\gamma \right(x\left({a}_{0}\right)),\gamma (y\left({a}_{0}\right))=\langle x({a}_{0}),y({a}_{0})\rangle .$$
By Proposition 2.14 we have
(3.4)
Mas
{
γ
(
D
)
,
γ
(
ker
(
A
*
+
a
I
)
)
;
a
∈
[
0
,
ε
]
}
=
−
∑
a
∈
(
0
,
ε
]
dim
ker
(
A
D
+
a
I
)
.
Combine equations ( 3.3 ), ( 3.4 ), and our lemma follows. □
Now our main result follows at once.

Proof of Theorem 1.5 .
By Lemma 3.2 , for each
${s}_{0}$
there exists an
$\epsilon \left({s}_{0}\right)>0$
such that the operators
${A}_{s}+aI$
satisfy weak inner UCP for all
$s,a$
with
$s{s}_{0},\lefta\right<\epsilon \left({s}_{0}\right)$
.
Here we use the continuity of the family
$\left\{{A}_{s}^{*}\right\}$
as bounded operators from
${D}_{max}$
to
$X$
. Since
$[0,1]$
is compact and
${A}_{s,{D}_{s}}$
are Fredholm operators for all
$s\in [0,1]$
, there exists an
$\epsilon >0$
such that the operators
${A}_{s}+aI$
satisfy weak inner UCP and
${A}_{s,{D}_{s}}+aI$
are Fredholm operators for all
$s\in [0,1]$
and
$\lefta\right<\epsilon $
.
We only need to prove the formula
1.12 in a small interval
$[{s}_{0},{s}_{1}]$
. We consider the twoparameter families
$$\{{A}_{s,{D}_{s}}+aI\}\text{and}\left\{\gamma \right({D}_{s}),\gamma (ker{A}_{s}^{*}+aI\left)\right\}$$
for
$s\in [{s}_{0},{s}_{1}]$
and
$a\in [0,\epsilon ]$
. Because of the homotopy invariance of spectral flow and Maslov index, both integers must vanish for the boundary loop going counter clockwise around the rectangular domain from the corner point
$({s}_{0},0)$
via the corner points
$({s}_{1},0)$
,
$({s}_{1},\epsilon )$
, and
$({s}_{0},\epsilon )$
back to
$({s}_{0},0)$
.
Moreover, for
${s}_{1}$
sufficiently close to
${s}_{0}$
we can choose
$\epsilon $
sufficiently small so that
$ker({A}_{s,{D}_{s}}+\epsilon I)=\left\{0\right\}$
for all
$s\in [{s}_{0},{s}_{1}]$
. Hence, spectral flow and Maslov index must vanish on the top segment of our box.
Finally, by the preceding lemma, the left and the right side segments of our curves yield vanishing sum of spectral flow and Maslov index. So, by additivity under catenation, our assertion follows. □
References

W. Ambrose, The index theorem in Riemannian geometry, Ann. of Math. 73 (1961), 49–86.

V.I. Arnol'd, Characteristic class entering in quantization conditions, Funkcional. Anal. i Priložen. 1 (1967), 1–14. (Russian; English translation: Functional Anal. Appl. 1 (1967), 1–13; French translation: Complément 1 to V.P. Maslov, “Théorie des perturbations et méthodes asymptotiques”, Dunod, GauthierVillars, Paris, 1972, 341–361).

M.F. Atiyah, V.K. Patodi, and I.M. Singer, Spectral asymmetry and Riemannian geometry. I, Math. Proc. Cambridge Phil. Soc. 77 (1975), 43–69.

B. Bojarski, The abstract linear conjugation problem and Fredholm pairs of subspaces, in: “In Memoriam I.N. Vekua”, Tbilisi Univ, Tbilisi, 1979, pp. 45–60 (Russian).

B. Booss–Bavnbek and K. Furutani, The Maslov index – a functional analytical definition and the spectral flow formula, Tokyo J. Math. 21 (1998), 1–34.

B. Booss–Bavnbek, K. Furutani, and N. Otsuki, Criss–cross reduction of the Maslov index and a proof of the Yoshida–Nicolaescu Theorem, Tokyo J. Math. 24 (2001), 113–128.

B. Booss–Bavnbek, M. Lesch, and J. Phillips, Unbounded Fredholm Operators and Spectral Flow, Preprint August 2001, Canad. J. Math. (in print), arXiv: math.FA/0108014.

B. Booss–Bavnbek, M. Lesch, and C. Zhu, Basic Analytic and Topological Properties of Formally SelfAdjoint Elliptic Differential Operators of First Order on Compact Manifolds with Smooth Boundary, in preparation.

B. Booss–Bavnbek, M. Marcolli, and B.L. Wang, Weak UCP and perturbed monopole equations, Internat. J. Math. 13/9 (2002), 987–1008.

B. Booss–Bavnbek and K.P. Wojciechowski, Elliptic Boundary Problems for Dirac Operators, Birkhäuser, Boston, 1993.

B. BoossBavnbek and C. Zhu, Weak Symplectic Functional Analysis and General Spectral Flow Formula, Preprint, Roskilde, December 2003, arXiv: math.DG/0406139.

S.E. Cappell, R. Lee, and E.Y. Miller, Selfadjoint elliptic operators and manifold decompositions Part II: Spectral flow and Maslov index, Comm. Pure Appl. Math. 49 (1996), 869–909.

A. Carey and J. Phillips, Spectral Flow in Fredholm Modules, Eta Invariants and the JLO Cocycle, Preprint 2003, arXiv: math.KT/0308161.

J.J. Duistermaat, On the Morse index in variational calculus, Adv. Math. 21 (1976), 173–195.

A. Floer, A relative Morse index for the symplectic action, Comm. Pure Appl. Math. 41 (1988), 393–407.

B. Himpel, P. Kirk, and M. Lesch, Calderón projector for the Hessian of the ChernSimons function on a 3manifold with boundary, Proc. London Math. Soc. (3) 89 (2004), 241–272, arXiv: math.GT/0302234.

T. Kato, Perturbation Theory for Linear Operators, SpringerVerlag, Berlin, 1966, 2d ed., 1976, corrected printing, 1980.

P. Kirk and M. Lesch, The
$\eta $
–invariant, Maslov Index, and Spectral Flow for Dirac–type Operators on Manifolds with Boundary, Preprint December 2000, to appear in Forum Math., arXiv: math.DG/0012123.

B. Lawruk, J. Sniatycki, and W.M. Tulczyjew, Special symplectic spaces, J. Differential Equations 17 (1975), 477–497.

Y. Long and C. Zhu, Maslovtype index theory for symplectic paths and spectral flow (II). Chinese Ann. of Math. 21B:1 (2000), 89–108.

M. Morse, The Calculus of Variations in the Large, A.M.S. Coll. Publ., Vol.18, Amer. Math. Soc., New York, 1934.

L. Nicolaescu, The Maslov index, the spectral flow, and decomposition of manifolds, Duke Math. J. 80 (1995), 485–533.

J. Phillips, Self–adjoint Fredholm operators and spectral flow, Canad. Math. Bull. 39 (1996), 460–467.

P. Piccione and D.V. Tausk, The Maslov index and a generalized Morse index theorem for nonpositive definite metrics, C. R. Acad. Sci. Paris Sér. I Math. 331 (2000), 385–389.

—, —, The Morse index theorem in semiRiemannian Geometry, Topology 41 (2002), 1123–1159, arXiv: math.DG/0011090.

A. Plis, A smooth linear elliptic differential equation without any solution in a sphere, Comm. Pure Appl. Math. 14 (1961), 599–617.

J. Robbin and D. Salamon, The Maslov index for paths, Topology 32 (1993), 827–844.

R.T. Seeley, Singular integrals and boundary value problems, Amer. J. Math. 88 (1966), 781–809.

K.P. Wojciechowski, Spectral flow and the general linear conjugation problem, Simon Stevin 59 (1985), 59–91.

Tomoyoshi Yoshida, Floer homology and splittings of manifolds, Ann. of Math. 134 (1991), 277–323.

C. Zhu, Maslovtype index theory and closed characteristics on compact convex hypersurfaces in
${\mathbb{R}}^{2n}$
, PhD Thesis (in Chinese), Nankai Institute, Tianjin, 2000.

C. Zhu, The Morse Index Theorem for Regular Lagrangian Systems, Preprint September 2001 (arXiv: math.DG/0109117) (first version); MPI Preprint 2003, no. 55 (modified version).

C. Zhu and Y. Long, Maslovtype index theory for symplectic paths and spectral flow. (I), Chinese Ann. of Math. 20B (1999), 413–424.
Institut for matematik og fysik, Roskilde University, 4000 Roskilde, Denmark Email address : booss@mmf.ruc.dk URL: http://imfufa.ruc.dk/
$\sim $
Booss Nankai Institute of Mathematics, Key Lab of Pure Mathematics and Combinatorics of Ministry of Education, Nankai University, Tianjin 300071, People's Republic of China Email address : zhucf@nankai.edu.cn