has a closed connection form.
(ii)
$\Gamma =0$
. If
${u}_{i}\in Symp$
are such that the product of commutators
$[{u}_{1},{u}_{2}]...[{u}_{2p1},{u}_{2p}]$
lies in
${Symp}_{0}$
, then there are elements
${g}_{1},...,{g}_{2p}\in {Symp}_{0}$
such that
$$[{u}_{1}{g}_{1},{u}_{2}{g}_{2}]...[{u}_{2p1}{g}_{2p1},{u}_{2p}{g}_{2p}]\in Ham.$$
The second condition in (ii) is not yet well understood. We show below that it is equivalent to the existence of a suitable extension of the flux homomorphism; see Proposition 1.13 and Remark 1.14 . However, we can prove that Flux extends only under very restrictive circumstances, for example if
$\left[\omega \right]$
vanishes on
$2$
tori and
${\pi}_{0}(Symp)$
acts on
${\pi}_{1}\left(M\right)$
by inner automorphisms; see Proposition 1.18 .
1.2 Extending the flux homomorphism
Flux is initially defined as a homomorphism from the universal cover
${\stackrel{~}{Symp}}_{0}$
of the identity component of the symplectomorphism group to the group
${H}^{1}(M;\mathbb{R})$
. For each element
$\left\{{g}_{t}\right\}\in {\stackrel{~}{Symp}}_{0}$
the value of the class
$\stackrel{~}{Flux}\left(\right\{{g}_{t}\left\}\right)\in {H}^{1}(M;\mathbb{R})$
on the
$1$
cycle
$\gamma $
in
$M$
is given by integrating
$\omega $
over the
$2$
chain
$(s,t)\mapsto {g}_{t}\left(\gamma \right(s\left)\right)$
. If we define the Flux group
$\Gamma $
to be the image of
${\pi}_{1}(Symp)\subset {\stackrel{~}{Symp}}_{0}$
under
$\stackrel{~}{Flux}$
, then
$\stackrel{~}{Flux}$
descends to a homomorphism
$$Flux:{Symp}_{0}(M,\omega )\to {H}^{1}(M;\mathbb{R})/\Gamma $$
that we shall call the flux homomorphism. Its kernel is the Hamiltonian group
$Ham$
.
When
$\Gamma =0$
the inclusion
$Ham\hookrightarrow {Symp}_{0}$
is a homotopy equivalence. Hence
$Ham$
trivially has the restricted extension property described in the next definition. However, when
$\Gamma \ne 0$
this is not true. Since it turns out to be hard to find subgroups of
$Symp$
with the extension property, we shall consider a modified property that is easier to satisfy. We also consider a restricted property, appropriate for symplectic bundles
$M\to P\to B$
that are trivial over the
$1$
skeleton of
$B$
and so are classified by a map into
$B{Symp}_{0}$
. We shall always suppose that
$M\to P\to B$
is a smooth bundle with base equal to a finite dimensional (possibly open) manifold with finite homotopy type and fiber a closed symplectic manifold.
Definition 1.4.
We shall say that a (possibly disconnected) subgroup
${\mathcal{\mathscr{H}}}_{0}$
of
${Symp}_{0}$
has the restricted extension property if the following condition holds: a
${Symp}_{0}$
bundle
$M\to P\to B$
has a closed connection form iff its classifying map
$B\to B{Symp}_{0}$
lifts to
$B{\mathcal{\mathscr{H}}}_{0}$
. Similarly a subgroup
$\mathcal{\mathscr{H}}$
of
$Symp$
has the modified extension property if the following condition holds: a
$Symp$
bundle
$M\to P\to B$
has a closed connection form iff the pullback of its classifying map
$B\to BSymp$
over some finite cover
$\rho :\stackrel{~}{B}\to B$
lifts to
$B\mathcal{\mathscr{H}}$
.
Thus to say that
${\mathcal{\mathscr{H}}}_{0}$
has the modified restricted extension property means that a
${Symp}_{0}$
bundle
$M\to P\to B$
has a closed connection form iff there is a homotopy commutative diagram
$$\begin{array}{ccc}\stackrel{~}{B}& \to & B{\mathcal{\mathscr{H}}}_{0}\end{array}$$  
$$\begin{array}{ccc}\rho \downarrow & & \downarrow \end{array}$$  
$$\begin{array}{ccc}B& \stackrel{\phi}{\to}& B{Symp}_{0},\end{array}$$  
where
$\rho :\stackrel{~}{B}\to B$
is some finite covering map and
$\phi :B\to B{Symp}_{0}$
classifies
$P\to B$
.
McDuff–Salamon [Thm 6.36][
10]
and Lalonde–McDuff [
7]
claim that the Hamiltonian group
$Ham(M,\omega )$
has the restricted extension property. But this is false: there are
${Symp}_{0}$
bundles
$P\to B$
that have a closed connection form but yet only acquire a Hamiltonian structure when pulled back over some finite covering
$\stackrel{~}{B}\to B$
. (See McDuff–Salamon [
12]
and the erratum to [
7]
.) The next proposition is proved in § 3.2 .
Proposition 1.5.
The Hamiltonian group
$Ham(M,\omega )$
has the modified restricted extension property. It has the restricted extension property iff
$\Gamma =0$
.
Our aim in this paper is to understand subgroups
$\mathcal{\mathscr{H}}$
of
$Symp$
that have the (possibly modified) extension property. Since every
$M$
bundle
$P\to {S}^{1}$
has a closed connection form, any such group
$\mathcal{\mathscr{H}}$
must intersect almost every component of
$Symp$
. The following proposition is proved in § 3.2 . We write
$Im\left({\pi}_{0}\right(\mathcal{\mathscr{H}}\left)\right)$
for the image of
${\pi}_{0}(\mathcal{\mathscr{H}})$
in
${\pi}_{0}(Symp)$
.
Proposition 1.6.
Let
$\mathcal{\mathscr{H}}$
be a subgroup of
$Symp$
with identity component equal to
$Ham$
. Then
$\mathcal{\mathscr{H}}$
has the modified extension property iff every finitely generated subgroup of
${\pi}_{0}(Symp)$
has finite image in the coset space
${\pi}_{0}(Symp)/Im\left({\pi}_{0}\right(\mathcal{\mathscr{H}}\left)\right)$
.
The previous results prompt the following question.
Question 1.7.
When does
$Symp$
have a subgroup
$\mathcal{\mathscr{H}}$
with the modified extension property and such that
$\mathcal{\mathscr{H}}\cap {Symp}_{0}=Ham$
?
Since
$Ham$
is the kernel of the Flux homomorphism, this question can be rephrased as a question about extending the Flux homomorphism to the whole group
$Symp$
. This problem arose (with rather different motivation) in the work of Kotschick–Morita [5] in the case when
$M$
is a Riemann surface of genus
$g>1$
or, more generally a monotone manifold, i.e. a manifold in which the symplectic class
$\left[\omega \right]$
is a multiple of the first Chern class. They showed that in this case Flux extends to a crossed homomorphism
$${F}_{KM}:Symp(M,\omega )\to {H}^{1}(M;\mathbb{R}),$$
that is, a map
$F:={F}_{KM}$
that instead of being a homomorphism, satisfies the identity
$$\begin{array}{c}F\left(gh\right)=F\left(h\right)+{h}^{*}F\left(g\right),\end{array}$$ 
(1.1)

where
${h}^{*}$
denotes the action of
$h$
on
${H}^{1}(M;\mathbb{R})$
via pullback.^{3 }
It is not always possible to extend Flux so that it takes values in
${H}^{1}(M;\mathbb{R})$
. For one thing,
$\Gamma $
may not vanish. Even if one looks for crossed homomorphisms with values in
${H}^{1}(M;\mathbb{R})/\Gamma $
there are more subtle problems, and it is unknown whether an extension must always exist: see Proposition 1.13 . However, the following result shows that this question is very closely related to our earlier considerations.
Proposition 1.8.
(i) If
$$\stackrel{~}{F}:Symp(M,\omega )\to {H}^{1}(M;\mathbb{R})/\Gamma $$
is a continuous crossed homomorphism that extends
$Flux$
, its kernel
$\mathcal{\mathscr{H}}$
intersects every component of
$Symp$
and has the modified extension property.
Moreover
$\mathcal{\mathscr{H}}$
has the extension property iff
$\Gamma =0$
.
(ii) Conversely, let
$\mathcal{\mathscr{H}}$
be a subgroup of
$Symp$
that intersects
${Symp}_{0}$
in
$Ham$
and denote by
${Symp}_{\mathcal{\mathscr{H}}}$
the union of the components of
$Symp$
that intersect
$\mathcal{\mathscr{H}}$
. Then there is a crossed homomorphism
$\stackrel{~}{F}:{Symp}_{\mathcal{\mathscr{H}}}\to {H}^{1}(M;\mathbb{R})/\Gamma $
that extends
$Flux$
.

Proof.
Given
$\stackrel{~}{F}$
, let
$\mathcal{\mathscr{H}}:={\mathcal{\mathscr{H}}}_{\stackrel{~}{F}}$
be the kernel of
$\stackrel{~}{F}$
. Then
$\mathcal{\mathscr{H}}\cap {Symp}_{0}=Ham$
. Further given any
$g\in Symp$
choose
$h\in {Symp}_{0}$
such that
$Fluxh=\stackrel{~}{F}\left(g\right)\in {H}^{1}(M;\mathbb{R})/\Gamma $
.
Then
$g$
is isotopic to the element
$gh\in \mathcal{\mathscr{H}}$
. Hence
$\mathcal{\mathscr{H}}$
has the modifed extension property by Proposition 1.6 . If
$\Gamma =0$
then the inclusion
$B\mathcal{\mathscr{H}}\to BSymp$
is a homotopy equivalence and every bundle has both a closed connection form and an
$\mathcal{\mathscr{H}}$
structure. If
$\Gamma \ne 0$
one can construct bundles that have a closed connection form but no
$\mathcal{\mathscr{H}}$
structure as in the proof of Proposition 1.5 .
To prove (ii) we define
$\stackrel{~}{F}$
on
${Symp}_{\mathcal{\mathscr{H}}}$
as follows: given
$g\in {Symp}_{\mathcal{\mathscr{H}}}$
denote by
${\sigma}_{g}$
any element in
$\mathcal{\mathscr{H}}$
that is isotopic to
$g$
and set
$$\stackrel{~}{F}\left(g\right):=Flux\left({\sigma}_{g}^{1}g\right).$$
This is independent of the choice of
${\sigma}_{g}$
. Further
$$({\sigma}_{g}{\sigma}_{h}{)}^{1}{\sigma}_{gh}\in \mathcal{\mathscr{H}}\cap {Symp}_{0}=Ham.$$
Hence
$$\stackrel{~}{F}\left(gh\right)=Flux\left(({\sigma}_{gh}{)}^{1}gh\right)=Flux\left({\sigma}_{h}^{1}\left({\sigma}_{g}^{1}g\right){\sigma}_{h}\right)Flux\left({\sigma}_{h}^{1}h\right)={h}^{*}\stackrel{~}{F}\left(g\right)+\stackrel{~}{F}\left(h\right).$$
Thus
$\stackrel{~}{F}$
satisfies ( 1.1 ) and so is a crossed homomorphism. □
Remark 1.9.
If Flux extends to
$\stackrel{~}{F}$
but
$\Gamma \ne 0$
then by part (i) of the above proposition the kernel of
$\stackrel{~}{F}$
does not have the extension property. On the other hand, the kernel
${\mathcal{\mathscr{H}}}_{\mathbb{Q}}$
of the composite map
$$\stackrel{~}{F}:Symp\to {H}^{1}(M;\mathbb{R})/\Gamma \to {H}^{1}(M;\mathbb{R})/(\Gamma \otimes \mathbb{Q})$$
does have the extension property by Remark 3.6 . This is the smallest group with this property. Note that it has countably many components in
${Symp}_{0}$
.
Although Flux may not always have an extension with values in
${H}^{1}(M;\mathbb{R})/\Gamma $
, its composite with projection onto a suitable quotient group
${H}^{1}(M;\mathbb{R})/\Lambda $
always can be extended. Below we define a continuous crossed homomorphism
$$\begin{array}{c}{\widehat{F}}_{s}:Symp(M,\omega )\to {H}^{1}(M;\mathbb{R})/{H}^{1}(M;\mathcal{P}\omega )=:\mathcal{A},\end{array}$$ 
(1.2)

where
$\mathcal{P}\omega :=\mathcal{P}\mathbb{Q}\omega $
is the rational period group of
$\omega $
(i.e. the values taken by
$\left[\omega \right]$
on the rational
$2$
cycles) and
$\mathcal{A}$
is given the obvious quasitopology. The map
${\widehat{F}}_{s}$
depends on the choice of a splitting
$s$
of a certain exact sequence. (See the definitions in § 2 .) However its restriction to the identity component
${Symp}_{0}$
is independent of this choice and equals the composite
$${Symp}_{0}\stackrel{Flux}{\to}{H}^{1}(M;\mathbb{R})/\Gamma \to {H}^{1}(M;\mathbb{R})/{H}^{1}(M;\mathcal{P}\omega ).$$
Recall that if a group
$G$
acts continuously on an
$R$
module
$\mathcal{A}$
(for suitable ground ring
$R$
) then the continuous group cohomology^{4 }
$${H}_{cEM}^{1}(G;\mathcal{A})$$
(defined using continuous Eilenberg–MacLane cochains) is the quotient of the module of all continuous crossed homomorphisms
$G\to \mathcal{A}$
by the submodule consisting of the coboundaries
$h\mapsto h\cdot \alpha \alpha ,\alpha \in \mathcal{A}$
. Therefore,
${\widehat{F}}_{s}$
defines an element
$$\left[{\widehat{F}}_{s}\right]\in {H}_{cEM}^{1}(Symp;\mathcal{A}).$$
Although there is no canonical choice for
${\widehat{F}}_{s}$
it turns out that the cohomology class
$\left[\widehat{F}\right]:=\left[{\widehat{F}}_{s}\right]$
is independent of the choice of
$s$
.
We now define
$$Hams(M,\omega ):=ker{\widehat{F}}_{s}.$$
These groups depend on the chosen representative
${\widehat{F}}_{s}$
for the class
$\left[{\widehat{F}}_{s}\right]$
, but they are all conjugate via elements of
${Symp}_{0}$
. Moreover their intersection with the subgroup
${Symp}_{H}$
of
$Symp$
that acts trivially on rational homology is independent of
$s$
. (See Lemma 2.5 .) This holds because any crossed homomorphism
$\widehat{F}:Symp\to \mathcal{A}$
restricts to a homomorphism on
${Symp}_{H}$
that depends only on the class represented by
$\widehat{F}$
in
${H}_{cEM}^{1}(Symp;\mathcal{A})$
.
The following result is proved in
§ 3 . It shows that a symplectic fibration
$P\to B$
has a closed connection form iff its structural group can be reduced to
$Hams$
.
Theorem 1.10.
$Hams$
has the extension property.
Because
$Hams$
is the kernel of a crossed homomorphism it follows from standard theory that one can use this homomorphism to define a class
$\mathcal{O}M$
that measures the obstruction to lifting a map
$\phi :B\to BSymp$
to
$BHams$
. Here
$$\mathcal{O}M\in {H}^{2}(BSymp;{H}^{1}(M;\mathcal{P}\omega \left)\right)={H}^{2}(BSymp;{\pi}_{1}(\mathcal{A}\left)\right),$$
where we think of
$\mathcal{A}={H}^{1}(M;\mathbb{R})/{H}^{1}(M;\mathcal{P}\omega )$
as a quasitopological group: see Remark 2.6 . Thus the local coefficient system
${\pi}_{1}(\mathcal{A})$
on
$BSymp$
has fibers isomorphic to the discrete group
${H}_{\mathbb{Q}}:={H}^{1}(M;\mathcal{P}\omega )$
.
The first part of the next result restates Theorem
1.10 .
Theorem 1.11.
(i) There is a class
$\mathcal{O}M\in {H}^{2}(BSymp;{H}_{\mathbb{Q}})$
such that the symplectic fibration
$P\to B$
classified by
$\phi :B\to BSymp$
has a closed connection form iff
${\phi}^{*}(\mathcal{O}M)=0$
. In particular,
$\mathcal{O}M\in {H}^{2}(BSymp;{H}_{\mathbb{Q}})$
is zero iff every symplectic
$M$
bundle has a closed connection form.
(ii) There is a crossed homomorphism
$\stackrel{~}{F}:Symp\to {H}^{1}(M;\mathbb{R})/\Gamma $
that extends Flux if and only if
$\mathcal{O}M$
lies in the image of
${H}^{2}(BSymp;\Gamma )$
in
${H}^{2}(BSymp;{H}_{\mathbb{Q}})$
.
Corollary 1.12.
The following conditions are equivalent:
(i) Every symplectic
$M$
bundle has a closed connection form.
(ii)
$\Gamma =0$
and there is a crossed homomorphism
$\stackrel{~}{F}:Symp\to {H}^{1}(M;\mathbb{R})$
extending Flux.
For example, when
$(M,\omega )$
is monotone, the first Chern class of the vertical tangent bundle of
$P\to B$
provides an extension of
$\left[\omega \right]$
. Therefore the obstruction class
$\mathcal{O}M$
must vanish. This is consistent with the corollary since the Kotschick–Morita homomorphism
${F}_{KM}$
extends Flux. Observe further that, because
$\mathcal{O}M\in {H}^{2}(BSymp)$
, Theorem 1.11 gives a second proof of Lemma 1.2 .
The next result clarifies the conditions under which Flux can be extended.
Proposition 1.13.
The following conditions are equivalent.
(i) There is an extension
$$\stackrel{~}{F}:Symp(M,\omega )\to {H}^{1}(M;\mathbb{R})/\Gamma $$
of Flux. (ii) For every closed Riemann surface
$\Sigma $
every representation of
${\pi}_{1}(\Sigma )$
in
${\pi}_{0}(Symp)$
lifts to a representation into the group
$Symp/Ham$
. (iii) For every product of commutators
$[{u}_{1},{u}_{2}]...[{u}_{2p1},{u}_{2p}]$
,
${u}_{i}\in Symp$
, that lies in
${Symp}_{0}$
, there are elements
${g}_{1},...,{g}_{2p}\in {Symp}_{0}$
such that
$$[{u}_{1}{g}_{1},{u}_{2}{g}_{2}]...[{u}_{2p1}{g}_{2p1},{u}_{2p}{g}_{2p}]\in Ham.$$
(iv) For every symplectic
$M$
bundle
$P\to \Sigma $
there is a bundle
$Q\to {S}^{2}$
such that the fiberwise connect sum
$P\#Q\to \Sigma \#{S}^{2}=\Sigma $
has a closed connection form.
Remark 1.14.
(i) It is immediate that (ii) is equivalent to (iii). It is also easy to see that (i) implies (ii): if we write
$\mathcal{\mathscr{H}}:=ker\stackrel{~}{F}$
, then the quotient group
$\mathcal{\mathscr{H}}/Ham$
is isomorphic to
${\pi}_{0}(Symp)$
because
$\mathcal{\mathscr{H}}\cap {Symp}_{0}=Ham$
. Hence any representation in
${\pi}_{0}(Symp)$
can be lifted to the subgroup
$\mathcal{\mathscr{H}}/Ham$
of
$Symp/Ham$
. Our proof of the converse involves some topology since it uses the relation between
$\stackrel{~}{F}$
and the obstruction cocycle
$\mathcal{O}M$
. On the other hand, if we restrict to the subgroup
${Symp}_{H}$
of
$Symp$
that acts trivially on
${H}^{1}(M;\mathbb{R})$
then (iii) is equivalent to saying that
$$\begin{array}{c}[{Symp}_{H},{Symp}_{H}]\cap {Symp}_{0}=Ham.\end{array}$$ 
(1.3)

and, as we show in Lemma 2.8 , there are immediate algebraic reasons why this is equivalent to the existence of an extension of
$Flux$
to
${Symp}_{H}$
.
We show in Corollary
4.12 that equation ( 1.3 ) holds when
$\left[\omega \right]$
vanishes on tori and lies in the subring of
${H}^{*}\left(M\right)$
generated by
${H}^{1}$
. However, it is so far unknown whether it always holds. If not, then
$\stackrel{~}{F}$
cannot always exist. On the other hand, there are indications that ( 1.3 ) might always hold. It seems that a large part of
${\pi}_{0}(Symp)$
can be generated by Dehn twists about Lagrangian spheres: cf. Seidel [
15,1.7]
. In dimensions
$>2$
these are well defined up to Hamiltonian isotopy and act trivially on
${H}^{1}\left(M\right)$
, and so one might be able to take
$\mathcal{\mathscr{H}}\cap {Symp}_{H}$
to be the group generated by Dehn twists. In any case, it does not seem that the methods used in this paper are sufficiently deep to resolve this question.
(ii) Recall that an
$M$
bundle
$P\to {S}^{2}$
has a closed connection form iff it is constructed from a loop of symplectomorphisms with trivial flux. Therefore the equivalence of (i) and (iv) in Proposition 1.13 is a geometric restatement of Theorem 1.11 (ii).
1.3 Further results and remarks
After discussing stability, we describe a few cases where it is possible to extend the Flux homomorphism. We end by discussing the integral case, and the question of uniqueness.
Stability under perturbations of
$\omega $
. It was shown in Lalonde–McDuff [
7]
that Hamiltonian bundles are stable under small perturbations of
$\omega $
. One cannot expect general symplectic bundles to be stable under arbitrary small perturbations of
$\omega $
since
${\pi}_{1}\left(B\right)$
may act nontrivially on
${H}^{2}(M;\mathbb{R})$
. Given a symplectic bundle
$(M,\omega )\to P\to B$
let us denote by
${V}_{2}\left(P\right)$
the subspace of
${H}_{2}(M;\mathbb{Q})$
generated by the elements
${g}_{*}\left(C\right)C$
, where
$C\in {H}_{2}(M;\mathbb{Z})$
and
$g$
is any symplectomorphism of
$M$
that occurs as the holonomy of a symplectic connection on
$P\to B$
around some loop in
$B$
. (Since
${g}_{*}\left(C\right)$
depends only on the smooth isotopy class of
$g$
, it does not matter which connection we use.) The subspace
${H}^{2}(M;\mathbb{R}{)}^{inv}$
, consisting of classes
$a\in {H}^{2}(M;\mathbb{R})$
that are fixed by all such
$g$
, is the annihilator of
${V}_{2}\left(P\right)$
. The most one can expect is that the existence of a symplectic structure on
$P\to B$
is stable under perturbations of
$\left[\omega \right]$
in this subspace. For example, if
$\omega $
is generic in the sense that it gives an injective map
${H}_{2}(M;\mathbb{Z})/Tor\to \mathbb{R}$
then
${V}_{2}\left(P\right)$
is torsion and
${H}^{2}(M;\mathbb{R}{)}^{inv\left(P\right)}={H}^{2}(M;\mathbb{R})$
.
Proposition 1.15.
Let
$(M,\omega )\to P\to B$
be a symplectic
$M$
bundle over a finite simplicial complex
$B$
. Then there is a neighborhood
$\mathcal{N}\left(\omega \right)$
of
$\omega $
in the space of all closed
$2$
forms on
$M$
that represent a class in
${H}^{2}(M;\mathbb{R}{)}^{inv\left(P\right)}$
such that for all
${\omega}^{\prime}\in \mathcal{N}\left(\omega \right)$
:
(i)
$P\to B$
has the structure of an
${\omega}^{\prime}$
symplectic bundle, and (ii) if there is a closed extension of
$\omega $
, then the same is true for
${\omega}^{\prime}$
.
Part (i) of this proposition follows by the arguments in [7,Cor. 2.5] . Part (ii) was also proved in [7] in the case when
$B$
is classified by a map into
$B{Symp}_{0}(M,\omega )$
. The proof of the general case is given at the end of § 3 . The next corollary is an immediate consequence of (ii).
Corollary 1.16.
If
$P\to B$
has a
${Ham}^{s}$
structure then the image of the restriction map
${H}^{2}(P;\mathbb{R})\to {H}^{2}(M;\mathbb{R})$
is the subspace
${H}^{2}(M;\mathbb{R}{)}^{inv\left(P\right)}$
of
${H}^{2}(M;\mathbb{R})$
that is invariant under the action of
${\pi}_{1}\left(B\right)$
.
This result implies that the differential
${d}_{2}^{2,0}$
in the Leray–Serre cohomology spectral sequence for
$P\to B$
vanishes,^{5 }
and so is a partial generalization of the vanishing results in [7] .
Remark 1.17.
Proposition 1.15 is proved using the Moser homotopy argument and so works only over compact pieces of
$Symp$
. This is enough to give stability for bundles over finite bases
$B$
but is not enough to allow one to make any statements about properties that involve the full group
$Symp$
. Hence even if
$Flux$
extends to
$\stackrel{~}{F}:Symp\to {H}^{1}(M;\mathbb{R})/\Gamma $
for the manifold
$(M,\omega )$
, it is not clear that it also extends for sufficiently close forms
${\omega}^{\prime}$
whose cohomology class is invariant under
$Symp(M,\omega )$
.
For one thing, however close
${\omega}^{\prime}$
is, there may be new components of
$Symp(M,{\omega}^{\prime})$
containing elements that are far from those in
$Symp(M,\omega )$
.
Manifolds with
$\Gamma =0$
. One expects that for most manifolds
$\Gamma =0$
. Whether
$\mathcal{O}M$
then vanishes is still not clear. We now discuss some special cases in which
$\Gamma =0$
and Flux extends to a crossed homomorphism defined either on the whole group
$Symp$
or on some large subgroup. Note that if Flux extends to
$Symp$
, then
$\mathcal{O}M=0$
by Theorem 1.11 (ii).
The first case is when
$(M,\omega )$
is (strongly) monotone, i.e. the symplectic class
$\left[\omega \right]$
is a multiple of the first Chern class. In this case
$\mathcal{O}M=0$
since one can always choose a closed connection form in the class of a multiple of the vertical first Chern class. Kotschick–Morita [
5]
observed that Flux always extends. We shall give an explicit formula for
$\stackrel{~}{F}$
in Theorem 4.8 .
Another somewhat tractable case is when
$(M,\omega )$
is atoroidal, i.e.
${\int}_{{\mathbb{T}}^{2}}{\psi}^{*}\omega =0$
for all smooth maps
$\psi :{\mathbb{T}}^{2}\to M$
. Note that
$\Gamma =0$
for such manifolds, because for each loop
$\left\{{f}_{t}\right\}$
in
${Symp}_{0}$
the value of the class
$Flux\left(\right\{{f}_{t}\left\}\right)$
on the
$1$
cycle
$\gamma $
is obtained by integrating
$\omega $
over the torus
${\cup}_{t}{f}_{t}\left(\gamma \right)$
. In the next proposition, we denote by
${Symp}_{\pi}$
the subgroup of
$Symp$
consisting of elements that are isotopic to a symplectomorphism that fixes the basepoint
${x}_{0}$
of
$M$
and induces the identity map
${\pi}_{1}(M,{x}_{0})\to {\pi}_{1}(M,{x}_{0})$
.^{6 }
Proposition 1.18.
If
$(M,\omega )$
is atoroidal then
$\Gamma =0$
and
$Flux$
extends to a homomorphism
$\stackrel{~}{F}:{Symp}_{\pi}\to {H}^{1}(M;\mathbb{R})$
.
We shall see in § 4.2 that in the above situation
$\stackrel{~}{F}$
can be extended to a crossed homomorphism defined on the whole of
$Symp$
but at the cost of enlarging the target group.
Proposition
1.18 gives a partial answer to Kedra–Kotschick–Morita's question [
3]
of whether the usual flux homomorphism
${Symp}_{0}\to {H}^{1}(M;\mathbb{R})$
extends to the full group
$Symp$
when
$\left[\omega \right]$
is a bounded class. This condition means that
$\left[\omega \right]$
may be represented by a singular cocycle that is uniformly bounded on the set of all singular
$2$
simplices.^{7 }
If
$\left[\omega \right]$
is bounded, then
$(M,\omega )$
is atoroidal since an arbitrary multiple of a toric class
$C$
can be represented by the sum of just two singular
$2$
simplices. Another interesting atoroidal case is that of symplectically hyperbolic manifolds. There are various possible definitions here.
We shall use Polterovich's definition from [
14]
in which
$(M,\omega )$
is called symplectically hyperbolic if the pullback
$\stackrel{~}{\omega}$
of
$\omega $
to the universal cover
$\stackrel{~}{M}$
of
$M$
has bounded primitive, i.e.
$\stackrel{~}{\omega}=d\beta $
for some
$1$
form
$\beta $
that is bounded with respect to any metric on
$\stackrel{~}{M}$
that is pulled back from
$M$
. For example,
$(M,\omega )$
might be a product of Riemann surfaces of genus
$>1$
with a product symplectic form. Because in the covering
${\mathbb{R}}^{2}\to {\mathbb{T}}^{2}$
the boundary of a square of side
$N$
encloses
${N}^{2}$
fundamental domains, it is easy to check that any
$2$
form on
${\mathbb{T}}^{2}$
whose pullback to
${\mathbb{R}}^{2}$
has bounded primitive must have zero integral over
${\mathbb{T}}^{2}$
. Hence we find:
Lemma 1.19.
Proposition 1.18 applies both when
$\left[\omega \right]$
is bounded and when
$(M,\omega )$
is symplectically hyperbolic.
§ 4.2 contains a few other similar results that are valid in special cases, for example when
$\omega $
vanishes on
${\pi}_{2}\left(M\right)$
. We end the introduction with some general remarks.
Remark 1.20.
(The integral case.) There is an analogous group
${Ham}^{s\mathbb{Z}}$
which is the kernel of a crossed homomorphism
${\widehat{F}}_{s}^{\mathbb{Z}}$
with values in
${H}^{1}(M;\mathbb{R}/\mathcal{P}\mathbb{Z}\omega )$
where
$\mathcal{P}\mathbb{Z}\omega $
denotes the set of values of
$\left[\omega \right]$
on the integral
$2$
cycles
${H}_{2}(M;\mathbb{Z})$
. In many respects the properties of this group are similar to those of
${Ham}^{s}$
. However, there are some interesting differences. If
$Tor$
denotes the torsion subgroup of
${H}_{1}(M;\mathbb{Z})$
, then the analog of the group
$\mathcal{A}$
occurring in equation ( 1.2 ) is
$${\mathcal{A}}^{\mathbb{Z}}:={H}^{1}(M;\mathbb{R}/\mathcal{P}\mathbb{Z}\omega )\sim =Hom(Tor,\mathbb{R}/\mathcal{P}\mathbb{Z}\omega )\oplus {H}^{1}(M;\mathbb{R})/{H}^{1}(M;\mathcal{P}\mathbb{Z}\omega ).$$
Hence Theorem 1.11 (i) does not immediately generalize; the proof of Lemma 3.4 shows that the obstruction to the existence of a
${Ham}^{s\mathbb{Z}}$
structure is twofold, the first coming from the finite group
${\pi}_{0}\left({\mathcal{A}}^{\mathbb{Z}}\right)$
(see Lemma 4.1 ) and the second an obstruction cocycle similar to
$\mathcal{O}M$
coming from
${\pi}_{1}\left({\mathcal{A}}^{\mathbb{Z}}\right)$
. Nevertheless, since every
$Hams$
bundle over a compact base
$B$
has a finite cover with a
${Ham}^{s\mathbb{Z}}$
structure, the latter group has the modified extension property.
The group
${Ham}^{s\mathbb{Z}}$
is most interesting in the case when
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
, i.e. when
$\left[\omega \right]$
is a primitive integral class. In this situation one might expect
$B{Ham}^{s\mathbb{Z}}$
to classify bundles
$P\to M$
that have a closed and integral connection form. We show in Proposition 3.3 that this is the case when
${\pi}_{0}(Symp)$
acts trivially on
${H}_{1}(M;\mathbb{Z})$
. Gal–Kȩdra [
1]
show that this remains true when
${\pi}_{0}(Symp)$
acts trivially on the torsion classes in
${H}_{1}(M;\mathbb{Z})$
. (This hypothesis is equivalent to saying that the action of
$Symp$
on the set of prequantum line bundles is trivial.) In general, however, we show in Example 4.2 that there may be no such classifying space. To get a well posed classification problem one must specify a particular integral lift of
$\left[\omega \right]$
; equivalently, one must choose a prequantum line bundle.
This question is pursued further in Gal–Kȩdra [
1]
. We show here only that the choice of
$s$
determines a prequantum line bundle and that
${Ham}^{s\mathbb{Z}}$
is homotopy equivalent to the automorphism group of this bundle. (See Proposition 4.5 .)
Remark 1.21.
(Issues of uniqueness) (i) Because we are interested in the algebraic and geometric properties of the symplectomorphism group we restricted ourselves above to the case when
$\mathcal{\mathscr{H}}$
is a subgroup of
$Symp$
. However, from a homotopy theoretic point of view it would be more natural to look for a group
$\mathcal{K}$
that classifies pairs consisting of a symplectic
$M$
bundle
$\pi :P\to B$
together with an extension
$\stackrel{~}{a}\in {H}^{2}(P;\mathbb{R})$
of the fiberwise symplectic class
$\left[\omega \right]$
. Here we should either normalize
$\stackrel{~}{a}$
by requiring
${\pi}_{!}\left({\stackrel{~}{a}}^{n+1}\right)=0$
(where
${\pi}_{!}$
denotes integration over the fiber) or consider
$\stackrel{~}{a}$
to be well defined modulo elements in
${\pi}^{*}{H}^{2}\left(B\right)$
. Then the homotopy class of
$B\mathcal{K}$
would be well defined and there would be a forgetful map
$\psi :B\mathcal{K}\to BSymp$
which is well defined up to homotopy (assuming that we are working in the category of spaces with the homotopy type of a CW complex). In general,
$\psi $
would not be a homotopy equivalence since the extension class
$\stackrel{~}{a}\in {H}^{2}(P;\mathbb{R})$
could vary by an element in
${H}^{1}(B;{H}^{1}(M;\mathbb{R}\left)\right)$
. Further, in this scenario,
$\mathcal{K}$
need not be a subgroup of
$Symp$
. (Cf. the discussion in Lalonde–McDuff [
7]
of the classification of Hamiltonian structures.) (ii) If we insist that
$\mathcal{K}$
be a subgroup of
$Symp$
then there are several possible notions of equivalence, the most natural of which is perhaps given by conjugation by an element in
${Symp}_{0}$
. With this definition equivalent groups would be isomorphic. We show in § 2 that the groups
${Ham}^{s}$
are equivalent in this sense, though when
${H}_{1}(M;\mathbb{Z})$
has torsion the integer versions
${Ham}^{s\mathbb{Z}}$
may not be. It is also not clear whether any two groups
${\mathcal{\mathscr{H}}}_{1},{\mathcal{\mathscr{H}}}_{2}$
that intersect each component of
$Symp$
and satisfy
${\mathcal{\mathscr{H}}}_{1}\cap {Symp}_{0}={\mathcal{\mathscr{H}}}_{2}\cap {Symp}_{0}=Ham$
must be isomorphic as abstract groups, although any such group must be isomorphic to an extension of
${\pi}_{0}(Symp)$
by
$Ham$
. However, there is no immediate reason why they should be conjugate. For example, suppose that the group
${\pi}_{0}(Symp)$
is isomorphic to
$\mathbb{Z}$
, generated by the component
${Symp}_{\alpha}$
of
$Symp$
. Then because
$Ham$
is a normal subgroup of
$Symp$
the subgroup
${\mathcal{\mathscr{H}}}_{g}$
of
$Symp$
generated by
$Ham$
together with any element
$g\in {Symp}_{\alpha}$
intersects
${Symp}_{0}$
in
$Ham$
and therefore has the required properties. Any two such groups
${\mathcal{\mathscr{H}}}_{{g}_{i}},i=1,2,$
are isomorphic, though they are conjugate only if there is
$h\in Symp$
such that
${g}_{1}h{g}_{2}^{1}{h}^{1}\in Ham$
. On the other hand, because
${g}_{1}$
and
${g}_{2}$
can be joined by an isotopy, there is a smooth family of injective group homomorphisms
${\iota}_{t}:{\mathcal{\mathscr{H}}}_{{g}_{1}}\to Symp$
,
$t\in [1,2]$
, that starts with the inclusion and ends with an isomorphism onto
${\mathcal{\mathscr{H}}}_{{g}_{2}}$
. Thus the homotopy properties of the inclusions
${\mathcal{\mathscr{H}}}_{{g}_{i}}\to Symp$
are the same.
(iii) Instead of looking for subgroups of
$Symp$
with the extension property one could look for covering groups
$\mathcal{\mathscr{H}}\to Symp$
with this property. Notice that if
$\Lambda $
is a discrete subgroup of an abelian topological group
$\mathcal{A}$
and if the continuous crossed homomorphism
$F:G\to \mathcal{A}/\Lambda $
extends the composite
${F}_{0}:{G}_{0}\to \mathcal{A}\to \mathcal{A}/\Lambda $
, where
$f:{G}_{0}\to \mathcal{A}$
is a homomorphism defined on the identity component of
${G}_{0}$
, then the fiber product
$$\stackrel{~}{G}:=\left\{(g,a)\in G\times \mathcal{A}\leftF\right(g)=a+\Lambda \right\}$$
of
$G$
and
$\mathcal{A}$
over
$\mathcal{A}/\Lambda $
is a covering group of
$G$
that contains a copy of
${G}_{0}$
, namely the graph of
$f$
. Moreover, the obvious projection
$\stackrel{~}{G}\to \mathcal{A}$
lifts
${F}_{0}$
. This approach is particularly relevant in the integral case mentioned in Remark 1.20 above, as well as the cohomologically symplectic case, where the analog of the Hamiltonian group is already a covering group of
${Diff}_{0}$
. For further discussion see § 4.3 and Gal–Kȩdra [
1]
.
Contents
Acknowledgements. The author thanks Swiatoslav Gal, Jarek Kȩdra, Jack Milnor, Leonid Polterovich and Zhigang Han for useful discussions. She also thanks Gal and Kȩdra for showing her early drafts of their paper [
1]
and making various helpful comments about earlier versions of this paper. In particular, they helped streamline the proof of Lemma 3.4 .
2 Definition and Properties of
${\widehat{F}}_{s}$
Define
$\mathcal{P}\mathbb{Z}\omega $
(resp.
$\mathcal{P}\omega :=\mathcal{P}\mathbb{Q}\omega $
) to be the set of values taken by
$\left[\omega \right]$
on the elements of
${H}_{2}(M;\mathbb{Z})$
(resp.
${H}_{2}(M;\mathbb{Q})$
). To define
${\widehat{F}}_{s}$
we follow a suggestion of Polterovich (explained in Lalonde–McDuff [
7]
). Define the homology group
$$S{H}_{1}(M,\omega ;\mathbb{Z})$$
to be the quotient of the space of integral
$1$
cycles in
$M$
by the image under the boundary map
$\partial $
of the integral
$2$
chains with zero symplectic area. Then there is a projection
${\pi}_{\mathbb{Z}}:S{H}_{1}(M,\omega ;\mathbb{Z})\to {H}_{1}(M;\mathbb{Z})$
and we set
$$S{H}_{1}(M,\omega ):=S{H}_{1}(M,\omega ;\mathbb{Z})\otimes \mathbb{Q}.$$
We shall consider
$S{H}_{1}(M,\omega )$
and
$\mathcal{P}\mathbb{Q}\omega $
as
$\mathbb{Q}$
vector spaces. Given a loop (or integral
$1$
cycle)
$\ell $
in
$M$
we denote its image in
${H}_{1}(M;\mathbb{Z})$
or
${H}_{1}(M;\mathbb{Q})$
by
$[\ell ]$
and its image in
$S{H}_{1}(M,\omega ;\mathbb{Z})$
or
$S{H}_{1}(M,\omega )$
by
$\langle \ell \rangle .$
We usually work over the rationals and shall omit the label
$\mathbb{Q}$
unless there is a possibility of confusion.
Lemma 2.1.
There are split exact sequences
$$\begin{array}{c}0\to \mathbb{R}/\mathcal{P}\mathbb{Z}\omega \to S{H}_{1}(M,\omega ;\mathbb{Z})\stackrel{{\pi}_{\mathbb{Z}}}{\to}{H}_{1}(M;\mathbb{Z})\to 0,\end{array}$$ 
(2.1)

and
$$\begin{array}{c}0\to \mathbb{R}/\mathcal{P}\mathbb{Q}\omega \to S{H}_{1}(M,\omega )\stackrel{\pi}{\to}{H}_{1}(M;\mathbb{Q})\to 0.\end{array}$$ 
(2.2)


Proof.
Choose a continuous family of integral
$2$
chains
${f}_{t}:D\to M$
for
$t\in \mathbb{R}$
with
${\int}_{D}{f}_{t}^{*}\omega =t$
. If
${\gamma}_{t}:={f}_{t}{}_{\partial D}$
denotes the boundary of
${f}_{t}$
, then the elements
$$\langle {\gamma}_{t}\rangle ,t\in \mathbb{R},$$
generate the kernel of the projection
${\pi}_{\mathbb{Z}}:S{H}_{1}(M,\omega ;\mathbb{Z})\to {H}_{1}(M,\mathbb{Z})$
. Moreover they represent different classes in
$S{H}_{1}(M,\omega ;\mathbb{Z})$
if and only if
$t{t}^{\prime}/\in \mathcal{P}\mathbb{Z}\omega $
. Hence the sequence
$$0\to \mathbb{R}/\mathcal{P}\mathbb{Z}\omega \to S{H}_{1}(M,\omega ;\mathbb{Z})\stackrel{{\pi}_{\mathbb{Z}}}{\to}{H}_{1}(M;\mathbb{Z})\to 0,$$
is exact. To see that it splits, we just need to check that each element
$\lambda =[\ell ]$
of finite order
$N$
in
${H}_{1}(M;\mathbb{Z})$
is the image of some element of order
$N$
in
$S{H}_{1}(M;\mathbb{Z})$
. But if
$W$
is an integral
$2$
chain such that
$\partial W=N\ell $
and if
$\mu :={\int}_{W}\omega $
then
$$N\left(\langle \ell \rangle \langle {\gamma}_{\mu /N}\rangle \right)=0\text{and}\pi \left(\langle \ell \rangle \langle {\gamma}_{\mu /N}\rangle \right)=[\ell ].$$
In fact every element of order
$N$
in the coset
${\pi}_{\mathbb{Z}}^{1}\left(\right[\ell \left]\right)$
has the form
$\langle \ell \rangle \langle {\gamma}_{\nu}\rangle $
where
$N\nu \in \mu +\mathcal{P}\mathbb{Z}\omega $
. The proof for ( 2.2 ) is similar. □
We explain in § 4.1 a natural way to understand splittings of
${\pi}_{\mathbb{Z}}$
in the case when
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
: cf. Definition 4.3 . Note also that in the previous lemma there is no need for
$\omega $
to be nondegenerate; it suffices for it to be closed. However if it were an arbitrary closed form it would not have many isometries, and so the next lemmas would have little interest.
Lemma 2.2.
The group
$Symp(M,\omega )$
acts on
$S{H}_{1}(M,\omega ;\mathbb{Z})$
and
$S{H}_{1}(M,\omega )$
.
The induced action of
${Symp}_{0}$
on the set of splittings of
$\pi $
is transitive. When
${H}_{1}(M;\mathbb{Z})$
has no torsion
$Symp$
also acts transitively on the splittings of
${\pi}_{\mathbb{Z}}$
.

Proof.
Again, we shall work with the sequence over
$\mathbb{Z}$
. The group
$Symp(M,\omega )$
acts on these spaces because it preserves
$\omega $
. To prove the transitivity statement, note first that any splitting
$s$
of
${\pi}_{\mathbb{Z}}$
has the form
$s{\lambda}_{i}=\langle {\ell}_{i}\rangle $
where
${\ell}_{1},...,{\ell}_{k}$
are loops (i.e. integral
$1$
cycles) in
$M$
that project to the basis
${\lambda}_{1},...,{\lambda}_{k}$
of
${H}_{1}(M;\mathbb{Z})$
. Suppose given two such splittings
$s,{s}^{\prime}$
corresponding to different sets
$L,{L}^{\prime}$
of representing
$1$
cycles for the
${\lambda}_{i}$
. Suppose also that
$dimM>2$
. Since Hamiltonian isotopies have zero flux, we may move the loops in
$L$
and
${L}^{\prime}$
by such isotopies, without affecting their images in
$S{H}_{1}(M,\omega ;\mathbb{Z})$
and so that no two intersect. Now choose
${T}_{1},...,{T}_{k}\in {\mathbb{R}}^{+}$
such that
$$\langle {\ell}_{i}^{\prime}\rangle =\langle {\ell}_{i}\rangle +\langle {\gamma}_{{T}_{i}}\rangle ,1\le i\le k,$$
where the
${\gamma}_{t}$
are as in Lemma 2.1 . For each
$i$
there is a symplectic isotopy
${h}_{i,t}$
such that for all
$t\in [0,{T}_{i}]$
,
$${h}_{i,t}{}_{{\ell}_{j}}=id,j<i,{h}_{i,t}{}_{{h}_{j,{T}_{j}}{\ell}_{j}}=id,j>i,{\int}_{{W}_{i}}\omega ={T}_{i},$$
where
${W}_{i}:={\cup}_{0\le t\le {T}_{i}}{h}_{i,t}\left({\ell}_{i}\right)$
. (Take the
${h}_{i,t}$
to be generated by closed
$1$
forms
${\alpha}_{i}$
that vanish near the appropriate loops and are such that
${\int}_{{\ell}_{i}}{\alpha}_{i}\ne 0$
. Here we are using the fact that
$\left[{\ell}_{i}\right]$
is not a torsion class.) Then
$h:={h}_{1,{T}_{1}}\circ \cdot \cdot \cdot \circ {h}_{k,{T}_{k}}$
takes
$s$
to
${s}^{\prime}$
.
To extend this argument to the case
$dimM=2$
, it is convenient to describe the splitting by its effect on a standard basis
${\lambda}_{i}$
of
${H}_{1}(M;\mathbb{Z})$
. Thus we may assume that
${\ell}_{i}$
and
${\ell}_{j}$
are disjoint unless
$(i,j)=(2k1,2k)$
in which case they intersect in a single point. If
${s}_{0}$
is the splitting defined by these loops, it suffices to show that for any numbers
${T}_{i}$
there are representatives
${\ell}_{i}^{\prime}$
for the
$\left[{\ell}_{i}\right]$
such that for each
$i$
there is a cylinder of area
${T}_{i}$
with boundary
${\ell}_{i}^{\prime}{\ell}_{i}$
. One achieves this by first isotoping the
${\ell}_{i}$
for
$i$
odd (fixing the other loops), and then adjusting the
${\ell}_{i}$
for even
$i$
. □
Choose a splitting
$s$
for
${\pi}_{\mathbb{Z}}$
. If
$h\in Symp$
and
$\lambda \in {H}_{1}(M;\mathbb{Z})$
, then the element
${h}_{*}\left(s\lambda \right)s\left({h}_{*}\lambda \right)$
lies in the kernel of
${\pi}_{\mathbb{Z}}:S{H}_{1}(M,\omega )\to {H}_{1}(M;\mathbb{Z})$
and one can define a map
$${\widehat{F}}_{s}^{\mathbb{Z}}:Symp\left(M\right)\to {\mathcal{A}}^{\mathbb{Z}}:=Hom\left({H}_{1}(M;\mathbb{Z}),\mathbb{R}/\mathcal{P}\mathbb{Z}\omega \right)$$
by setting
$$\begin{array}{c}{\widehat{F}}_{s}^{\mathbb{Z}}\left(h\right)\left(\lambda \right):={h}_{*}\left(s\lambda \right)s\left({h}_{*}\lambda \right)\in \mathbb{R}/\mathcal{P}\mathbb{Z}\omega ,\lambda \in {H}_{1}(M;\mathbb{Z}).\end{array}$$ 
(2.3)

Explicitly, if we denote by
$\overline{\lambda}$
the image
$s\left(\lambda \right)$
of
$\lambda \in {H}_{1}\left(M\right)$
, then
$$\begin{array}{c}{\widehat{F}}_{s}\left(h\right)\left(\lambda \right)=a\left(h\overline{\lambda}\overline{h\lambda}\right),\end{array}$$ 
(2.4)

where
$a(\langle {\ell}^{\prime}\rangle \langle \ell \rangle )$
is the symplectic area of any cycle with boundary
${\ell}^{\prime}\ell $
. Similarly, for each splitting
$s$
of ( 2.2 ) we define
$${\widehat{F}}_{s}:Symp\left(M\right)\to \mathcal{A}:=Hom\left({H}_{1}\left(M\right),\mathbb{R}/\mathcal{P}\mathbb{Q}\omega \right)={H}^{1}(M;\mathbb{R})/{H}^{1}(M;\mathcal{P}\mathbb{Q}\omega )$$
by
$${\widehat{F}}_{s}\left(h\right)\left(\lambda \right):={h}_{*}\left(s\lambda \right)s\left({h}_{*}\lambda \right)\in \mathbb{R}/\mathcal{P}\mathbb{Q}\omega ,\lambda \in {H}_{1}(M;\mathbb{Q}).$$
Proposition 2.3.
(i)
${\widehat{F}}_{s}^{\mathbb{Z}}$
is a crossed homomorphism that equals the composite
$${Symp}_{0}\stackrel{Flux}{\to}{H}^{1}(M;\mathbb{R})/\Gamma \to {\mathcal{A}}^{\mathbb{Z}}:=Hom\left({H}_{1}(M;\mathbb{Z}),\mathbb{R}/\mathcal{P}\mathbb{Z}\omega \right)$$
on
${Symp}_{0}$
. Moreover the class
$$\left[{\widehat{F}}^{\mathbb{Z}}\right]:=\left[{\widehat{F}}_{s}^{\mathbb{Z}}\right]\in {H}_{cEM}^{1}\left(Symp,{\mathcal{A}}^{\mathbb{Z}}\right)$$
is independent of the choice of
$s$
.
(ii) The analogous statements hold for
${\widehat{F}}_{s}$
.

Proof.
${\widehat{F}}_{s}^{\mathbb{Z}}$
is a crossed homomorphism because for all
$g,h\in Symp$
$$\begin{array}{ccc}{\widehat{F}}_{s}^{\mathbb{Z}}\left(gh\right)\left(\lambda \right)& =& a\left(gh\overline{\lambda}\overline{gh\lambda}\right)\end{array}$$  
$$\begin{array}{ccc}& =& a\left(gh\overline{\lambda}g\overline{h\lambda}\right)+a\left(g\overline{h\lambda}\overline{gh\lambda}\right)\end{array}$$  
$$\begin{array}{ccc}& =& a\left(h\overline{\lambda}\overline{h\lambda}\right)+a\left(g\overline{h\lambda}\overline{gh\lambda}\right)\end{array}$$  
$$\begin{array}{ccc}& =& {\widehat{F}}_{s}^{\mathbb{Z}}\left(h\right)\left(\lambda \right)+{\widehat{F}}_{s}^{\mathbb{Z}}\left(g\right)\left(h\lambda \right)\end{array}$$  
$$\begin{array}{ccc}& =& {\widehat{F}}_{s}^{\mathbb{Z}}\left(h\right)\left(\lambda \right)+{h}^{*}{\widehat{F}}_{s}^{\mathbb{Z}}\left(g\right)\left(\lambda \right).\end{array}$$  
The rest of the first statement in (i) is immediate from the definition.
To prove the second statement in (i) observe that two choices of splitting
$s,{s}^{\prime}$
differ by the element
$\alpha \in {\mathcal{A}}^{\mathbb{Z}}:=Hom\left({H}_{1}(M;\mathbb{Z}),\mathbb{R}/\mathcal{P}\mathbb{Z}\omega \right)$
given by
$$\alpha \left(\lambda \right):={s}^{\prime}\left(\lambda \right)s\left(\lambda \right)\in \mathbb{R}/\mathcal{P}\mathbb{Z}\omega ,\lambda \in {H}_{1}(M;\mathbb{Z}).$$
It follows easily that
$$\begin{array}{c}{\widehat{F}}_{{s}^{\prime}}^{\mathbb{Z}}\left(h\right){\widehat{F}}_{s}^{\mathbb{Z}}\left(h\right)=\alpha {h}^{*}\alpha ,\end{array}$$ 
(2.5)

and so is a coboundary in the Eilenberg–MacLane complex.
The proof of (ii) is similar. □
Definition 2.4.
Given a splitting
$s:{H}_{1}(M;\mathbb{Q})\to S{H}_{1}(M,\omega )$
we define the enlarged Hamiltonian group
$Hams(M,\omega )$
to be the kernel of
${\widehat{F}}_{s}$
. Similarly, we define
${Ham}^{s\mathbb{Z}}(M,\omega )$
to be the kernel of the integral crossed homomorphism
${\widehat{F}}_{s}^{\mathbb{Z}}$
.
Lemma 2.5.
Let
$s,{s}^{\prime}$
be two splittings and define
${Symp}_{H}$
to be the subgroup of
$Symp$
that acts trivially on rational homology.
(i)
${Symp}_{H}\cap Hams={Symp}_{H}\cap {Ham}^{{s}^{\prime}}$
.
(ii) The map
${\pi}_{0}\left({Ham}^{s}\right)\to {\pi}_{0}(Symp)$
is surjective.
(iii) The subgroups
$Hams$
and
${Ham}^{{s}^{\prime}}$
are conjugate in
$Symp$
by an element in
${Symp}_{0}$
.
(iv) When topologized as a subspace of
$Symp$
, the path component of
$Hams$
containing the identity element is
$Ham(M,\omega )$
.

Proof.
(i) is an immediate consequence of the identity ( 2.5 ). (iii) follows from the fact that
${Symp}_{0}$
acts transitively on the set of splittings and the description of
$Hams$
as the subgroup of
$Symp$
whose action on
$S{H}_{1}(M,\omega )$
preserves the image of
$s$
. To prove (ii), we must show that any element
$h\in Symp$
is homotopic to an element in
${Ham}^{s}$
. This holds because the splittings
${s}^{\prime}={h}_{*}\left(s\right)$
and
$s$
are conjugate by an element in
${Symp}_{0}$
. To prove (iv) consider a continuous path
${h}_{t}\in Symp$
that starts at the identity and is such that
${\widehat{F}}_{s}\left({h}_{t}\right)\left(\lambda \right)=0\in \mathbb{R}/\mathcal{P}\omega $
for all
$t$
. By Proposition 2.3 , the path
$t\mapsto {\widehat{F}}_{s}\left({h}_{t}\right)\left(\lambda \right)\in \mathbb{R}/\mathcal{P}\omega $
has the continuous lift
$t\mapsto Flux\left({h}_{t}\right)\left(\lambda \right)\in \mathbb{R}$
. Since
$\mathcal{P}\omega $
is totally disconnected this lift must be identically zero; in other words the path
${h}_{t}$
is a Hamiltonian isotopy. □
Part (iv) of Lemma 2.5 holds for the group
${Ham}^{s\mathbb{Z}}$
, and (i) holds if one replaces
${Symp}_{H}$
by the group that acts trivially on
${H}_{1}(M;\mathbb{Z})$
. However, one must take care with the other two statements. For further details see § 4.1 .
Remark 2.6.
(Topologies on
$Hams$
and
$\mathbb{R}/\mathcal{P}\omega $
.) The intersection
$Hams\cap {Symp}_{0}$
is disconnected. In fact it is everywhere dense in
${Symp}_{0}$
. Hence the subspace topology
${\tau}_{s}$
on
$Hams$
is rather counterintuitive and it is better to give
$Hams$
a finer topology in which its path components are closed. Therefore, although we give the group
$Symp$
the usual
${C}^{\infty}$
topology (which is the subspace topology it inherits from the diffeomorphism group), we give
$Hams$
the topology
${\tau}_{c}$
that it inherits from the Hamiltonian topology on
$Symp$
. Then the identity map
$(Hams,{\tau}_{c})\to (Hams,{\tau}_{s})$
is continuous and is a weak homotopy equivalence. Thus this change in topology does not affect the homotopy or (co)homology of the space.
Correspondingly we shall always think of
$\mathcal{P}\omega $
as a discrete group. Further we think of quotients such as
$\mathbb{R}/\mathcal{P}\omega $
as quasitopological spaces, i.e. we specify which maps
$f:X\to \mathbb{R}/\mathcal{P}\omega $
are continuous, where
$X$
is a finite simplicial complex. This gives enough structure so that we can talk of homotopy groups. In the present situation we say that
$f$
is continuous iff
$X$
has a subdivision
${X}^{\prime}$
such that the restriction of
$f$
to each simplex in
${X}^{\prime}$
has a continuous lift to
$\mathbb{R}$
. Hence
$${\pi}_{1}(\mathbb{R}/\mathcal{P}\omega )\sim =\mathcal{P}\omega ,{\pi}_{j}(\mathbb{R}/\mathcal{P}\omega )=0,j>1.$$
Thus
$\mathbb{R}/\mathcal{P}\omega $
is (weakly homotopic to) the Eilenberg–MacLane space
$K(\mathcal{P}\omega ,1)$
. (Another way to deal with this technical problem — that also arises when one deals with spaces of germs — is to replace
$\mathbb{R}/\mathcal{P}\omega $
by an appropriate semisimplicial complex. But then one has to replace all spaces and groups by their semisimplicial analogs.)
$\square $
Remark 2.7.
If
$g\in Symp$
and
$h\in {Symp}_{0}$
then it is easy to check that
$Flux\left({g}^{1}hg\right)={g}^{*}(Fluxh)$
. Hence, if
${Symp}_{H}$
denotes the subgroup of
$Symp$
acting trivially on
${H}^{1}(M;\mathbb{R})$
then
$$[{Symp}_{0},{Symp}_{H}]=Ham.$$
On the other hand
$[{Symp}_{0},Symp]=Ham$
only if
$Symp={Symp}_{H}$
. Hence when
$Symp\ne {Symp}_{H}$
the flux homomorphism does not extend to a homomorphism
$Symp\to {H}^{1}\left(M\right)/\Gamma $
.
There is another relevant subgroup, namely
${Symp}_{{H}^{\mathbb{Z}}}$
, consisting of elements that act trivially on
${H}_{1}(M;\mathbb{Z})$
. Note that
$[{Symp}_{{H}^{\mathbb{Z}}},{Symp}_{{H}^{\mathbb{Z}}}]\cap {Symp}_{0}$
lies in
${Ham}^{s\mathbb{Z}}$
because
${\widehat{F}}_{s}^{\mathbb{Z}}$
restricts to a homomorphism on
${Symp}_{{H}^{\mathbb{Z}}}$
and so vanishes on the commutator subgroup
$[{Symp}_{{H}^{\mathbb{Z}}},{Symp}_{{H}^{\mathbb{Z}}}]$
. But this is the best we can say; in particular, it is not clear whether
$[{Symp}_{H},{Symp}_{H}]\cap {Symp}_{0}$
must always equal
$Ham$
.
Lemma 2.8.
The following statements are equivalent.
(i)
$[{Symp}_{H},{Symp}_{H}]\cap {Symp}_{0}=Ham$
; (ii) For every product of commutators
$y:=[{u}_{1},{u}_{2}]...[{u}_{2p1},{u}_{2p}]$
,
${u}_{i}\in {Symp}_{H}$
, that lies in
${Symp}_{0}$
, there are elements
${g}_{1},...,{g}_{2p}\in {Symp}_{0}$
such that
$$f:=[{u}_{1}{g}_{1},{u}_{2}{g}_{2}]...[{u}_{2p1}{g}_{2p1},{u}_{2p}{g}_{2p}]\in Ham.$$
(iii) The flux homomorphism
$Flux:{Symp}_{0}\to {H}^{1}(M;\mathbb{R})/\Gamma $
extends to a continuous homomorphism
$F:{Symp}_{H}\to {H}^{1}(M;\mathbb{R})/\Gamma $
.

Proof.
Clearly (iii) implies (i), which in turn implies (ii). To see that (ii) implies (i), note the identity
$$[ug,vh]={g}^{u}{h}^{uv}\left({g}^{1}{)}^{uv}\right({h}^{1}{)}^{uv{u}^{1}}[u,v],$$
where
${g}^{a}:=ag{a}^{1}$
. It follows that
$f{y}^{1}$
may be written as a product of terms of the form
${g}_{2i1}^{\prime}{g}_{2i}^{\prime}\left({g}_{2i1}^{\prime \prime}{)}^{1}\right({g}_{2i}^{\prime \prime}{)}^{1}$
where
${g}_{j}^{\prime}$
and
${g}_{j}^{\prime \prime}$
are conjugate to
${g}_{j}$
by products of the
${u}_{i}$
.
Since the
${u}_{i}$
lie in
${Symp}_{H}$
,
$Flux\left({g}_{j}^{\prime}\right)=Flux\left({g}_{j}^{\prime \prime}\right)=Flux{g}_{j}$
. Hence
$Fluxy=Fluxf=0$
, and
$y\in Ham$
.
It remains to show that (i) implies (iii). As in the proof of Proposition
1.8 given in §1, it suffices to find a section
$$\sigma :{\pi}_{0}\left({Symp}_{H}\right)\to {Symp}_{H},\alpha \mapsto {\sigma}_{\alpha}\in {Symp}_{\alpha},$$
such that
$$\begin{array}{c}{\sigma}_{\alpha \beta}{\sigma}_{\beta}^{1}{\sigma}_{\alpha}^{1}\in Ham,\alpha ,\beta \in {\pi}_{0}(Symp).\end{array}$$ 
(2.6)

We first define
$\sigma $
on the commutator subgroup
$\left[{\pi}_{0}\right({Symp}_{H}),{\pi}_{0}({Symp}_{H}\left)\right]$
. When
$\alpha $
lies in this group then the component
${Symp}_{\alpha}$
contains elements that are products of commutators. We define
${\sigma}_{\alpha}$
to be such an element. Then
${\sigma}_{\alpha}$
is well defined modulo an element in
$Ham$
because
$[{Symp}_{H},{Symp}_{H}]\cap {Symp}_{0}=Ham$
by assumption.
Hence (
2.6 ) holds for these
$\alpha $
. Now we extend by hand, defining a lift on the abelian group
${\pi}_{0}\left({Symp}_{H}\right)/\left[{\pi}_{0}\right({Symp}_{H}),{\pi}_{0}({Symp}_{H}\left)\right].$
This is easy to do on the free part, and on the torsion part one uses the divisibility of
${H}^{1}(M;\mathbb{R})/\Gamma $
. Note that
$F$
is necessarily continuous since it is continuous on
${Symp}_{0}$
. □
3 Bundles with structural group
$Hams$
This section contains the proofs of the main results about the group
${Ham}^{s}$
and the obstruction class. In § 3.1 we give a simple proof that
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
has the extension property (Proposition 1.1 ). Because
$Hams$
is geometrically defined, a similar argument shows that
$Hams$
has the extension property when restricted to bundles
$P\to B$
where
${\pi}_{1}\left(B\right)$
acts trivially on
${H}_{1}(M;\mathbb{Q})$
(Proposition 3.3 (i)). It also shows that bundles with structural group
${Ham}^{s\mathbb{Z}}$
have an integral closed connection form (Lemma 3.2 ), and that conversely, any bundle with a closed integral connection form and such that
${\pi}_{1}\left(B\right)$
acts trivially on
${H}_{1}(M;\mathbb{Z})$
has a
${Ham}^{s\mathbb{Z}}$
structure (Proposition 3.3 (ii)).
We start
§ 3.2 by defining the obstruction cocycle
$\mathcal{O}M$
and then use it to prove that
$Hams$
has the extension property in all cases. For completeness, we shall give most of the details of the proof. (As pointed out by Tsemo [
16]
, this is a special case of a more general theory that can be nicely expressed in the language of gerbes.) We then prove Propositions 1.6 and 1.5 . In § 3.3 we discuss what it means for the obstruction class to vanish and prove Theorem 1.11 and Proposition 1.13 . The section ends with a proof of the stability result Proposition 1.15 .
3.1 Groups with the extension property
We begin by proving Proposition 1.1 which states that the group
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
has the extension property.
Proof of Proposition 1.1 . Suppose first that a smooth
$M$
bundle
$P\to B$
has a closed connection form
$\Omega .$
Because the holonomy of the corresponding connection is Hamiltonian round all contractible loops, it defines a continuous map from the space of based loops in
$B$
to the group
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
. This deloops to a lift
$B\to B{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
of the classifying map for
$P\to B$
. Therefore the classifying map of any bundle with a closed connection form does lift to
$B{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
.
Conversely, consider the universal
$M$
bundle
$${M}_{{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}}\to B{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}.$$
It suffices to show that the fiberwise symplectic class
$a=\left[\omega \right]$
extends to a class
$\stackrel{~}{a}\in {H}^{2}({M}_{{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}};\mathbb{R})$
. If not, there is a map of a finite CW complex
$X\to B{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
such that the fiberwise symplectic class in the pullback bundle
$M\to {P}_{X}\to X$
does not extend to
${P}_{X}$
. By embedding
$X$
in Euclidean space and replacing it by a small open neighborhood, we may assume that
$X$
is an smooth (open) manifold. Hence we may suppose that
$M\to {P}_{X}\to X$
is smooth. Since the structural group is
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
this bundle has a symplectic connection with holonomy in
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
. The holonomy round contractible loops lies in the identity component of
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
and hence is Hamiltonian. Therefore the Guillemin–Lerman–Sternberg construction provides a closed connection form
$\tau $
on
${P}_{X}$
that defines this connection: see [
10,Thm 6.21]
. Since
$\left[\tau \right]\in {H}^{2}\left({P}_{X}\right)$
extends
$\left[\omega \right]$
, this contradicts our initial assumption.
$\square $
Corollary 3.1.
Let
$\mathcal{\mathscr{H}}$
be any subgroup of
$Symp$
whose identity component is contained in
$Ham$
. Consider the universal
$M$
bundle
$$M\to {M}_{\mathcal{\mathscr{H}}}\to B\mathcal{\mathscr{H}}.$$
Then the fiberwise symplectic class
$a:=\left[\omega \right]$
extends to
$\stackrel{~}{a}\in {H}^{2}({M}_{\mathcal{\mathscr{H}}};\mathbb{R})$
.

Proof.
The hypothesis on
$\mathcal{\mathscr{H}}$
implies that the inclusion
$\mathcal{\mathscr{H}}\to Symp$
factors continuously through
${Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}$
. Therefore the class
$\stackrel{~}{a}\in {H}^{2}({M}_{{Symp}^{\mathit{H}\mathit{t}\mathit{o}\mathit{p}}};\mathbb{R})$
constructed above pulls back to
${H}^{2}\left({M}_{\mathcal{\mathscr{H}}}\right)$
. □
Lemma 3.2.
If
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
then the universal
$M$
bundle over
$B{Ham}^{s\mathbb{Z}}$
carries an integral extension of
$\left[\omega \right]$
. In other words, we may choose
$\stackrel{~}{a}$
to lie in the image of
${H}^{2}({M}_{\mathcal{\mathscr{H}}};\mathbb{Z})$
in
${H}^{2}({M}_{\mathcal{\mathscr{H}}};\mathbb{R})$
.

Proof.
The universal
$M$
bundle over
$B{Ham}^{s\mathbb{Z}}$
carries a connection with holonomy in
${Ham}^{s\mathbb{Z}}$
. Since this has Hamiltonian holonomy round closed loops, the GLS construction shows that it is given by a closed connection form
$\Omega $
. We claim that there is
$b\in {H}^{2}(B;\mathbb{R})$
such that
$[\Omega ]+{\pi}^{*}\left(b\right)$
is integral.
To see this, it suffices to consider the pullback bundle
$P\to \Sigma $
over any map
$\phi :\Sigma \to B{Ham}^{s\mathbb{Z}}$
, where
$\Sigma $
denotes a Riemann surface. There are three kinds of integral classes in
${H}_{2}(P;\mathbb{R})$
: cycles lying entirely in the fiber, cycles that project nontrivially to
${H}_{2}(\Sigma )$
and cycles
$C(\gamma ,\delta )$
formed as follows. Suppose that
$\gamma $
is a closed path in the base with holonomy
${m}_{\gamma}:M\to M$
that fixes the class
$\delta \in {H}_{1}(M;\mathbb{Z})$
. Choose a loop
${\ell}_{\delta}$
in
$M$
such that
$\langle {\ell}_{\delta}\rangle =s\left(\delta \right)$
, and define
$C(\gamma ,\delta )$
to be the union of the cylinder
${C}^{\prime}$
formed by the parallel translation of
${\ell}_{\delta}$
around
$\gamma $
with a chain
${C}^{\prime \prime}$
in
$M$
with boundary
${\ell}_{\delta}{m}_{\gamma}\left({\ell}_{\delta}\right)$
.
Since
$\Omega =0$
on
${C}^{\prime}$
and
${m}_{\gamma}\in {Ham}^{s\mathbb{Z}}$
, it follows from equation ( 2.4 ) that
$${\int}_{C(\gamma ,\delta )}\Omega ={\int}_{{C}^{\prime \prime}}\omega ={\widehat{F}}_{s}^{\mathbb{Z}}\left({m}_{\gamma}\right)\in \mathcal{P}\mathbb{Z}\omega \subset \mathbb{Z}.$$
Thus
$[\Omega ]$
takes integral values on cycles of the first and third kinds.
The class
$[\Omega ]+{\pi}^{*}\left(b\right)$
is represented in the
${E}_{2}$
term of the Leray–Serre spectral sequence for the cohomology of
$P\to \Sigma $
by a sum whose
${E}_{2}^{02}$
entry is
$\left[\omega \right]$
and whose
${E}_{2}^{11}$
entry is the unique element of
${H}^{1}(\Sigma ,\{{H}^{1}(M;\mathbb{R})\left\}\right)$
that equals
$[\Omega ]$
on the cycles
$C(\gamma ,\delta )$
. Both these entries are integral. By suitable choice of
$b\in {H}^{2}(B;\mathbb{R})$
we can also arrange that the
${E}_{2}^{20}$
entry is integral. Hence result. □
Proposition 3.3.
(i) Let
$P\to B$
be a symplectic bundle over a finite simplicial complex
$B$
such that
${\pi}_{1}\left(B\right)$
acts trivially on
${H}_{1}(M;\mathbb{R})$
. Then
$P$
has a closed connection form iff the classifying map for
$P\to B$
lifts to
$BHams$
.
(ii) Let
$P\to B$
be a symplectic bundle over a finite simplicial complex
$B$
such that
${\pi}_{1}\left(B\right)$
acts trivially on
${H}_{1}(M;\mathbb{Z})$
. Then
$P$
has a closed and integral connection form iff the classifying map for
$P\to B$
lifts to
$B{Ham}^{s\mathbb{Z}}$
.

Proof.
Corollary 3.1 shows that every
$Hams$
bundle has a closed connection form.
Conversely, suppose that
$P\to B$
has a closed connection form. Then the restriction map
${H}^{2}(P;\mathbb{R})\to {H}^{2}(M;\mathbb{R})$
contains
$\left[\omega \right]$
in its image. Because
$\mathbb{Q}$
is a field, the restriction map
${H}^{2}(P;\mathcal{P}\mathbb{Q}\omega )\to {H}^{2}(M;\mathcal{P}\mathbb{Q}\omega )$
also contains
$\left[\omega \right]$
in its image. Choose a class
$a\in {H}^{2}(P;\mathcal{P}\mathbb{Q}\omega )$
that extends
$\left[\omega \right]$
. Thurston's construction (cf. [
10,Thm 6.3]
) provides a closed extension
$\Omega $
in class
$a$
. We claim that the holonomy of
$\Omega $
round loops
$\gamma $
in the base
$B$
lies in
$Hams$
. Granted this, one can use the local trivializations given by
$\Omega $
to reduce the structural group to
$Hams$
.
To prove the claim, observe that because
${\pi}_{1}\left(B\right)$
acts trivially on
${H}_{1}(M;\mathbb{R})$
one can use the connection defined by
$\Omega $
to construct for each loop
$\gamma $
in
$B$
a
$2$
cycle
$C(\gamma ,{\delta}_{i})$
as in Lemma 3.2 , where
$\left[{\delta}_{i}\right]$
runs through a basis of
${H}_{1}(M;\mathbb{Q})$
. Then, the
$\Omega $
holonomy
${m}_{\Omega}\left(\gamma \right):M\to M$
round the loop
$\gamma $
in
$B$
satisfies the identity:
$${\widehat{F}}_{s}\left({m}_{\Omega}\right(\gamma \left)\right)\left({\delta}_{i}\right)={\int}_{C(\gamma ,{\delta}_{i})}\Omega \in \mathcal{P}\mathbb{Q}\omega .$$
Hence
${m}_{\Omega}\left(\gamma \right)\in Hams$
. This completes the proof of (i).
The proof of (ii) is similar and is left to the reader. □
We show in the next section that part (i) of this proposition extends to arbitrary bundles. However the integral case is more subtle. Example 4.2 shows that even if
$\left[\omega \right]$
is integral there might be manifolds for which there is no group that classifies symplectic bundles with integral closed connection form.
3.2 The obstruction class
Denote by
$\mathcal{A}$
the quasitopological abelian group
$${H}^{1}(M;\mathbb{R})/{H}^{1}(M;\mathcal{P}\omega ),$$
where
$\mathcal{P}\omega :=\mathcal{P}\mathbb{Q}\omega $
. As explained in Remark 2.6 , this is a
$K(\pi ,1)$
with
${\pi}_{1}$
isomorphic to the free (discrete) abelian group
${H}^{1}(M;\mathcal{P}\omega )$
. To simplify notation we shall often write
$${H}_{\mathbb{R}}:={H}^{1}(M;\mathbb{R}),{H}_{\mathbb{Q}}:={H}^{1}(M;\mathcal{P}\omega ),$$
so that
$\mathcal{A}:={H}_{\mathbb{R}}/{H}_{\mathbb{Q}}$
.
Lemma 3.4.
There is an obstruction class
$$\mathcal{O}M\in {H}^{2}(BSymp;{H}^{1}(M;\mathcal{P}\omega \left)\right),$$
such that the classifying map
$\phi :B\to BSymp$
of a symplectic bundle lifts to
$BHams$
iff
${\phi}^{*}(\mathcal{O}M)=0$
. Moreover
$\mathcal{O}M=0$
if there is a continuous crossed homomorphism
${\stackrel{~}{F}}_{s}:Symp\to {H}^{1}(M;\mathbb{R})$
that lifts
${\widehat{F}}_{s}$
.

Proof.
Consider the fibration sequence
$$Hams\to Symp\stackrel{{\widehat{F}}_{s}}{\to}\mathcal{A},$$
that identifies
$\mathcal{A}$
as the homogeneous space
$Symp/Hams$
. There is an associated homotopy fibration
$$\mathcal{A}\to BHams\to BSymp.$$
Because
$\mathcal{A}$
is a
$K(\pi ,1)$
, there is a single obstruction to the existence of a section of this fibration, namely a class
$\mathcal{O}M\in {H}^{2}(BSymp;{\pi}_{1}(\mathcal{A}\left)\right)$
. Since
${\pi}_{1}(\mathcal{A})\sim ={H}^{1}(M;\mathcal{P}\omega ))$
, this proves the first statement.
To prove the second, denote by
$\mathcal{\mathscr{H}}$
the kernel of
${\stackrel{~}{F}}_{s}$
. Then
$\mathcal{\mathscr{H}}$
is a subgroup of
$Hams$
and the composite map
$\mathcal{\mathscr{H}}\to Hams\to Symp$
is a homotopy equivalence. Hence every map
$\phi :B\to Symp$
lifts to
$B\mathcal{\mathscr{H}}$
and hence to
$BHams$
. □
Proposition 3.5.
Suppose that the symplectic fibration
$(M,\omega )\to P\stackrel{\pi}{\to}B$
that is classified by
$\phi :B\to BSymp$
has a closed connection form
$\Omega $
. Then the obstruction class
${\phi}^{*}(\mathcal{O}M)$
vanishes.

Proof.
Consider the change of coefficients exact sequence
$$...{H}^{1}\left(B;{H}_{\mathbb{R}}/{H}_{\mathbb{Q}}\right)\stackrel{\delta}{\to}{H}^{2}\left(B;{H}_{\mathbb{Q}}\right)\stackrel{\iota}{\to}{H}^{2}\left(B;{H}_{\mathbb{R}}\right)...$$
(where
${\pi}_{1}\left(B\right)$
may act nontrivially on the coefficients.) The map
$\iota $
is injective because
${H}_{\mathbb{Q}}$
is a
$\mathbb{Q}$
vector space. (Here it is essential to work over
$\mathbb{Q}$
, not
$\mathbb{Z}$
.) We shall use the connection
${\tau}_{\Omega}$
given by
$\Omega $
to construct a
$Hams$
structure over the restriction
${P}_{1}\to {B}_{1}$
of
$P$
to the
$1$
skeleton of
$B$
and a cocycle
$\alpha \in {C}^{1}\left(B;{H}_{\mathbb{R}}/{H}_{\mathbb{Q}}\right)$
such that the vanishing of
$\delta \left(\right[\alpha \left]\right)\in {H}^{2}\left(B;{H}_{\mathbb{Q}}\right)$
implies that the
$Hams$
structure extends over
${P}_{2}\to {B}_{2}$
. But
$\delta \left(\right[\alpha \left]\right)$
always vanishes since
$\iota $
is injective. Hence the classifying map
$\phi :{B}_{2}\to BSymp$
lifts to
$BHams$
. It follows that
${\phi}^{*}(\mathcal{O}M)=0$
.
The construction of the cocycle
$\alpha $
is essentially tautological. We give
$B$
the structure of a CW complex with one vertex
$v$
and with
$1$
cells that are embedded loops. Then it suffices to define the values
$\alpha \left(\gamma \right)\in \mathcal{A}:={H}_{\mathbb{R}}/{H}_{\mathbb{Q}}$
for the
$1$
cells
$\gamma $
in this structure, where we identify
$M$
with the fiber at the vertex
$v$
. The
$1$
cells are loops in
$B$
and we set
$$\alpha \left(\gamma \right)={\widehat{F}}_{s}\left({m}_{\Omega}\right(\gamma \left)\right),$$
where
${m}_{\Omega}\left(\gamma \right)\in Symp$
is the holonomy of the connection
${\tau}_{\Omega}$
round
$\gamma $
. To see that
$\alpha $
is a cocycle we need to check that
$\alpha (\partial e)=0$
where
$\partial e$
is the boundary of a
$2$
cell. Because
$\Omega $
is closed, the
${\tau}_{\Omega}$
holonomy round any contractible loop (such as
$\partial e$
) is Hamiltonian and so lies in the kernel of
${\widehat{F}}_{s}$
. This immediately translates into the fact that
$\alpha (\partial e)=0$
.
To be more precise, we may write the boundary of
$e$
as a word in generators
${\gamma}_{i}$
corresponding to the
$1$
cells in
$B$
:
$\partial e={\gamma}_{{i}_{p}}^{{\varepsilon}_{p}}*\cdots *{\gamma}_{{i}_{1}}^{{\varepsilon}_{1}}$
where
${\varepsilon}_{j}=\pm 1$
and we think of going first around
${\gamma}_{{i}_{1}}$
in some direction and then around
${\gamma}_{{i}_{2}}$
and so on. Denote by
${f}_{j}$
the
$\Omega $
holonomy round
${\gamma}_{{i}_{j}}^{{\varepsilon}_{j}}$
. Then the identity
${\widehat{F}}_{s}\left(gf\right)={\widehat{F}}_{s}\left(f\right)+{f}^{*}\left({\widehat{F}}_{s}\right(g\left)\right)$
implies that
$${\widehat{F}}_{s}({f}_{p}\circ \cdots \circ {f}_{1})={\widehat{F}}_{s}\left({f}_{1}\right)+({f}_{1}{)}^{*}\left({\widehat{F}}_{s}\left({f}_{2}\right)\right)+\cdots +({f}_{1}{)}^{*}\cdots ({f}_{p1}{)}^{*}\left({\widehat{F}}_{s}\left({f}_{p}\right)\right).$$
Because the total holonomy round
$\partial e$
is Hamiltonian, this must vanish. But this is precisely the cocycle condition in the twisted cohomology group
${H}^{1}(B;\mathcal{A})$
.
This defines the class
$\left[\alpha \right]\in {H}^{1}(B;\mathcal{A})$
. Putting a
$Hams$
structure over the
$1$
skeleton
${B}_{1}$
is equivalent to putting a flat connection
$\psi $
over each
$1$
cell
$\gamma $
whose holonomy
${m}_{\psi}\left(\gamma \right)$
lies in
$Hams$
. For each
$\gamma $
we choose
${\stackrel{~}{g}}_{\gamma}\in {\stackrel{~}{Symp}}_{0}$
so that
${m}_{\Omega}\left(\gamma \right){g}_{\gamma}\in Hams$
. Then we define
$\psi $
on
$\gamma $
so that it differs from
${\tau}_{\Omega}$
by the path
${\stackrel{~}{g}}_{\gamma}$
. In particular,
$$\begin{array}{c}{m}_{\psi}\left(\gamma \right):={m}_{\Omega}\left(\gamma \right){g}_{\gamma}.\end{array}$$ 
(3.1)

(This step uses the fact that
$Hams$
intersects every component of
$Symp$
.) The next step is to describe the obstruction to extending this
$Hams$
connection over a given
$2$
cell
$e$
. Denote by
$\pi :{P}_{1}\to {B}_{1}$
the pullback of
$P$
over
${B}_{1}$
. This fibration is extended over
$e$
by attaching the product
${D}^{2}\times M$
to
${\pi}^{1}\left({B}_{1}\right)$
along the boundary loop
$\lambda \equiv \partial e$
in
${B}_{1}$
by a family
$\Phi $
of symplectomorphisms
$${\phi}_{t}:\left\{t\right\}\times M\to {\pi}^{1}\left({\lambda}_{t}\right),t\in {S}^{1}.$$
The connection
${\tau}_{\Omega}$
provides a map
$${\iota}_{\Omega}:[0,1]\times M\to {\pi}^{1}\left(\lambda \right),(t,x)\mapsto {\iota}_{\Omega ,t}\left(x\right)$$
such that
${\iota}_{\Omega ,1}={m}_{\Omega}(\partial e)\circ {\iota}_{\Omega ,0}.$
To say that
$\Omega $
extends over
${\pi}^{1}\left(e\right)$
is equivalent to saying that the path
$t\mapsto ({\iota}_{\Omega ,t}{)}^{1}{\phi}_{t}$
is isotopic with fixed endpoints to a Hamiltonian path from the identity to
${m}_{\Omega}(\partial e{)}^{1}$
. Similarly, if
${\iota}_{\psi}$
denotes the trivialization along
$\lambda =\partial e$
provided by
$\psi $
, the
$Hams$
structure extends over
${\pi}^{1}\left(e\right)$
iff
$t\mapsto ({\iota}_{\psi ,t}{)}^{1}{\phi}_{t}$
is isotopic rel.
endpoints to a Hamiltonian path from the identity to
${m}_{\psi}(\partial e{)}^{1}$
. Since
$\Omega $
does extend by hypothesis, this happens iff the path
$${p}_{e}:t\mapsto ({\iota}_{\Omega ,t}{)}^{1}{\iota}_{\psi ,t}$$
is isotopic rel. endpoints to a Hamiltonian path from the identity to
${m}_{\Omega}(\partial e{)}^{1}{m}_{\psi}(\partial e)$
.
As above, denote the
$\Omega $
holonomy round
$\partial e$
by
${m}_{\Omega}(\partial e)={f}_{p}\circ \cdots \circ {f}_{1}$
. In obvious notation the
$\psi $
holonomy is
$${m}_{\psi}(\partial e):={f}_{p}{g}_{p}\circ \cdots \circ {f}_{1}{g}_{1},{g}_{i}\in {Symp}_{0}.$$
Since each
${f}_{i}{g}_{i}\in Hams$
,
${m}_{\psi}(\partial e)\in Hams$
. Now observe that
$$m(\partial e):={\left({m}_{\Omega}(\partial e)\right)}^{1}{m}_{\psi}(\partial e)={\left({f}_{p}\circ \cdots \circ {f}_{1}\right)}^{1}{m}_{\psi}(\partial e)$$
is a product of elements of the form
$({f}_{k1}...{f}_{1}{)}^{1}{g}_{k}({f}_{k1}...{f}_{1})$
. Lift
$m\left(\delta \right)$
to an element
$\stackrel{~}{m}(\partial e)\in {\stackrel{~}{Symp}}_{0}$
by replacing the elements
${g}_{i}$
by
${\stackrel{~}{g}}_{i}\in {\stackrel{~}{Symp}}_{0}$
. It is not hard to check that the obstruction path
${p}_{e}$
mentioned in the previous paragraph represents the element
$\stackrel{~}{m}(\partial e)\in {\stackrel{~}{Symp}}_{0}$
. Since
$\stackrel{~}{m}(\partial e)\in {\stackrel{~}{Symp}}_{0}$
,
$Flux\left(\stackrel{~}{m}\right(\partial e\left)\right)$
is defined as an element in
${H}^{1}(M;\mathbb{R})$
. It lies in
${H}_{\mathbb{Q}}:={H}^{1}(M;\mathcal{P}\omega )$
since its endpoint lies in
$Hams$
.
Moreover, the remarks in the previous paragraph imply that
$Flux\left(\stackrel{~}{m}\right(\partial e\left)\right)=0$
iff the
$Hams$
structure extends over
${\pi}^{1}\left(e\right)$
.
We now relate this information to the cocycle
$\alpha $
. To find a representative for
$\delta \left(\right[\alpha \left]\right)\in {H}^{2}(B;{H}_{\mathbb{Q}})$
, we first choose a lift
$\stackrel{~}{\alpha}\in {C}^{1}(B;{H}_{\mathbb{R}})$
of the cocycle
$\alpha $
. Then we set
$\delta \alpha \left(e\right):=\stackrel{~}{\alpha}(\partial e)$
. This takes values in
${H}_{\mathbb{Q}}$
because it vanishes in the quotient
${H}_{\mathbb{R}}/{H}_{\mathbb{Q}}$
.
Because, for each
$1$
cell
$\gamma $
$$0={\widehat{F}}_{s}\left({m}_{\psi}\right(\gamma \left)\right)={\widehat{F}}_{s}\left({m}_{\Omega}\right(\gamma \left)\right)+{\widehat{F}}_{s}\left({g}_{\gamma}\right),$$
we may choose
$$\stackrel{~}{\alpha}\left(\gamma \right):=Flux\left({\stackrel{~}{g}}_{\gamma}\right)\in {H}^{1}(M;\mathbb{R}),$$
where
${\stackrel{~}{g}}_{\gamma}$
is as above. It is then immediate that
$$\stackrel{~}{\alpha}(\partial e)=Flux\left(\stackrel{~}{m}\right(\partial e\left)\right)\in {H}_{\mathbb{Q}}.$$
Hence this value determines whether or not the chosen
$Hams$
connection extends over
${\pi}^{1}\left(e\right)$
. Clearly, changing the lift
$\stackrel{~}{\alpha}$
corresponds to changing the choice of
${\stackrel{~}{g}}_{\gamma}$
satisfying ( 3.1 ). In particular, since
$\delta \left[\alpha \right]=0\in {H}^{2}(B;{H}_{\mathbb{Q}})$
we may choose
$\psi $
over
${B}_{1}$
in such a way that it extends to a
$Hams$
structure over
${B}_{2}$
. This completes the proof. □
Proof of Theorem 1.10 . Corollary 3.1 shows that every
$Hams$
bundle has a closed connection form. The converse follows from Lemma 3.4 and Proposition 3.5 .
$\square $
Remark 3.6.
In the above arguments we did not use the specific form of the crossed homomorphism
${\widehat{F}}_{s}$
but just that its restriction to
${Symp}_{0}$
may be lifted to a (possibly discontinuous) homomorphism
$\stackrel{~}{{Symp}_{0}}\to {H}^{1}(M;\mathbb{R})$
whose kernel is the universal cover of
$Ham$
. In particular, the analog of Lemma 3.4 holds if we replace
${\widehat{F}}_{s}$
by any crossed homomorphism
$\stackrel{~}{F}:Symp\to {H}^{1}(M;\mathbb{R})/\Gamma $
that extends Flux. Thus if
$\mathcal{\mathscr{H}}:=ker\stackrel{~}{F}$
, the obstruction to the existence of an
$\mathcal{\mathscr{H}}$
structure on the bundle
$P\to B$
is an element of
${H}^{2}(B;\Gamma )$
. Therefore it vanishes if
$\Gamma =0$
. This is to be expected since the inclusion
$\mathcal{\mathscr{H}}:=ker\left(\stackrel{~}{F}\right)\to Symp$
is then a homotopy equivalence. As we shall see in the proof of Proposition 1.6 below, most of Proposition 3.5 also goes through. All of it goes through if we consider a crossed homomorphism into a divisible group, such as
${H}^{1}(M;\mathbb{R})/\Gamma \otimes \mathbb{Q}$
.
Proof of Proposition 1.6 . Let
$\mathcal{\mathscr{H}}$
be a subgroup of
$Symp$
with identity component
$Ham$
and consider the image
$Im\left({\pi}_{0}\right(\mathcal{\mathscr{H}}\left)\right)$
of
${\pi}_{0}(\mathcal{\mathscr{H}})$
in
${\pi}_{0}(Symp)$
. If
$\mathcal{\mathscr{H}}$
has the modified extension property then we must show that the intersection with
$Im\left({\pi}_{0}\right(\mathcal{\mathscr{H}}\left)\right)$
of every finitely generated subgroup
$G$
of
${\pi}_{0}(Symp)$
has finite index in
$G$
. But otherwise there would be a map of a finite wedge
$V$
of circles into
$B\left({\pi}_{0}(Symp)\right)$
such that no finite cover
$\stackrel{~}{V}$
of
$V$
lifts into the image of
$B\mathcal{\mathscr{H}}$
in
$B\left({\pi}_{0}(Symp)\right)$
. Since any bundle over a
$1$
complex has a closed extension form, this contradicts our assumption on
$\mathcal{\mathscr{H}}$
.
Conversely, assume that the cokernel of
$Im\left({\pi}_{0}\right(\mathcal{\mathscr{H}}\left)\right)$
in
${\pi}_{0}(Symp)$
has the stated finiteness properties and let
${Symp}_{\mathcal{\mathscr{H}}}$
be the subgroup of
$Symp$
consisting of elements isotopic to
$\mathcal{\mathscr{H}}$
. If
$P\to B$
is classified by a map into
$B\mathcal{\mathscr{H}}$
then it has a closed connection form by Corollary 3.1 . Therefore we just need to see that if
$\phi :B\to BSymp$
classifies a bundle with a closed extension form its pullback over some finite cover
$\stackrel{~}{B}\to B$
lifts to
$B\mathcal{\mathscr{H}}$
. Observe first that the composite map
${\pi}_{1}\left(B\right)\to {\pi}_{0}(Symp)/{\pi}_{0}\left({Symp}_{\mathcal{\mathscr{H}}}\right)$
has finite image by hypothesis. (Recall that we always assume
${\pi}_{1}\left(B\right)$
is finitely generated.) Therefore we may replace
$B$
by a finite cover such that the pullback bundle
$\stackrel{~}{P}\to \stackrel{~}{B}$
is classified by a map into
$B{Symp}_{\mathcal{\mathscr{H}}}$
. Note that this bundle still has a closed connection form. Set
$\Lambda \subset {H}^{1}(M;\mathbb{R})$
equal to the (discrete) group
$Flux(\mathcal{\mathscr{H}}\cap {Symp}_{0})$
, and then define a crossed homomorphism
$$F:{Symp}_{\mathcal{\mathscr{H}}}\to {H}^{1}(M;\mathbb{R})/\Lambda $$
as in the proof of Proposition 1.8 given in § 1.2 . By Remark 3.6 there is an obstruction class
$\mathcal{O}M\Lambda \in {H}^{2}(BSymp;\Lambda )$
whose pullback
${\phi}^{*}(\mathcal{O}M\Lambda )\in {H}^{2}(B;\Lambda )$
vanishes iff
$\phi $
lifts to
$B\mathcal{\mathscr{H}}$
. As in the proof of Proposition 3.5 , consider the change of coefficients exact sequence
$$\cdot \cdot \cdot \to {H}^{1}\left(B;{H}_{\mathbb{R}}/\Lambda \right)\stackrel{\delta}{\to}{H}^{2}\left(B;\Lambda \right)\stackrel{\iota}{\to}{H}^{2}\left(B;{H}_{\mathbb{R}}\right)\to ...$$
As before, we may construct an element
$\alpha \in {H}^{1}\left(B;{H}_{\mathbb{R}}/\Lambda \right)$
such that
$\mathcal{O}M\Lambda =\delta \left(\right[\alpha \left]\right)$
.
But now
$\iota $
need not be injective: its kernel consists of torsion elements that will vanish when pulled back over a suitable finite covering map. Therefore there is an
$\mathcal{\mathscr{H}}$
structure on the pullback of
$\stackrel{~}{P}\to \stackrel{~}{B}$
over yet another finite cover.
$\square $
Proof of Proposition 1.5 . The first claim is that
$Ham$
has the modified restricted extension property. This is a corrected statement of the conclusions that one can draw from the proof of Theorem 1.1 in [
7]
. The claim also follows by arguing as in the proof of Proposition 1.6 using
$Flux$
instead of
${\widehat{F}}_{s}$
; the argument can be greatly simplified because the group
${Symp}_{0}$
acts trivially on the coefficients. Here one should also note that if the cover
$\stackrel{~}{B}\to B$
is chosen so that
${H}_{1}(\stackrel{~}{B};\mathbb{Z})$
has no torsion, then the boundary map
$\delta :{H}^{1}(B;{H}_{\mathbb{R}}/{H}_{\mathbb{Q}})\to {H}^{2}(B;{H}_{\mathbb{Q}})$
vanishes.
The second claim is that when
$\Gamma \ne 0$
the group
$Ham$
does not have the extension property. To see this choose a nonzero element
$\beta \in {H}^{1}(M;\mathbb{R})\mathbb{\backslash}\Gamma $
such that
$2\beta \in \Gamma $
and then choose
$g\in {Symp}_{0}$
with
$Flux\left(g\right)=\beta $
. Consider the bundle
$P\to \mathbb{R}{P}^{2}$
that is formed from the mapping torus bundle
$${M}_{g}:=M\times [0,1]/(x,1)\sim (gx,0)\u27f6{S}^{1}$$
by attaching
$M\times {D}^{2}$
by the map
$(x,{e}^{2\pi it})\mapsto \left({g}_{t}\right(x),2t)$
where
${g}_{t}$
is a path in
$Ham$
from the identity to
${g}_{1}:={g}^{2}\in Ham$
. The flat connection on
${M}_{g}$
pulls back to a Hamiltonian connection on the boundary
$M\times \partial D$
and so extends over the rest of
$P$
to a connection with the property that its holonomy around contractible loops in the base is Hamiltonian. Hence the GLS construction provides a closed connection form
$\Omega $
.
(Alternatively one can directly construct such a form: see the proof of Proposition
3.7 .) We claim that this bundle has no Hamiltonian structure. To see this consider the classifying map
$\phi :\mathbb{R}{P}^{2}\to B{Symp}_{0}$
. Just as in the discussion before Lemma 3.4 the homomorphism
$Flux:{Symp}_{0}\to {H}_{\mathbb{R}}/\Gamma $
defines an obstruction class
$\mathcal{O}M\Gamma \in {H}^{2}(B{Symp}_{0};\Gamma )$
such that
${\phi}^{*}(\mathcal{O}M\Gamma )$
vanishes iff the bundle
$P\to \mathbb{R}{P}^{2}$
has a Hamiltonian structure. Since
$B({H}_{\mathbb{R}}/\Gamma )$
is a
$K(\Gamma ,2)$
, this class is the pullback to
$B{Symp}_{0}$
of the canonical generator of
${H}^{2}\left(K\right(\Gamma ,2);\Gamma ).$
We claim that the composite map
$$\mathbb{R}{P}^{2}\to B{Symp}_{0}\to B({H}_{\mathbb{R}}/\Gamma )=K(\Gamma ,2)$$
is not null homotopic. Since
$\mathbb{R}{P}^{2}$
is the
$2$
skeleton of
$\mathbb{R}{P}^{\infty}=K(\mathbb{Z}/2\mathbb{Z};1)$
and
$K(\Gamma ,2)$
is homotopy equivalent to a product of copies of
$B{S}^{1}$
, this assertion is equivalent to saying that under the map
$B(\mathbb{Z}/2\mathbb{Z})\to B{S}^{1}$
induced by the obvious inclusion
$\{\pm 1\}\to {S}^{1}$
the generator of
${H}^{2}(B{S}^{1};\mathbb{Z}/2\mathbb{Z})$
pulls back to a nonzero element of
${H}^{2}(\mathbb{R}{P}^{2};\mathbb{Z}/2\mathbb{Z})$
.
This is well known. For a direct proof identify the
$2$
skeleton of
$B{S}^{1}=\mathbb{C}{P}^{\infty}$
with the quotient
${S}^{3}/{S}^{1}$
and observe that the
$\mathbb{Z}/2\mathbb{Z}$
equivariant map
$${S}^{2}\to {S}^{3},(r,s,t)\mapsto \left(r+is,\frac{1}{\sqrt{2}}(t+it)\right)\in {S}^{3}\subset {\mathbb{C}}^{2},$$
descends to a map
$\mathbb{R}{P}^{2}\to {S}^{2}$
of (mod 2) degree
$1$
.
$\square $
3.3 Vanishing of the obstruction class
The next proposition explains what it means for the obstruction class
$\mathcal{O}M$
to vanish.
Proposition 3.7.
Suppose that
$(M,\omega )$
is a symplectic manifold. Then the following conditions are equivalent.
(i) Every symplectic
$M$
bundle over a closed Riemann surface
$\Sigma $
has a closed connection form.
(ii) Every symplectic
$M$
bundle has a closed connection form.
(iii) The obstruction class
$\mathcal{O}M\in {H}^{2}(BSymp;{H}^{1}(M;\mathcal{P}\omega \left)\right)$
vanishes.
(iv) The flux subgroup
$\Gamma $
vanishes. Further, for every closed Riemann surface
$\Sigma $
every representation of
${\pi}_{1}(\Sigma )$
in
${\pi}_{0}(Symp)$
lifts to a representation into a subgroup of
$Symp/Ham$
.

Proof.
If
$\mathcal{O}M\in {H}^{2}(BSymp;{H}_{\mathbb{Q}})$
does not vanish, then it evaluates on some integral
$2$
cycle to be nonzero. Since a multiple of this
$2$
cycle can be represented by a Riemann surface, Proposition 3.5 implies that (i) will not hold. Therefore (i) implies (iii).
Conversely if (iii) holds then the classifying map of any a symplectic bundle
$P\to \Sigma $
lifts to
$BHams$
by Lemma 3.4 . Therefore (ii) holds by Corollary 3.1 . Since (ii) obviously implies (i) the first three conditions are equivalent.
Suppose now that (iv) holds and consider any symplectic bundle
$P\to {\Sigma}_{g}$
. Then one can decompose the bundle
$P$
into the union of a trivial bundle
$M\times D\to D$
over a
$2$
disc with a symplectically flat bundle
${P}^{\prime}\to {\Sigma}_{g}\mathbb{\backslash}D$
whose holonomy round the generators of
${\pi}_{1}\left({\Sigma}_{g}\right)$
is given by elements
${u}_{i}\in Symp$
such that^{8 }
$[{u}_{1},{u}_{2}]\cdots [{u}_{2g1},{u}_{2g}]\in {Symp}_{0}$
. The
${u}_{i}$
determine a representation
$\rho $
of
${\pi}_{1}(\Sigma )\to {\pi}_{0}(Symp)$
. By hypothesis, we may homotop the
${u}_{i}$
to elements
${f}_{i}\in Symp$
such that
$f:=[{f}_{1},{f}_{2}]\cdots [{f}_{2g1},{f}_{2g}]\in Ham$
. Correspondingly there is a new flat connection on
${P}^{\prime}$
with boundary holonomy
$f$
. In other words we may identify the boundary
$\partial {P}^{\prime}$
with
$M\times [0,1]/(x,1)\equiv \left(f\right(x),0)$
. Note also that the connection form
${\stackrel{~}{\omega}}^{\prime}$
on
${P}^{\prime}$
determined by this second connection vanishes in the horizontal directions and is closed because the connection is flat.
The bundle
$P$
is reconstructed from this data by gluing the boundary
$$M\times [0,1]/(x,1)\equiv (x,0)$$
of
$M\times \partial D$
to
$\partial {P}^{\prime}=M\times [0,1]/\sim $
by a map
$$\phi :M\times {S}^{1}\to \partial {P}^{\prime},(x,t)\mapsto \left({g}_{t}\right(x),t)$$
where
${g}_{1}={f}^{1}$
. Since
$f\in Ham$
and the flux group vanishes, the path
$\{{g}_{t}{\}}_{t\in [0,1]}$
has zero flux and we may homotop it (fixing its endpoints) into
$Ham$
. It is now easy to check that
${\phi}^{*}\left({\stackrel{~}{\omega}}^{\prime}\right)p{r}^{*}\omega $
is exact on
$M\times {S}^{1}$
(where
$pr$
is the obvious projection onto
$M$
). Hence
${\stackrel{~}{\omega}}^{\prime}$
extends to a closed connection form on
$P\to \Sigma $
. Thus (iv) implies (i). To see that (i) implies (iv) note first that the flux subgroup
$\Gamma $
vanishes iff every symplectic bundle over
${S}^{2}$
has a closed connection form. (This is well known and may be proved by explicit calculation as in the last step in the previous paragraph.) Now consider the second condition in (iv). A representation
$\rho :{\pi}_{1}\left({\Sigma}_{g}\right)\to {\pi}_{0}(Symp)$
is determined by a collection of elements
${u}_{1},...,{u}_{2g}\in Symp$
such that
$[{u}_{1},{u}_{2}]\cdots [{u}_{2g1},{u}_{2g}]\in {Symp}_{0}$
. Given such a collection, one can, as described above, build a symplectic fibration
$P\to {\Sigma}_{g}$
with fiber
$M$
and holonomy
$\left[{u}_{i}\right]\in {\pi}_{0}(Symp)$
round the
$i$
th element
${\gamma}_{i}$
in a standard basis for
${\pi}_{1}\left({\Sigma}_{g}\right)$
. If (i) holds this must have a closed connection form.
The holonomy of this connection round two homotopic loops differs by an element in
$Ham$
and hence determines a lift of
$\rho $
to
$Symp/Ham$
. Thus (i) implies (iv). □
The second part of condition (iv) above is rather hard to understand. The next results spell out some of its implications.
Lemma 3.8.
The following conditions are equivalent.
(i) For every closed Riemann surface
$\Sigma $
every representation of
${\pi}_{1}(\Sigma )$
in
${\pi}_{0}(Symp)$
lifts to a representation into
$Symp/Ham$
.
(ii) for every symplectic
$M$
bundle
$P\to \Sigma $
there is a bundle
$Q\to {S}^{2}$
such that the fiberwise connect sum
$P\#Q\to \Sigma \#{S}^{2}=\Sigma $
has a closed connection form.
(iii) The class
$\mathcal{O}M$
takes values in
$\Gamma $
, i.e. it lies in the image of
${H}^{2}(BSymp;\Gamma )$
in
${H}^{2}(BSymp;{H}_{\mathbb{Q}})$
.

Proof.
We argue as in the proof of Proposition 3.7 (iv). Consider a bundle
$P\to \Sigma $
classified by
$\phi :\Sigma \to BSymp$
, decompose it as a union of
${P}^{\prime}\to \Sigma \mathbb{\backslash}D$
with
$M\times D\to D$
and construct a flat symplectic connection on
${P}^{\prime}$
as before. Denote its connection form by
${\stackrel{~}{\omega}}^{\prime}$
and boundary holonomy by
$f\in {Symp}_{0}$
. Choose a path
${f}_{t}$
from
$f:={f}_{1}$
to the identity. Then the obstruction
$${\phi}^{*}\mathcal{O}M\left(\right[\Sigma \left]\right)\in {H}^{1}(M;\mathcal{P}\omega )$$
to extending
${\stackrel{~}{\omega}}^{\prime}$
over the rest of
$P$
is the sum of
$Flux\left\{{f}_{t}\right\}\in {H}^{1}(M;\mathbb{R})$
with an element of
$\Gamma $
given by the flux of the attaching map of
$M\times D$
. If (i) holds, then as explained earlier we may assume that
$f\in Ham$
. Hence in this case
${\phi}^{*}\mathcal{O}M\left(\right[\Sigma \left]\right)\in \Gamma $
for all
$2$
cycles
$\phi :\Sigma \to BSymp$
, i.e. (iii) holds. Conversely, if (iii) holds, we may replace
$P$
by its fiber sum with a bundle
${P}^{\prime \prime}\to {S}^{2}$
so that the obstruction for
$P\#{P}^{\prime \prime}\to \Sigma $
vanishes.
Then there is a closed connection form on
$P\#{P}^{\prime \prime}$
. We saw earlier that this implies that (i) holds. The equivalence of (ii) and (iii) is obvious. □
Proof of Theorem 1.11 . Part (i) is an immediate consequence of Lemma 3.4 and Theorem 1.10 . If there is an extension
$\stackrel{~}{F}:Symp\to {H}^{1}(M;\mathbb{R})/\Gamma $
of
$Flux$
, then Remark 3.6 implies that
$\mathcal{O}M$
takes values in
$\Gamma $
. Hence it remains to prove the converse, i.e. that
$Flux$
extends if
$\mathcal{O}M$
takes values in
$\Gamma $
.
As in the proof of Proposition
1.8 given in §1, it suffices to find a section
$$\sigma :{\pi}_{0}(Symp)\to Symp,\alpha \mapsto {\sigma}_{\alpha}\in {Symp}_{\alpha},$$
such that
$$\begin{array}{c}({\sigma}_{\alpha \beta}{)}^{1}{\sigma}_{\alpha}{\sigma}_{\beta}\in Ham,\alpha ,\beta \in {\pi}_{0}(Symp).\end{array}$$ 
(3.2)

To do this, consider the fibration sequence
$\mathcal{A}\to BHams\stackrel{\pi}{\to}BSymp$
of Lemma 3.4 .
By assumption the obstruction to the existence of a section
$s:BSymp\to BHams$
is an element of
${H}^{2}(BSymp;\Gamma )$
where
$\Gamma $
is identified with its image in
${H}_{\mathbb{Q}}={\pi}_{1}(\mathcal{A})$
.
This means that for any compatible CW structures put on
$BHams$
and
$BSymp$
one can choose a map
$s:(BSymp{)}_{1}\to (BHams{)}_{1}$
(where
${B}_{1}$
denotes the
$1$
skeleton of
$B$
) so that
$\pi \circ s\sim id$
and so that the corresponding obstruction cocycle takes values in
$\Gamma $
. Choose a CW structure on
$BSymp$
with one vertex, and one
$1$
cell
$I\times {g}_{\alpha}$
for each component
$\alpha \in {\pi}_{0}(Symp)$
. (This is possible because
${\pi}_{0}(Symp)$
is countable.) Then for each pair
$\alpha ,\beta $
in
${\pi}_{0}(Symp)$
there is a
$2$
cell
${c}_{\alpha ,\beta}$
with boundary
$(I\times {g}_{\alpha \beta}{)}^{1}(I\times {g}_{\alpha}\left)\right(I\times {g}_{\beta})$
.
(There are other
$2$
cells in
$BSymp$
coming from the
$1$
skeleton of a CW decomposition for
$Symp$
, but these are irrelevant for the current argument.) We define a CW structure on
$BHams$
in a similar way. Then the map
$s$
takes each
$1$
cell
$I\times {g}_{\alpha}$
in
$BSymp$
to a loop in
$(BHams{)}_{1}$
. This loop is given by a word
${w}_{\alpha}$
in the elements of
$Hams$
that represents an element
$h\left({w}_{\alpha}\right)$
in
$Hams\cap {Symp}_{\alpha}$
.
The obstruction to extending
$s$
over the
$2$
cell
${c}_{\alpha ,\beta}$
is the homotopy class in
$BHams$
of the loop corresponding to the word
${w}_{\alpha \beta}{)}^{1}{w}_{\alpha}{w}_{\beta}$
. This can be identified with the homotopy class
$$\left[h\right({w}_{\alpha \beta}{)}^{1}h\left({w}_{\alpha}\right)h\left({w}_{\beta}\right)]\in {\pi}_{0}(Hams\cap {Symp}_{0})\sim ={H}^{1}(M;\mathbb{Q})/\Gamma $$
of the element
$h\left({w}_{\alpha \beta}{)}^{1}h\right({w}_{\alpha}\left)h\right({w}_{\beta})$
. To say the obstruction
$\mathcal{O}M\left({c}_{\alpha ,\beta}\right)$
takes values in
$\Gamma $
means that this class lies in
${\pi}_{0}(Hams)=Ham$
. Hence it is always possible to extend
$s$
over these
$2$
cells (though it may not extend over the other
$2$
cells in
$BSymp$
). Further if we define the section
$\sigma :{\pi}_{0}(Symp)\to Symp$
by
$\sigma \left(\alpha \right):=h\left({w}_{\alpha}\right)$
then the identity ( 3.2 ) holds. This completes the proof.
$\square $
Proof of Proposition 1.13 . This follows by combining Lemma 3.8 with part (ii) of Theorem 1.11 .
$\square $
3.4 Stability
We finally discuss the question of stability.
Proof of Proposition 1.15 .
Let
$\mathcal{N}\left(\omega \right)$
be a path connected neighborhood of
$\omega $
in the space of forms annihilating
${V}_{2}\left(P\right)$
such that
$P\to B$
has an
${\omega}^{\prime}$
symplectic structure for all
${\omega}^{\prime}\in \mathcal{N}\left(\omega \right)$
. Our aim is to shrink
$\mathcal{N}\left(\omega \right)$
so that each such
${\omega}^{\prime}$
has a closed extension to
$P$
. We claim that for each map
$\psi :\Sigma \to B$
of a Riemann surface into
$B$
there is a homologous map
${\psi}^{\prime}:{\Sigma}^{\prime}\to B$
such that the pullback bundle over
${\Sigma}^{\prime}$
admits a closed extension of
${\omega}^{\prime}$
, provided that
${\omega}^{\prime}$
is sufficiently close to
$\omega $
. Granted this, we may choose
$\mathcal{N}\left(\omega \right)$
so that this holds for a finite set of
${\psi}_{i}$
that represent a set of generators for
${H}_{2}(M;\mathbb{R})$
and all
${\omega}^{\prime}\in \mathcal{N}\left(\omega \right)$
.
It follows that the obstruction class
${\mathcal{O}}_{{\omega}^{\prime}}^{M}$
must vanish when pulled back to
$B$
, i.e. that
$(M,{\omega}^{\prime})\to P\to B$
has a closed connection form when
${\omega}^{\prime}\in \mathcal{N}\left(\omega \right)$
.
To prove the claim, consider a map
$\psi :\Sigma :={\Sigma}_{p}\to B$
. By Proposition 3.7 we may assume that the pullback bundle
${\psi}^{*}P\to \Sigma $
has a flat
$\omega $
symplectic connection over
$\Sigma \mathbb{\backslash}{D}^{2}$
whose holonomy
$y$
around the boundary of the disc
${D}^{2}$
may be expressed as:
$$y:=[{u}_{1},{u}_{2}]\cdots [{u}_{2p1},{u}_{2p}]\in Ham(M,\omega ),{u}_{i}\in Symp(M,\omega ).$$
Since
$Ham(M,\omega )$
is a perfect group, we may, by increasing the genus of
$\Sigma $
and choosing the flat connection on the extra handles to have Hamiltonian holonomy, assume that
$y=id.$
By hypothesis on the deformation
${\omega}^{\prime}$
, we can choose:
$\bullet $
a path
${\omega}_{t}$
from
${\omega}_{0}:=\omega $
to
${\omega}_{1}:={\omega}^{\prime}$
in
$\mathcal{N}\left(\right[\omega \left]\right)$
, and
$\bullet $
${C}^{1}$
small paths^{9 }
${g}_{it}\in {Diff}_{0}\left(M\right)$
such that
${u}_{i}{g}_{it}\in Symp(M,{\omega}_{t})$
for all
$i$
and
$t\in [0,1]$
.
Since
$y=id$
the smooth path
$${y}_{t}=[{u}_{1}{g}_{1t},{u}_{2}{g}_{2t}]\cdots [{u}_{2p1}{g}_{(2p1)t},{u}_{2p}{g}_{\left(2p\right)t}]$$
is
${C}^{1}$
small and lies in
${Symp}_{0}(M,{\omega}_{t})$
for all
$t$
. If we could arrange that
${y}_{t}\in Ham(M,{\omega}_{t})$
for each
$t$
then the connection could be extended to a Hamiltonian connection over the disc for all
$t$
and the proof would be complete.
We show below how to modify
$\psi :\Sigma \to B$
to a map
${\psi}^{\prime}:{\Sigma}^{\prime}\to B$
so that
${y}_{t}\in {Ham}^{s}(M,{\omega}_{t})\cap {Symp}_{0}(M,{\omega}_{t})$
for all
$t$
. But this finishes the proof, because the
${y}_{t}$
must then be in
$Ham(M,{\omega}_{t})$
for all
$t$
by the following continuity argument. Observe that for each
$[\ell ]\in {H}_{1}\left(M\right)$
and
$t\in [0,1]$
, the number
$$\Phi \left(t\right)[\ell ]:={\int}_{[0,t]\times {S}^{1}}{\phi}_{t,\ell}^{*}{\omega}_{t}\in \mathbb{R},\text{where}{\phi}_{t,\ell}(r,s):={y}_{r}(\ell (s\left)\right),$$
projects to
${\widehat{F}}_{{\omega}_{t}}^{s}\left({y}_{t}\right)\left(\right[\ell \left]\right)\in \mathbb{R}/{\mathcal{P}}_{{\omega}_{t}}$
. Since, by assumption,
${y}_{t}\in {Ham}^{s}(M,{\omega}_{t})$
, we find that
$\Phi \left(t\right)[\ell ]\in {\mathcal{P}}_{{\omega}_{t}}$
for all
$t$
. But
$\Phi \left(t\right)$
varies continuously with
$t$
and
$\Phi \left(0\right)=0$
. Hence the fact that
${V}_{2}\left({\omega}_{t}\right)\supseteq {V}_{2}\left(\omega \right)$
implies that
$\Phi \left(t\right)=0$
for all
$t$
. It remains to check that
$\Phi \left(t\right)$
projects to
${Flux}_{{\omega}_{t}}\left({y}_{t}\right)\in {H}^{1}(M;\mathbb{R})/{\Gamma}_{{\omega}_{t}}$
for all
$t\in [0,1]$
. But because the
${y}_{t}$
are
${C}^{1}$
small, for each fixed
$t\in [0,1]$
the path
$\{{y}_{r}{\}}_{r\in [0,t]}$
may be canonically homotoped to a path
$\{{y}_{rt}^{\prime}{\}}_{r\in [0,t]}$
in
$Ham(M,{\omega}_{t})$
by a Moser process that fixes its endpoints.
${Flux}_{{\omega}_{t}}\left({y}_{t}\right)$
is given by integrating
${\omega}_{t}$
over the corresponding chain
${\phi}_{t,\ell}^{\prime}:[0,t]\times {S}^{1}\to M$
. Since this is homotopic to
${\phi}_{t,\ell}:[0,t]\times {S}^{1}\to M$
mod boundary, we find that
$\Phi \left(t\right)={Flux}_{{\omega}_{t}}\left({y}_{t}\right)$
mod
${\Gamma}_{{\omega}_{t}}$
, as required.
To complete the proof we must construct an appropriate modification of
$\psi :\Sigma \to B$
.
Consider the subspace
${V}^{1}$
of
${H}^{1}(M;\mathbb{R})$
generated by the elements
${u}_{i}^{*}\alpha \alpha $
, where
$i=1,...,2p,$
and
$\alpha $
runs through the elements of
${H}^{1}(M;\mathbb{R})$
. If the elements
$a,b\in Symp(M,\omega )$
are each homotopic to some
${u}_{i},i=1,...,2p$
then
$$\begin{array}{ccc}{\widehat{F}}^{s}\left(\right[a,b\left]\right)& =& {\widehat{F}}^{s}\left({b}^{1}\right)+\left({b}^{1}{)}^{*}{\widehat{F}}^{s}\right({a}^{1})+({a}^{1}{b}^{1}{)}^{*}{\widehat{F}}^{s}\left(b\right)+\left(b{a}^{1}{b}^{1}{)}^{*}{\widehat{F}}^{s}\right(a)\end{array}$$  
$$\begin{array}{ccc}& =& {b}^{*}{\widehat{F}}^{s}\left(b\right)+\left({a}^{1}{b}^{1}{)}^{*}{\widehat{F}}^{s}\right(b)({b}^{1}{)}^{*}{a}^{*}{\widehat{F}}^{s}\left(a\right)+\left(b{a}^{1}{b}^{1}{)}^{*}{\widehat{F}}^{s}\right(a),\end{array}$$  
which is easily seen to lie in
${V}^{1}/\left({V}^{1}\cap {H}^{1}(M;\mathcal{P}\omega )\right)$
. Hence
$${\widehat{F}}_{{\omega}_{t}}^{s}\left({y}_{t}\right)\in {V}^{1}/\left({V}^{1}\cap {H}^{1}(M;{\mathcal{P}}_{{\omega}_{t}})\right),t\in [0,1].$$
By compactness we can therefore find a finite collection of smooth families
$({v}_{jt},{\alpha}_{jt}),j=1,...,m$
, such that
$${\sum}_{j=1}^{m}{v}_{j}^{*}{\alpha}_{jt}{\alpha}_{jt}\in {\widehat{F}}_{{\omega}_{t}}^{s}\left({y}_{t}\right)+{H}^{1}(M;{\mathcal{P}}_{{\omega}_{t}}),t\in [0,1],$$
where each
${v}_{jt}\in Symp(M,{\omega}_{t})$
is a product of the elements
$({u}_{i}{g}_{it}{)}^{\pm 1},i=1,...,2p,$
and
${\alpha}_{jt}$
is a path in
${H}^{1}(M;\mathbb{R})$
with initial point
${\alpha}_{j0}=0$
.
For each pair
$({v}_{jt},{\alpha}_{jt})$
choose a path
${\stackrel{~}{h}}_{jt}$
in
${\stackrel{~}{Symp}}_{0}(M,{\omega}_{t})$
starting at
$id$
such that
${Flux}_{{\omega}_{t}}{\stackrel{~}{h}}_{jt}={\alpha}_{jt}$
. Then
$${Flux}_{{\omega}_{t}}[{v}_{jt}^{1},{\stackrel{~}{h}}_{jt}^{1}]={\alpha}_{jt}{v}_{jt}^{*}{\alpha}_{jt}=:{\beta}_{jt}.$$
Therefore there is a fibration
$M\to {Q}_{j}\to {\mathbb{T}}^{2}$
that admits an
${\omega}_{t}$
symplectic structure for each
$t$
and a flat connection over
${\mathbb{T}}^{2}\mathbb{\backslash}{D}^{2}$
whose boundary holonomy has flux
${\beta}_{jt}$
. As an
${\omega}_{t}$
symplectic bundle,
${Q}_{j}$
is pulled back from a bundle over
${S}^{1}$
with holonomy
${v}_{jt}$
.
Our choice of
${v}_{jt}$
implies this bundle is a pullback of
$P\to B$
by some map
${\psi}_{j}:{S}^{1}\to B$
that we can assume to be independent of
$t$
(since the holonomy
${\beta}_{jt}$
depends only on the homotopy class of
${v}_{jt}$
.) However the connection varies smoothly with
$t$
. Therefore we can change the flux
${\widehat{F}}_{{\omega}_{t}}^{s}\left({y}_{t}\right)$
of the boundary
${\omega}_{t}$
holonomy of the chosen flat connection on
${f}^{*}P\to \left(\Sigma \mathbb{\backslash}{D}^{2}\right)$
to
${\widehat{F}}_{{\omega}_{t}}^{s}\left({y}_{t}\right)+{\beta}_{jt}$
by replacing
$\psi :\Sigma \to B$
by the homologous map
$$\psi \#{\psi}_{j}:\Sigma \#{\mathbb{T}}^{2}\to B.$$
Repeating this process for
$j=1,...,m$
allows us to perform the required modification.
$\square $
4 Further considerations
We begin by collecting together various observations about the groups
${Ham}^{s\mathbb{Z}}$
in the case when
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
. We then explain some situations in which
${\widehat{F}}_{s}$
lifts to a crossed homomorphism with values in
${H}^{1}(M;\mathbb{R})/\Gamma $
. This is followed by a short discussion of
$c$
Hamiltonian bundles and covering groups.
4.1 The integral case
We shall assume throughout this section that
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
. Many (but not all) of our results have some analog in the general case.
We begin by considering the integral analog of Lemma
2.5 . If
${H}_{1}(M;\mathbb{Z})$
is torsion free, then Lemma 2.2 applies and the whole of this lemma extends. But if this group has torsion then it is possible that (ii) does not hold.
Lemma 4.1.
Suppose that
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
and set
$Tor:=Tor\left({H}_{1}\right(M;\mathbb{Z}\left)\right)$
. Then: (i)
${\widehat{F}}_{s}^{\mathbb{Z}}$
induces a crossed homomorphism
${C}_{s}:{\pi}_{0}(Symp)\to Hom(Tor,\mathbb{R}/\mathbb{Z})$
, whose kernel equals the image of
${\pi}_{0}\left({Ham}^{s\mathbb{Z}}\right)$
in
${\pi}_{0}(Symp)$
.
(ii) The image
$\left[{C}_{s}\right]$
of
${C}_{s}$
in
${H}_{EM}^{1}(Tor;\mathbb{R}/\mathbb{Z})$
is independent of the choice of
$s$
. In particular, if
$Symp$
acts trivially on
$Tor$
then the kernel of
${C}_{s}$
is independent of the choice of splitting
$s$
.
(iii) There is a splitting such that
${C}_{s}=0$
iff
$\left[{C}_{s}\right]=0$
.

Proof.
We saw in Lemma 2.5 that if
$[\ell ]$
has order
$N$
there are precisely
$N$
distinct elements of order
$N$
in the coset
${\pi}_{\mathbb{Z}}^{1}\left(\right[\ell \left]\right)$
, namely
$\langle \ell \rangle \langle {\gamma}_{(i+\mu )/N}\rangle $
for
$i=0,...,N1$
, where
$\mu $
is the area of a chain
$W$
that bounds
$N\ell $
. Since
${Symp}_{0}$
is a connected group it must act trivially on these elements. Therefore, for each
$g\in Symp$
the restriction of
${\widehat{F}}_{s}^{\mathbb{Z}}\left(g\right)$
to the torsion elements in
${H}_{1}(M;\mathbb{Z})$
depends only on the image of
$g$
in
${\pi}_{0}(Symp)$
.
This shows that
${C}_{s}$
exists. Its kernel obviously contains the image of
${Ham}^{s\mathbb{Z}}$
. To complete the proof of (i) we must show that if
${\widehat{F}}_{s}^{\mathbb{Z}}\left(g\right)$
vanishes on
$Tor$
then
$g$
may be isotoped to an element in
${Ham}^{s\mathbb{Z}}$
. But this holds by the proof of Lemma 2.2 . (Note that we may assume that
$dimM>2$
here since otherwise
$Tor=0$
.) Statement (ii) holds by the argument of Proposition 2.3 : given a splitting
$s$
of
${\pi}_{\mathbb{Z}}$
over
$Tor$
, any other splitting
${s}^{\prime}:Tor\to {\pi}_{\mathbb{Z}}^{1}(Tor)$
has the form
$s+\langle {\gamma}_{\beta \left(\right[\ell \left]\right)}\rangle $
where
$\beta \in Hom(Tor;\mathbb{R}/\mathbb{Z})$
. (iii) is an immediate consequence of (ii). □
Example 4.2.
(i) We again assume that
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
. Suppose that for some
$g\in Symp$
there is a loop
$\ell $
such that
$[\ell ]$
has order
$N>1$
in
$Tor$
and there is a
$2$
chain
$W$
with boundary
$g(\ell )\ell $
and area
$i/N$
, where
$0<i<N$
. Then
${C}_{s}\left(g\right)\ne 0$
for all splittings
$s$
, and there is no splitting such that
$g$
is isotopic to an element in
${Ham}^{s\mathbb{Z}}$
. Note that the corresponding mapping torus bundle
$${M}_{g}:=M\times [0,1]/(x,1)\sim (gx,0)\u27f6{S}^{1}$$
has a closed connection form but not one that is integral.^{10 }
Equivalently,
$g$
does not fix any integral lift
$\tau \in {H}^{2}(M;\mathbb{Z})$
of
$\left[\omega \right]$
.
(ii) The following yet more intriguing situation cannot be ruled out in any obvious way. Suppose that
$Tor=\mathbb{Z}/2\oplus \mathbb{Z}/2$
is generated by the elements
$[\ell ]$
and
$\left[{\ell}^{\prime}\right]$
which are interchanged by two symplectomorphisms
${h}_{1},{h}_{2}$
. Suppose further that
$$({h}_{1}{)}_{*}\langle \ell \rangle \ne ({h}_{2}{)}_{*}\langle \ell \rangle ,({h}_{1}^{2}{)}_{*}\langle \ell \rangle =({h}_{2}^{2}{)}_{*}\langle \ell \rangle =\langle \ell \rangle .$$
Then
$g:={h}_{1}{h}_{2}$
fixes
$[\ell ]$
but acts nontrivially on
$\langle \ell \rangle $
and so has the properties assumed in (i) above. Now consider the splittings
${s}_{i}$
defined by
$${s}_{i}[\ell ]=\langle \ell \rangle ,{s}_{i}\left[{\ell}^{\prime}\right]=\langle {h}_{i}(\ell )\rangle ,i=1,2.$$
Then
${h}_{i}\in {Ham}^{{s}_{i}\mathbb{Z}}$
by construction. The corresponding mapping tori
${M}_{{h}_{i}}$
have
${Ham}^{{s}_{i}\mathbb{Z}}$
structures and so Lemma 3.2 implies that each supports a closed integral connection form. But their fiber connect sum
$P\to V:={S}^{1}\vee {S}^{1}$
does not, since one of its pullbacks is the bundle
${M}_{g}\to {S}^{1}$
considered in (i). Indeed, the torsion in
${H}_{1}(M;\mathbb{Z})$
creates new terms in
${H}_{2}(P;\mathbb{Z})$
on which any closed connection form is nonintegral. But if every
$M$
bundle with closed integral connection form is pulled back from some universal bundle
$\mathcal{P}\to \mathcal{\mathcal{B}}$
with this property, then
$P\to V$
would also be such a pullback, and hence would also have a closed integral connection form. Thus, if
$Symp(M,\omega )$
contains elements
${h}_{1},{h}_{2}$
as above there is no universal
$M$
bundle
$\mathcal{P}\to \mathcal{\mathcal{B}}$
with closed integral connection form.
A similar argument applies whenever there are two splittings
${s}_{i},i=1,2,$
such that the images of
${Ham}^{{s}_{i}\mathbb{Z}}$
in
${\pi}_{0}(Symp)$
are different. It follows from Proposition 4.5 that this happens iff there are integral lifts
${\tau}_{i},i=1,2$
, of
$\left[\omega \right]$
that are stabilized by different subgroups of
${\pi}_{0}(Symp)$
. To get around this difficulty, one must reformulate the classification problem: see Gal–Kȩdra [
1]
.
We next explain a very natural way to think of a splitting
${s}^{\mathbb{Z}}$
of
${\pi}_{\mathbb{Z}}$
when
$\mathcal{P}\mathbb{Z}\omega =\mathbb{Z}$
.