20.07.2013 Views

Gauge theory for embedded surfaces, II

Gauge theory for embedded surfaces, II

Gauge theory for embedded surfaces, II

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Gauge</strong> <strong>theory</strong> <strong>for</strong> <strong>embedded</strong> <strong>surfaces</strong>, <strong>II</strong> 51<br />

the natural inclusion O⊂O (c) gives an equality µ(O)=µ O (c) . This situation<br />

is a local model <strong>for</strong> the natural inclusion<br />

µ ϕ ∗ ( ¯ F) ⊂ µ ϕ ∗ ( ¯ F) ⊗O (c) ,<br />

which must there<strong>for</strong>e be an equality also. So (8.9) is indeed isomorphic to ¯ F.<br />

Now consider the other direction: let F be an orbifold line bundle of weight<br />

b, define ¯ F by (8.5) and then apply the construction (8.8) to ¯ F to <strong>for</strong>m the line<br />

bundle<br />

G = ϕ ∗<br />

µ F⊗O ([ν/2])<br />

⊗O (b) .<br />

In general, ϕ∗ (µ(F)) can be identified with the OX-module which is generated,<br />

on each Ũ, by the invariant sections of F. There is there<strong>for</strong>e a natural inclusion<br />

ϕ∗ (µ(F)) →F, and hence in our case a map<br />

ɛ : G→F⊗O (b+[ν/2]) .<br />

Since b +[ν/2] is positive, F is a subsheaf of the orbifold sheaf on the righthand<br />

side. This subsheaf is precisely the image of ɛ, as one can verify locally by<br />

considering the case F = O (b) ,whereϕ ∗ µ F⊗O ([ν/2]) is equal to ϕ ∗ (O ¯ X),<br />

which is OX. This completes the verification that the two constructions are<br />

mutually inverse.<br />

The correspondence defined by (8.7) is functorial, as is the inverse construction.<br />

This is manifest when the line-bundles involved all have the same weight.<br />

When the weights are different, suppose (to take just one of two directions) that<br />

θ :( ¯ F1,β1)→( ¯ F2,β2) is a map between parabolic line bundles, with βi = bi/ν.<br />

If b1 ≤ b2 then O (b1) ⊂O (b2) ,andsoθinduces a natural map<br />

ϕ ∗ ( ¯ F1) ⊗O (b1) θ ′<br />

−→ ϕ ∗ ( ¯ F 2) ⊗O (b1) ↩→ϕ ∗ ( ¯ F2)⊗O (b2) ,<br />

as required. If b1 >b2 then the definition of map in the parabolic setting requires<br />

that θ should vanish along Σ in ¯ X. The first map θ ′ above is still there in this<br />

case, but its image lies in the subsheaf<br />

ϕ ∗ ( ¯ F2) ⊗O (b1−ν) ;<br />

and this is contained in ϕ ∗ ( ¯ F2)⊗O (b2) now, because b2 >b1−ν.Soineithercase<br />

we obtain a map between the corresponding orbifold bundles as defined by (8.8).<br />

To summarize, then:<br />

Proposition 8.10. The correspondence (8.7) is invertible and functorial, from<br />

the category of orbifold line bundles to the category of parabolic line bundles<br />

whose weights are rational with denominator ν. ⊓⊔<br />

We can now repeat these constructions with vector bundles in place of line<br />

bundles. Let E be an orbifold SL(2, C) bundle: a locally free orbifold OX module<br />

of rank 2, with Λ 2 E ∼ = OX. Locally, such bundle is isomorphic to O (a) ⊕O (−a)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!