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.

40 P. B. Kronheimer and T. S. Mrowka<br />

A technical device of a slightly different sort is our use of cone-like orbifold<br />

metrics throughout. The exposition could have been streamlined somewhat if<br />

Conjecture 8.2 from [KrM] had been proved in complete generality, rather than<br />

the untidy approximation (2.12). Recall that in our arguments above <strong>for</strong> the<br />

proof of the key Proposition 5.1, we first needed to choose a suitably small value<br />

of α, smaller than a quantity α0 determined by n and g alone, and then ν had<br />

to be chosen so that the conclusion of (2.12) held <strong>for</strong> some interval I containing<br />

α.<br />

(iv) Other remarks.<br />

We gather together here a few remarks and questions related to Theorem 1.1.<br />

First of all, <strong>for</strong> an oriented 4-manifold X with non-zero invariants qk, the main<br />

theorem gives a lower bound <strong>for</strong> the genus of an <strong>embedded</strong> surface in terms of<br />

the self-intersection number; but this lower bound is trivial if the self-intersection<br />

number is negative. This contrasts, <strong>for</strong> example, with the lower bound obtained<br />

in [HS], which is independent of the sign. The point here is that the hypothesis<br />

on qk is sensitive to orientation; at present, no 4-manifold is known <strong>for</strong> which<br />

the polynomial invariants are non-zero <strong>for</strong> both orientations.<br />

To put some flesh on this remark, consider <strong>for</strong> X a K3 surface. As was<br />

mentioned in [KrM], the lower bound provided by Theorem 1.1 is sharp in this<br />

case, <strong>for</strong> all classes of positive square:<br />

Proposition 6.6. If ξ is a class of positive square in a K3 surface, then the<br />

smallest possible genus of an <strong>embedded</strong> surface representing ξ is given by 2g−2=<br />

ξ·ξ.<br />

If the class ξ has negative square however, it may well be represented by a sphere.<br />

The usual description of the Kummer surface exhibits 16 disjoint spheres of<br />

square −2 inK3; by joining any number of these together by piping, we obtain<br />

<strong>embedded</strong> spheres of square −4, −6,andsoondownto−32. With a little more<br />

ingenuity (as was pointed out to the authors by Ruberman), it is possible to go<br />

quite a bit further down, but it is not clear whether there is eventually a lower<br />

bound. Up to the action of the diffeomorphism group, there is in K3 a unique<br />

primitive class of square −2n, <strong>for</strong> every n, and we do not know the smallest<br />

genus surface which can represent this class. So we ask:<br />

Question 6.7. Can every primitive class in K3 with negative self-intersection be<br />

represented by an <strong>embedded</strong> sphere?<br />

The phrasing of this question is not meant to imply that the authors conjecture<br />

that the answer is ‘yes’.<br />

Some results about immersed (rather than <strong>embedded</strong>) <strong>surfaces</strong> are easily<br />

deduced from Theorem 1.1. For example, let Σ now be an immersed sphere in a<br />

4-manifold X, having d+ double points of positive sign and d− double points of

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

Saved successfully!

Ooh no, something went wrong!