Put . This surface with boundary deformation retracts onto a wedge of circles, so has dimension and . If were a retraction, would make injective. Its image would have dimension .
Every pair of classes in has , since . Naturality gives
Thus is an isotropic subspace of a symplectic vector space for the nondegenerate skew Poincare duality pairing on the -dimensional space . The stated linear-algebra bound gives . Hence
The bound is sharp. Double along its boundary: is the closed oriented surface of genus . Identify each copy with and fold them onto one copy. The two maps agree on the joining circle, so they give a continuous retraction fixing the first copy pointwise. This includes , where the double of a disc is a sphere.
More generally, for , add handles in the interior of the second copy. Pinch those extra handles onto their connecting point, keeping the boundary fixed, and then fold onto . This is still a retraction. In fact the construction works for any embedding in the question: the connected complement has one boundary component and genus by Euler-characteristic additivity, and the classification theorem for surfaces identifies it, relative to that boundary, with a copy of having additional handles. Thus the exact existence criterion is , as expressed by retraction onto a punctured oriented surface.

Articles by others on the same topic (0)

There are currently no matching articles.