Full Idea
Kripke's first (good) route to the necessary a posteriori is based on the idea that certain properties of objects that they can be known to have only a posteriori, may be known a priori to be essential properties of anything that has them.
Gist of Idea
Kripke claims that some properties, only knowable posteriori, are known a priori to be essential
Source
report of Saul A. Kripke (Naming and Necessity lectures [1970]) by Scott Soames - Significance of the Kripkean Nec A Posteriori p.180
Book Reference
Soames,Scott: 'Philosophical Essays 2:Significance of Language' [Princeton 2009], p.180
A Reaction
Interesting, and a key issue. I think this is precisely where I disagree with the Kripkean view of necessities. Logicians want to know a priori what is essential for identity, but scientists want to know what makes things tick. See Kripke on pain.