I'm interested in the following paper http://www.computer.org/csdl/trans/tc/1981/02/06312176.pdf
In particular i'm interested in the construction Valiant describes to prove that it is possible to embed a $4$-planar graph in a grid of size $9n^{2}$ (on the top right of page 2). He says that the construction is described in fig 2, but i do not understand it completely. I was hoping someone may be able to explain it more clearly. One of my confusions with the diagram in fig 2 is how do we know the neighbours of the next vertex to be added namely the $(I+1)$st vertex lie on the border of the currently constructed embedding, what's to say stopping one it's neighbours being in the middle of the embedding?