Hi Rowan,
Do you mean "grafting" small networks onto a larger "backbone" network? Do I understand correctly that you want to replace 1 taxon in the backbone network by 1 "subclade" network? If that's what you need, then:
1. I don't recall any dedicated function to do this in PhyloNetworks, unfortunately. But I would be happy to write one and share it as a "gist" on github to share it quickly.
2. For this, each small network needs to be correctly rooted. Can you do this confidently?
3. If you had edge lengths, then some external information is needed for the final network to also have lengths on all branches. The grafting function would need to know how long the "stem" edge should be, between the node in the backbone network and the root of the small network.