Self pollinating narcissus flowers

[annew]

I've just been reading the very interesting article on the inheritance of mutations in narcissus by Peter Brandham in the RHS yearbook. There is one particular part that puzzles me. He says 'The genus is known to be almost completely self-incompatible..'
I'm pretty sure that I've had plenty of seeds off flowers of several different divisions that had been 'selfed' Sometimes I only had one bulb of a particular plant so had no choice. Also, some flowers eg of the Apodanthi seem to be designed to self-pollinate with the stamens and style close together within the tube.
So what is happening here? Are the selfed plants producing seeds by apomixis?
How 'almost completely' is the genus self-incompatible?

[Maggi Young]

Links suggested by  Fernando Ureña Plaza......

http://onlinelibrary.wiley.com/doi/10.1111/j.1420-9101.2004.00767.x/full 
and   http://ebd06.ebd.csic.es/pdfs/Medrano.et.al.2005.Ann.Bot.pdf

[Maggi Young]

Link from Daffnet:   http://daffnet.org/the-inheritance-of-recessive-and-dominant-mutations-in-daffodil-cultivars/#comment-17946

[annew]

Thank you, Maggi. The pdf in the last link is excellent. Prof Theo gives the case for hybridisers for whom even a couple of seeds from a cross are welcome!

[Ian Y]

I have also had seed set on Narcissus and other bulbs often described as not self compatible.

Ideally Narcissus (and others) are most fertile when cross pollinated by another clone and there are mechanisms within the structure of the flower to inhibit self pollination - commonly the stigma is receptive before the pollen is ripe, or the stigma and anthers are at different levels. I believe that the inhibitor is often on the stigmatic surface which  if no cross pollination occurs  breaks down as it ages allowing the flower to be pollinated by its own pollen.