Hard to find images of cryptocarpic cactus fruits (name = hidden fruit), the fruits being embedded in the body and not protruding from it. Whether the seeds are gradually expelled from the old stems, or have to wait until the plant actually dies to germinate is conjecture. But probably older seed germinates better than fresh. Therefore if you set any don't throw the seed pan away for a few years, just let it dry out at the end of the season and water again for the next few years and you may still get germination?
http://www.cactus-art.biz/note-book/Dic ... tocarp.htm
This Japanese site shows cryptocarpic fruits for Mammillaria theresae if you scroll right down.
http://blog.sina.com.cn/s/blog_7db6105c010126n2.html
If you scroll to the bottom of this Mammillaria luethyi link you will see a section through the plant showing the fruit and seed embedded in the plant.
http://www.llifle.com/Encyclopedia/CACT ... ia_luethyi
The M. saboae group, M. theresae, M luethyi and M. hernandezii all produce such fruit, which is why they are often propagated by grafting since seed is hard to collect commercially, unlike more normal fruits. That means most available are in fact clones of one original plant. Whether these are self fertile I do not know, but if self sterile you would obviously require a non related clone to produce seed, something difficult to find with clones propagated by grafting and not marked as to which close they are.
With the offsetting species like the M. saboae group however an offset can be taken off and rooted, but degrafting some like M. luethyi seems to be very difficult since they usually dry up before rooting, though I understand seedlings on their own roots (if you can get them) do grow quite well.
but it looks like it had about 6 or 7 flowers on it at some point
