Self-incompatibility is a complicated and confusing matter. Sometimes it can be weakened or eliminated by environmental conditions. There is also variation among specimens of a single species.
For instance, Stout (1922), discussing sterility in lilies, wrote, “The phenomena of incompatibility are well illustrated by the results obtained with L. regale, a splendid and beautiful lily recently introduced from China. Of the ten plants tested, nine were self-incompatible and one was self-compatible. Cross pollination between plants sometimes succeeded and sometimes failed.”
http://bulbnrose.x10.mx/Heredity/Stout_lily_hybrids.html
Leaky self-incompatibility was also responsible for the rapid development of cultivated Primulas. Ordinarily the plants result from crosses between “Pin” and “Thrum” breeding types. But “illegitimate” self-pollinations can occur, and be continued generation after generation, exposing recessive traits that would normally be hidden.
Weather can have a profound influence on fertility – both self- and cross. Plants that seem sterile (or nearly) may suddenly become fertile. It would be interesting to know how the weather in England in 1887 differed from normal.
The Garden, 1887
p. 194: Liliums seeding.—I have never known the different species of Lilies to bear seed so freely as this year. I have one seed pod on a plant of L. testaceum. As this Lily is supposed to be a hybrid, I wish to know if it often bears seed, as I have never known it do so before.—W. SHOOLBRED, St. Ann’s, Chepstow.
p. 394: Lilium testaceum seeding.—In answer to W. Shoolbred in THE GARDEN (p. 194), I once got a good seed-pod of this. It was exactly between Lilium candidum and L chalcedonicum. L candidum and L. testaceum both seeded under the same circumstance: the stems were cut off in flower and put into a pot of water in the open air. It is said if candidum is cut and the stems hung topsy-turvey they will seed .—F. MILES
Withers (1975) dealt specifically with Lilies, but also discussed research into self- and cross-incompatiblity in other plants. One note that may be of some interest: “Dr. Peter Valder of Sydney University has successfully crossed evergreen with deciduous rhododendrons by cutting off the style of the seed parent just in front of the ovary and then applying pollen from the pollen parent to the cut surface of the style. The pollen has germinated, and has had a much shorter journey to the ovules in the ovary, and successful fertilization and seed production has taken place.”
http://bulbnrose.x10.mx/Heredity/improbab.html
Lucien Reychler (1926) also experimented with pollen applied to the stumps of cut styles, and introduced directly into the ovary.
http://bulbnrose.x10.mx/Heredity/Reychler/reyctrau.htm
The shorter journey would reduce or eliminate whatever incompatibilty reactions may be occurring in the style.
Karl