Chronology of major discoveries in gametophytic SI
I did not necessarily read all the articles in the chronology so
some earlier works may not be cited in the references section.

There are many types of blocks to self-fertilization: heteromorphic (dioecious 
plants, stamen and pistil maturation), homomorphic (sporophytic & gametophytic 
self-incompatibility). SI is associated with single a polymorphic S-locus 
(having many S-alleles.

In GSI plants, pollen begins to grow pollen tube, but growth is arrested due 
       to same S-allele. 
Need allele, recognition and process of halting for SI to work. Therefore, 
       3 things to look for (tripartite model of S-locus: a common allele, 
       a pollen part that expresses it, and a receptor to recognize it from the
       pistil end. 

Looked for and found proteins that co-segregate with S-alleles => S-proteins.
As it turns out, these S-proteins are expressed along the path of pollen-tube 
       growth (Cornish, 1987).
These S-proteins are not just any old proteins: they have RNase abilities. 
       (McClure et al., 1989).
2 histidines which are thought to be particularly important in RNase function 
       in fungi are highly conserved, i.e. present in all S-proteins known 
       (Kawata et al., 1990). 
GSI involves rRNA degradation (McClure et al., 1990). RNases degrade RNA. 
       Pollen tubes do not produce new rRNA .
These proteins are probably the style-part.

Hypervariable regions of the S locus probably confer specificity (Tsai, 1992)

The pollen part must look pretty different from style part since no one has 
       been able to find the former yet. Some people looked for S-protein 
       homologues in pollen. Found some, but only in trace amounts (Dodd et al.,
       1993) that are practically undetectable even with PCR. 

Knock-out and gain-of-function experiments show that S-proteins are enough for 
       the pistil to recognize and reject self-pollen (Leeet al., 1994; Murfett
       et al., 1994).
Some evidence that RNase function is important to style part SI function 
       (Huang et al., 1994; Royo et al., 1994).

2 models of how degradation by S-RNase works: receptor, inhibitor. Both 
       involve the degradation of pollen tube rRNA, which subsequently halt 
       pollen tube growth. 
There are problems with receptor model based on self-compatibility of 
       tetraploid mutants.

More research into the structure of S-proteins results in the discovery of 
       4 S=S bridges (Ishimizu et al., 1996).

A naturally occuring mutant of a variety of Japanese pear, Osa-Nijisseiki, is 
       discovered and the S-allele is characterized (Ishimuzu et al., 1996).
Based on research with variety Osa-Nijisseiki, S-proteins are not important in 
       the pollen-part side of SI (Sassa et al., 1997).

Some believe the S-proteins do not necessarily work primarily by degrading rRNA.
In incompatible pollination, pollen tube growth is arrested but it can resume 
       if the style is grafted to a compatible style (Lush and Clarke, 1997).  
There is no evidence of a general degradation of ribosomes (Walles et al., 
       1998). 
Either the rRNA is not getting degraded or the pollen tube has manufactured 
       more ribosomes. The S-protein substrate must be more super-specific than
       at first thought.