I am planning the planting of new orchards for this winter and considering the importance of cross-pollination. I have been able to find two late-blooming varieties in the nursery: Guara and Filippo Cea. They are self-compatible, so they do not require another variety to pollinate. However, it is claimed that including several varieties in the orchard can improve fruit set and yields. After checking some papers, I substantiate here that this is not a wise option for the two available varieties.
Most almond varieties are self-incompatible. The pistil recognizes pollen from the same variety according to a set of genes: S genotype. Every almond plant has a pair of S-types alleles. Flowers are diploid, but in the anther they produce haploid pollen grains. When these grains reach the pistil, they are rejected if their allele coincides with one of the alleles of the plant. For example, the flowers of Ferragnes variety, with S1S3 genotype, will reject all pollen grains with S1 or S3 genotype.
The so called self-compatible varieties have one S-allele that does not work, so when a grain of pollen with this unusual gene reaches the pistil of the same plant, they are not rejected. This allele is named as Sf. As the draw shows for the variety Lauranne, it happens with half pollen grains from the same plant.
Cross pollination in self-compatible varieties is always possible. If the normal functioning S-alleles are different, all pollen grains are able to pollinate the other variety. If S alleles are the same, only half of the pollen will do so: in this case cross pollination does not improve fruit set.
Filippo Cea and Guara have the same S-genotype (S1Sf), so there is no benefit from mixing them in the orchard. An option that could have been considered is a mix of Lauranne and Tuono.