Fig. 6.

A population-dependent, multilocus genetic basis underlies the A. lyrata × A. arenosa reciprocal hybridization barriers. (A and B) A. lyrata × A. arenosa F₁ backcrosses to determine the genetic basis of A. lyrata × A. arenosa incompatibility. F₁A and F₁B correspond to two different A. lyrata × A. arenosa F₁ hybrids generated using A. arenosa individuals from two different populations. Seed abortion rate is shown for A. lyrata × F₁ (A) and F₁ × A. arenosa (B) backcrosses. (C and D) F₁ backcrosses to determine the genetic basis of A. arenosa × A. lyrata incompatibility. (Left) Seed size distribution of A. arenosa × F₁ (C) and F₁ × A. lyrata (D); (Right) expected and obtained rate of abnormal seeds. The range of abnormal seed size (lower than the 2A × 2L average seed size of 0.21 mm²) (Fig. 1) is indicated with red range bar. The average size of parental A. arenosa, A. lyrata, and A. arenosa × A. lyrata hybrid seeds is shown by black arrows. Experimental values are represented by purple and blue bars for A. lyrata × A. arenosa and A. arenosa × A. lyrata crosses, respectively. Gray bars represent theoretical values based on the number of indicated involved loci (38). Error bars represent SD between backcross replicates (4 backcrosses for each F₁ type). **P < 0.01 (χ² test) compared with the expected abortion rate for the number of indicated loci. N.S: nonsignificant.