Abstract
Natural hybrid populations in Opuntia, Iris, Aquilegia, Gilia, and Oxytropis are examined or reexamined for character coherence. Coherence is present in some but not in other populations. The differences between plant groups and between populations are greater than expected, according to the standard theory. A reassessment of the standard theory reveals three flaws leading to overestimates of the amount of coherence: (1) failure to exclude parental-type individuals from hybrid population samples, (2) use of non-quantitative pictorialized scatter diagrams rather than correlation coefficients, and (3) the assumption that selection will always favor the parental types in natural hybrid populations. An alternative framework recognizing more variable factors is proposed. Population-to-population differences in character coherence in natural hybrid populations can be accounted for with this alternate approach.

This publication has 1 reference indexed in Scilit: