Genetic similarity matrices obtained from morpho-agricultural
descriptions and molecular markers were correlated in order
to establish a correspondence relationship between them,
according to Mantel (1967). The calculated correlation value
was not significant, for 10.000 permutations (r = 0.07, p = 0.099).
According to Emygdio et al. (2003), the lack of correspondence
between morphological characteristics and similarity coefficients
may be attributed to the pressure of artificial selection to which
the accessions are submitted, by breeding programs or by the
growers themselves. Vera et al. (1999) propose that biochemical
and molecular markers have no direct effect on the plant
phenotype and, are, therefore, not directly affected by selective
pressure.