Different types of non-genetic variation were compared in order to inv
estigate the relationship between phenotypic plasticity and developmen
tal instability. The following types of variation were estimated for 2
2 quantitative traits in Iris pumila: variation within the same flower
, variation between ramets on the same genet, variation between replic
as of the same genotype within a macroenvironment and also between two
types of macroenvironmentals, spatial and temporal. The first type of
variation is asymmetry between replicated parts, and, therefore, is a
measure of developmental instability. The second and third types are
also measures of instability but are confounded by microenvironmental
variation, and the last two types of variations are measures of phenot
ypic plasticity. For analysis of spatial variability, I chose two type
s of environments (open and shaded) in natural habitats of I, pumila.
Two successive flowering seasons were used to estimate temporal variab
ility. Asymmetries of different traits were poorly integrated and had
a small number of significant correlations with other types of variabi
lity. Significant correlations varied in sign. Other types of within-e
nvironment variation were better integrated, having a limited number o
f significant positive correlations with measures of phenotypic plasti
city. However, such positive correlations were observed mostly with on
e kind of plasticity (temporal) and only for some of the less plastic,
floral traits. Plasticities of vegetative traits were not correlated
with measures of developmental instability. Implications of these resu
lts on different models explaining the genetic basis of phenotypic pla
sticity, as well as on hypotheses concerning differences in plasticity
between species and populations, are discussed.