Some previously reported heritabilities are significantly lower than the
values reported here. There are a number of factors which may account
for the discrepancy, relatedness assumptions are a likely candidate.
There was no consistency across studies in assumptions regarding
relatedness, as they have differing levels of knowledge of their
populations and species. A number of the previous studies were multisite
studies and may include growth restricted sites, high mortality etc.
which likely contribute to the comparatively low values. A single
nursery site study gives much more control over environmental variation,
and hence higher discernable genetic influences (for the particular
environment). \citet{Davies2017} was a
single nursery site study, which produced similar values to the ones
reported here. Genetic parameters could be expected to reduce when
considered over multiple sites with harsher growing conditions. Given
the large variation and limited previous studies, significant
conclusions with regard to the previously reported genetic correlations
and those reported in this study cannot be drawn. Measurement error
reduces calculated heritability as it adds random error to all results,
it may have a significant influence on all of the reported traits here,
however is likely quite pronounced for strain, as it will be in all
studies measuring growth-strain.
Genetic correlations require more data and complex statistical modeling
than heritability estimates. As a result fewer have been reported.
Relevant genetic correlations in various eucalypt species are reported
in Table \ref{221635}.
Investigating multiple populations to gain an understanding of genotype
x environment interaction information is common practice among modern
tree breeders. Our study dealt with a single site; however, the trial
represented multiple provenances, some of which were partially
geographically isolated from the remainder of the population. No
previous studies have investigated divergent evolution in E.
bosistoana. While no clear differences existed between provenances
within this experiment, further work is needed to identify where true
provenance boundaries lie, as a better understanding of the genetic
isolation of various sub populations of E. bosistoana may help
explain the comparatively high heritabilities in this study.