Íslenskar landbúnaðarrannsóknir - 01.03.1979, Side 91

TRAITS IN ICEL. TOELTER HORSES I. 89
any trend is to be distinguished, it points
downwards.
The peaks of the estimated influence of
years seem to some extent to be associated
with the location of the horse shows in
their four year cycle around the country.
Looking at this and considering that the
effects of district of origin seem unimpor-
tant, one gets the impression that a genetic
difference between regions is negligible,
but environmentai diflerence (in training
and quality of riders) might exist.
Before running model (2) the data were
additively corrected by the least-squares
constants for the various sex and age
groups. The least-squares constants were
estimated as the difference between m and
the least-squares means of the appropriate
subclass. Obviously the adjusting factors
are estimated irrespective of any possible
genetic trend.
2. Estimates of heritability and correlations
The estimates of heritability, genetic and
phenotypic correlations obtained from
runs 1, 2 and 3 on model (2) are listed in
table 5.
The heritability estimates of the four
body measurements indicate that the en-
vironment (or the non-additive genetic
eflect) is no more than 50% responsible
for the phenotypic variation observed. It
should also be noted how strongly correl-
ated these measurements are, especially
genetically. These results are in good ag-
reement with previous studies on body
measurements in horse and cattle popul-
ations.
Estimated correlations between body
measurements and the traits are noticably
low in most cases, but none significantly
negative. Increased selection pressure for
the four body measurements, would
therefore not be expected to hinder res-
ponse in the other traits, except as a result
of a reduced selection intensity.
Apart from ,,pace“ the subjectively
scored traits have substantially lower
heritability estimates, some of them not
even differing significantly from zero. It
should be stressed that within the square
(in table 5), all estimates of parameters
concerning pace, are estimated after the
exclusion of non-pacers (i. e. those witfi
pace score below 6.0). It is possible that
this has not been a sufficiently accurate
way of reaching normal distribution of
pace scores. If this is true, the heritability
estimate of pace might be somewhat bias-
ed.
The heritability estimates for the traits
are on average somewhat lower than the
heritability ofperformance traits obtained
in other studies (see table 1). Heritability
estimates of character are of similar mag-
nitude though, but that is the only trait
which by its nature is strictly comparable
with the traits of the Icelandic ponies used
in this study.
DISCUSSION
Validity of heritability estimates obtained
by the procedure followed in this study
relies on several assumptions. Amongst
them are assumptions of no selection of
parents nor oflspring measured and the
absence of environmental correlations
between the relatives being used to estim-
ate the amount of additive genetic variat-
ion present in the population. There is
reason to believe that neither of these ass-
umptions are completely fulfilled in the