Íslenskar landbúnaðarrannsóknir - 01.03.1979, Qupperneq 93

TRAITS IN ICEL. TOELTER HORSES I. 91
to be expected. Although far from conc-
lusive, this might therefore indicate that
mares judged on and in connection with
the horse shows are more selected for ga-
iting ability and less selected for „legs“
score than the males. Although the selec-
tion of the offspring used in this material,
or their parents, is more intensive on
stallions, when compared to the even sex
ratio of newborn foals, this is totally re-
versed when looked at the selection of al-
ready adult horses to be performance test-
ed in connection with the horse-shows.
Almost every stallion broken and trained
is tested at 4 to 8 years of age, but has
undergone a preselection when it was
decided whether the colt should be ca-
strated or not at the age of 2 to 3 years. As
Icelandic ponies are unbroken at that age
one would expect selection on the indivi-
dual performance at this stage to be more
eífectively based on body conformation
than the traits of riding ability.
It should be noticed, however, that the
preselection is based on some characters,
whose parameters remain totally unest-
imated. Therefore there is a need for a
defmition of the characters on which it is,
or should be, based. The desired charac-
ters must be sufficiently heritable and
have high genetic correlations with the
economically important traits of adult
horses. Such information would probably
most efficiently be combined with infor-
mation on parents and sibs, so as to
maximise the correlation between such an
index and the true breeding value of the
individual (Henderson, 1963). When the
second stage selection occurs, based on
the traits, whose parameters were estim-
ated in this study, an inclusion of infor-
mation, on the nearest kin in the index,
should increase the accuracy of the selec-
tion further still.
From table 5 it is obvious that the phen-
otypic values of the body measurements
are good guides to the genotypic values, in
contrast to the phenotypic values of most
of the subjectively scored traits. The low
correlations between the body measurem-
ents and the traits do not indicate that
much correlated response is to be expec-
ted in the traits from selection of the four
body measurements. Thus it is desirable
to investigate and develop another objec-
tive body measurements, more related to
the mechanism of movement and the des-
ired body conformation. Inclusion of a
highly heritable trait genetically correlat-
ed with valuable traits, although of no
economic importance itself, can increase
the efficiency of a selection index substan-
tially (Gjedrem, 1967). Sales and Hill
(1976) point out that for these kinds of
indices the loss in efficiency from poor
estimation of parameters can be substan-
tial. Reasonably secure use of such indices
therefore requires quite precise parameter
estimates.
A very great deal of the environmental
(or non-additive genetic) variation which
modifies the phenotypic value of the traits,
is undoubtedly due to differences in the
skill of riders and trainers. This could be
reduced, and the efficiency of selection
consequently increased, by more educat-
ion and standardization of horsemanship.
That would certainly be a quicker means
of improvement than the slow process of
breeding, but also speeds up the latter
one.