JOURNAL OF THE SCIENTIFIC AGRICULTURAL SOCIETY OF FINLAND

517

Maataloustieteellinen Aikakauskirja
Voi. 47: 517—532, 1975

Experiences in the evaluation of carcass value of live
pigs by ultrasonic photography

Mikko Varo, Marjatta Perälä, Matti Ojala and Hannu Varo
University of Helsinki, Department of Animal Breeding

Abstract. In the present study it was found that the mean cross-sectional lean area
of three ultrasonic photographs has a closer correlation than any one of the three individ-
ual areas to the cross-sectional lean area measured from the carcass. The correlation of
this mean value with the lean + bone amount of the most valuable part of the half-
carcass approached that of the cross-sectional area of the lean measured from the carcass.
The correlation of the ultrasonic area measurements with the carcass amounts of fat
and lean -f- bone and with the fat-lean ratio calculated from these was closer than that
of the ultrasonic thickness measurements. The fat-lean ratio calculated from the areas
was also slightly better correlated to carcass quality than was the ratio calculated from
the thickness measurements. In other respects the fat thickness measurements appeared
to be better than area measurement. Although the measurement of lean correlated to
the carcassa value of the animal more poorly than the fat measurements, the highest
correlations were nevertheless obtained by the combined use of the fat and lean measure-
ments.

Introduction
The Department of Animal Breeding of the University of Helsinki lias,

with the financial support of the Finnish National Research Council for Agricul-
ture and Forestry, procured a Scanogram ultrasonic photographic device
developed in the United States by Professor Stouffer (Stouffer 1965) with
which pictures of cross-sectional areas of fat and lean can be obtained in live
animals. This instrument was used for photographic measurements of litters
reared for progeny testing of boars done in form of Danish at the Pig Husbandry
Experiment Station in Hyvinkää in 1970 and 1971. Quite satisfactory results
were obtained already at the training stage during the first year. The results
here presented are based on the measurements made in 1971.

Material and methods
The material studied comprised ten test litters of Yorkshire and ten of

Finnish Landrace, totalling 80 animals. One week prior to slaughter ultrasonic
photographs were taken at three sites: the middle picture B was obtained
behind the last thoracic vertebra, picture A 10 cm anterior to B, and C the

https://www.c-info.fi/en/info/?token=B2wMFZOeOa2MPvow.ZNzhGA3VwVKUBe_ZQNLvFw.A4I3oLyFcDGF2TCA647_9zv10jEVX32VIGdSHeMo32CozE4ePyYRRmkhQwKO2sjmNWbFGz44zRJvm8qn4EYEowvAx7D_8lP6BWoHBIJtiTjItotHauuMgbgp7XVoF8FsRJ0vzhk80Z9AsuumV-no5_peIMCN1ysuE9Q50Ij3vjJp9cumcvlKHJMZR3tIlGBKxgQGue2N8ccrZOpigjTf5x_fHE4fH4oNJ3yS5Um3iA


518

same distance posterior to B. From these cross-sectional pictures of the back
the thickness and area of lean and the thickness and area of fat were measured
and the fat-lean ratio was calculated. Since, in addition to the ultrasonic meas-
urements, the normal test results and carcass measurements were also obtained
for these animals, it was possible to compare the carcass values estimated
from the live animals with the actual carcass values obtained on dissection.

The comparisons between the ultrasonic photography results and the
carcass measurements were made by means of correlation coefficients calculated
individually within breeds and within sexes and by test litters within breeds.
The combining of the results of the small material in this manner was considered
meaningful because the breeds did not differ significantly, as will be seen below.
This simplified the treatment of the subject. The sexes differed significantly
with respect to several characteristics. Such differences could have caused
errors in the correlation coefficients computed on individual basis had they not
been calculated within sexes.

Results
The mean values and standard deviations, the coefficients of heritability

of the ultrasonic measurements, and the significances of differences between
the test litters and the sexes are presented in Table 1. A statistically significant
difference between the breeds was found only in the thickness of the lean at
site A. It is a noteworthy finding that, with the two exceptions of lean thickness
and area at A, the standard deviations of values were greater in the Yorkshire
than in the Finnish Landrace. The differences were even quite significant
with respect to fat and the fat-lean ratio. The coefficients of heritability were
estimated from variations between the test litters, since with one or two excep-
tions each boar had only one litter in the test. Although the numerical readings
of heritability coefficients should be accepted with some reservation because
of the small material, it is possible on the basis of their relative magnitudes to
compare the reliability of the different measurements. Investigators have
generally found a greater heritability of thickness of fat than of thickness of
lean. The coefficient of heritability of the fat lean ratio, again, approaches
that of fat. Certain coefficients of heritability of ultrasonic measurements
obtained in large pig materials are cited below:

Fat Lean Fat-lean
thickness thickness ratio n Author

0.59 0.35 0.48 1096 Glodek (1964)
0.25 0.24 0.26 1673 Rittler (1966)
0.40 0.10 0.40 7353 Lauprecht et al. (1967)

0.34-0.38 0.26-0.34 0.33-0.38 1332 Rittler (1967)
0.33-0.35 0.30-0.40 0.37-0.43 3437 * »

In the light of the correlations between ultrasonic measures presented in
Table 2 it appears that the reliability of measurements was greater for the
Yorkshire breed than for the Landrace. This may have been due to the difference
in standard deviations of values of the two breeds in the present material.



519

Table 1. Means, standard deviations and heritability coefficients of ultrasonic measurements
and significance of differences between test litters (L) and sexes (S).

Landrace Yorkshire Total

Characteristic q q C+ G C G C+ G materia!

x x a xx a h 2 L S

Fat thickness, mm A 21 19 2.65 22 21 4.120.43 *

B 22 20 2.45 23 22 3.740.54 •

C 25 23 3.32 25 24 4.580.86 *•

x 22 20 2.45 23 22 4.000.66 • •

Lean thickness, mm A 38 39 3.16 36 37 2.450.48
B 38 39 2.45 37 38 2.650.55 *

C 38 40 3.46 39 40 3.870.53
x 38 39 2,65 37 38 2.650.83 »* *

Fat-lean ratio A 0.550.48 0.100.63 0.570.14 0.36 * *

of thicknesses B 0.580.51 0.080.63 0.580.12 0.55 * *

C 0.640.58 0.110.66 0.620.17 0.89 •»

X 0.590.53 0.09 0,64 0.590.14 0.62 * *

Fat area, cm 2 A 26.525.2 2.8229.1 27.14.86 0.37
B 29.727.5 2.9730.9 29.85.30 0.69 *

C 34.431.9 3.5734.5 33.95.67 1.17 ••• *

x 30.228.2 2.7231.5 30.3 4,88 0.82 ••

Lean area, cm 2 A 24.426.6 2.7624.8 26.52.42 0.11 ••

B 27.328.3 1.9727.3 29.12.24 0.04 ••

C 28.629.7 2.19 28,6 30.32.83 0.31 *

x 26.828.2 1.8526.9 28.62.02 0.25

Fat-lean ratio A 1.10 0,95 0.171.20 1.040.26 0.45 • ••

of areas B 1.100.98 0.141.14 1.030.22 0.76 ** •*

C 1.211.08 0.171.22 1.140.27 1.05 ••• »»

x 1.141.00 0.151.19 1.07 0,24 0.79 ** *•

C = Castrates
G = Gilts
A, B, C = Sites of measurements
x = Average A, B, C

The correlations demonstrate further the importance of measuring the fat and
the lean separately because the connection between them is relatively small.
Similar correlations between the thickness of fat and lean have been calculated
by the following workers:

Phenotypic Genetic n Author

-0.27 -0.43 1673 Kittler (1966)
- 0.18 —0.45 7353 Lauprecht et al. (1967)
-0.16 -0.04 0.10 3437 Rittler (1967)

- 0.02 0,24 412 Langlet et al. (1968)



Table 2. Correlation coefficients between dimensions measured from ultrasonic photographs.
Calculations are based on the mean values of three photographs (ABC).

Fat thickness Fat areaMaterial
lean thickness lean area

Landrace: castrates, I —0.19 —0.46*
gilts, I -0.42 -0.03
total breed, I —0.35* —0.37*
total breed, L —O.OB —0.17

Yorkshire: castrates, I —0.55* —o.69***
gilts, I -0.38 -0.38
total breed, I -0.48** -o.s2***
total breed, L —0.43 —0.63*

Total material, I —o.44*** —o.43***
Total material, L —0.37 —0.36

I = calculated from individual values
L = calculated from test litter means

Table 3 shows the properties that are the most important and in the present
connection the mos interesting characteristics obtained by carcass evaluation.
Lean -)- bone is the combined weight of lean and bone in the most valuable part
of the left half of the carcass. Fat is the corresponding weight of fat, and the
fat-lean ratio is the ratio between the two. The most valuable part of half
carcass consists of ham, loin, back, foreback and choulder. The area of m.l.dorsi
was measured from the carcass broken behind the last rib, i.e. at the site of
ultrasonic picture B. The side points are based on sensory evaluation. For the
posterior points (II) the amounts of fat and lean were estimated from the
cross-section surface of the carcass broken behind the last rib. The points (I—s)
increase with increasing relative amount of lean. The middle points (I) are
evaluated similarly from about the middle of the rib piece. The remarks

Table 3. Means, standard deviations and heritability coefficients of some characteristics de-
termined by carcass evaluation, and significance of differences between litters (L) and sexes (S).

Landrace Yorkshire Total

Characteristic q GC+ G C G CJ-G material

xx a xxa h 2 L S

Carcass weight, kg 65.365.1 2.1665.4 65.52.30
Lean + bone, kg 14.3314.65 0.8914.49 14.120.94 0.90 •• *

Fat, kg 4.934.49 0.635.12 4.930.78 0.33
Fat lean ratio 0.350.31 0.060.37 0.340.07 0.61 • •

M.l. dorsi area, cm2 29.331.5 3.22 27,2 30,6 3.690.63 •*

Back fat thickenss, mm: ... 28.927.3 3.8329.9 29.04.82 0.67 **

Side fat thickness, mm 25 22 3.87 25 23 5.570.86 • •

Side points I (middle) 2.62.9 0.442.6 3.10.66 0.64 ***

Side points II (posterior) ... 3.2 3.2 0.46 3.1 3.5 0.64 0.65 ***

520



521

Table 4. Correlation coefficients between some carcass evaluation characterisics.

o

.2 £

_ ’C 4J tr to C

0+«O"| 5 S ftrt V tj «-• '*■* c ”

U Co -r O MM rtlSc -m +1 -S —fl OO ® o "

o 5 « « t! k-isi !2'So ft xi ft ft 2 Sc B J t/)£ m
1 2 3 4 5 6 7

1 —.s2*** —.74*** .72*** —.4l*** —.s9*** .50***
2 -.67** - .96*** -.s9*** .67*** .80*** -.4B***
3 —.B3*** .97*** —.7o*** .66*** .84*** —.s4***
4 .72*** —.6o** -.7o*** _.39*** _.6o*** .53***
5 -.48* .80*** .75*** .35 - .71*** -.47***
6 —.62** .82*** .82*** —.s7** .79*** —.sB***
7 .58** -.51* -.56** .60** -.50* -.sB**

The correlations calculated from individual values are presented in the uppert par, and
these calculated from test litter means in the lower part of table.

in Table 1 on standard deviations and coefficients of heritability are valid
also here.

Table 4 presents the correlations between the characteristics included in
carcass evaluation. The coefficients in the upper half of the matrix were
calculated from the individual values and those in the lower part from the test
litter means. With a few exceptions the latter coefficients are higher than
those individually calculated.

The correlation coefficients presented in Table 5 show that the connection
between the individual measurements (A, B, C) and the area of ml.dorsi
actually measured on the carcass was usually poorer than the correlation between
the mean values |

A + +

0f the ultrasonic measures and the lean area.

Table 5. Correlation coefficients between ultrasonic and carcass measurements of lean area.

Material A B C x

Landrace: castrates, I 0.60** 0.58** o.6B*** o.Bo***
gilts. I 0.66** 0.51* 0.57*» o.76***
total breed, I o.6B*** o.sB*** o.66*** o.Bl***
total breed, L 0.77»* 0.52 0.70* o.BB***

Yorkshire: castrates, I 0.65** 0.06 o.69*** 0.64**
gilts, I 0 17 o.B3*** o.69*** o.73***
total breed, I o.so*** o,s9*** o.72*** o.7s***
total breed, L 0.56 0.46 0.72* 0.69*

Total material, I o.s7*** o.ss*** o.66*** o.74***
Total material, L 0.59** 0.38 0.63** 0.67**

I = calculated from individual values
L = calculated from test litter means



522

Table 6. Correlation coefficients between ultrasonic and carcass measurements of lean area
and amount of lean + bone in half-carcass.

Area measured from ultrasonic photograph Area measured
Material ,

ABC x from carcass

Landrace: castrates. I 0.55* 0.64** 0.67** o.79*** o.77***
gilts, I 0.56** 0.06 0.23 0.38 0.69*»*
total heed, I o.s7*** 0.37* o.so*** o.62*** o.73***
total heed, L 0.67* 0.09 0.43 0.57 0,63*

Yorkshire: castrates, I 0.61** 0.36 0.64** o.6B*** o.77***
gilts, I 0,35 0.26 0.37 0.42 0.52*
total breed, I 0.55** 0.35* o.ss*** o.6l*** o.69***
totalbreed, L .... 0.58 0.41 0.65* 0.65* 0.74*

Total material, I o.s4*** 0.33** o.sl*** o.sB*** o.72***
Total material, L 0.55» 0.22 0.51* 0.54* o.72***

I = calculated from individual values
L = calculated from test litter means

The same is true also of the correlations between the ultrasonic measurements
and the carcass lean + bone amounts, as is seen in Table 6. Reliability, however,
is not notably increased by using the mean of more than three pictures (An-
derson & Wahlström 1969 and Lauprecht et al. 1965). The coefficients in
Table 5 agree well with values reported in the literature, e.g. 0.74 (Price et
al. 1960); 0.70 (Stouffer et al. 1960); 0.66—0.85 (Blende 1969); 0.68 (An-
dersen et al. 1970); 0.57 (Horst 1971); 0.55—0.70 (Busk & Pedersen 1972).
From Table 6 it can be noted further that the mean of ultrasonically obtained
values very satisfactorily measures the amount of carcass lean -(- bone as
compared to measurement of the m.l.doresi area in the carcass.

In pig breeding aiming at improvement of the carcass value the main objec-
tives are an increased proportion of lean and a decreased proportion of fat in
the carcass, or in terms of the two goals combined; changing of the fat-lean
ratio in a direction more favourable for lean meat production. The carcass
values estimated from live animals are therefore compared below with the
amounts of fat and of lean + bone measured from the carcass and with their
ratio. These correlation coefficients are presented in Table 7: The coefficients
were calculated on the base of the means of the three pictures (A, B, C).
It will be noted from the figures in this table that fat thickness is on the
average better indicator of the carcass value of the slaughtered animal than fat
area, whereas the opposite is the case with lean. The betterness of lean area
reflected in the fat-lean ratio, which on the average estimates the quality of
the carcass slightly better when calculated from the area dimensions. When
the calculations were based on the test litter means the lean measurements
obtained from live animals of landrace were clearly better than the fat
measurements as indicators of the true amount of carcass lean -f- bone. The
situation is not as clear with respect to the amount of fat and the fat-lean



523

Table
7.
Some
correlations
of
ultrasonic
thickness
and

area
measurements
with
the

amounts
of
fat,
lean
+

bone
and
fat-lean

ratio
in
the
most
valuable

part
of
the

half-carcass.

Landrace

Yorkshire

Characteristic

Castrates
Gilts

Total
breed

Castrates
Gilts

Total
breed

Total
material

II
IL

I

I

I

L

IL

Fat
in
carcass

-
fat
thickness

0.61**
0.61**
o.66***
0.52

o.Bl***
o.Bl***
o.Bl***
o.93***
o.76***
o.B4***

-
lean
thickness

-0.19
-0.43
-0.34*
-0.36
-0.31
-0.38
-0.37*
-0.28

-o.39***
-0.39

-multiple
correlation
of

thicknesses

0.61*
0.64*
o.67***
0.61

o,B2***
o.Bl***
o.Bl***
o.94***
o.76***
o.B4***

-
fat-lean
ratio
of

thicknesses
.0.58**

o.6B***
o.66***
0.60

o.72***
o.79***
o.7s***
0.82**

o.73***
o.79***

-fat
area

0.53*
0.59**
o.6l***
0.28

o.Bo***
o.Bl***
o.Bl***
o.9s***
o.7s***
o.B2***

-lean
area

-o.73***
-0.41

-o.6s***
-0.68*
—o.6l**
-0.35
-0.47**
-0.60

—o.s2***
-0.53*

-
multiple

correlation
of

areas
0.76**
0.71**
o.76***
0.74

o.Bo***
o.Bl***
o.Bl***
o.9s***
o.7B***
o.B6***

-
fat-lean

ratio
of
areas
o.73***
o.72***
o.76***

0.73*

o.Bo***
o.77***
0.78*»*
o.93***
0.77*»*
o.B6***

Lean
+

bone
in
carcass

-
fat
thickness

-0.45*
-0.51*
-o.so***
-0.05
-o.79***
-0.60**
-o.7o***
-0.84**
-o.63***
-o.69***

-
lean
thickness

0.50*
0.51*
o.so***
0.35

0.54*
0.38

o.sl***
0.61

o.s3***
0.55*

-multiple
correlation
of

thicknesses

0.61*
0.61*
0.61**
0.35

o.Bo***
0.62*
o.73***
o.BB**

o.69***
0.76**

-
fat-lean

ratio
of
thicknesses

.

—o.s9**
—o.62**
—o.s9***
—0.21
—o.Bl***
—o.64**
—o.73***
—o.B9***
—o.6B***
—o.77***

-fat
area

-0.46*
-0.37
-0.47**
0.06

-o.72***
-0.59**
-o.6s***
-0.78**
-o.s9***
-0.65**

-lean
area

o.79***
0.38

o.62***
0.57

o.6B***
0.42

o.6l***
0.65*

o.sB***
0.54*

-
multiple

correlation
of

areas
0.79**
0.53

o.67***
0.57

0.77**
0.62**
o.72***
o.Bl*

o.69***
0.73»*

-
fat-lean

ratio
of
areas
-o.73***
-0.54*
-0.65*»*

-0.37
-o.74***
-0.63**
-o.72***
-0.82**
-o.69***
-o.73***

Fat-lean
ratio
in
carcass

-
fat
thickness

o.6B***
o.B3***
o.7o***
0,43

o.9l***
o.92***
o.B7***
o.97***
o.B2***
0.87»**

-
lean
thickness

-0.33
-o.6B***
-0.44**
-0.36
-0.43
-0.48*
-0.46**
-0.44

-o.4B***
-0.47*

-multiple
correlation
of

thicknesses

0.71**
o.9l***
o.73***
0.54

o.92***
o.93***
o.B7***
o.9s***
o.B3***
o.B9***

-
fat-lean

ratio
of
thicknesses
.

0.65**
0.83**»
o.73***
0.54

o.Bl***
o.BB***
o.B3***
o.92***
o.Bl***
o.B6***

-fat
area

0.64**
o.77***
o.66***
0.24

o.BB***
o.93***
o.Bs***
o.9s***
o.79***
o.B3***

-lean
area

-o.BB***
-0.54*
-o.74***
-0.74*

-o.72***
-0.44*

-o.sB***
-0.67*
-o.6o***
-0.58**

-
multiple
correlation
of

areas
.o.92***

o.92***
o.Bs***
0.83*

o.9o***
o.93***
o.B7***
o.9s***
o.B4***
o.BB***

-
fat-lean

ratio
of

areas

o.Bs***
o.B4***
o.Bs***
0.75*

o.B7***
o.B7***
o.Bs***
o.9s***
o.Bs***
o.B9***

I

=

Calculated
from
individual
values

L
=

Calculated
from
test
litter

means



524

ratio in carcass. In Yorkshires the fat measurements gave in all cases higher
correlations than the lean measurements. Although the measurements of lean
evaluate the slaughter value less satisfactorily than the measurements of fat,
it will be seen from the table that the best estimate on carcass value is obtained
when the fat and lean measurements are used in combination.

The correlation of the ultrasonic values of fat and lean combined, i.e. the
multiple correlation, with the actual carcass value is generally higher than the
simple correlation calculated from their ratio. In calculating the fat-lean ratio
the effect of the measurements on the carcass value were combined without
weighting. In multiple correlation, on the other hand, they have been weighted
with the weight coefficients leading to maximum correlation.

Generelly it has been found that fat measurements obtained ultrasonically
evaluate the amounts of fat and lean + bone with the same reliability as do fat
measurements from the carcass (Appendices A,B and C). In also the evaluation
of the carcass fat-lean ratio the ultrasonic fat measurements are equal to the
carcass fat measurements. On the other hand, the results of ultrasonic lean
measurements evaluate the carcass value less well than the respective
measurements made from the carcass. The measurement of lean by means
of ultrasonic is also less reliable in general than that of fat.

Fom Table 8 will be seen in which extent the ultrasonic measurements
give an estimate of the carcass value or general breeding value. As an estimate
of the percentage of lean in the carcass the ultrasonic measurements proved
in the present tests to be of quite equal value as in the studies of Bech Ander-
sen et al. (1970) and Busk & Pedersen (1972). The figures presented in the
table, which are correlation coefficients calculated by test litters within breeds,
show that the fat-lean ratio best describes the various characteristics of carcass
value and also the feed efficiency and the index of test results. This means
that the information obtained from the fat concerning not only the carcass
value but also other breeding values is supplemented and improved by the
information obtained from the lean. Nevertheless, the figures indicating
growth, one of which is calculated as the average daily growth and the other
as age at slaughter, are indicated by the fat-lean ratio no better than by fat
alone. On comparison of the determining measures in the table it can be said
that the lean measure emphasizes directly the main objective of breeding, i.e.
meat content of the carcass, and the test result index, which latter additionally
takes growth into consideration. Evaluation of the live animal solely on the
basis of fat reduces most effectively the amount of fat and improves growth
and feed efficiency, but it does not focus sufficient attention on increasing
the amount of lean. The best result is evidently obtained also here by using
the fat and the lean measurements together.

Table 9 presents the correlation matrix of varoiusly compiled indices and
feed consumption. The figures are within-breed correlations calculated by test
litters. Feed consumption has been included in the comparisons because it is
not included in the indices although it is an important factor in working towards
the maximum economic endresult. It is therefore interesting to know how
selection for breeding based on these indices takes feed consumption into consid-
eration.



Table 8. Means of ultrasonic measurements of live animals as estimates of certain litter test
results.

Independent variables
Dependent variables Fat area Lean area Fat-lean ratio

r ? r r* r r 2
Percent lean -0.51* 0.26 o.73*** 0.54 -o.77*** 0.59
Lean, gm/day -0.60** 0.36 0.48* 0.23 -o.73*** 0.54
Fat thickness x o.7s*** 0.57 -0.44» 0.19 o.Bl*** 0.66
Slaughter loss -0.29 0.09 -0.10 0.01 -0.17 0.03

Growth, gra/day -o.6B*** 0.46 0.29 0.09 -o.6B*** 0.46
Age at 88 kg 0.56** 0.31 -0.22 0.05 0.56** 0.31
Feed efficiency, feed

units/kg 0.66** 0.44 -0.39 0.16 o.72*** 0.52

Index -0.62** 0.38 0.59** 0.35 0.80**» 0.65

Table 9. Correlation matrix of feed efficiency and indices compiled in various ways.

Feed units/kg Indices
Variables

12 3 4

Feed efficiency, feed units/kg —0.87 —0.82 —0,82 0.87
Test result index 1 —0.87 0.78 0.86 —0.90
Old test result index using ultrasonic

measurements 2 —0.82 0.78 0.89 —0.96
Old test result index using carcass

measurements 3 —0.82 0.86 0.89 —0.93
Ranking numbers index using ultrasonic

measurements 4 0.87 —0.90 —0.96 —0.93

All correlations are highly significant

The indices are:

1. The test result index currently in use, based on the carcass dissection
and calculated from the following formula;

Lean + bone (kg) . Lean + bone (%)
I ' ' 1

Days in experiment station + age on arrival at station : 2

2 and 3. The old test result index calculated as the sum of standard
values of the included characteristics weighted with the heritability coefficient
and with the economic weight factor. The characteristics were thinness of fat,
crossectional area of m.l. dorsi, length of side, daily weight gain, and points for
ham. Index 3 was calculated from carcass measurements of the characteristics,
and index 2, except the ham points, was calculated on the base of the

6 525



526

ultrasonic photographs of fat and lean areas. It was also necessary to use
length of carcass because the length of live animals was not mesured.

4. The ranking numbers index used in the progeny testing of boars in field
conditions in the district of Salpausselkä Artificial Insemination Society
(Varo 1967) is based on the characteristics used in index 2. To compose the
index the ranking numbers of the test litters with respect to each characteristic
added up without weighting.

Table 9 shows that the correlations between all the indices are very close,
as also is their correlations with the feed consumption. The superiority of the
ranking numbers index over the old test result index can probably be accounted
for by the fact that the economic weight factors, which lower the heritability,
are not used in its calculation. The close correlation between the various indices
and feed consumption indicates that the present mode of selection favours
good feed utilizers. This finds its explanation both in the above presented
correlations between feed consumption and carcass value and in the very high
correlation between growth and feed consumption, which also in the present
material was 0.97.

Discussion

A very small material was availeble for the present investigation and
therefore the reliability of the heritability coefficients based on measurements
from ultrasonic photographs is doubtful. Nevertheless the inter-litter differences
on which the heritability evaluations are based were, with the exception of lean
areas, significant in most cases. Indeed, the poorest heritability coefficients
were obtained for cross-sectional areas of the lean. This may of course have
been due to difficulty in interpretation of the ultrasonic pictures, but it also
seems possible that certain interactions disturb the ranking order of the test
groups. An attempt will be made to study the subject more fully by analysing
the combined materials from the two years, which are in principle similar.

The fat-lean ratio measured from the ultrasonic picture gives a quite good
estimate of the carcass fat-lean ratio. In a critical consideration of the correla-
tions it is well to remember that the ultrasonically determined ratio expersses
the fat-lean ratio in the cross-sectional area of the back piece, while the ratio
obtained from the dissected carcass is calculated from the amounts of fat +

hide and lean -f- bone in the most valuable part of the half-carcass. This
difference may to some extent have a lowering effect on the correlation
coefficients.

In evaluating the practical value of ultrasonic photography in selection
for breeding the present results justify the conclusion that for the estimation
of carcass value the information from the live animal on the lean area will
usefully supplement the information obtained on fat. Phenotypic selection
based solely on fat will undoubtedly reduce the amount of fat most effectively
and will indirectly promote growth and feed utilization, but it does not focus
sufficient attention on lean meat. The possibility to estimate also the leanness
of the live animal has a direct effect on the main objective of breeding, i.e. an
increased amount of lean meat in the carcass. The positive correlation between



527

the lean content and the test result index warrants one to expect that selection
for leanness also favours rapid growth, which is one of the components of the
test result index. The ultrasonic photography of boars which currently is in
progress at the South Ostrobothnia (Etelä-Pohjanmaa) phenotype testing
station will in due time demonstrate whether or not these conclusions are
correct.

Summary
Ten test litters of Finnish Landrace and ten of Yorkshire breed pigs were

studied, each litter consisting of two castrates and two gilts. One week before
slaughter of the animals ultrasonic photographs were taken at three sites:
A, B and C. Photograph B was taken posterior to the last thoracic vertebra,
A at 10 cm anterior to B, and C at 10 cm posterior to B. From the cross-sectional
pictures were measured the thickness an areas of the lean and of the overlying
fat. The fat-lean rations were then calculated from these measurements.
The measurements so obtained from live animals were compared with the
carcass value determined from the dissected carcass. The comparisons were
made between individuals, using within-breed and within-sex correlation
coefficients, and between test litters using within-breed correlation coefficients.
The heritability coefficients of the ultraxsonic measurements were also estimated
on the basis of variation between the test litters.

In the present study it was found that the mean cross-sectional lean area
of three ultrasonic photographs has a closer correlation than any one of the
three individual areas to the cross-sectional lean area measured from the
carcass. The correlation of this mean value with the lean -f- bone amount
of the most valuable part of the half-carcass approached that of the cross-
sectional area of the lean measured from the carcass. The correlation of the
ultrasonic area measurements with the carcass amounts of fat and lean + bone
and with the fat-lean ratio calculated from these was closer tahn that of the
ultrasonic thickness measurements. The fat-lean ratio calculated from the
areas was also slightly better correlated to carcass quality than was the ratio
calculated from the thickness measurements. In other respects the fat thickness
measurements appeared to be better than the area measurements. Although
the measurements of lean correlated to the carcass value of the animal more
poorly than the fat measurements, the highest correlations were nevertheless
obtained by the combined use of the fat and lean measurements.

Acknowledgements

We are grateful to Mr. J. Partanen, Director of the Pig Husbandry Experiment Station,
and to the other staff members and personnel of the experimet station for the assistance ren-
dered by them, which made it possible for us to carry out this investigation.



528

REFERENCES

Andersen, B. 8., Pedersen, O. K., Busk, H., Lund, S. A. & Jensen, P. 1970. Ultraschall-
messungen bei Rinder und Schwein. Sonderdruck aus Die Fleischwirtschaft. 50,6.

Anderson, L. M. & Wahlström, R. C. 1969. Ultrasonic Prediction of Swine Carcass Composi-
tion. J. Anim. Sci. 28: 593—600.

Aulstad, D. 1969. In Vivo Estimation of Carcass Composition in Young Boars. II The Use
of Ultrasonic Measurements of Backfat Thickness. Acta Agr. Scand. 19: 189 196.

Blendl, H. M. 1968. Eigenleistungspriifung im Feld mittels Echolotverfahren beim Schweinin
Bayern. Z-kunde 40: 264 283.

Busk, H. & Pedersen, O. K. 1972. Mälning af arealet af m. longissimus dorsi og sides-
pa'ktykkelse pä levende svin ved hjaelp af scanningsapparat. Saertryk af Forsogslabora
toriets ärbog 1972: 102 107.

Ettala, E. 1973. Boars for breeding: a study of methods of evaluation used at testing stations.
J. Sci. Agr. Soc. Fin. 45, 1:17 119.

Glodek, P. 1964. t)ber ein unter praktischen Verhältnissen anwendbares Verfahren zur
Abschätzung des Fleischbildungsvermögens lebender Schweine mit Ultraschall. Z-
kunde 36:365-370.

Halkonen, M.-L. 1970. Elävän sian teuraslaadun arvioiminen. B. Sc. -thesis. University of
Helsinki, Department of Animal Breeding.

Hazel, L. N. & Kline, E. A. 1959. Ultrasonic Measurement of Fatness in Swine. J. Anim.
Sci. 18: 815-819.

Hofmann, F. & Peter, W. 1964. Die Anwendungsmoglichkeiten der Speckdickenmessung
mit Ultraschall bei der Selektion in der Fleischschweinezucht. Arch. Tierzucht7: 255 287.

Horst, P. 1971. Ultrasonic and X-Ray Measuring Techniques in Pig Performance Tests. I
The Ultrasonic Technique. Z. Tierz. Ziicht. biol. 88: 89 98.

—» 1971. Erste Untersuchungsergebnisse liber den Einsatz des »Vidosom-Schnittbildgerätcs
beim Schwein, Z-kunde 43: 208 218.

Langlet, J. F., Ernst, E. & Glaner, H. D. 1968. Die Möglichkeit einer Selektion beim Schwei-
nen mit Hilfe von Echolotmessungen unter den Verhältnissen der Praxis. Z. Tierz.
Ziicht. biol. 85:27 35.

Lauprecht, E. 1960. Bericht iiber Ergebnisse von Messungen mit dem Ultraschall-Echolot
an Schweinen, Z-kunde 32:441 449.

—» Kirsch, W., Rittler, A., Munzer, W. & Fewson, D. 1965. Untersuchungen iiber
die Abschätzung des Anteiles der fleisch- und fettreichen Teilstiicke aufgrund von
Echolotmessungen am lebenden Schwein. Z. Tierz. Ziicht. biol. 81: 89 99.

—» Schutzbar, W. von & Flock, D. 1967. Untersuchungen iiber die Heritabilität der
mit dem Ultraschall-Echolot an Jungebern ermittelten Speck- und Muskeltiefe. Z.
Tierz. Ziicht, biol. 84: 80 95.

* Walter, E. & Saathoff, F. 1960. Beitrag zur Messung der Querschnittsfläche des
langen Riickenmuskels (musculus longissimus dorsi) und der dariiberliegenden Fett-
schicht von lebenden Schweinen mit dem Ultraschall-Echolot. Z-kunde 32:450—464.

Price, J. F., Pearson, A, M. & Emerson, J. A. 1960. Measurement of the Cross-sectional
Area of the Loin Eye Muscle in Live Swine by Ultrasonic Reflections. J. Anim. Sci.
19: 786-789.

Rittler, A. 1967. Echolotmessungen als Hilfsmittel fiir die Zuchtwertschatzung beim Schwein.
2. Mitteilung: Weitere Untersuchungen iiber die genetische Fundierung der Echolot-
messungen an Jungebern. Z. Tierz. Ziicht. biol. 84: 179 187.

—» 1968. Echolotmessungen als Hilfsmittel fiir die Zuchtwertschatzung beim Schwein. 3.
Mittcilung: Abschätzung des Fleisch- und Fettanteiles auf Grund von Echolotmessungen
am lebenden Schwein. Z. Tierz. Ziicht. biol. 85: 78 89.

Schoen, P. 1964. Untersuchungen iiber die Eignung von Echolotmessungen am lebenden
Schwein zur Abschätzung des Anteiles der fleisch- und fettreichen Teilstiicke. Z-kunde
36: 206-216.

Schutzbar, W. von & Werhahn, E. 1972. Beitrag zur Schätzung des Anteils wertvoller Teil-
stiicke am Schlachtkörper mit Hilfe von Messungen mit dem Ultraschall-Echolot am
Riicken unde am Schinken lebender Schweine. Z-kunde 44: 323 330.



529

Stouffer, J. R., 1965. Objective technical methods for determining carcass value in live
animals with special emphasis on ultrasonics. E. A. A. P,, Nordwijk.

—» Wallentine, M. V., Wellington, G. H. & Diekmann, A. 1961. Development and
Application of Ultrasonic Methods for Measuring Fat Thickness and Rib-Eye-Area in
Cattle and Hogs. J. Anim. Sci. 20:759—767.

Uusisalmi, U. 1969. Vorläufige Ergebnisse iiber das Messen der Schlachteigenschaften beim
Schwein. J. Sci. Agr. Soc. Fin. 41:50—59.

Werhahn, E. 1968. Untersuchungen fiber die Eignung verschiedener Echolotmessverfahren
zur Abschätzung des Fleisch- und Fettanteiles in Schlachtkörper von Schweinen. Z.
Tierz. Ziicht. biol. 85:135 142.

Varo, M. 1967. Preliminary results of the »field testing» of A. I. boars. Acta Agr. Fenn. 109,
1: 105-109.

» 1971. Sikojen ultraäänikuvauksen tuloksia. Sika 2: 57.

SELOSTUS

Kokemuksia elävien sikojen teuraslaadun määrittämisestä
ultraäänikuvauksella.

Mikko Varo, Marjatta Perälä, Matti Ojala ja Hannu Varo
Kotieläinten jalostustieteen laitos, Helsingin Yliopisto, 00710 Helsinki.

Tutkimuksen aineistona oli 10 maatiaisrodun ja 10 yorkshirerodun kahdesta leikko- ja
kahdesta emakkoporsaasta muodostettua kantakoeryhmää. Eläimet kuvattiin viikkoa ennen
teurastusta kolmesta kohdasta: A, B ja C. B-kuva otettiin viimeisen rintanikaman takaa,
A-kuva 10 cm tämän kohdan etu- ja C-kuva takapuolelta. Saaduista kyljysselän poikkileik-
kauskuvista mitattiin kyljyslihaksen ja sen päällä olevan silavan paksuudet japinta-alat. Näis-
tä mitoista laskettiin edelleen silava/liha-suhteet. Elävistä eläimistä näin saatuja mittoja ver-
rattiin ruhojen leikkelyssä todettuun teurasarvoon. Vertailut tehtiin yksilöittäin rotujen ja
sukupuolen sisäisin ja koeryhmittäin rotujen sisäisin korrelaatiokertoimin. Myös ultraääni-
mittojen heritabiliteettikertoimet arvioitiin koeryhmien välisen muuntelun perusteella.

Tässä tutkimuksessa ilmeni, että kolmeen ultraäänikuvaan perustuva kyljyslihaksen poik-
kileikkauspinta-alojen keskiarvo oli korreloinut ruhosta todettuun kyljyslihaksen poikki-
leikkauspinta-alaan kiinteämmin kuin yksikään pinta-ala erikseen. Tämän keskiarvon yhteys
ruhon puolikkaan arvokkaimman osan liha + luumäärään oli lähes yhtä hyvä kuin ruhosta
todetun kyljyslihaksen poikkileikkauspinta-alan. Pinta-alamittojen yhteys leikkelyssä todet-
tuun silavan ja lihan + luun määrään sekä näistä laskettuun silava/liha-suhteeseen oli
kiinteämpi kuin paksuusmittojen. Myös pinta-aloista laskettu silava/liha-suhde oli hieman
voimakkaammassa korrelaatiossa ruhon laatuun kuin paksuusmitoista laskettu suhde. Muu-
ten näyttivät silavan paksuusmitat pinta-alamittoja paremmilta. Vaikka lihasmitat korreloi-
tuivat ruhon teurasarvoon huonommin kuin silavan mitat, saatiin kuitenkin suurimmat korre-
laatiot silava- ja lihasmittojen yhteiskäytöllä.



Appendix A. Some correlation coefficients between amount of lean + bone in cut carcass and ultrasonic and
carcass measurements.

ru . ■ .■ Ultrasonic Carcass . ~Characteristic n Author
measurements measurements

Lean + bone in carcass
fat thickness 54 —0.52 —0.73 —0.53 —0.74 Aulstad (1969) 4

108 -0.12 0.43 -0.13 0.55 Blendl (1968) 3

285 -0.20 0.26 -0.35 0.48 Ettala (1973) 2

205 -0.42 0.56 -0.38 0.60 Gerwig (1965 a)
191 -0.42 0.71 -0.49 0.60 Halkonen (1970) 2
56 -0.76 0.90 -0.85 Hazel & Kline (1959) 2

132 -0.27 -0.31 Lauprecht et ai. (1965) 2

101 -0.23 0.37 -0.35 0.54 Mennerich (1967)
1332 -0.61 0.63 -0.46 0.50 Rittler (1968) 2
1332 -0.83 0.861 -0.65 0.74 1 *

100 -0.65 0.69 -0.49 0.68 Schoen (1964)2

190 -0.15 0.43 -0.20 0.48 Sundgren (1969)
190 -0.55 0.57 -0.43 0.49 *

97 -0.18 0.41 -0.31 0.46 Uusisalmi (1969) 2

126 -0.58 0.76 -0.68 0.70 Werhahn (1968) 3

fat area 134 -0.51 0.63 -0.39 0.55 Blendl (1968) 3

111 -0.19 0.36 -0.33 0.35 Horst (1971)2

100 -0.49 -0.52 Schoen (1964) 2
126 -0.68 -0.64 Werhahn (1968) 3

lean thickness 132 0.36 0.42 Lauprecht et ai. (1965) 2

1332 0.31 - 0.35 - Rittler (1968) 2

1332 0.55 0.57 1 0.61 1 *

100 0.12 0.16 Schoen (1964)2

126 0.34 - Werhahn (1968) 3

lean area 78 0.50 0.53 0.57 0.60 Anderson & Wahlström (1969) 2

134 0.46 - 0.66 0.56 - 0.71 Blendl (1968) 3
102 0.41 Dickmann (1960)
111 0.17 - 0.42 0.50 - 0.52 Horst (1971)2
132 0.36 - 0.42 0.66 Lauprecht et ai. (1965) 2

101 0.16 0.69 Mennerich (1967)
91 0,53 Pfeiffer (1970)

100 0.32 0.52 Schoen (1964)2
126 0.46 0.70 Werhahn (1968) 3

fat-lean ratio of thicknesses 132 —0.55 —0.59 Lauprecht et ai. (1965) 2

1332 -0.59 0.62 - Rittler (1968) 2

1332 -0.81 0.901
- »

100 -0.65 -0.75 Schoen (1964) 2

126 -0.73 - Werhahn (1968) 3

fat-lean ratio of areas 111 —0.40 0.49 —0.53 —0.55 Horst (1971) 2

100 -0.65 -0.73 Schoen (1964) 2
126 -0.70 -0.80 Werhahn (1968) 3

Multiple correlation; a 132 0.59 Lauprecht et ai. (1965) 2
b 132 0.56 - 0.58 »

c 132 0.60 - »

d 105 0.72 Schutzbar & Werhahn (1972) 2

Multiple correlations;
1 = Genetic correlation a, 3 fat thickness measurements
2 = Lean + bone of most valuable part of carcass
3 = Lean -f bone of half carcass

b, 2 fat and 1 lean thickness measurements
c, 4 fat and 1 lean thickness measurements
d, 2 fat and 2 lean thickness measurements4 = Lean + bone of whole carcass

530



ppcndix B. Some correlation coefficients between amount of fat in cut carcass and ultrasonic and carcass
easurements.

Ultrasonic Carcass
laracteristic n Authormeasurements measuremets

it in carcass
fat thickness 54 0.51 0.74 Aulstad (1969) 4

108 0.14 - 0.59 0.28 - 0.60 Blendl (1968) 3

285 0.45 - 0.46 0.62 - 0.64 Ettala (1973) 2

191 0.39 - 0.68 0.49 0.59 Halkonen (1970)2

132 0.47 - 0.66 0.50 - 0.54 Lauprecht et ai. (1965)2

101 0,51 - 0.68 0.54 - 0.72 Mennerich (1967)
1332 0.70 - 0.73 0.57 0.59 Rittler (1968) 2

1332 0,74 - 0.78 1 0.63 - 0.81 1 »

100 0.64 - 0.69 0,72 - 0.73 Schoen (1964) 2

190 0.15 - 0.66 0.23 - 0.66 Sundgren (1969)
97 0.18 - 0.61 0.52 - 0.72 Uusisalmi (1969) 2

126 0.62 0,79 0.68 - 0.74 Werhahn (1968) 3

101 0,33 - 0.68 - Weniger et al. (1967)

-fat area 134 0.55 - 0.77 0.67 - 0.78 Blendl (1968) 3

100 0.46 0.57 Schoen (1964) 2

126 0.71 0.69 Werhahn (1968) 3

lean thickness 132 —0.29 —0,42 Lauprecht et ai. (1965)2

1332 -0.24 0.28 - Rittler(l96B) 2

1332 -0.26 0.37 1
- »

100 -0.07 -0.24 Schoen (1964) 2

126 -0.28 - Werhahn (1968) 3

-lean area 134 -0.35 0.64 -0.35 0.55 Blendl (1968) 3

132 —0.29 0.42 —0.50 Lauprecht et ai. (1965) 2

101 —0.34 —0.51 Mennerich (1967)
100 -0.28 -0.52 Schoen (1964) 2

126 -0,43 -0,66 Werhahn (1968) 3

fat-lean ratio of thicknesses 132 0.60 0.67 Lauprecht et ai. (1965) 2

1332 0.62 - 0.65 - Rittler (1968 b) 2

1332 0.62 - 0.78 1 - *

100 0.56 0.70 Schoen (1964) 2

126 0.72 - Werhahn (1968) 3

fat-lean ratio of areas 100 0.60 0.72 Schoen (1964) 2

126 0.71 0.81 Werhahn (1968) 3

multiple correlation; a 132 0.67 Lauprecht et ai. (1965) 2

b 132 0.64 - 0.69 - *

c 132 0,67 - »

= Genetic correlation Multiple correlations;
= Fat of most valuable part of carcass a, 3 fat thickness measurements
= Fat of half carcass b, 2 fat and 1 lean thickness measurements
= Fat of whole carcass c, 4 fat and 1 lean thickness measurements

531



532

Appendix C. Some correlation coefficients between fat-lean ratio in cut carcass and ultrasonic and carca
measurements.

Ultrasonic Carcass
Characteristic n ,

, Authormeasurements measurements

Fat-lean ratio in carcass
fat thickness 54 —0.49 0.72 1 Aulstad (1969) 4

108 0.14 - 0.56 0.31 - 0.64 Blendl (1968) 3

285 0.43 0.64 Ettala (1973) a

53 0.57 0.79 Hofmann & Peter (1964)
126 0.60 - 0.78 0.69 - 0.74 Werhahn (1968) 3

fat area 134 0.59 - 0,78 0.51 - 0.75 Blende (1968) 3

126 0.71 0.65 Werhahn (1968) 3

lean thickness 126 —0.27 0.28 *

lean area 134 —0.42 0.70 —0.46 0.65 Blende (1968) 3

126 -0.41 -0.70 Werhahn (1968) 3

fat-lean ratio of thicknesses 126 0.50 0.68 »

fat-lean ratio of areas 126 0.71 0.82 »

1 =Lean-fat ratio in carcass
2 = Fat-lean ratio of most valuable part of carcass
3 = Fat-lean ratio of half-carcass
4 =Fat-lean ratio of whole carcass