Last update: September 2, 2011 02:29:40 PM E-mail Print




Animal Production Research

Grootfontein College of Agriculture

Private Bag X529

Middelburg Cape 5900


1SA Fleece testing Centre, Middelburg, 5900

2Merino Stud Breeders’ Society of SA, Graaff-Reinet, 6280



Type classification systems, which refer to systems of subjective scoring, were developed by breed societies in the hope that they might serve as phenotypic and/or genetic indicators of production (White, 1973). Owing to the categorical nature of older type classification systems, special statistical methods had to be used to analyse the data (Snell, 1964).

Since 1977 linear type scoring of conformation became popular in dairy cattle (Thompson, et al., 1981; Meyer, et al., 1987), and even goats (Dickerson, et al., 1984).  Some of the advantages of linear type scoring were mentioned by Thompson, et al., (1983) and Lucas, et al., (1984). One of the advantages of linear type scoring is the fact that traits are scored individually rather than in combination, which increases the score's accuracy. However, it appears that some of the traits in dairy breeds were difficult to score consistently (Meyer, et al., 1987). Another main advantage of linear type scoring is that traits are scored over a large range (0-50, 0-100), which makes the probability of a normal distribution more likely. A normal distribution is favoured because the data can be analysed with normal statistical procedures. Sire evaluation for type scored traits is also more feasible.

Many type classification systems for Merino sheep have been developed and used over the years (Grootfontein College of Agriculture, 1953). The shortcomings of these scoring systems were discussed by Roux (1961). Body and wool type traits played an important role in selection systems of many Merino stud breeders (Jordaan, 1966). This is further demonstrated by the fact that only 19% of the 785 Merino breeders associated with the Merino Stud Breeders' Society are members of the National Woolled Sheep Performance and Progeny Testing Scheme (Watermeyer, 1986). It is of importance to determine the heritability and genetic/phenotypic correlations of the various type traits and their correlation with production traits. For this purpose a suitable score card has to be developed. In the light of the storage of ram semen of superior rams by the Merino Stud Breeders' Society, it is of importance to evaluate type traits of sires accurately. This paper evaluates a proposed linear type scorecard for the Merino sheep. This scorecard was tested on a number of rams and ewes of the Grootfontein Agricultural College and results are given.


Materials and Methods

A linear score card of 23 body traits and 9 wool traits was drawn up (Table 1). Most of the traits were scored from 0-100 with 50 as the average. However, with some of these traits an intermediate value was regarded as ideal and to maintain linearity a score of 50 was in these cases regarded as the ideal. Deviations from the ideal were then scored from 0-50 or from 50-100. For example sheep with excessive wool yolk were scored from 50-100 and those tending towards dry wool from 0-50.


This scorecard was tested 118 two-tooth rams and 120 two-tooth ewes from the Grootfontein Merino Stud. Two judges, with extensive knowledge of judging Merinos, scored the sheep. The body traits were scored after the sheep had been shorn. On average it took about 3 minutes to score the body traits and 2 minutes for the wool traits.


Results and Discussion

The average, standard deviation of the mean, and minimum and maximum of each trait are presented in Table 2.


All the traits scored were above average. Hoof colour had the highest score, which implied that the hoofs were free of any colour. Front legs had the lowest average score and also the lowest minimum value. The above average scores of most of the traits could be an indication that this specific stud may be above average for the scored traits in comparison to the rest of the country's Merino flocks. The high average could also be due to the reluctance or the judges to score a specific trait below 30 points. This is clear from the fact that the minimum scores for most of the traits were 30. A sheep with a score of 30 or below was considered to be a cull. The scoring range of 0-100 was therefore not used properly by the judges. It was further clear that the judges were not able to make use of the fineness of the scale because they scored most of the animals as either 40, 50, 60 etc.

As mentioned earlier, linear type scoring makes the probability of a normal distribution more likely. With statistical tests for normality it was found that 14 of the 32 traits were normally distributed. Some of the traits such as mouth, jowels and hoof colour had an extremely skew distribution. This was due to the fact that in most cases there was an absence of mouth and jowel faults and colour on the hoofs. It appears therefore that most of the traits can be analysed by normal statistical procedures.

An analysis of variance was done to obtain an indication of the factors influencing the different linear traits. The effects of sire of the lamb, sex, birth status and age of the dam on the different traits were investigated. The F values found for the different traits are presented in Table 3.


Birth status (born as a single or a twin) and age of dam (born from a maiden or mature ewe) had no significant effect on the scores for the different traits, except the traits chest and hind leg pasterns. On the other hand, sex and sire had a significant (P < 0,01) effect on the scores of most of the traits (Table 3). It was difficult to determine whether the significant effect of sex on most of the traits was due to sex differences or to the inconsistency of scoring by the judges (ewes were scored 6 weeks later than the rams). Of the 32 traits scored, the sire had a significant (P< 0,01) effect on 14 traits. This is an indication of breeding value differences between rams for different traits. It also indicates that some of the traits may have a higher heritability than others.



From these results it was clear that the proposed linear score card can be used to quantify body and wool type traits in Merino sheep. It appears that this scorecard can be used for the determination of breeding values and for variance and covariance analysis. Although the methods are quite time-consuming, it can be improved by omitting some of the less important traits. In general, the judges did not make use of the whole range of the scorecard, suggesting further investigation of the optimal range required (viz. 0-50). The consistency of the judges in scoring the different traits also needs further investigation.

According to the above-mentioned, this specific linear trait scorecard should be changed to eliminate some of the disadvantages mentioned. Firstly, the number of traits to be scored can be reduced to 13 body and 7 wool traits. However this will need the comment of the Merino Stud Breeders' Society. Secondly, the scoring range may be reduced to 0-50 which would probably force the judges to score the below average animals correctly. The proposed range is as follows:


  0 - very poor

12 - below average

25 - average

37 - above average

50 – excellent


This scorecard should however be tested extensively before it is used in practice. It is suggested that the proposed score card be used in the Grootfontein stud to determine the correlations between the various type and production traits.



The authors wish to thank Mr Louw Zandberg, BKB, Colesberg, for his assistance in scoring the sheep during this trial.



DICKERSON, F.N. & NORMAN, H.D., 1984. Linear type appraisal for dairy goats. International Goat & Sheep Research. 2,189 - 192.

JORDAAN, T.B., 1966. Merinoteling. In: Die skaap en sy vag. Red. J .C. Swart. Nas. Boekhandel Beperk. Kaapstad.

LUCAS, J.L., PEARSON, R.E., VINSON, W.E. & JOHNSON, L.P., 1984. Experimental linear descriptive type classification. Journal of Dairy Science. 67,1767 - 1775.

MEYER, KAREN., BROTHERSTONE, SUSAN & HILL, W.G., 1987. Inheritance of linear type traits in dairy cattle and correlations with milk production. Animal Production. 44,1-10.

ROUX, C.Z., 1961. Oorwegings by die opstel en uitvoer van geskikte teelplanne vir wolskape. M.Sc.-tesis. Universiteit van Stellenbosch.

SNELL, E.J., 1964. A scaling procedure for ordered categorical data. Biometrics 20, 592.

THOMPSON, J.R., FREEMAN, A.E., WILSON, D.J., CHOPIN, C.A. & BERGER, P.J., 1981. Evaluation of a linear type program in Holsteins. Journal Dairy Science., 64, 1610 - 1617.

THOMPSON, J.R., LEE, K.L., FREEMAN, A.E. & JOHNSON, L.P.,1983. Evaluation of a linearized type appraisal system for Holstein cattle. Journal Dairy Science. 66, 325 - 331.

WHITE, J.M., 1973. Genetic parameters of conformational and managemental traits. Journal of Dairy Science. 57,1267- 1281.

WATERMEYER, A.J., 1986. Die vordering en ontwikkeling van die Nasionale Wolskaapprestasie- en nageslagtoetsskema. Karoo Agric 3 (8), 32.



Karoo Agric 3 (9), 1-4