THE RAM CIRCLE AS A PROGENY TESTING METHOD

,

Finnish statistics for 1968 show that there were only 12.7 ewes per farm on an average.The progeny group that is necessary for progeny testing can consequently hardly be obtained for the average Finnish flock for more than one ram, and thus no selection can be made.Conversely, no comparisons can be made directly between the rams on the various farms which have produced progeny, for the effect of the flock cannot be distin- guished from the effect of the sire.
To make field evaluations possible in the conditions outlined above, a system of ram circle has been in use in Norway since 1956.Gjedrem (1969) states that the method is employed to ensure that each of the rams will have progeny on the various farms, in order that the effects of the flocks on the progeny groups might be distributed as evenly as possible.Random selection of mate is attempted.In practice the sheep breeders within a specified area use the same rams for servicing.The servicing is started simultaneously on the farms, and the rams are circulated daily during the breeding season from farm to farm.Thus in every flock the rams will service those ewes that come into heat on the day concerned and the lambs will be born close to one another in time.
The Sheep Breeding Society began to study the possibilities of applying the ram circle system in Finland by setting up a ram circle in 1968-1969.Analyses were con- currently made of the variation and hereditability of meat production features of sheep.
This experiment was set up on four farms in the vicinity of Pieksämäki.The experiment included four rams which were circulated from one farm to another at intervals of five days.On three of the farms the servicing began on October 18th 1968.On one of these one of the rams had already been in the flock for a few days.The rams were trans- ferred three times, and after the third transfer they remained with the same flock to the end of the breeding season.The lambs born were weighed not only at the ordinary testing but also at the age of c. 120 days.Likewise, 40 lambs were selected at random among the male lambs and were slaughtered at the slaughterhouse of Osuusteurastamo Karjaportti at Mikkeli.The live- weight of the selected lambs was also taken about one week before slaughter.The car- casses were cut up acording to commercial practice.The left side of the carcass was dis- sected in order to establish the proportions of meat and bone.Table 1 shows the means and standard deviations of the characteristics.Table 2 shows the size of the flocks on the various participating farms and the frequ- ency distributions of lambs born.It appears that the rams produced progeny on the farms somewhat unevenly, for only one of them had progeny on every farm.This factor makes it difficult to interpret the results and may have made it impossible reliably to separate the variation between rams from that existing between flocks.
The results were analysed at the National Computer Centre by means of a covariance analysis programme prepared for Elliot 503 and based on the method of least squares presented by Harvey (1960).
In order to ascertain the significance of factors affecting the weights taken at various ages and the respective growth rate, use was made of the model Yijki = a+Bi + s u + ak + bA ijki + dB ijki + fC ijki + hD ijki + e ijki where y = the estimate of the respective characteristic of the lamb born on the i th farm as progeny of the j th ram of k th sex, a = the theoretical mean of the population, assuming that the observations are distributed evenly on the various levels of the classes, and that -®ijkl -^ijki g. = effect of i ,h farm, s., = effect of j th sire with in i th farm, b, d, f, h = partial regressions of the y jjkl on the equivalent regression variables, e..., = random error, distribution of which is assumed to be normal. ijkl

Variation in growth characteristics
In order to ascertain the effect of the dam, the dam's age and weight as regression vari- ables were included in all analyses.It was found that initially the progeny of the older ewes grew faster than did the progeny of the younger ewes, but that later they grew more slowly.A similar result has been arrived at among others by Hazel and Terril    (1945).According to them, the high initial growth rate in the progeny of older ewes is due to the fact that the milk production capacity of ewes is at its maximum at the age of 5-6 years.The weight of the dam had no significant effect on the growth of the progeny.This is evident from Table 3.
Table 3 also shows that the effect of the type of rearing on the growth of the progeny is very significant.The effect, however, did weaken in the same manner as did that of the age of the dam.According to Rendel (1954) the rapid initial growth of lambs born in small litters is due to the fact that at this stage they get more milk than do lambs born in large litters.Later, when they all begin to eat fodder, the differences disappear and the lambs of large litters tend to catch up in weight.Table 3 also shows that the regression between the type of rearing and the growth during the period from 120-150 days was no longer significant.
Nor does the time of lambing seem to have any effect on the growth rate.In contrast, the differences between sexes were clear during all the periods of growth.The farm of birth likewise had a definite effect on the growth of the young lambs, but no longer on the growth occurring during days 120-150.It seems that if the rams are evaluated on the basis of living progeny, the effects of at least the type of rearing and the sex of the lamb and, among younger lambs, of the dam's age, must be equalized before progeny testing.Account should then also be taken of the effect of the farm of birth.

Factors affecting variation in slaughter characteristics
When the factors affecting the variation in the carcass characteristics of 40 slaughtered male lambs were analysed in accordance with the above model, it was found that the variation between farms in the various characteristics was generally smaller than the variation in error.Also the variation between sires within farms was greater than was the variation between farms.Consequently, a final model explaining the slaughter char- acteristics was drawn up without separation of the component of variation between farms.The effect of sex, of course, was not separated, as the slaughtered lambs were all males.
The effect of the characteristics selected as regression variables was not significant, with the exeption of the time of birth.This connection too must be regarded as being chiefly due to the age, as all the male lambs were slaughtered on the same day, and the time of birth was expressed by the number of days from the turn of the year to the date of birth.Calculating in accordance with the model, the weight of the half-carcass of a lamb that was one day older would be 45 g greater ifother regression variables are assumed to remain unchanged.Between sires the differences were significant in four characterisrics only when a total of 20 characteristics were examined, but is should be noted that only 4 sires were evaluated.
Table 4 shows that the parameters for the various characteristics calculated for the sires regulary vary in a single direction from the theoretical mean of the respective char- acteristic.At a cautious estimate, the progeny of the 111 sire seem to be the poorest and those of sire II the next poorest, although the differences were generally not statistically significant.
Table 4. Estimates of the least-squares constants for the sires from the most important characteristics of the carcass; and the significance of differences.The observations were made on the left half of the carcass.

Estimates of heritabilities
The heritabilities of liveweights and growths were calculated from paternal half-sibs by means of covariance analysis, from values corrected in accordance with the above model.The results are shown in Table 1.Gjedrem (1967) has estimated the herit- ability of weaning weight in a Norwegian lamb material to be 0.26, Varo (1968) has calculated the heritability of liveweight as being 0.37, and Maijala (1967) gives the heritability of weaning weight as 0.25.All these correspond closely with the heritability of the weight corrected to the age of 150 days here estimated.
The heritabilities of slaughter characteristics were likewise calculated on the basis of half-sibs correlation.It should be noted that the share of variance between sires in this case was not calculated between farms, because the variation between farms was smaller than the variation due to error.
It can be seen from Table 1 that the heritability had highly differing values depending on the characteristic concerned.
The estimates of the heritability of the half carcass at 0.40 and that of the total weight ofmeat at 0.41 is probably of greater interest.Botkin et al. (1969) calculated the respective values is English races to be 0.33 and 0.39, and Varo (1968) calculated those in a Finnish material to be 0.37 and 0.34.Thus, at a certain age the weight of the carcass and the meat are at least as clearly heritable as is the growth tendency in general.As the weighing of the carcass and the assessment of the fullness cannot be accurately performed before the slaughterhouse is reached, at least a number of the progeny produced in the ram circle have to be slaughtered in order to make the evaluation certain.

Discussion
One of the purposes of a ram circle is to have the progeny of different rams distributed into circumstances in order to equalize the effect of the environment, i.e. chiefly the effect of the rearing farm.As was shown in Table 2, however, this was not completely successful on this occasion.One reason may be that on one of the farms the ram had been with the flock before October 18th, the date on which the experiment was initiated.And the success of the method seems to require simultaneous commencement of servicing on all the farms in the circle.Also, the rams should be circulated more frequently than at inter- vals of five days, for almost all the ewes may come on heat in a small flock within a period of that length.
The variation between farms proved to be significant for growths and liveweights other than those occuring during the age of 120-150 days.The reason for the difference may lie in differences in feeding, or in an uneven distribution of the progeny groups of the sires by farms, because variation was also found between the sires.In any event, it is necessary to attempt to implement the ram circle fully on every farm, so that the effect offarm on progeny group will be levelled out.
The effect offarm of birth on the characteristics of the lambs decreased with increasing age.There was no very clear connection between the carcass characteristics and the investigated environmental factors.In view of progeny evaluation the most important factor requiring any correction may be the age of the animal, for its effect was sinificant on the weight of the meat of the leg, the weight of the loin and the percentage of meat on the leg when the relationship was measured by means ofpartial regression.Yet 40 slaughtered male lambs must be regarded as being far too small a sample for the evaluation of all factors that require correction.
In the analysis of slaughter results it was found that the variation between farms was generally smaller than the variation due to error.This suggests that the breeding value of the slaughter characteristics of sires can be evaluated by means of a ram circle directly within the circle without the necessity of giving any attention to the effect of the farms.
Differences between sires were generally not significant where growth characteristics were concerned, but became increasily evident with the increasing age of the lambs and concurrent decrease in the variation between farms.Differences between sires in slaughter characteristics at a 5 % risk were significant in four of the characteristics investigated.

Table 1 .
Means, standard deviations, variation percentages and heritabilities of characteristics.
N.B.All slaughter characteristics measured from left half of carcass.

Table 2 .
Numbers of ewes and the size of the progeny groups of rams and ewes per farm and in total.

Table 3 .
Results of covariance analysis of daily growths.