The progress of our upbreeding program depends heavily on a number of factors. Among these are conception rate, birthing (or twinning) rate, gender ratio, and death rate. In addition, the progress depends sharply on whether a given generation of lambs can be bred in their first year or cannot be bred until their second year.
Conception rate is the probability that a ewe will become pregnant. Rams are very successful in getting ewes pregnant, but the success rate with artificial insemination (AI) is considerably lower. These rates vary with the breed of the sheep (both ewe and ram), but a typical value is 70%. That is, if 100 ewes are artificially inseminated, on average, about 70 will become pregnant.
The second factor is the birthing or twinning rate. That is, how many lambs can we expect the pregnant ewes to give birth to. This varies greatly with the breed of the ewe. For example, Finn sheep almost always have multiple lambs frequently exceeding two in number. For Shetland sheep, the birthing rate is about 1.5; on average, each pregnant ewe will give birth to 1.5 lambs. The birthing rate for Ouessant sheep is close to 1; twins are rare.
The third factor is the gender ratio. Natural insemination produces close to a 50:50 ratio of ewe lambs and ram lambs. However, with AI, the ratio is closer to 30:70 ewe lambs to ram lambs. The reason for this is not fully understood, but is probably because the process of collecting, storing, and administering the semen adversely affects the sperm producing female lambs more that it affects the sperm producing ram lambs.
The fourth factor is the rate at which the pregnant ewes fail to produce viable lambs. There are many causes. Fortunately, the rate is not too high amounting to a few percent. We will ignore this factor.
So what do these numbers mean? A simple calculation for the case of cross-breeding Shetland ewes with Ouessant rams (using AI) will help. If we start with 40 Shetland ewes (as we did in 2008) and inseminate them with Ouessant semen, we would expect 70% or 28 to get pregnant. We would expect those 28 ewes to give birth to 1.5 lambs each or 42 lambs of which 30% or 12 would be female. Since only the female lambs can be used for the next generation, our usable flock is now down to 12. If we then repeat this process (optimistically assuming 1.5 lambs per pregnant ewe), the next generation would only have 3.6 ewe lambs and the generation after that would be down to just one (and a fraction) ewe lamb. We solve this problem by breeding each generation of lambs (including the original Shetland ewes) for multiple (two or more) years.
The final factor — the age at which a lamb can be bred — is the biggest factor in the rate of progress. If lambs cannot be bred until their second year, the rate of progress will be slowed by a factor of two.
While we have found that some lambs are ready to breed in their first year, the difficulty of reliably determining which ones are ready coupled with the expense of performing AI makes it difficult to take advantage of these cases. Our present plan is to expose the new lambs to rams of the same generation in our flock. If the ewe lambs are ready, the rams will know and will do their job. Since the rams available will never be of higher percentage than the corresponding ewe lambs, this process will not increase the Ouessant percentage, but it will increase the numbers.