Dr Malcolm B Willis

We have been scoring GSD since mid 1978 so the scheme has been operating for almost 22 years though most other breeds came in largely in late 1983. For a long time the GSD ranked first in numbers scored but it has now been ousted by the Labrador and is going further behind each week as more Labradors are being scored than GSD at a rate of about 2 of the former to each GSD.

Thus far we have scored 30,451 GSD up to the last reading that I have (16/08/00) whereas we have assessed 34,100 Labradors. The GSD covers the full range from 0 to 106 with an arithmetic mean of 18.65 which makes it 13th worst breed (among 75 breeds with at least 40 scored). Otterhounds with 114 dogs at 42.96 are the worst in terms of high score. Progress is minimal but that is not a reflection of the merits of the scheme but simply illustrates that selection is not being practiced to any degree. In countries where schemes involve compulsion i.e. that one cannot breed from (and register) specific levels of hip status (as in Scandinavia) or where there is peer and club pressure such that untested dogs cannot be shown (as with the SV) there is evidence of progress. However in the UK and USA where breeders are free to do as they please progress is not usually seen and there is minimal advance. Nevertheless some 70+ % of GSD score 15 or better and thus would experience no problem with their hips.  

Kennels which stay out of the scheme do not affect it but among scoring kennels selective scoring has an adverse effect. This applies to early (puppy) screening where poor ones are never scored as well as failure to submit specific animals because they look poor on X-ray. Some argue that it is unimportant as long as these dogs are not bred from but if poor dogs are deliberately not submitted the net result is that specific sires (dams) may look better than they are and this will affect German breeding values as much as ours. In Germany up to a third of dogs registered are hip graded but in Britain we are operating at <10% in many breeds though BMD have scored 24% of all registrations since the breed started in 1969 and currently are closer to 30%. 

In Newfoundlands (2,574 scored) there is evidence of consistent progress of about 0.73 points per year. This may seem small but it means that the breed has declined from about 35 to 21 in 20 years. The rolling mean for Newfoundlands is 27.33 but that is a cumulative value as is the GSD mean of 18.65. Means by year of birth show the Newfoundland progress.  

If we regard HD as 40% heritable, which it is in GSD and Newfoundlands among others, and we had a breed mean of 20 and used parental stock that averaged 9 then we would be using parents that were 11 points (20-9) better than breed mean. Some 40% of this 11 would be transmitted (i.e. 4.4 points) so the breed mean in the next generation would be 20-4.4 or 15.6. However a generation is about 5 years in dogs so that progress per year would be 4.4/5 or 0.88 points which is only slightly higher than Newfoundland actual progress but better than GSD. Nevertheless it can be seen that annual progress is not huge and cannot be unless rigid selection criteria are enforced (and I am not advocating that).

In GSD we have a situation in which many breeders use scores higher than breed mean and remember that the mean at which 50% of the breed is obtained is about 4 points lower than the 18.65 or closer to 14 than 19. In addition sires that are not good producers or average producers are often widely used. I am not disputing the fact that features other than hips must be considered and that there are occasions when dogs that are not the best hip animals may have to be used. Nevertheless, HD will only improve by selection for it. Failure to select on a breed scale is a feature of the breeders not the scheme. Most breeders blame the tools but it is their use of the tools which is at fault. 

The SV has shown good progress with their scheme but in 1998 I was talking to a Dutch breed warden who pointed out that most Dutch GSD breeders sent their hip plates to Germany because the chance of a Normal (A) was higher than in Holland where A's were handed out very sparingly. The number of imports with "a" stamps which have had very high UK scores is not legion but that they exist at all is cause for alarm. The new breeding techniques will help but if selective submission is used breeding values will obviously be affected. Moreover breeding values with low reliabilities (<80%) are not very meaningful. 

The sire tables give the sires (* indicates scored) and the number of progeny scored along with the number of dams these progeny were from. The best and worst progeny score are followed by the mathematical average of the progeny. The percentage of progeny in each score group is given and is rounded to the nearest whole number so may not total exactly 100. 

Select sires with plenty of progeny for reliability and with low progeny means and with high percentages in the 0-5, 6-10 and even 0-20 categories and with low percentages above 20. Deceased sires are not listed if I know that they are dead. 

M B Willis, 2000


Please click on the following links to see the records