100 Scientific Proceedings, Royal Dublin Society. 



it likely that the gaps between two-year-old and three-year-old, three-year-old and 

 four-year-old, and even four-year-old and five-year-old yields should be as small as 

 these scales indicate. Indeed, something was happening among Ayrshire cattle at 

 the time the records which have been relied on were being taken which made 

 them inappropriate for determining how milk yield varies with age. 



As already stated, the late John Speirgot a small number of Ayrshire breeders 

 to begin keeping milk records in 19U3. One of the first results was that, when 

 the good cows came to be distinguished from the bad by an accurate method, 

 some which would have been retained before were now discarded, while others 

 which would have been discarded were now retained ; and the yields in milk- 

 recording herds began to rise. It would be diflticult to say whetlier the breeders 

 were readier to discard younger or older cows which had been shown to be of 

 poor capacity, and thus whether the average yields of older and younger cows 

 were affected equally. 



Another result was that breeders began to select sires by their dams' yields 

 rather than by the old and generally worthless tests. This did not affect the 

 yields at once, however, for, siuce sires cannot be used till they are yearlings and 

 must be four years old when their eldest daughters are two, it could scarcely have 

 affected the breed yieids at all before 1906 or 1907. And, when it did have effect, 

 it must have affected only two-year-olds the first year, two- and three-year-olds 

 the second year, two-, three-, and four-year-olds the third year, and so on. Thus, 

 though the average yields in milk-recording herds were increased by both the new 

 way of discarding cows and the new way of selecting sires, the increase was not 

 distributed equally over all ages, and the records after 1907 or so cannot be used 

 to say how milk yield varies with age. All they can say is how yields varied 

 with age in the year or years in which they were taken. 



The shifting of the relative positions of the average yields at different ages is 

 indicated in the following diagram, in which the averages for two- to eight-year- 

 old cows at four different periods are plotted out and traced together. The 

 averages for each period have been multiplied by the figures which bring their 

 eight-year-old averages to 100. The periods chosen are 1903 to 1907, 1908 and 

 1909, 1913 and 1920. The basal figures for the first period were extracted by 

 Mr. Speir hiraself,^ who used them to find how yield varied with age ; those for the 

 second are Dr. Pearl and Mr. Miner's ■; those for the third and fourth were extracted 

 shortly after the reports for 1913 and 1920 were published. 



Thus, since scales based upon the Ayrshire records cannot represent the normal 

 rise in milk yield fairly — even Mr. Speir's was constructed too late — the scales 

 based upon the London Dairy Show and Lord Eayieigh's cows, or a combination of 

 these, must stand in the meantime. Since Mr. Gavin's scale is probably too low 

 for three-year-olds, the combined scale would indicate that the yields of three-, 

 four-, five-, six-, and seven-year-old cows should stand to their yields at eight 

 years old as approximately 67, 80, 90, 95, and 98 to 100. As yet, little can be 

 said about two-year-old yields. The early Ayrshire records give no clear help, 

 because, while the average ages are approximately 3-J- years for three-year-olds, 4J 

 for four-year-olds, and so on, the average age of two-year-olds is undoubtedly more 

 than 2^ years. If the combined figures got from the London Dairy Show and 

 Lord Rayleigh's cows were plotted to scale and produced to the left, the produced 

 line would indicate that the yields of two-year-old cows whose ages average about 

 2| years should be to their yields at eight years old as about 50 to 100. 



^ Report on Milk Records for Season 1908, p. 23. 



