whether the advance is to be considered as a cumulative effect of the selection of fluctuations or the gradual purification by the selection, of mutants which occurred during the selection or possibly even before the first selections were made. The purification of a type even when the character concerned is easily observable is known to require a number of years unless both parents are carefully followed. Whether these qualities will segregate as unit characters after hybridization has not been determined so far as the writer is informed.
Very many cases of increases obtained in quantitative characters could be cited, but the majority of the experiments were undertaken primarily to obtain practical results, and whether such apparently new characters would stand the test of unit characters is doubtful.
The improvement of the sugar beet by selection forms a typical and instructive case of this kind. The careful selection of the sugar beet was started over sixty years ago by Louis Vilmorin, at which time a range of variation in sugar content of from 5 per cent. to 21 per cent. was known to exist. Since that time the industry has grown extensively until hundreds of thousands of beets are examined annually and the richest in sugar content selected for seed production. The process of selecting the beets richest in sugar content for mothers has now been continued for sixty years and is practised extensively every year, and yet there is no evidence that the maximum sugar content has been increased, and it is certain that the character of richness in sugar content has not been rendered permanently heritable, as sugar beet growers well know that their success depends upon the continuance of the selection. Here it is certain that no distinct unit character has been added by the continuous selection.
The strongest evidence as to the method of origin of new characters is derived naturally from our knowledge of known cases of the origin of such typical new characters. When we view the evidence critically, I think it must be admitted that in practically all, if not all, of the cases of new characters appearing, they have come into existence suddenly. The cut-leaved Celedonium, the cupid sweet pea, Bursa heegeri, the Otter sheep, the muley cow, are illustrations familiar to all and doubtless each of us could add several such illustrations from our own knowledge. Such new characters appearing suddenly are heritable and maintain themselves as unit characters in hybridization. We can not but admit that the evidence of these known cases counts against the origin of characters by gradual cumulative selection.
In summarizing this part of our discussion, we can only state that at present it appears that far the greatest weight of evidence is opposed to the origin of a new unit character through the cumulative action of selection.
Are we, then, to conclude that the practise of breeders in continually selecting from the best for propagation is useless, and must we advise practical breeders to discontinue their selection? How can we do this in the light of the success of the sugar beet breeders? Have not Sea Island cotton growers increased and maintained the length and fineness of their staple by continuous selection? Have not corn growers maintained high productiveness of different strains by selection? Are not the Jersey and the Holstein maintained at a high degree of efficiency by selection? Has not the speed of our trotting and pacing horses been increased and maintained at a high rate by the most careful selection? To one familiar with the history of agriculture and breeding these questions arise fast and are
