The common cockscomb or _Celosia cristata_, one of the oldest and most widely cultivated fasciated varieties may be used to illustrate the first point. In beds it is often to be seen in quite uniform lots of large and beautiful crests, but this uniformity is only secured by careful culture and selection of the best individuals. In experimental trials such selection must be avoided, and in doing so a wide range of variability at once shows itself. Tall, branched stems with fan-shaped tops arise, constituting a series of steps towards complete atavism. This last [412] however, is not to be reached easily. It often requires several successive generations grown from seed collected from the most atavistic specimens. And even such selected strains are always reverting to the crested type. There is no transgression, no springing over into a purely atavistic form, such as may be supposed to have once been the ancestor of the present cockscomb. The variety includes crests and atavists, and may be perpetuated from both. Obviously every gardener would select the seeds of the brightest crests, but with care the full crests may be recovered, even from the worst reversionists in two or three generations. It is a double race of quite the same constitution as the twisted teasels. My second point is a direct proof of this assertion, but made with a fasciated variety of a wild species. I took for my experiment the rough hawksbeard. In the summer of 1895 I isolated some atavists of the fifth generation of my race, which, by ordinary selection, gave in the average from 20-40% of fasciated stems. My isolated atavists bore abundant fruit, and from these I had the next year a set of some 350 plants, out of which about 20% had broadened and linear rosettes. This proportion corresponds with the degree of inheritance which is shown in many years by the largest and strongest [413] fasciated stems. It strengthens our conclusion as to the innermost constitution of the double races or ever-sporting varieties. Twisted stems and fasciations are very striking monstrosities. But they are not very good for further investigation. They require too much space and too much care. The calculation of a single percentage requires the counting of some hundreds of individuals, taking many square meters for their cultivation, and this, as my best races are biennial, during two years. For this reason the countings must always be very limited, and selection is restrained to the most perfect specimens.