Ever sporting Varieties with Heritable F (isolation. 509 
therefore we can not directly see whether both races or 
only the first grow in the particular locality; but their 
great rarity points to the mixed condition. 
The heritable races which have hitherto been found 
and isolated in this way, behave like eversporting vari- 
eties inasmuch as each generation consists both of fasci- 
ated individuals and of atavists, even under conditions 
of the most stringent selection. Moreover the proportion 
of these two types appears to be pretty constant, at least 
under similar conditions of life. As a rule, there are 
about 40% fasciated individuals and 60% atavists. Higher 
percentages of the former occur only under favorable 
circumstances, whilst the proportion of the latter very 
easily increases under unsuitable conditions of culture, in 
spite of selection. 
The first instance that I shall describe was afforded 
by C re f>is bicnnis, an exclusively biennial plant, fasciated 
stems of which have been frequently observed in various 
localities in Holland. The starting point of my culture 
consisted in two fasciations, which I found in May 1886, 
in a meadow near Hilversum, amongst hundreds of 
normal plants of Crepis. The broadening of their stems 
was small and limited to the top. I collected ripe seed in 
this meadow in June, but from normal plants only. 
Whether all or only some of these belonged to the ever- 
sporting race I was in search of could, of course, not be 
determined then. 
This seed furnished in the following year about one 
hundred plants, of which three were already fasciated 
in the rosette stage, whilst in the following year nine 
more of them developed more or less flattened stems or 
branches. The total proportion, therefore, was 12%. 
In order to make perfectly sure, I retained only the first 
