512 The Inconstancy of Fasciatcd Races. 
generation. In the fasciated races, on the other hand, 
we are obliged to limit our selection to the best represen- 
tatives of the anomaly; but there is no further reason 
to suppose that these also possess the highest hereditary 
values. Thus, an essential part of the selective process 
as applied to the tricotyls is omitted in this case. This 
is mainly due to the impossibility of calculating the hered- 
itary values in the seed pans, and the fact that these 
would need cultivation on a very large scale in the gar- 
den. In order to calculate the hereditary values for only 
20 seed-parents from lots of only 100 offspring each 
and even this would hardly give reliable results 80 
square meters of the garden would have to be devoted 
to Crcpis, and this can scarcely be done in an ordinary 
garden. It is to be hoped that institutions will .soon 
be erected where such determinations can be carried out. 1 
Besides Crcpis bicnnis I discovered one or two other 
species behaving in the same way and succeeded in rais- 
ing eversporting varieties from them. 2 The first to be 
mentioned is Aster Tripolium, of which I obtained a 
splendid fasciated example with ripe fruits in the autumn 
of 1900, from this neighborhood. At first I grew the 
plant as an annual and reached only a low proportion of 
fasciated individuals as a result of this. The figure was 
7% for the fourth generation. In the fifth generation, 
however, in the summer of 1894, the plants were sub- 
jected to better treatment, and more than half of them 
produced fasciated stems, amongst which many were 
more than 3-4 centimeters broad. I shall deal with 
My experimental garden contains 75 beds of about 4 square 
meters each. 
2 Botanisch Jaarbock, Gent, 1894, an d Bull. Scicntifiqtie de la- 
France ct de la Bclgiqnc, public par A. GIARD, XXVIT, 1896. p. 402. 
