368 Tricotylous Races. 
a simple method of searching for such anomalies. Fas- 
ciations are so common in nature and in the garden that 
special means for obtaining them are not required; but 
twisting is much rarer and ordinarily it is only by a lucky 
chance that we meet with an instance of it. 1 If we wish 
to become independent of this chance we must have re- 
course to the culture of variations of cotyledons, because 
such will offer a greater likelihood of furnishing the de- 
sired anomaly than other examples of the same species. 
In the first, or at least in the second, generation we may 
count on finding them, if the extent of the experiment 
is sufficiently great, and once obtained, they can easily 
be further improved by ordinary selection. A single in- 
stance will suffice. MORREN found a very fine specimen 
of twisting in Dracocephalum spcciosnm in a meadow not 
far from Liege, 2 and when I read his description I be- 
came extremely anxious to investigate such a case of 
torsion in this species, or at any rate in this genus. For 
this purpose I selected Dracocephalum moldavicum, 
which, being an annual, seemed more suitable. In the 
spring of 1892 I selected a single hemi-tricotylous seed- 
ling found in a sowing of commercial seed (a little less 
than 20,000 seedlings), and from this bred a race which 
in the first year exhibited nothing remarkable but pro- 
duced fasciations in the second year and traces of twist- 
ing in the third, and finally, in the fourth, some very 
fine instances of spiral torsion. One of these had the 
whole main stem transformed into a screw (Fig. 74). 
Fortunately in such experiments, the aim can be attained, 
as a rule, in a much smaller number of years. 
1 Monographic dcr Zwan^sdrehun^en. PUINGSHEIM'S Jahrb. f. 
wiss. Bot, Vol. XXIII, p. 116. 
z Bull. Acad. Roy. Bclg., Vol. XVIII, p. 37. 
