350 
C. H. Ostenfeld. 
III. 
Having reviewed the sexual conditions in Hieracium , we may 
now turn to the question of the origin of new forms. These must 
arise either (i.) by hereditary variations of already existing species, 
or (ii.) by crossing between them, or (iii.) by a combination of both 
these methods. My experiments show that new forms arise by 
hybridisation and seem also to allow the first possibility. Thus in 
Hieracium new species arise both by hybridisation and by hereditary 
(single) variations (so-called mutations), and the continued existence 
of the new forms is favored by prevailing apogamy which prevents 
the blending and ultimate effacing of new characters through 
repeated crossings back with the parents. 
Hitherto my hybridisation experiments concern only species of 
the Pilosella-sab-gemis. The few exact experiments with Archie- 
racia have not yet given positive results. But Mendel made two 
actual crosses with Archieracia, both having H. limhellatum (which 
we now know is sexual) as the one parent; and a plant of H. virga 
aurea (also a sexual species) in my cultures gave without isolation a 
heterogeneous offspring with hybrid-looking individuals. 
Mendel also made many hybridisation experiments with 
Pilosella- species, and I can confirm his results. The most salient 
matter is that the first generation, F x , is heterogeneous, in contrast 
to the ordinary rule for hybrids. The following generations, F 0 , 
F 3 , etc., are homogeneous and like their individual parent in Fj. 
In other words, the Hieracium hybrids behave in just the opposite 
way from other hybrids, in which F x is homogeneous, whilst the 
later generations segregate. The explanation of the non-segregation 
is to be found in apogamy. As was said above, the Pilosella species 
are only partly apogamous and it is possible to cross two such 
species with each other. We get then an offspring which consists 
of (i.) some individuals like the mother and of (ii.) some individuals 
which are evidently hybrids, but each of which differs from the 
others. The individuals which have the appearance of the mother 
have, probably, been produced apogamously, while the others are 
the result of crossing. The hybrids obtained are often wholly or 
nearly sterile, and thus segregation in the later generations cannot, 
by the nature of the case, be observed. However some of the 
hybrids are partially or wholly fertile and these are apogamous, as 
are one or both parents; and consequently the later generations 
remain like their offspring of F x . In other words, by means of 
