250 DUPLICATION OF GENERATIVE NUCLEI BY MEANS OF 
lar nucleus of the character of a vegetative one, followed by two or mo- 
re smaller, almost globular nuclei, ressembling greatly generative nu- 
clei. Moreover the two polar nuclei which fuse at the formation of the 
endospermnucleus are also globular. 
We are not the only ones, who thought it advisable to start by trying 
to produce triploid seedlings. 
ERNST (1918 p. 360) writes, in reference to NEMEC’s observations: 
Auf ein positives Resultat solcher Untersuchungen mit Gagea oder 
anderen dazu besonders geéigneten Pflanzen, wurde allerdings wohl erst 
bei Ausführung hunderter van Kreuzungen und Heranzucht der aus 
den Samen erhaltenen Bastarde gerechnet werden können’. We see, 
that ERNST does not underestimate the difficulties of such an experi- 
ment; but he possibly did not suspect, that a change in external condi- 
tions could cause diploid plants to form so many diploid and plurinu- 
clear pollennuclei; he writes at least (lc. p. 349): „Nach den bis jetzt 
vorliegenden Beobachtungen ist zu erwarten, dass solche diploidkernige 
Pollen-körner und Embryosäcke bei sonst normalgeschlechtlichen 
Pflanzen, wenn ueberhaupt, nur in sehr kleine Zahl gebildet werden. 
c. Our experiments have clearly shown, that even when diploid and 
plurinuclear pollengrains are in the majority, pollination still causes the 
progeny to be chiefly diploid. 
VAN OVEREEM (1921 p. 25) also found that the number of triploid 
descendants is by no means proportional to the number of diploid pol- 
lengrains. 
In the case of O. Lamarckiana semigigas he found that, among the 
fertile pollengrains, the triangular (haploid) ones were in the majority, 
but that a considerable number of quadrangular (diploid) ones were 
present. One could therefore suppose that fertilisation of the haploid 
eggcells would give a large number of triploid plants. „Das ist aber 
nicht der Fall, Pflanzen mit 21 chromosomen treten in den verschiede- 
nen Kreuzungen selten, manchmal gar nicht auf’’. It should be remar- 
ked that in the case of Oenothera, it is possible (VAN OVEREEM 1921 p. 
31) to conclude whether pollengrains are haploid or diploid: „Bei den 
diploiden Formen sind die Pollenkörner dreieckig und besitzen 7 Chro- 
mosomen. Bei den tetraploiden Formen, ist die Körnerform viereckig 
und ihr Pollen führt 14 Chromosomen’. | 
If this is always the case, one should watch whether diploid Oenothe- 
ras could not form quadrangular pollengrains under abnormal conditions. 
