Szabo .— The Development of the Flower of the Dipsacaceae. 333 
as this stele (g) has been called by van Tieghem (8), is to be seen. Van 
Tieghem did not know the structure and origin of this bundle; he has 
only distinguished it as a bundle much larger than the others and running 
into the ovule. I described it in like manner in dealing with the 
Knautiae ( 7 ), and Fodor ( 3 ) has done the same in his work on the Cepha- 
lariae. The structure of this vascular bundle, however, shows its hadro- 
centric nature ; ft is not a collateral bundle as are all the eight others; while 
tracing its development it is possible to determine that it originates in four 
collateral bundles with their woody portions turned inwards ; they actually 
form the central stele of the axis. The eight peripheric bundles join this 
stele (Text-fig. 4, 5) at the bottom of the ovary. The monocyclic eustele 
(Text-fig. 4, 6 ) is seen restored in the stalk and shows an arrangement 
identical with that figured in Text-fig. 3, 6 . The final result is a collateral 
ectophloic siphonostele. If the ^-bundles would go in the wall of the ovary, 
surrounding its cavity in median-transverse direction, before the ^-bundles, 
then one might suppose that the tissue which lines the inside of the ovary 
has a carpellary origin. But it is not so ; the carpels do not take part in the 
building of the wall of the ovary. On the other hand, the independence of 
the ^-stele, even the fact that it often parts from the wall of the ovary, shows 
directly that this stele must be a rudiment of the separating wall of the 
ovary, which is a general formation in the series of the Rubiales. Its 
connexion with the wall of the single cavity of the ovary may be explained 
by the fact that other cavities do not develop at all. It might be concluded 
from this circumstance that the ^-stele is a direct lengthening of the top of 
the axis. 
IV. Summary. 
From the investigation of the development of the floral leaves it may be 
concluded that the wall of the ovary as a whole is an axile structure, 
originating in that part of the axis which corresponds with the internode 
between the insertions of the calyx and corolla. In the wall of the ovary 
as in an axis there are the eight vascular bundles of the flower whorls, the 
bundles of the gynoecium still forming for a while a separate stele. From 
the construction and development of this separate stele one may conclude 
that it is a rudiment of the ancestral diaphragm (partition wall) or centric 
column, and it actually shows its ancestral tetramerous structure. Of the 
four original carpels the median-anterior alone forms the ovule ; the other 
three (the one median-posterior and the two diagonal) form the simple 
style. The change in the situation of the stele, which is pushed back¬ 
wards in the median plane, makes it appear that there is only one median- 
posterior carpel which gave origin to both the style and the ovule. 
