350 
L. H. MACDANIELS 
plates, or lattices, on the side walls are not so well developed as in the 
other types. This condition is illustrated by Lactuca scariola (text- 
fig- 3). 
In these three types there is a correlation between the type of side 
wall and the type of end wall; type i has well-developed lateral plates, 
whereas type 3, the highest type, has the lattices poorly developed. 
Thirty species of the lower woody dicotyledons were studied and 
found to have the same general sieve tube structure as have the gym- 
nosperms and vascular cryptogams. In general there is no wide varia- 
tion in type of sieve tube even in different genera of the same family, 
except when there are both herbaceous and woody genera within the 
family, in which case the herbaceous plants show the higher type. 
No woody dicotyledons studied have sieve tubes of the third type. 
In the list of species arranged in groups according to type, the woody 
dicotyledons are placed almost exactly in the phylogenetic order of the 
Engler and Prantl system. The herbaceous dicotyledons and the 
monocotyledons are all placed above the second type. 
Companion cells are rare if not wanting in many of the lower 
dicotyledons. 
The evolution of the sieve tube parallels that of the vessel, there 
being a gradual transition from the gymnosperm type with oblique 
end walls and well developed lateral sieve plates to the so-called dicoty- 
ledonous type with transverse sieve plates and poorly developed 
plates or lattices upon the side walls. 
The study of sieve tube types adds an argument in favor of the 
view that herbaceous plants are more advanced in evolutionary de- 
velopment than woody plants. 
The initial purpose of the present research was to make a study of 
the phloem of seedlings of a number of woody dicotyledons selected to 
represent a series from those phylogenetically lowest to those highest 
in the Engler system. Such a study was expected to reveal any dif- 
ferences between the structure of the phloem in the seedling and in 
the mature plant, both root and stem, and to supply evidence of re- 
capitulation. If the latter were present, it was hoped that light would 
be thrown on what constitutes the primitive type of sieve tube among 
angiosperms. As the work progressed, however, discrepancies of a 
rather startling nature began to appear between the results of the 
present research and those of Hemenway. Further, examination of 
