HISTOLOGY OF PHLOEM IN WOODY ANGIOSPERMS 373 
where no correlation exists is perhaps within the Hmits to be expected 
for morphological characters. It can also be said that the distribu- 
tion of sieve-tube type in the phylogenetic sequence is apparently 
just as consistent with the idea that type i is primitive and type 3 
advanced, as is the distribution of vessel type with the more widely 
accepted theory that the scalariform vessel is primitive, and the porous 
advanced; and if the types of vessel are given phylogenetic significance, 
sieve-tube types may also be given equal value. It cannot be stated, 
however, that there is a gradual advance in type of sieve tube without 
great variation or many exceptions in ascending the phylogenetic tree, 
as indicated by Hemenway (6). 
The above-named author also appears to be unjustified in the 
statement that the sieve tubes of the Juglandaceae are like those of 
the gymnosperms and vascular cryptogams (5). The reason given for 
such an assertion is that the sieve tubes of the Juglandaceae have well- 
developed sieve plates upon the side walls which do not differ in struc- 
ture and appearance from those upon the end walls, a condition similar 
to that present in the above-mentioned groups. In a later paper (6) 
it is stated that thirty species of the lower woody dicotyledons show the 
same sieve-tube structure as that found in the Juglandaceae. 
These statements regarding the sieve tubes of the Juglandaceae and 
the "thirty species of lower dicotyledonous trees" are confirmed 
neither by the present research nor by literature. In the first place, 
as stated by Hill (7), the sieve plates of the gymnosperms, as illus- 
trated by Pinus, have multiperforate sieve fields between the meshes 
of the cellulose framework, whereas in the angiosperms, as illustrated 
by Cucurbita, Wisteria, Vitis, and others, the sieve fields between the 
meshes of the sieve are perforated by a single large pore which takes 
up nearly the whole width of the mesh. In the present research, 
examination of the sieve plates and the lattice of the sieve tubes of 
Juglans, Salix, Populus, and others under high magnification, shows 
that there is such a discrepancy in the size of the pores on the lateral 
and terminal wall that they cannot be said to have the same structure. 
In all cases the pores on the sieve plates of the end walls are large 
enough to permit accurate measurement, but on the side walls the 
pores, though distinct, are so minute as to prevent such measurement 
even under the oil immersion. This condition is well brought out in 
the photomicrographs of Populus deltoides (figs. 2, 3). In fig. 2 the 
pores in the sieve plate on the end wall are plainly visible, but on the 
