684 MINNESOTA BOTANICAL STUDIES. 



raphides lying- parallel with the radial diameter of the cells. 

 These are, however, far less numerous than the usual variety 

 of crystal sacs. 



Anatomically no sharp line can be drawn between pericycle 

 and phloem; they shade imperceptibly into one another. 



Phloem. — In cross section the phloem is not definitely marked 

 off from the pericycle. Its cells, however, are usually some- 

 what smaller, and when crystal sacs are present they too are 

 smaller than are those in the pericycle. In Kneiffia linearis 

 and Kneiffia pumila no crystal sacs were observed. 



The cells of the normal phloem are mostly parenchymatous. 

 In cross section they are circular or elliptical. The long dia- 

 meter of the cells of the phloem parenchyma is seldom over 

 25mik., as a rule not over 20mik. The cell walls are quite thin, 

 and any investigation of the tissue in herbarium material is alto- 

 gether unsatisfactory. Sieve tubes could not be demonstrated 

 with certainty except in fresh material. In cross section they 

 appear scattered in small groups through the parenchyma. 

 There are no bast fibres. 



Some of the species which grow to a considerable size have 

 the phloem area fairly well developed, however, in most cases, 

 it is but a few cells broad, usually six to twelve, occasionally 

 even narrower. It may be forty cells or even more in breadth, 

 (Onagra biennis, Megapterium missouriensis.) The cells of 

 the phloem, though sometimes, for a part of the way irregularly 

 placed are more often arranged in radiating rows. The med- 

 ullary rays can often be distinctly traced as single rows of cells 

 radially elongated. 



The lack of phloem tissue in the usual position is perhaps 

 compensated for by the presence of phloem elements in the 

 pith, and in some species in small patches in the xylem. 



Medullary phloem. — Apparently the vascular bundles of all 

 the species may be described as bicollateral, whether or not 

 this bicollaterality exists from the first or is the result of a sec- 

 ondary growth of tissue within the medulla can, of course, only 

 be determined by an examination of very young material, show- 

 ing the developmental stages. In Onagra biennis, the medul- 

 lary phloem is developed in the hypocotyl of quite young seed- 

 lings. 



The cells of the medullary phloem usually form distinct groups 

 arranged with greater or less regularity in a circle toward the 

 outside of the pith. Usually the cells are quite small, and are 



