RESEARCH ON XYLEM AND PHLOEM II3 



colleagues, she has broadened further her studies of phloem structure over a 

 wide range of plants. These studies have chiefly made clear how vague and 

 unsatisfactory is our total understanding of variation in the phloem. 



One of the latest papers has shown, for example, the differences in some 

 of the divisions within the phloem derivatives of the vascular cambium. The 

 divisions [in planes more or less at right angles (anticlinal) to that of the 

 cambium] may be vertical to almost transverse. The derivatives of the divi- 

 sions that are nearly transverse obviously are shorter than the fusiform 

 cambial initial from which they were indirectly derived. Some of these deriva- 

 tives become sieve elements. Since they do not elongate before maturation, 

 it is clear, accordingly, that length of sieve elements may have no obvious 

 relation to the length of fusiform cambial initials. Length — the cornerstone 

 of evolutionary studies of xylem elements — thus can be used only when one 

 knows the developmental history of the sieve elements; this is a crucial 

 obstacle to overcome in similar studies of phloem elements. 



Another interesting aspect of divisions within the cambial derivatives 

 (phloem initials) is related to the number of divisions that may occur. For 

 example, it has often been considered that a given phloem initial in the 

 dicotyledons acts as a sieve-tube mother cell, and by one or more divisions 

 produces one or more small companion cells [which may undergo transverse 

 divisions themselves (fig. 18)]. It is now clear that the phloem initials may 

 undergo several longitudinal (anticlinal) divisions prior to those which 

 produce companion cells. The products of these divisions may be sieve ele- 

 ments or parenchyma cells (fig. 17, P) or both. If they are to be sieve-tube 

 members (fig. 17, STM), they in turn undergo divisions resulting in the 

 formation of companion cells (fig. 17, CC). A widely varying assemblage of 

 cells may thus arise from what has heretofore been considered as a sieve-tube- 

 member mother cell. In addition to companion cells, accordingly, there are 

 at least two kinds of parenchyma cells in many species of the dicotyledons, 

 those which are derived as noted above and those which arise directly (usu- 

 ally after one or more transverse divisions) from a phloem initial. Those 

 arising in the latter manner are almost invariably larger than those indirectly 

 derived in the same species. 



The significance of these divisions in relation to food conduction in sieve 

 elements is yet unknown, but they must be considered in any theories con- 

 cerning the mechanism of conduction. Reduction in size of the potential con- 

 duits, together with the insertion of additional end walls, must make some 

 impact on the efficiency of translocation. 



The problems for study in phloem anatomy are numerous and varied. 

 They concern, for example, structure not only as related primarily to con- 

 duction but also as related to toughness of the bark. Some problems of special 

 importance to physiology concern the sieve areas, and particularly the minute 

 composition of connecting strands; the functional duration of sieve ele- 



