Using 4 methods (association, correlation, exceptions, and sequences) in the study of the evolutionary development of characteristic dicotyledon woods, the author proceeds to determine the characteristics and origin of the primitive vessel segments. The investigations are based upon the following assumptions: "(1) That the theory of evolution is correct; (2) that specialization results in a series of graded forms; (3) that the tracheid is a more primitive structure than the vessel segment; (4) that the vessel segment is genetically related to and derived from the tracheid." Using the method of association it follows that the truly primitive vessel segment will resemble tracheids in general in: "(1) great length, (2) small cross-sectional area, (3) angularity of outline, (4) thin walls, (5) constant width of walls in transverse section, (6) absence or very slight development of an end wall." Vessel segments having the above tracheid characteristics have also "scalariform perforations similar to those found in Pteris and totally different from the Gnetalean type." The scalariform perforation type of vessel segment is therefore the "most primitive type found in the dicotyledons."[long dash]The origin of the vessel segment in dicotyledons is traced from the tracheid "in precisely the same manner as it originated in Pteris by the loss of membranes from the scalariform pits at each end of the tracheid and the formation of an end wall." The author was unable to find transition stages from scalariform tracheids to scalariform segments in the secondary wood of primitive dicotyledons. However, such transitions were found in the primary wood, and "in positional sequence from the pith outward." The transitions are shown in great detail. There is also "striking evidence to show that the evolution of the vessel proceeded along the same lines" in the protoxylem, metaxylem and in the secondary xylem of the first annual ring.[long dash]The wood of Trochodendron, Tetracentron, and Drimys (Magnoliaceae) is believed to throw additional light on the origin of the vessels in secondary wood. The primary and secondary wood is composed entirely of tracheids and parenchyma. The secondary tracheids are exceptionally long with either scalariform or circular pits. The author predicts that fossil secondary wood will be found showing "all the stages in the formation of vessel segments from long scalariform tracheids.".