432 BOTANICAL GAZETTE [December 



the structure, and probably led to the rejection of all other criteria, 

 making the bar "an infallible test for tribal affinities" (Holden 5), 

 both in fossil and living Conifers. 



, In 1919 Bailey (i) studied the origin and development of bars 

 of Sanio, and concluded that those in transition regions are merely 

 normal middle lamellae left between thinned pit areas in the 

 primary wall. He states that when the pits are opposite the bars 

 go smoothly from side to side of the tracheid, because the pits are 

 formed on one primary scalariform pit area, and the bar is the 

 thickening of the lamellae between this area and the next. Thi 

 theory, however, will not explain the bars figured in the writer's 

 paper of 191 5. Those shown in Araucaria are connected with pits 

 in regular horizontal rows, but still fork, following round the cir- 

 cumference of each pit, so as to leave clear diamond-shaped areas 

 bounded by opposing forks. These small areas cannot be con- 

 sidered to be other primary pits. The same fact holds for the 

 bars in the Cycas petiole, which fork, and are even split into two 

 separate rims in some cases, although the pits are not far enough 

 apart to make it possible to attribute the thin space to another 

 primary pit area. 



Bars of Sanio have now been found in other portions of Cycads 

 than the transitional primary xylem. Figs. 6 and 7 illustrate them 

 in the stem wood of Dioon spinulosum. These bars often extend 

 beyond the margins of the pits with which they are in contact, as 

 in the middle tracheid of fig. 7 near the bottom, and so are of a 

 higher type than those figured in the former paper. They are still 

 much more primitive than those of the Abietineae, however, lying 

 in close contact with the pits, whenever such are present. Fig. 6 

 shows the ordinary type of bar in this plant. Between the pits 

 of the single row on the right are bars of the regular Araucarian 

 type. Their length is not greater than the borders to which they 

 cling, and they spread slightly at the ends. The pitting of this 

 tracheid is conspicuously of the opposite type, so that if Bailey's 

 theory of the origin of the bars is entirely correct, they should in 

 this case pass beyond the pits to the limit of the tracheids. In 

 the two left-hand tracheids of fig. 6 is shown a condition which 

 is quite common, namely, the presence of these bars in connection 



