[sifton] bar of SANIO IN GYMNOSPERMS 95 



tendency toward enlargement of pitting even in forms where elimina- 

 tion has taken place to a slight extent or not at all. In the right hand 

 tracheid of Fig. 20 the pits are more scattered, and retain their rounded 

 contours. The only bars or rims of Sanio in the figure are of the 

 short, narrow type found in the Araucarian stem. 



Fig. 21, Plate V, is from the short shoot of Ginkgo. Medullary 

 ray cells at the upper right indicate that in the right half of the figure 

 we are dealing with secondary wood, but the pitting at the left is 

 that of the metaxylem. The three right hand tracheids have alternate 

 pitting. The pits are not flattened by mutual contact, and above and 

 below them are bars of Sanio of the type seen in the cone axis of 

 Araucaria (Fig. 4, Plate I). The forking at the ends is plain, especially 

 in the third tracheid from the right. 



Fig. 22, Plate V, shows the ends of two stem tracheids, and is 

 included because it illustrates a stage often found in Ginkgo. The 

 two pits at the bottom of the right hand tracheid have what appears 

 on casual view to be a straight bar of Sanio between them, but close 

 examination shows two separate rims with ends overlapping but not 

 fused, each rim being connected with its own pit. The rest of the 

 figure has the compound bars rather well fused. 



Fig. 23, Plate V, which is from the root, presents a special form 

 of rim resembling that of higher forms more closely than the usual 

 Ginkgo type. There has here been a considerable enlargement of the 

 primordial pits and the rims are curved, and more heavily marked 

 than usual. In parts where the pitting is somewhat crowded, how- 

 ever, their compound origin is still in evidence. See, for example, 

 the thinning out at the centre of the second and third bars from the 

 bottom, and the diamond-shaped, clear area in that fifth from the 

 bottom. Fig. 24 shows bars similarly thinned out at the centre al- 

 although here the pitting is uniseriate and the thinned place comes 

 opposite the bordered pit. Thus it cannot be said that the gaps are 

 due to a tendency of scalariform bars to break up and adhere to the 

 bordered pits. In the light of the evidence from primitive regions 

 and primitive forms, their presence is easily explained as a relic of 

 the condition in which short bars, each composed of the rims of two 

 vertically adjacent, small primordial pits, were arranged in a series, 

 end to end, across the tracheid. 



The three remaining figures illustrate common types of bar in 

 Ginkgo. Fig. 25 is from the same region as Fig. 20, and here the short 

 bars have fused end to end, though the enlargement of primordial pits 



