INTKRCELLULAK CANALS IN DICOTYLKDONOUS WOODS 431 



In different woods with radial canals there is wide variation in the 

 width of the rays. When very narrow, the presence of a duct causes it 

 to enlarge and become fusiform, as in the conifers. In wider rays the 

 ducts, unless very large, may have little or no effect on the shape of the 

 ray. A single ray may contain one or two, occasionally three, and 

 rarely four ducts, and in no case does there appear to be any regularity 

 as to the location within the ray. Usually the cells about a duct are 

 very much smaller than the other ray cells, the greatest disparity being 

 noted in Rhus lauriiia, from southern California, and R. metopium, 

 from southern Florida. At the other extreme, in the Araliaceai exam- 

 ined there is not nuich contrast in size and shape of the cells, the canals 

 appearing simply as large intercellular spaces among the ordinary ray 

 cells. In the conifers a ray usually contains a single duct, centrall}! 

 located, but occasionally two are found, especially in Picea. 



In section, a radial canal is roughly circular or elliptical. The limit- 

 ing cells are fully as thick-walled as the other cells, and sometimes 

 thicker, though in some instances the wall adjacent to the cavity is thin- 

 ner than the remainder. In Odina wodicr (Anacardiaceae) there is a 

 single distinct and very regularly disposed layer of epithelial cells, with 

 large, simple pits facing the cavity. Viewed longitudinally, each epi- 

 thelial cell bears a row of these pits which remind one of the pits on the 

 lateral walls of the ray parenchyma cells in Pinus strobus. In other 

 cases studied pits into the cavity were not visible, though in most in- 

 stances the secondary epithelial cells were copiously pitted. Not infre- 

 quently these cells are very short, and sometimes very irregular in 

 shape. The canals may be open or more or less completely filled with 

 resinous or gmnmy material, amber-colored, red. or dark brown. In 

 one case the ducts, tiiough highly resinous, were traversed by my- 

 celium. 



Radial canals in the wood are commonly associated with vertical 

 ducts or cavities in the cortex, and (a) may end blindly in the ray with- 

 out reaching the pith (Anacardiaceae) ; (b) extend to the pith and unite 

 with vertical ducts there (noted in a twig of Spondias sp.), or (c) con- 

 nect with vertical ducts in the wood (Styrax bcttcoin and sometimes 

 in Liquidamhar). Serial tangential sections made by the writer in a 

 twig of Scliinus ^cioiinaiuiifolius revealed first the few large ducts in 

 the outer cortex, and then the anastomosing character of the numennis 

 smaller ones deej^cr in. .\t places in the latter there were branches 

 which turned at right angles and entered the rays. In some rays there 

 was a single duct ; in others two or three. Xear their end in the wood, 

 which in this instance was in wound parenchyma resembling a pith- 



