1915 Kibbe; on Alaria 49 



this explanation of chamber formation in / fit,talosa. Where the chains 

 of rapidly dividing cells of the inner cortex come in contact with the 

 now inactive sieve-tubes, the resistance to elongation offered by the passive 

 cells naturally draws the outer cells out of proper alignment, and at the 

 same time subjects the inactive cells to a severe strain. Hence the 

 walls of the midrib bulge out to 'relieve the strain on the longitudinal 

 inner cells. The bulging tears asunder the hyphal chains extending 

 transversely across the chamber. Long continued severe strain eventually 

 breaks the longitudinal chains of cells at their weakest points, and the 

 tension is slightly relieved at the points between the breaks in the chains. 

 These relaxed regions thus form the septa, the chambers representing 

 the regions where the chains broke. Therefore, there is no definite 

 chamber length, the length being determined by the location of these 

 weakest regions in the cell chains. The later elongation of the cham- 

 bers is due to continued transverse division and enlargement of the 

 cortical cells. 



In the cross-sectional view of the rifts and chambers a further 

 cause of chamber formation is found. The cortical cells continue to 

 divide radially and tangentially (4) as well as transversely, after the 

 inner cells cease active division. The radial divisions increase the cir- 

 cumference of the chamber. This stretches the inactive medullar cells 

 laterally along the sides of the medullary area. The lateral strain 

 naturally bulges out the walls of the cortical portion, thus decreasing 

 the strain on the inner cells parallel to the plane of the lamina. This 

 tears the pith asunder along the median line, thus forming a chamber. 

 This is made an easier task on account of the still greater longitudinal 

 strain to which these tissues are being subjected. At the septa the 

 longitudinal strain is reduced, hence the tissues are able to withstand the 

 lateral strain as well. 



As A. valida forms no chambers, both cross and longitudinal sec- 

 tions of the midrib were made for comparison with A. fistulosa. The 

 pith area of the former was much smaller in cross section. If sieve-tubes 

 had ever been produced b)^ the inner cortex, they had all disappeared. 

 It is probable that there had been none, since the pith region was so 

 small it could have contained few if an3\ Even the primary pith cells 

 which had been formed into sieve-tubes had broken down, leaving a 

 rather loose pith web. The small number of either or both of these 

 types of sieve-tubes would make it possible for them to produce a 

 bulging of the walls to form chambers. 



As already noted in A. fistulosa, the original pith area of the stipe 

 had many more sieve-tubes at the ends than at the center of the pith 

 region. In sections of the upper portion of the stipe as well as the 



