MacMillan : Ouservatioxs ox Nereocvstis 287 



. cells have not yet developed the characteristically thick walls of 

 mature stages, but mucilage canals arc now in evidence, a circle of 



u 



them appearing subepidcrmally. I am able to confirm Will's and 

 Guignard*s account of their development. They originate by 

 vertical radial partitions of certain cortical cells in the cambium re- 



i 



gion. Just outside of the two cells thus formed a cleft appears and 

 this cleft increases in size while the two small cells within divide b}' 

 other vertical and transverse partitions finally coming to line the 

 half of the canal w^hich faces the central cylinder. These small 

 cells are the secretion-cells. They do not persist but after a time 

 break down so that in mature portions of the stipe the mucilage- ' 

 duct in cross section appears merely as a circular-outlined inter- 

 cellular space. - i 

 It is in material of the same age or younger that one finds 

 well-developed trumpet-hyphae intermingled with the ordinary 

 anastomosing filaments of the pith-web. While these cells do not 

 become nearly so attenuated as the sieve -tubes, they are nev^erthe- 

 less slenderer and three or four times as long as the ordinary cells 



i 



of the pith-web. Their ends, where two come in contact, are 

 much swollen. I have not been able to discover' whether frae- 



o 



mentation of the nucleus precedes the formation of a trumpet- 

 hypha as it does that of the sieve-tube. Neither my Russow's 

 callus reagent nor corallin-soda gave results such as those which 

 Oliver obtained, nor have I been able by microchemical methods 

 to demonstrate the presence of true callus, such as has been an- 

 nounced for the sieve -tubes. This is possibly owing to the pre- 

 servatives which ha\'e been applied to the tissues, or their age, but 

 I am inclined to accept rather the \"Icws of Wille regarding the 



callus than those of Oliver. No evidence of protoplasmic connec- 

 tions between adjacent sieve-tubes or trumpet-hyphae has been 

 obtained. At this point it is well to state definitely that the 

 material I have examined seems to show clearly that there are 

 two very different kinds of tubular cells with perforated end parti- 

 tions. -Wille, in criticising the results of Oliver, who made a dis- 

 tinction between trumpet-hyphae and sieve-tubes, seems to suggest 

 that there is only one category of such cells and that differences 

 are of degree and not of kind. According to my obser\^ations the 

 trumpet-hypha does not become extremely attenuated nor does it 



