284 MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



figure 9 shows that the mother-wall in Crepis is about one four- 

 teenth of the diameter of the cell, just before quadrlpartitlon 

 occurs. In Ipomoea Beer (5a) shows the mother- wall very much 

 thickened after the homoeotypic division. He says that this 

 wall gives a reaction to tests for callose and pectose, but does not 

 discuss the cytokinesis. In Oenothera he says the mother-wall 

 dissolves in i per cent NaOH only after about 24 hours. It is 

 not soluble in strong phosphoric acid. It stains with Bismarck 

 brown, methylene blue, and fuchsin. He therefore concludes it 

 is a modified form of callose or pectose. He says of division that 

 "Septa are developed which form an extension of the mucilaginous 

 mother-cell-wall." Gates (23), however, diagrams a cell-plate in 

 his figure 47, though he does not discuss the process in his paper 



on Oenothera. 



III. Material and methods 



The greater part of the work here reported has been done upon 

 the pollen-mother-cells of Nicotiana Tabacum. In addition ob- 

 servations were made for comparison upon the pollen-mother-cells 

 of Primula sinensis, Tropaeolum majus, Ambrosia artemisiaefolia, 

 Chrysanthemum frutescens, and Helianthus annuus. The root- 

 tips of the last-named were also studied and compared with the 

 root-tips of the onion and the pollen-mother-cells of Larix and 

 Lilium. 



Both living and fixed preparations of the pollen-mother-cells 

 were studied. With the exception of the Ambrosia material, the 

 flowers were collected from individuals growing in the Columbia 

 University greenhouse. In most cases the stamens were fixed 

 individually to insure good penetration of the reagents; however, 

 in the Compositae it was found more practicable to treat the entire 

 flower. 



The material was fixed in Flemming's chromic-acetic-osmic solu- 

 tions in the weak, medium, and strong proportions. The following 

 reagents were also employed: Merkcl's; picro-formol ; Wilson's 

 sublimate acetic; and Benda's solution. The stron g Flemming's 

 formula was found to be the most satisfactory, and had incidentally 

 the additional advantage of being the same reagent which Timber- 

 lake and Mottier used, thus affording means of more direct com- 

 parison. The same might be said of Flemming's triple stain. 

 While iron-haematoxylin was also used, it contributed little of 

 value to the study. 



