2 ARKIV FÖR BOTANIK. BAND 17. N:0 11. 



from the Bahamas, West-Indies and Florida. I myself cannot 

 remember having ever seen any seeds in pine-apples. On 

 this species I have made some cytological observations which, 

 though very fragmentary, may be of some importance for 

 the solution of the problem referred to above. The material 

 for my investigation was collected in Ecuador where several 

 races of Ananas sativus are cultivated, though I do not know 

 their horticultural or vulgar names. Two races were fixed, 

 one at Tenguel on the coastal lowland, south of Guayaquil, 

 and the other at the village of Bahos at a height of 1800 

 m., in the valley of Pastaza. They will be referred to in 

 this work as the Tenguel-race and the Banos-race resp. Both 

 Flemming's, Zenker's and Carnoy's liquids were tried, but 

 only the one last-mentioned gave satisfactory results. 



The material of the Tenguel-race proved to be too old, 

 the reduction division in the pollen mother-cells having 

 already been carried out. All the tetrads formed are, how- 

 ever, quite normal (Fig. 4), indicating that the meiosis too 

 has been normal, without any irregularities whatever. The 

 pollen-grains contain much plasma and are apparently capable 

 of germination. I have not been able to determine the 

 somatic chromosome number. 



In the Banos-race the reduction division was found 

 to be much delayed as compared with the Tenguel-race. Buds 

 of equal si7.e contain pollen tetrads in the case of the latter, 

 but only mother-cells in early prophase in that of the former. 

 I do not know if this slow development is due to the colder 

 climate at Batios, or to some germinal quality of the race. 

 Perhaps, it depends on disturbances brought about by de- 

 fective chromosome conjugation, or on some other factor 

 connected with the irregular meiosis that takes place. 



In the heterotypic metaphases a great many chromosomes 

 are found, and both bivalents and univalents can 

 be distinguished (Figs. 1, 2). The former are dumbbell-shaped 

 and arranged to an equatorial plate, the latter are spherical 

 and partly scattered irregularly in the spindle outside the 

 plate. Hence the aspect of the metaphases is just that which 

 characterizes many hybrids. Unfortunately, the chromosomes 

 are so numerous and so minute as to make an exact counting 

 impossible, but from several metaphases the number of 

 bivalents could approximately be estimated at 25—30 and 



