12 ARKIV FÖR BOTANIK. BAND 17. N:0 12. 



rise to an almost diploid embryo-sac, which perhaps would 

 have been capable of further development without fertili- 

 zation. 



The later stages of the seed-development are lacking ir 

 my material of C. pentagona, so I have not been able to 

 decide whether apomixis takes place or not, neither have 1 

 seen any pollen-tubes (of C. candamarcensis!). As indicating 

 a beginning apomictical development, there might, perhapsj 

 be regarded two embryo-sacs — one in the Banos-race, on« 

 in the Quito-race — which contained 10 and 9 nuclei res- 

 pectively, instead of the 5 that are generally found. This 

 conclusion is, of course, very uncertain, especially because 

 the extra nuclei might have arisen in consequence of irre^ 

 gular distribution of the chromosomes in meiosis. If, however 

 apomixis really takes place, it should he characterized as 

 apospory in the Banos-race, whereas in the Quito-race wc 

 cannot with certainty decide between apospory and apO' 

 gamy. It must be pointed out that beside the mother-eel 

 with the irregular meiosis, described above, remains were 

 found af a collapsed cell, which, perhaps, had representee 

 a normal embryo-sac mother-cell, and had been forced asid( 

 by another, aposporical one. Such collapsed cells are, b^^ 

 the bye, often found in the archesporial region of the ovules 

 of both C. pentagona and C. chrysopetala. A priori aposporj 

 must be regarded as the more probable, because the seeds 

 actually produced in C. pentagona, are very few, whereas 

 apogamous species generally produce plenty of seeds. Or 

 the other hand, in the above-mentioned case, we may ai 

 well speak of a bicellular archesporium, the two mother-celL 

 showing a different degree of chromosome conjugation. Cross 

 fertilization with C. candamarcensis is not excluded, at leas 

 not in the Banos-race, where, as stated above, sexual em- 

 bryo-sacs with reduced chromosome number are developed 

 — As in C. chrysopetala many ovules in this species too d( 

 not contain any archesporium at all. In many others th< 

 embryo-sac degenerates in an early stage, but even well 

 developed, mature embryo-sacs are often found. 



Summarizing we may state that the few seeds producec 

 in the fruits of C. chrysopetala and C. pentagona are eithe; 

 of apomictical or of hybrid origin. In the latter case th< 

 fertilizing pollen must have come from G. candamarcensis 



