10 



ARKIV FÖR BOTANIK. BAND 17. N:0 12. 



tion must have taken place with pollen of C. candamarcensis. 

 This latter species is generally found planted together with 

 C. chrysopetala. 



The investigation of C. pentagona gave more interesting] 

 results. In an individual from Bafios I happened to find 

 both the heterotypic and the homotypic division of the 

 embryo-sac mother-cell. They are quite normal and 9 chro- 

 mosomes can be clearly counted (Figs. 12, 13). In the same 

 ovule, however, in which the homotypic metaphase of Fig. 13 

 is found, another cell with dividing nucleus is distinguished 

 (Fig. 10, 14), in which 15 chromosomes can be counted. 



Fig. 8. Mature embiyo-sac of C. candamarcensis. 9. Two embryo-sac 



mother-cells in an ovule of C. pentagona (from Baiios). 10. Ditto, bot} 



dividing (cf. the text!). — Reich, hom. imm. ^/12, oc. 2. Red. at reprod. 



These chromosomes, seen in polar view of the metaphase 

 are short and thick, as is usual in reduction divisions (thougl 

 a little smaller than in normal mother-cells), and quite dif 

 ferent from somatic chromosomes. Obviously this ovule ha; 

 developed two mother-cells, one normal with complete re 

 duction, and one supernumerary, with the conjugation of th( 

 chromosomes almost totally absent. The latter cell most cer| 

 tainly contains 3 bivalent and 12 univalent chromosomes 

 From this it seems obvious that the physiological process o 

 chromosome conjugation is in some way localized to a spe 

 cial part of the ovule. A similar case is found in anothe 

 ovule (Fig. 9) with both mother-cells still in prophase. 



This leads us at once to make comparisons with th« 

 aposporical mother-cells, embryo-sacs and embryos, found b] 



