324 Belling. 



If we consider onlj^ the first five ovules iu the ovaries, we should 

 have, I think, on the average, Tvith a random distribution of abortion 

 and non-abortion of embryo-sacs, out of 32 ovaries: 

 1 ovarj- TAith 5 aborted embryo-sacs; 

 5 ovaries with 4 aborted embryo-sacs; 

 10 ovaries with 3 aborted embryo-sacs; 

 10 ovaries with 2 al)orted embryo-sacs; 

 5 ovaries with 1 aborted embryo-sac; and 

 1 ovary with aborted eml)ryo-sacs among- the first 5 ovules. 

 Also there should be, on the average, an equal number of aborted 

 embryo-sacs and of perfect embiyo-sacs in each position, first, second, 

 third, &c., in the ovaries. For 18 ovaries we have: 



Numbers of atorted Observed numbers Calculated numbers 



ovules of ovaries of ovaries 



5 1 0-6 



4 1 2-8 



3 7 5-6 



2 .5 5-6 



(I ()■(! 



The results of observation agree fairly with the calculated numbers. 

 There are 9 ovaries above, and 9 below the mean. 



The numbers of perfect and of aborted embryo -sacs for each po- 

 sition in the 18 ovaries are: 



1 2 .3 4 .5 6 Totals 



Perfect embryo-sacs ...12 — 9 — 10— 6 — 9 — 4 50 



Aborted embryo-sacs .. fj — 9 — 8 — 12 — 9 — 5 49 



The balance is even. Hence it seems that there is a random distri- 

 bution of abortion of embryo-sacs: and such random distribution corres- 

 ponds to the distribution of Mendeliau factors among the megaspores, 

 but to no other cause known to me. 



Fresh pods were opened from 8 plants, and the seeds and abortive 

 ovules counted on the following dates. 



Date Pods Seeds Abortive ovules 



August 2.i, 191;^ .... 7 27 1.5 



September 1, 191.S ... (1 25 U 



September Id, 19i:i ... 12 41 :{1 



September 2». 191:h . . . 2.') 79 71 



Totals .50 172 12S 



On the whole, 43 per cent of ovules were aborted; and 41 per cent were 

 aborted in the distal halves of the pods. More flowers and young pods 



