352 INTRODUCTION TO EMBRYOLOGY OF ANGIOSPERMS 



made the same interpretation but later (1935) considered it more 

 probable that the development of more than one embryo sac within 

 an ovule could also account for the origin of the twins. Harland 

 (1936) agreed with this view, adding that the fertilization of the 

 egg in one embryo sac might stimulate a parthenogenetic develop- 

 ment of the egg in the second and adjacent embryo sac. The re- 

 maining authors mentioned above fall in line with one or the other 

 of these explanations. Briefly then, in a case of haploid-diploid 

 seedlings, the haploid member is derived from an unfertilized cell 

 belonging either to the same embryo sac or to an adjacent embryo 

 sac. 



Diploid-triploid twins have been reported in Triticum vulgare 

 (Yamamoto, 1936), Secede cereale (Kostoff, 1939), and a few other 

 plants (see especially Skovsted, 1939). According to Kostoff and 

 Yamamoto, the triploids observed by them arose from a part of the 

 endosperm, but this is merely a supposition without any positive 

 evidence in its favor. It is more likely that the triploid embryo 

 originated by the fertilization of an unreduced (aposporic) embryo 

 sac or by the fusion of a cell of a haploid embryo sac with two male 

 gametes or one unreduced male gamete. 



Haploid-triploid seedlings are of comparatively rarer occurrence. 

 Nissen (1937) recorded one such case in Phleum. It is probable 

 that the haploid member arose from an unfertilized cell of the 

 embryo sac and the triploid member by one of the methods men- 

 tioned in the preceding paragraph. 



While these are the probable ways in which twins and triplets 

 arise, it is often impossible to be sure of the exact origin of the 

 aberrant member or members of the combination, and it is unsafe 

 to make any categorical statements without taking all the possi- 

 bilities into consideration. 



A particularly careful cytogenetic study of the multiple seedlings 

 of Asparagus officinale has recently been made by Randall and Rick 

 (1945). Of 405 multiple seedlings, 97 per cent were twins, 11 were 

 triplets, and one was a quadruplet. Diploids (2n = 20) were the 

 most frequent, but a few showed other chromosome numbers: 30 

 (triploid), 21 (trisomic), 10 (haploid), and 40 (tetraploid) . In 

 haploid-diploid pairs the haploid member was always much smaller 

 than its diploid partner, but aside from this combination the degree 

 of difference in size seldom gave any clue to the chromosome number 

 or the origin of the polyembryonate condition. 



