462 Allopolyploids 



ratio of lAA : 4:Aa : laa. If such a plant is selfed, the offspring 

 will be lyl^Ai4 : SAAAa : ISAAaa : SAaaa : laaaa, or 35A : la 

 phenotypically. This will be true not only if the configuration 

 of the chromosomes bearing this gene is a quadrivalent, as in an 

 autotetraploid and some allotetraploids that are not ''pure," 

 but also if there are only bivalents, provided that either auto- 

 syndesis or allosyndesis occurs at random. Yarnell has given 

 us a good example of a 35 : 1 ratio in an amphidiploid. A cross 

 between Fragaria bracteata (n = 7) and F. vesca rosea {n = 7) 

 resulted in some tetraploid plants. At meiosis, quadrivalents 

 were occasionally found at diakinesis, but there were usually 

 fourteen bivalents at first metaphase. The genetic ratios of a 

 gene for pink flower color indicate that, so far as the chromo- 

 some which bears this gene is concerned, apparently chromosome 

 pairing may be by allosyndesis or autosyndesis with equal fre- 

 quency. Pink (P) is dominant over white (p). The bracteata 

 parent was white-flowered (pp) whereas the vesca rosea parent 

 had pink flowers and was apparently PP, so that the tetraploid 

 hybrid was duplex for the P gene. When seven F2 plants were 

 tested for their genotypic constitution by crossing with recessive 

 diploids, it was apparent that one w^as triplex or quadriplex for 

 P, two were duplex, and four were simplex. The population is 

 small but it indicates that the chromosomes bearing the P and 

 p genes paired allosyndetically or autosyndetically at random. 



Allotriploids 



Allotriploids, like autotriploids, naturally represent an unbal- 

 anced condition. An allotriploid involving two species has two 

 genomes from one parent and one from the other. In a "pure" 

 allotriploid in which the odd genome is sufficiently distinct 

 from the two which were introduced by the other parent, the two 

 identical genomes form bivalents, whereas the other genome 

 normally behaves as a set of univalent chromosomes. Hollings- 

 head found some triploid hybrids between Crepis capillaris [n = 

 3) and C. tectorum (n = 4) that behaved in that manner. The 

 triploids contained two C. capillaris genomes, which paired to 

 form three bivalents, and one genome of C. tectorum, which con- 

 sisted of four univalents (Fig. 134). A different result has been 

 reported by Yarnell in triploids produced by crossing a Fragaria 



