258 AN INTRODUCTION TO MODERN GENETICS 



Physcomitrium (n = 36) which can be shown to consist of two sepa- 

 rately viable sets of chromosomes (n = 18). In most cases, the analysis 

 of a polyploid species has to be based on the pairing of its chromosomes 

 in hybrids with the constituent diploids or with other related polyploids. 

 Thus if we cross a tetraploid with a diploid and obtain a triploid in 

 which two-thirds of the chromosomes pair^ at the reduction division, 

 we may argue that the tetraploid contained a diploid set of chromo- 

 somes homologous with those of the diploid used in the cross. One of 

 the first observations of this kind was made by Rosenberg,^ who found 

 pairing in ten bivalents and ten univalents in a hybrid between Drosera 

 longifolia (n = 10) and D. rotundifolia {n = 20). He argued that in the 

 hybrid the 10 longifolia chromosomes paired with 10 of the rotundi- 

 folia ones, and that the ten impaired chromosomes were the rest of the 

 rotundifolia set; and concluded that rotundifolia is an allotetraploid in 

 which longifoha is one of the constituents diploids. But clearly this 

 argument in itself is not conclusive; we know that in some cases auto- 

 syndesis may occur and it is therefore possible that the twenty rotundi- 

 folia chromosomes may pair among themselves, leaving the longifolia 

 set as univalents. If this were the case, one would have to assmne that 

 the two diploids making up rotundifoUa are more related to each other 

 than either is to longifoha. There is no immediate way of deciding 

 between these two hypotheses. 



Arguments from chromosome pairing must therefore be accepted 

 with great caution unless the series of crosses is so complete as to 

 exclude one of the possibilities. For instance, in the compUcated rela- 

 tions which Goodspeed and Clausen^ have demonstrated between the 

 Nicotiana species sylvestris, tomentosa, Rusbyi, Tabacum, glutinosa 

 and the artificial digluta, the fact that no bivalents are formed in the 

 Tabacum-glutinosa hybrid indicates that autosyndesis cannot take place 

 in the Tabacum set, and this entitles one to argue that the bivalents in 

 the other Tabacum crosses are not formed in this way. (Fig. 113.) 



The different species of wheat^ (Triticum) form a polyploid series 

 with the basic number 7. The diploids (2n =^14) include T. mono- 

 coccum, T. aegilopoidesj etc., and are known as the Einkorn group; the 

 tetraploids or Emmer group include T. durum, T. dicoccum, T. turgidum , 

 etc., and are often grown for macaroni; the normal bread wheats are 

 hexaploids, known as the Vulgare group after the best known species 



^ Usually the occurrence of prophase pairing is only an inference from 



observations of metaphase association. - Rosenberg 1909. 



^ Cf. Clausen 1928, Sansome and Philp 1932. * Rev. Watkins 1930 



