220 THE BEHAVIOUR OF POLYPLOIDS 



Following is a list of species of this type with the numbers of 

 somatic chromosomes found or inferred in each. 



Table 33 



Polyploid Species with Variable Numbers 



Cochlearia anglica, 37-44, 49, 50 . Crane and Gairdner, 1923. 



Salix triandra, 38, 44 . . Blackburn and Harrison, 1924. 



Prunus laurocerasus (22a; = 176), Meurman, 1929. 



ca, 176 



Crepis biennis {8x = 40), ;^g-4^ . ) HoUingshead and Babcock, 



C. ciliata, 40-42 . . . f 1930 ; Babcock, 1935. 



Viola canina, 40-47 . . -J- Clausen, 1931 b. 



V, Battaudieri, 52-60 . -J. Clausen, 1931 c. 



A£;avesisalana(^x = i<,o),i^S-i4^) T^ ,^ , 



/ -7 J, . r J' ^o ^^ Doughty, unpub. 



A. Zapiipe (4% — 120), ca.\\o.\ 



Chrysanthemum ornaium (8x=y 2), Shimotomai, 1933. 

 72-74. 



6. THE DEVELOPMENT OF THE POLYPLOID SPECIES 



(i) Introduction. The increase of chromosome number by which 

 a polyploid arises has been reversed in experiment, first, by crossing 

 an autopolyploid with its diploid relatives and selfing the triploid 

 progeny, or back-crossing them to the diploid (as in Solanum, 

 Campanula, Primula sinensis) ; or, secondly, by parthenogenesis, 

 in which an autotetraploid gives its fertile diploid parent (as in 

 Campanula, Gairdner and D., 1931), and an allotetraploid gives its 

 sterile diploid parent (as in Digitalis, Buxton and D., 1931). Thus 

 when the allopolyploid reverts to the diploid, sterility is produced, 

 which can only be got rid of by another doubling of the chromosome 

 number. Chromosome doubling in a hybrid must therefore be an 

 irreversible process in nature. On these and other grounds it is to 

 be supposed that polyploid species are often of great antiquity, and 

 this suggests two problems : How is a polyploid species to be 

 recognised as such ? And how do polyploid species change after 

 their origin ? The second problem will be considered first. 



(ii) Variation in Cell Size. External and developmental factors 



