The Sigiiificaucc of the Ataz'isls. 561 



were saved separately from each plant. In the spring ij\ 

 1901 I sowed the seeds of two decussate, three ternary 

 and four twisted seed-parents and recorded the main 

 stems produced by each parent in 1902, just before the 

 flowering period ; but since the numbers from the indi- 

 vidual seed-parents of each group did not differ to any 

 considerable extent amongst themselves, I shall give the 

 total result only. 



OFFSPRESTG, IN PERCENTAGES. 

 SEED-PARENTS Nl'MBER 



DECUSSATE TERNARY TWISTED 



Decussate 48 8 44 201 



Ternary 39 24 37 136 



Twisted 45 14 41 155 



Total 492 



We see that the offspring of the afaz'isfs produced 

 just as large a proportion of twisted steins as the offspring 

 of the tzvisted individuals. On the other hand the choice 

 of decussate or of ternary atavists seems to have exerted 

 some inHuence in the direction of ternary offspring. 



The twisted stems of Dipsacus are partly right-hand 

 and partly left-hand spirals; and about equal numbers 

 of the two sorts are always found. -^ 



A very curious question thus presents Itself, viz., 

 whether this equilibrium can be disturbed by selection 

 and whether the balance can be upset in a given direc- 

 tion.^ According to the view laid down in this book, 

 that it is not true that with selection all things are pos- 

 sible, we might expect that this would not be a fixablo 

 character and that selection would have no influence on 



"^Annals of Botany, he. cit., p. 404. 



'R. M. Yerkes, Variation in the Fiddler Crab, GeJasimus pup- 

 lator, Proceed. Amer. Ac. of Arts and Sciences, 1901. Vol. XXX\'I. 

 No. 24, p. 417. On page z^4T this autlior states tliat rigiit- and left- 

 handed animals occur in approximately equal numbers. 



