No. 517] A BIMODAL VARIATION POLYGON 27 



dition of division of the leaf. All that we can conclude 

 from the present data is that : (a) There is a strong tend- 

 ency to the production of either simple or ternately com- 



pound leaves— those divided into two leaflets are rare; 

 (b) a larger proportion of the leaves are simple in inflo- 

 rescences with three or four leaves than in those with but 

 two leaves. 



To determine whether the distribution for number of 

 leaf lamina 1 per involucre in this series forms a bimodal 

 distribution similar to that found in 1905, I extract the 

 totals from the two tables of data. To ascertain more 

 accurately the real source of the bimodal condition, 

 should it be found to exist in this series, the combinations 

 which give rise to the different total numbers are given 

 in Table V. 



From these figures appear several points which could 

 not be determined at all on the lumped data as collected 

 in 1905. 



