172 LOCK : 



or drooping." It is possible, therefore, to suppose tliat there 

 is practically complete correlation between the character, 

 roughness, and the erect habit of the leaves. The word 

 " erect " in this case denotes that the leaf droops for less 

 than one-third of its length, as opposed to a two-thirds droop, 

 which is characteristic of I\Iaha-pengiri. 



So far nothing has been found which is obviously incon- 

 sistent with i^fendelian theory. The remaining characters, 

 however, show results vvhicii cannot easily be fitted into any 

 existmg Mendelian scheme. It is tentatively suggested that 

 the former set of characters have more in common with 

 what Mould commonly be called varietal characters : whilst 

 the latter arc more of the sort upon which specific differences 

 are usually founded by systematists. In other words, the 

 results so far as they go lend some support to the opinion 

 of Be Vries, as expressed in the " Mutationstheore," that 

 specific and varietal characters are distinct in kind. 



Thus, Glume I. is winged in Cymho'pogon conf ertiflorus , 

 wingless in ]\Iaha-pengiri, and usually wingless or intermediate 

 in Lena-batu. The seedlings show : — 



No. 4 

 No. 12 



No ex])lanation nf these figures in terms of Mendelian 

 theory is apparent. 



Glume II. is keeled in ('. conferti/Joni.'i, not keeled in 

 Maha-pengiri, slightly keeled in Lena-batu. Seedlings : — 



Kcoled. Sliglitly kcolod. Not keeled. 



No. 4 .. 8 .. f) .. 



No. 12 7 . . 9 .. 3 



To suppose the keeled character dominant with intensifi- 

 cation of this character in llie second generation seems to 

 exceed the legitimate use of hypothesis. 



The sj)ikelets are dissimihii- in C. ronf ertiflorus ; in Maha- 

 jK'iigiri and l^-na-batu they are similar. In Ihe seedlings : — 



Kiiuiliir. Nearly. Dissimilar. 



No. 4 I . . 1 . . 8 



No. 12 :j :{ .. 13 



