Kleinere Mitteilungen. 



125 



principally a normal curve with a top with 4: yet a very strong indivi- 

 duallity of "property 5" becomes the cause that after the top with 4 the 

 curve line no longer descends but continues to rise: through which a 

 half-curve arises. 



We see this remarkably in Coffea Launiitii. In this variety of coffee 

 I have never found complete normal curves for the number of capitulae 

 per leaf-axil, nor yet correct half-curves, but always something intermediate 

 between these two. This is exemplified in the here following curve VI. 



So we see here "property 6" very strongly developed so that the curve 

 is intermediate between a normal curve and a half one. We can now 

 imagine that, under definite circumstances which are favourable to "pro- 

 perty 6", these becomes so active that a halt-curve arises. 



t 2 3 4 5 

 number ofeapilfl/a ptr leaf- aril 



C.UAyirt.~V 



2 3 4 5 



It is of course also possible that, on the grounds of our experience 

 about the suddenly becoming latent of one or more properties, curve V 

 originally represented a normal curve with the top in 5, but that "pro- 

 perty 6" became latent. So it can be with curve VI in C. Laureutii the 

 "property 6" "becomes latent" after having previously given rise to a 

 half-curve. Now I believe that this question is indeed experimentally to 

 be established. With Cancphora I have only on a single occasion met 

 with 6 capitulae per axil, so that thus "property 6" does not appear to 

 be foreign to Canephoni: so in the case of curve V the half-curve should 

 arise through the becoming latent of "property 6". 



§ 5- 

 In connection with the preceding paragraph I should to make finally 

 some observations about a few correlative tables refering to the connection 

 between the number of flowers per capitulum and the number of capitulae 

 per axil. 



