Kleinere Mitteilungen. I77 



do not see the genetic bond of the numbers themselves. They work 

 with masses of numbers, which singly take their places of signification 

 after having been, by way of one or more formulae, brought back to a 

 single number. The final numbers are the working material ; the substance 

 of a regularly arising complexity of numbers is thereby shelved. In this 

 respect according to the critical standpoint of modern statistics, it may 

 be said that "the nut is thrown away with the shell'. That "life" is no 

 mathematical formula is subscribed to by a legion of physiologists, and 

 consequently one has generally been averse to using statistical productions 

 physiologically. But the cause of this is that one proceeded to formulate 

 those statistical data. Hence we are annoyed at the fruitless mathematical 

 onesidedness of the Biometricians which Johannsen has wished to save us 

 from. Johannsen has however gone too far on the other hand with his 

 sceptical impressions, and has thereby made statistics nothing else than a 

 mere science of control. 



In reality one has to contemplate numbers just as they are given to 

 us by nature without reducing the numbers directly to formulae. In 

 measuring, one has to follow up the numbers by themselves as well as in 

 connection with others just as the physiologist considers the numbers he has 

 obtained, and at the farthest, builds them up into a curve for general survey. 

 The condensing of complex numbers to a single number tells us nothing. 

 In my former publications I had the opportunity in this way to criticise the 

 correlative notion. Now what does a so-called correlative coefficient actually 

 tell us? Something to the mathematician perhaps; yet it tells the physiologist 

 nothing at all. 



On this the difficulty of the present statistics hangs with which even 

 the extremely critical Johannsen has indeed also met; that "life" is not 

 mathematics cannot be too sufficiently and too consequently recognized ; but 

 "ciphering" not being an element of life, yet "ciphers" may tell us an 

 astonishing amount. 



And it is so that I easily an-ive at the point at which my new re- 

 searches will attach themselves. They here are concerned with some corre- 

 lation tables which exhibit in a very striking way such deviations, that we 

 obtain a new insight into the manner in which such correlation tables must 

 be comprehended. 



The correlation tables refer to the connection between the number of 

 "bloomheads" per leaf-axil and the number of blooms per "head". All have 

 reference to Coffea Quillou. I comprehend the correlation-concept in the 

 same sense as it has been explained in the second communication'). From 

 out of this standpoint lies the criterion of the correlation- concept of the 



*) Vol 11. — of this perioilieal. 



