Kleinere Mitteilungen. 



183 



table VI 



and again consider a horizontal row: then we clearly see that anew maxima 

 occur over and over again, and that marked depressions lie between the 

 maxima. The maxima are therefore very pronounced points in the tables. 

 I have now only traced the diagonals through those maxima which lie in 

 a very pure diagonal row. It however seems to me also possible to trace 

 still more of the lines between those diagonals which have been already 

 traced, although indeed they are not so striking. Yet the results are not 

 generally so convincing as in my tables. However it is now obvious 

 how the correlation tables may be analysed and the possibility is not 

 excluded that the principle occurs to a greater extent than has hitherto been 

 thought po.ssible. Each of these tables is clear enough and speaks for 

 itself. It is therefore superfluous to offer further comment. The principle 

 of analysis of the correlation tables in this treatise is completely brought 

 to light. In conclusion I wish then to once more call to mind that the 

 presence of multiple correlation-diagonals indicates the ]iresence of multiple 

 factors. This is therefore again a new method of bringing this phenomenon 

 to light in close connection with all the cases in which it appeared from the 

 curve analysis of my three previous treatises. Moreover I must from my 

 heart once more earnestly point out that most of these tables, reckoned 

 according to the correlation-coefficient, would not illustrate any correlation, 



