On the Logarithmic Frequency Curve and its 
Biological Importance 
by 
C. J. BAART DE LA FAILLE. 
When a study in biological statistics has advanced so 
far, that its results are expressed in the form of a frequency 
table or a variation polygon, we must by no means think 
that we have attained our end. It is desirable that these 
results should give us a deeper insight into the biology 
of our material. 
In this connection I will not here treat of the coefficients 
expressing the degree of variability, but only of the con- 
clusions to be drawn from the way in which the frequen- 
cies are distributed. 
The fact that the broken line of a polygon does not 
satisfy our mind, and that we feel intuitively inclined to 
smoothen it into a curve, is justified by the following two 
considerations. First, there are irregularities due to the 
restricted number of individuals; with a similar but more 
numerous material these would gradually disappear. Secondly, 
the necessity of dividing the line of abscissae into a con- 
venient number of intervals, gives to the line connecting 
the summits of the ordinates a broken character that is 
not inherent in the measured quality. In general we 
may suppose that the ideal distribution of frequencies 
would show a fluent line. 
Now we may assume the existence of an analytical 
curve which would completely coincide with the ideal 
Recueil des trav. bot. Néerl. Vol. XII. 1915. 23 
LIER 4 
NEW Y{ 
BOT ANR 
UARHLH 
