134 



THE RATE OF GROWTH 



[CH. 



Let us sum the same figures up, so as to show the whole number 

 above or below the respective sizes. 



Our first set of figures, the actual measurements, would give us 

 the '*courbe en cloche," in the formrof an unsymmetrical (or "skew") 

 Gaussian curve: one, that is to say, with a long sloping talus on 



Extreme 

 Decile 



Quartile 



Median 



Quartile 



Decile 

 Extreme 



Fig. 25 B. 



18 



33 



100% 

 90 



75 



.50 



36 39 



21 24 27 30 

 Length in mm. 

 'Curve of distribution" of a population of minnows. 



one side of the hilj. The other gives us an "S-shaped curve,'' ap- 

 parently hmited, but really asymptotic at both ends (Fig. 25 B) ; and 

 this S-shaped curve is so easy to work with that we may at once divide 

 it into two halves (so finding the ''median" value), or into quarters 

 and tenths (giving the "quartiles" and "deciles"), or as we please. 

 In short, after drawing the curVe to a larger scale, we shall find that 

 we can safely read it to thirds of a miUimetre, and so draw from it 

 the following somewhat rough but very useful tabular epitome of 

 our population of minnows, from which the curve can be recon- 

 structed at aiiy time: 



mm. 



13 



210 



25-3 



28-6 



30-6 



32-3 



Extreme 

 First decile 

 Lower quartile 

 Median 

 Upper quartile 

 Last decile 

 Extreme 



39 



