KUNOL. SV. VET. AKADEMIENS HANDLINGAR. BAND 58- N:0 3- 



Table VII. 



45 



X 



2/x(Pearson) 



2/*((35)I) 



Observation 



3,5 



110,90 



117,oo 



115,25 



4,5 



117,66 



117,72 



116,96 



5,5 



118,42 



118,49 



117,47 



6,5 



119,24 



119,30 



119,10 



7,5 



120,08 



120,12 



120,30 



8,5 



120,93 



120,96 



121,63 



9,5 



121,78 



121,79 



121,72 



10,5 



122,62 



122,6 1 



122,82 



11,5 



123,42 



123,39 



123,14 



12,5 



124,18 



124,15 



123,89 



13,5 



124,88 



124,84 



124,86 



14,5 



125,52 



125, is 



125,71 



15,5 



126,07 



126,05 



126,16 



16,5 



126,52 



126,53 



126,53 



17,5 



126,87 



126,92 



126,91 



18,5 



127,09 



127,19 



127,02 



19,5 



127,18 



127,35 



129,56 



20,5 



127,1 1 



127,38 



123,82 



21,5 



126,88 



127,20 



126,50 



22,5 



126,48 



127,00 



125,25 



Example 7. Pearson's illustration C. On the skew correlation between size 

 of cell and size of body in 2010 Daphnia magna. This is a case where no regression 

 formulae that are given by vvhole rational functions can give a ny very good fit. 

 Indeed, the regression, for certain reasons mentioned by Pearson, varies periodically. 

 Hence, neither Pearson's nor our curves will give a really good description of 

 the data. 



For the moments and correlation table we refer to Pearson's memoir. We 

 have (x size of body; y size of cell) 



m l = 8,5025 



er, = 3,8648 



/?30 = + 0,0148 

 fi 2l = +0,1594 



r 30 = + 0, 1 4 I 



171, = 9,2686 

 V, =2,5417 



r = +0,3963 



Po* 



0,0441 

 0,0535 



r i0 = + 0,0164 



Now, here, it will make very little difference whether (35) I or II is used. 

 We give here the curve (35) II. It is 



Vx = 9,6 2 90 + 0,2 4 4 6 (x — »i,) — 0,02 4 1 ('X — Wi,) 2 + 0,0001 7 (x — mj 1 '. 



