5» 



Records of the Indian Museum. [Vol. XXIII, 



detail. I calculated the grouped moment- coefficients directly 

 upto m« with 50 mm. as the unit of grouping. I find 



n 2 = 1*82 10 42 



H = " °'47 73 43 77 

 p 4 = ii"94 16 22 08 



^ 5 =- 77 8 23 



fi 6 = 12974 38 42 48 



Thus 



/3 a = 3-601 



and &= 21-474 



Since ft is greater than 3, it is necessary that 3ft should be 

 greater than 5ft. Actually we find 



3ft =64-42 20 

 while 5ft' 2 = 64-83 60, so that 5ft' 2 is > 3ft 



Thus no real solution is possible in this case. But we must 

 note that there is some tendency towards a solution of this type. 

 I do not propose to draw any inference from this result. I have 

 not yet analysed the other frequency curves and so I am not in 

 a position to either confirm or refute this tendency towards a very 

 special type of splitting up. 1 



Goodness of fit with Stun of Dissected Components. 



First component : — 



Mean stature = 16 5679 mm. 

 S.D. =<r, = 74-12 mm. 

 Total = w, = 150 



Second component : — 



Mean stature =16 56-79 mm. 



S.D. =<* = 40-89 32 mm. 



Total = n : = 50 





1 



First 

 Component. 



I5-97 53 

 31 "31 



33 82 

 17-26 74 



723 61 



II 

 Second 

 Component. 



0.04 71 



[•46 22 



: 129 66 



22*93 l() 



12-42 61 



177 15 



•06 47 



I + 11 (Total) 



Observed 





(m—m'y 2 



1 heoretical 





m u> m . 





7)1 . 



;;/. 





m' 



6-58 18 



8 



1 *4 1 8 



•30 55 



I7-43 75 



14 



3-43 75 



77 



42*61 18 



45 



2-38 82 



•13 07 



62-36 45 



Oo 



2-36 45 



•08 90 



44-06 43 



48 



V33 57 



•24 91 



19-03 89 



20 



•96 1 1 



•04 85 



730 08 



5 



2-30 08 



•72 40 



»' = /. 





**=2'22 57 



J Since going to press, I have obtained expressions lor the Probable Errors 

 <>t the Component Frequency ConstantSj which confirms the non-significant 

 character oi the dissection in tin- present case. I hope to publish these new 



lulae foi Probable Errors at an early date. 



