Porter — The Growth of St. Louis Children. 277 
TABLE No. 1. 
THE Seay eis oF 2,000 MEASUREMENTS OF THE SAME QUANTITY WHEN 
TION OF THE INDIVIDUAL OBSERVATIONS FROM THE TRUE 
Vann es THE Muiainluns QUANTITY IS DUE TO PURELY ACCIDENTAL 
Caus 
d=the Probable Deviation. Number of Observations. 
way th 
> 
os 
PP LEE LLP tae +et+4+t++4+ 
Om Oo DORR O 
a a 
ne 
RL 
SaOmMONomonNs HOUSNSUNOHD 
AAAGAGARAA 
« 
nn 
a nn 
aan 
Om boo 09 £9 
VMAAVUAAIAAa 
O bo bo 0 Co I OF OF 
SHAOROMOUSAS HOUSUSUOUDS 
AARAAAAAAR ASQAQAAAAgkaRrs® 
bo 
So 
~ 
cal 
s 
EFEREE ERT EL B+++4+4+4+4+4++4+ 
om mE WN Ne HO 
Greater than 
This result is quite independent of the nature of the thing 
measured, provided that no unvarying influences are at work. 
An examination of Table No. 1 shows (1) that the distribu- 
tion of the observations is symmetrical about a median point, 
in other words that equal deviations to one and the other side 
of this point are equally probable; (2) that half the whole 
number of observations fall within + 1.0 d of the most prob- 
able value, indicating that a small deviation is more probable 
than a large one; and (3) that there is a limit beyond which 
no deviation occurs. Such are the peculiarities of a series of 
observations in which the deviations from the true value 
are due to purely accidental causes. It is evident that the 
most probable value of the true measurement is typical of the 
whole series, and that the degree of probability of any devia- 
* After Thoma: Untersuchungen iiber die Grésse der anatomischen Be- 
standtheile des menschlichen K6érpers im gesunden und kranken Zustande. 
Leipzig, 1882. Page 28 
