Porter — The Growth of St. Louis Children. 279 
If the individual heights of a group of adults are found to 
be approximations of a middle, typical height, it would seem 
that the heights of children of the same sex, age and class 
must show a similar relation to a type; and this inference is 
justified by observation. 
| TABLE No. 3. 
OBSERVED DISTRIBUTION OF THE HEIGHTS OF 2192 St. Louis ScHOOL 
GIRLS, AGED 8. 
Heights at Intervals of 2 Centimetres. Number of Observations. 
141 and 142 cm. 1 
166: 4 24007! 
22 Giles SAE os he Ay 1 
IG tea SE 5 
TRS RE 10 
TSE LBs SF 21 
129 “© 180 *6 28 
120 Soee ss 79 
12644: ARG ot * 138 
123! ye ts 183 
ASL 19227 $8 243 
EIS 120 tt 342 
BET St 338 321 
£15 16. 297 
118 fF 114. 222 
PEP O18 *t 137 
JOD. fA 
107: **= 108; -* 42 
105: 5 105. 27 
108.4 104.2 8 
8a) MRR raie 11 ee, 3 
10D 26 L 
Tothliiv<scvcus cies bane Peery 3th 
The characteristics of a series in which the individual ob- 
servations are accidental deviations from a typical middle 
value are established by methods which have long been used 
by astronomers and mathematicians. These methods seek to 
determine (1) the middle or typical value and (2) the way in 
which the individual observations are dispersed on either side 
of this value. Some statisticians take the average to be the 
nearest approach to the typical value, others prefer the mean, 
i. e., median value.* It will be necessary to consider at some 
length the relative value of average and mean in anthropomet- 
rical studies of the growth of children, but for the present a 
* Throughout this work the word MEAN is used as the synonym of MEDIAN 
VALUE, and not in the sense of ARITHMETICAL MEAN, which is called here 
the AVERAGE. The MIDDLE VALUE will be employed as meaning either 
Median or Average value. 
