284 Trans. Acad. Sci. of St. Louis. 
ber of series in which the distribution follows a common 
type. It will suffice for our purposes therefore to compare 
the distribution of the 2192 heights in Table No. 3, page 279, 
with the distribution of an equal number according to the 
calculus of probabilities. The method by which this is done 
permits the construction of a curve from the observations 
actually made which shall be the probable form of a curve 
representing the distribution of the entire class from which 
the observations have been drawn. 
The number of observation (Z) which should be included 
between the average (A) and any deviation from the average, 
in other words any multiple (m) of the probable deviation 
(d@), is obtained from the equation.* 
t 
ee (4) 
The labor of calculating deviations with the aid of this 
equation is avoided by the use of such tables as Stieda’s, 
reproduced below. 
TABLE No. 6. 
Strepa’s TaBLE ror CALCULATING THE NUMBER OF OBSERVATIONS AT ANY 
DISTANCE FROM THE MEAN OR AVERAGE WITHIN THE LIMITS: M+ 
Pp Percent. 1) Percent. 
0.1 5.4 1.8 77.5 
0.2 10.7 1.9 80.0 
0.3 16.0 2.0 82.3 
0.4 21.3 9.1 84.3 
0.5 26.4 2.2 86.2 
0.6 31.4 as 87.9 
0.7 36.3 2.4 89.5 
0.8 41.1 2.5 90.8 
0.9 45.6 2.6 92.1 
1.0 50.0 7 93.1 
ti 54.2 2.8 94.1 
1.2 58.2 2.9 95.0 
1.3 61.9 3.0 95.7 
1.4 65.5 3.5 98.2 
1.5 68.8 4.0 99.3 
1.6 71.9 4.5 99.8 
17 ‘74.8 5.0 99.93 
Sesaeperenmameremme ee 
| Kramp. L’ Analyse des réfractions astronomiques. 
