169 
when m is uneven, of = when m is even; in the latter 
case however the coéfficient (5) assumes a negative sign, thus 
compensating the effect of the quantity = which remains 
between the brackets of the sine. 
When applying Honxsrzix's method the quantities under con- 
sideration are arranged according to the numbers 24—28; in 
the most unfavourable case R is equal to 25.5; then, for 
K= 26, the coëfficient (5) assumes the following values: 
after one year т == 14, coéflicient — 0.881 
» two years m= 28 » == 0.575 
» three » m= 42 » = 0.204 
» four » т-- 56 » = 0.087 
by which it is shown that, when the operation s extended 
lo a period of only two years, the amplitude has been reduced 
to about half its original value. When, in confirmity with the 
facts, we assume for R a value equal to 25.8, the values of 
the coëfficient are for K= 26: 
after one year 0.976 
» two years 0.920 
» three » 0.851 
» four » 0.715 
» five » 0.580 
У » six » 0.459 
and, when a six-yearly series of observations is employed, 
lor the different values of Ж: 
X 44 m = 91 coéfficient = 0.045 
25 в. 0.088 
26 84 0.459 
2 81 0.045 
98 78 0.045 
A second objection to which this method is open , is that after 
arrangement according to the numbers 24 — 28, the am- 
Plitudes of the resulting curves are calculated in the usual way 
