114 U. S. COAST AND GEODETIC SURVEY 
319. This form is also adapted for use with the long-period con- 
stituents. Assuming that the daily means have been seer of the 
effects of the short-period constituents (p. 89), and that these means 
have been assorted into 24 groups to cover the constituent period, 
the 24 group means may then be entered in form 194 in place of the 
24 hourly means used for the short-period constituents. Then, treat- 
ing the constituents Mm and Sa the same as the diurnal tides and 
the constituents Mf, Msf, and Ssa as the semidiurnal tides, the form 
may be followed except that the log-augmenting factor must be taken 
from table 20 and then combined with items (46) and (48) to obtain 
item (49), striking out item (47). 
320. To obtain Sa and Ssa from the monthly means of sea level, or 
tide level, the following process may be used: Enter the monthly 
means beginning with that for January in alternate spaces provided for 
the hourly means in form 194, placing the value for January in the 
space for the 0 hour. For convenience consider all the intermediate 
blank spaces as being filled with zero values and make the computa- 
tions indicated by (8) to (12) and (18) to (21). Correct the co- 
efficients of s,; and c, from 12 to 6, at top and foot of columns (9), (12), 
(19), and (21). In bottom of form enter Sa in column having sub- 
script 2 and Ssa in column with subscript 4 in order to obtain correct 
augmenting factors and strike out numerals indicating subscripts. 
For (38) and (89) take the logarithm of twice the values of 6s and 6c 
as obtained above. The ¢’s as obtained from (40) must have the 
following corrections applied in order to refer them to 0 hour of the 
first day of January—common years, Sa correction=-+14.07°, Ssa 
correction= + 28.14°; leap years, Sa correction=-+14.94°, Ssa cor- 
rection= +29.88°. For convenience in recording the results Lgels 
suggested that the ¢ as directly obtained from (40) be entered (in 
its proper quadr Sa in the space just below the logarithm from which 
it is obtained, and that the ¢ corrected to the first day of January 
be entered in the same line in the vacant column just to the right. 
The V-+u, computed to the first day of January, may then be entered 
immediately under the corrected ¢’s and the «’ of (43) readily obtained. 
For (49) the combination (46)-+ (47) will be used. 
321. Form 452, R, x, and ¢ from analysis and inference (figs. 17 and 
18)—This form provides for certain computations preliminary to the 
regular elimination process. The constants for constituents K, and 
S, as obtained directly from form 194 may be improved by the appli- 
cation of corrections from tables 21 to 26; and constants for some of 
the smaller constituents, which have been poorly determined or not 
determined at all by the analysis, may be obtained by inference. If 
the series of observations is very short, the inferred values for the 
constants of some of the constituents may be better than the un- 
eliminated values from form 194. 
322. Form 452 is based upon paragraphs 229 to 248. It is designed 
to take account of the diurnal constituent on one side (fig. 17) and the 
semidiurnal constituents on the other side (fig. 18). The amplitudes 
and epochs indicated by the accent (’) are io be taken from form 194 
and the quantities indicated by the asterisk (*) from form 244 or 244a. 
If the series is less than 355 days, values for S,; and 285M may be 
omitted. 
323. For all short series the values in columns (4) and (8) are to be 
computed in accordance with the equivalents and factors in columns 
