(*} 
For the period treated (Dec. 1902— April 1905) holds: 
J=18°33\ v = 0°.31 
with which we find' with the aid of the wellknown tables of 
Borgen for the astron. coefficient of the K tide composed of (1') 
and (2) 
q" = 0.4732, 
from which follows for the theoretical value 
K x = 0".005037, P = 0".001869 
4. Already 15 years ago I have pointed out in “Studien over 
Getijden in den Indischen Archipel”, that for the determination of 
the constants of the tides K x , P and K 2 the trouble of an arrange¬ 
ment of the hourvalues according to the angular velocities of these tides 
is superfluous and that we can deduce these constants with equal 
accuracy and little labour directly out of the monthly means. 
The application of this method has furnished good results not only 
for the Indian tides but also for the determination of the tidal 
constants on the Dutch coasts where all three tides are very 
small. 1 ) Especially if as in this case hour-observations are at hand, 
the calculation is exceedingly simple, for the diurnal variation can 
be represented in its variability in the course of the year with great 
approximation by the expression: 
+ + . (3) 
where 
= Xi — F 0 , ■ V 0 = h 0 — v' — ~ — 204°. 71, 
= X,- v \ , v \ = - h 0 + ^ = 155° 16 ; 
/io = length of the sun at the commencement of the time (epoch) 
(294°.84) i. e. in this case January 16; v' is a small correction, 
caused by the inclination of the orbit of the moon with respect to 
the ecliptic. It is clear that the inaccuracies committed here, namely 
I the angular velocities holding for one day being all taken equal 
instead of respectively 
15° , 15° -}- 7i , 15° — 7] 
And 2 the monthly means being regarded as 12 equidistant points, 
can have no perceptible influence on the result of the calculation. 
Omitting the first term we can bring (3) into the form : 
ftudes des phenomenes de maree 
sur les cotes neerlandaises I. Utrecht, 1904, 
