Tidal Plienomena at Inland Borelioles near Cradock. 81 
the level records for purposes of general comparison. Accordingly the 
barograph curve for each week was traced under the water-level curve. I 
have also indicated by insertion of the usual conventional signs the times 
of the moon's phases during the fifteen weeks. 
A comparison of the barograph and water-level curves shows very 
clearly and satisfactorily that (if we ignore the semi-diurnal fluctuations) 
the mean water-level line in general rises when the barometric pressure 
falls, and vice versa. Thus in the beginning of the first week the baro- 
meter falls and the mean water-level rises, while about the end of the 
week the barometer rises and the mean water-level falls. 
In the second week, during March 10, 11, and 12, the barometer falls 
and the mean water-level rises. On March 13 and 14 the barometer is 
higher and the mean water-level correspondingly lower. During March 15 
and 16 the barometer is lower and the mean water-level higher. This 
connection between the two lines obviously continues throughout most of 
the fifteen weeks, any marked rise or fall of the barometer being accom- 
panied by a corresponding fall or rise of the mean water-level. The week 
from June 8 to June 15 appears to me to be noteworthy as being the only 
week in which the above rule seems to fail. 
During June 10 and 11 the barometer shows a 24-hour period of high 
pressure without any obvious corresponding depression of the mean water- 
level, and on June 14 and 15 the mean water-level seems to be falling 
without any appearance of a barometric rise. It must be borne in mind 
that the barograph record is to be interpreted as indicating in a general 
way only the time of marked variations. I cannot regard its quantitative 
indications as being reliable within inch. This opinion is based on 
several weeks of daily comparisons which I have made of the indications 
of this barograph instrument with readings of a Kew mercurial barometer. 
I conclude, however, that the series of correspondences indicated on these 
fifteen weeks records is sufficiently notable and consistent to justify the 
belief that no theory of fortuitous coincidence will satisfactorily account 
for them, and that the records must be reasonably held to establish 
the general truth of the conclusion — 
" Increase of barometric pressure at Tarka Bridge is accompanied 
by a lowering of the general ivater-level m Borehole No. V., and vice 
versa." 
An examination of the semi-diurnal fluctuations indicated on the water- 
level record shows that these fluctuations attain a maximum amplitude 
during the few days around each of the following dates : March 3, 
March 18, April 1, April 16, April 29, May 15, May 30, June 13, and that 
the amplitudes as recorded attain minimum and practically vanishing 
values in the neighbourhood of each of the following dates : March 8, 
March 25, April 8, April 23, May 8, May 22, June 6. 
6 
