84 
Transactions of the Boyal Society of South Africa. 
and low water turning-points for the periods, during which I obtained 
simultaneous records at Tarka Bridge in 1905. 
On these tide-gauge records the vertical scale is 1 inch to the foot of 
water-level movement, and the horizontal or time scale is 1 inch to the 
hour. The records are in monthly lengths of 60 or 62 feet each. 
On Schedules II., III., IV., and V. are presented the results of my 
measurements on these curves for the period extending from June 3 to 
June 20, and these figures may be compared with those obtained from my 
June records at Tarka Bridge, and shown on Schedule I. In this connec- 
tion it may be well to remark that the determination of the exact time 
of any individual high or low water point on a tide curve is by no 
means easy. 
The curve in the neighbourhood of a turning-point is as a rule so flat 
that two observers might differ in their estimate of the time in some cases 
by as much as a quarter of an hour or even more. G. H. Darwin remarks 
(on page 223 of the work already cited) that if the water rises only about 
a foot from low to high water in the course of five hours it is almost 
impossible to say with accuracy when it was highest, and two observers 
might differ by half an hour or even by an hour. It must be borne in 
mind that whereas the difference in the coastal tide curves amounted to 
several feet the fluctuations of level on the Tarka Bridge records amounted 
only to an inch or two. Hence, while I am of opinion that my determina- 
tions of the time of turning-points on the coastal tide curves are in most 
cases accurate within a quarter of an hour, I cannot hope to have attained 
similar accuracy in the determination of the times of high and low water 
on the Tarka Bridge curve. Any individual determination may easily be in 
error by half an hour, and occasionally by an hour or even more. These 
errors, however, tend to disappear in the average periods calculated over 
the entire fortnight. 
On Diagram No. 8 the data contained in the Schedules I., II., III., 
IV., and V. have been plotted in parallel lines to the same time scale, and 
this diagram exhibits the general similarity of the Tarka Bridge curve to 
the coastal tide curves. 
The principal dissimilarities would appear to be — 
(1) The scale of the semidiurnal amplitudes. These amplitudes 
in the Tarka Bridge curve range through a couple of inches, while in 
the coastal curves they range through several feet. 
(2) The amplitude of the diurnal inequality seems to be relatively 
greater in the Tarka Bridge curve than in the coastal curves. 
The following averages are easily calculated from the scheduled 
data. 
