86 Transactions of the Boyal Society of South Africa. 
The difference in these two averages seems too great to be due to 
observational errors and raises a strong suspicion that the Tarka Bridge 
lag at low water is really greater than the lag at high water. Now if this 
is really so the curve records at Tarka Bridge will be slightly steeper 
when rising from low to high water than when falling from high to low 
water, and this is a matter which is easily investigated, the necessary data 
being contained in Schedule I. 
If we take the difference of each time of low water as given in 
Schedule I. and the time of the succeeding high water we find that the 
average time interval between low water and the succeeding high water at 
Tarka Bridge is 5 hrs. 53 min. Then taking the time interval between 
each high water and the succeeding low water, we find the average 
interval to be 6 hrs. 35 min. 
A comparative inspection of two sides of each individual wave repre- 
sented in these lists of half-wave intervals makes the conviction stronger 
that we are here dealing with no imaginary phenomenon, for in the list of 
27 ascending half-waves and 26 descending half-waves we see that in 20 
individual cases the ascending half has a shorter interval than the 
descending half immediately following. In two cases the intervals are 
practically equal, and in the remaining four cases the ascending interval 
is greater than the descending interval. In each of the anomalous cases 
an inspection of the adjacent intervals strongly suggests that an unusually 
large error has occurred in the determination of the culminating point of 
the curve, and this would entirely account for the apparent anomalies. 
I conclude that the inference is a fair one that in general a shorter 
time elapses between low water and high water than between high water 
and the succeeding low water of the well. In other words, at Tarka 
Bridge the water rises more quickly than it falls, and thus the rise and 
fall of this well water is to that extent analogous to the tidal fluctuations 
in an estuary or tidal river rather than to the tides in the open ocean. As 
will be seen later, this result is important as having a bearing on the 
questions arising as to the nature of natural mechanism by which the 
tidal influence is transmitted from the ocean to the well, if such a theory 
of the origin of Tarka Bridge tides be admissible. 
The correlation of a particular high water on the coastal tide curve 
with a particular high water on the Tarka Bridge curve cannot be regarded 
as satisfactorily accomplished, and although in the foregoing discussion 
I have stated the lag of Tarka Bridge behind East London to be for 
high waters 14 hrs. 27 min., all I have really proved is that the lag is 
14 hrs. 27 min. + p . 12| hrs., in which expression _p is a constant 
integer of undetermined positive or negative value. 
It is indeed open to question whether such attempts at correlation are 
not from their very nature wholly illusory, since there is as yet no clear 
