356 
Part III. — Ninth Annual Be'port 
perature is observed and water collected for specific gravity from ilie light- 
house ladder at high tide, and in the deepest water at the side of the 
boat jetty at low tide. The observations refer to the surface water only 
and at 9 a.m. to air temperature as well. This commenced on December 
24, 1889, and with the exception of a fortnight in September 1890, the 
month of December 1890, and part of February 1891, they have been 
carried out with regularity. 
Making allowance, as far as this is possible, for the months in which 
the observations are defective, the mean surface temperature of the water 
for 1890 was 8°-0 at 9 a.m., 8°-4 at 3 p.m., and for the first three months 
of 1891 4°'l at 9 a.m., 4°'6 at 3 p.m. The highest temperature observed 
w^as 13° '8, on September 8, 1890, at 3 p.m. ; the lowest observed was 
2° -4, on January 10, 1891, at 9 a.m. The lowest temperature recorded 
in 1890 was 3°-6, reached on March 3, at 9 a.m. 
The results for this station are a little puzzling. The mean tempera- 
ture is nearly a degree"lower than that at any other station, and on 
examining the curve, comparing the temperature at Oxcar with that at 
Bell Rock (I., Plate XVII.), one is struck by the remarkably low maxi- 
mum, and the early date at which it is recorded. The diagram is ruled 
with red lines, to show temperatures at each degree on the centigrade 
scale, and for convenience of comparison is also ruled in blue at each 
5 degrees Fahrenheit. It summarises Table V., where the actual mean 
temperature of each ten-day period is given. Table YI. gives the monthly 
mean temperature, which is taken as the mean of the three ten-day periods 
comprising the month. The curve of monthly means is shown on a large 
scale in Diagram IV., on which each degree centigrade is ruled in red 
lines, and each degree Fahrenheit in blue lines. This diagram gives the 
change of monthly mean temperature at each station for the whole period 
of observation, and the singular character of the Oxcar curve is very pro- 
nounced. As might be expected from the position of the station, much ex- 
posed to the influence of fresh water from the Forth and Almond, the 
minimum in February and January, when the rivers are at their coldest, 
are lower than those of any other station. But the same argument would 
lead us to expect that in August and September, when the rivers are at 
their warmest, the maximum of the Oxcar surface temperatures would be 
higher than that of any other station. The reverse is indicated by the 
record. The maximum monthly mean, IT'O, occurred in June, and was 
lower than that found anywhere else. In the absence of any record 
for September, it is unsafe to say anything decided as to the trustworthi- 
ness of the observations in July and August, but it seems quite possible 
that the readings in these months were too low. The air temperatures 
recorded show that there was no cold weather sufficiently pronounced to 
account for the low temperature of the water ; and since the low readings 
affected the afternoon as well as the morning observations, it is possible 
that the thermometer may have been temporarily deranged. 
In the average of ten-day periods the temperature was always a fraction 
of a degree higher in the afternoon than in the forenoon, an effect evi- 
dently due to the heat of the sun. It is obvious, however, that with 
observations taken six hours apart, the tide is at opposite phases, and any 
effect which might be produced by tidal difference would be masked in 
ten days, and entirely neutralised in a month, supposing that the effects 
produced by flood and ebb tide are equal and opposite. 
In order to discriminate the tidal effect the whole record was gone 
over carefully, and all the observations within an hour and a half of high- 
water and of low-water were copied out for special consideration. The 
result was as follows :— 
