362 
Part III. — Ninth Annual Beport 
by means of another line from the gallery. The thermometer dips into 
a considerable depth of water even at low tide, and at high water it is 
phiced in the most advantageous position possible for recording tempera- 
ture, unaffected by disturbing causes. 
Observations were commenced on October 26, 1889, and have been 
carried on with great care, the record being one of the most trustworthy. 
Unfortunately, however, the exigencies of fog-signal duty caused many 
gaps in the observations during the winter months, and, as appears from 
the record in Table V., many periods of ten days are without observations, 
and in several the mean is that of a few days only. The 3 p.m. observa- 
tions are less frequent and regular than the 9 a.m. It happens that the 
position of the Bell Kock, remote from land influences, prevents any 
sudden or great vaiiations in physical conditions from making themselves 
felt, and the uniformity of the curve of mean ten-day temperatures in 
Diagram I. is of itself a testimony to the correctness of the work as a 
whole. 
The average temperature of the water at 9 a.m. for 1890 was 9°*0, and 
at 3 P.M. 9° '2 — figures which accord closely with the readings at the Carr 
Rock and at Abertay. For the three first months of 1891 the mean 
temperature at 9 a.m. was 5^T, and at 3 p.m. 5°-2. The highest reading 
recorded was 14°'6 at 3 p.m. on September 8, 1890, and the lowest 4°*4- 
at 9 A.M. on March 27, 1891, the lowest in 1890 being 5°T, showing a 
range of 10° "2, slightly greater than at the North Carr. 
The curves of change of temperature show well-marked minima in 
March and a distinct maximum early in September, the warming and 
cooling of the water taking place between these dates with remarkable 
regularity. The record may be taken as giving a fairly accurate view of 
the variations in temperature of the North Sea within 10 miles of land, 
and those for Avinter are practically the only data on the subject that have 
ever been obtained. It is difficult to over-estimate the value of the 
physical work done here in a scientific sense, yet greater regularity in 
observing, had that been practicable, would have further enhanced its 
importance. 
As far as can be judged from the comparison of isolated cases, there 
was no difference perceptible as a result of tidal phase, but the gaps in 
the record made it appear inadvisable to undertake the labour of a com- 
plete tidal classification, such as was carried out in the other cases. 
The specific gravity observations are equal in value to those of tem- 
perature, and are unique in their importance as the only record of the 
kind vet obtained. The average specific gravity of the whole period 
from October 1889 to December 1890 was 1026'0 (corresponding to 
3*37 per cent, of salts) ; that for the first three months of 1891 was 
1025*8. No difference appears in the monthly averages between fore- 
noon and afternoon observations. The highest specific gravity recorded 
Avas 1026-7, on August 9, 1890 ; the lowest, 10247, on February 1st of 
the same year. This shows remarkable constancy of salinity. The 
curve of monthly mean specific gravity at the Bell Rock, shown in 
Diagram III., shows a distinct seasonal variation. It is below the 
average from December to June, and above the average from June to 
December, the minimum appearing to occur in February and the 
maximum in August. Tlie salinity is higher, the range less, and the 
variations much more uniform than at any other station. There is no 
apparent influence produced by high and low tide, the salinity be'ng 
sometimes less and sometimes greater at low tide than at high tide. 
This indicates that the comparatively fresh surface-water of the Tay and 
the less fresh and more thoroughly mixed water of the Firth of Forth are 
