CLYDE SEA AREA. 685 



salinity of the seaward portion of the Area was considerably reduced — a sudden increase 

 of rainfall having occurred over that region, which in September was rainier than 

 the landward portion. The wind was blowing up Loch Fyne during the time it was 

 examined, and the salinity diminished steadily and uniformly from mouth to head. The 

 surface density at Otter was unusually high, being equal to that in the south of the 

 Arran Basin ; it was observed shortly after half-flood, when the tidal race was at a 

 maximum. In all the other lochs, except Loch Goil, the wind was blowing strongly 

 down, and the maximum surface density occurred at the head of Loch Strivan, Loch 

 Long, and the Gareloch, in all of which densities were found greater than any occurring 

 on the surface of the rest of the landward portion. The bottom density at Inveraray 

 was lower (1*02404) than had ever previously been found, and, as at stations farther 

 down Loch Fyne, it was above the average, and at the stations nearer the head of the 

 loch attained an absolute maximum, it appears possible that some mistake may have been 

 made in collecting the sample or reading the hydrometer. Another anomalous bottom 

 density of this trip is that at Kilfinan (1*02544), which is greater than was found any- 

 where else in the Area, and equal to that of the channel south of Sanda. 



General Result of the Salinity Trips. — Table XVIII. summarises the actual results 

 for each station, as calculated in the Observation-Book, and printed in detail in the 

 Appendix. In order to present a fairer idea of the average conditions the data were 

 corrected in one or two cases where there seemed to have been an error of observation 

 or a mistake with regard to labelling a sample, and by careful comparison of neighbour- 

 ing observations on the maps for each trip the densities for stations which for any reason 

 had been omitted were supplied. These numbers are printed in Tables XIX. and XX., 

 the interpolated or otherwise manipulated figures being printed in a special type. In 

 the case of July 1887 for the landward division, and May and July 1887 for the seaward 

 division, the interpolated data are so numerous as to deprive the column read vertically 

 of much of its value ; but read horizontally it adds considerably to the probability of 

 accuracy in the station means for the whole period. 



[Table 



