G74 DR HUGH ROBERT MILL ON THE 



like the previous discussion, that the densities may, in dealing with large averages, be 

 trusted to the fourth decimal place, and have some meaning in the fifth. 



The hydrometer work done by Dr Gibson* with an instrument in every way similar 

 to that employed in the work under discussion, when examined with reference to his 

 halogen determinations and Dittmar's tables, showed a mean deviation of 0'000157 

 for 22 determinations made at sea, but a considerably better result — deviation of 

 0*000087 — for 46 determinations in the laboratory. In both sets, when the sign of 

 the deviation is taken into account, the average showed the density determined by 

 hydrometer to be 0*000047 above the calculated. 



This in a general way confirms my results, and shows that the hydrometer is a 

 thoroughly satisfactory instrument, provided the reasoning based on its use does not 

 assume accuracy beyond the fourth decimal place. 



Observers and Observing Stations. — The determinations of density and calculation of 

 results were made by myself personally in most cases, but I have to acknowledge the 

 assistance of Mr J. T. Morrison, M.A., on the trips of April and June 1886, while the 

 observations in August 1886 were made by him alone. Mr H. N. Dickson assisted me 

 on the trip of September 1886, and many of the calculations were subsequently worked 

 out by Mr J. B. Clark. To those three gentlemen I have to record my thanks. 



The observing stations for salinity were not quite so numerous as those at which 

 temperature was ascertained, nor were the observations so regular and systematic, as it 

 was impossible to determine the densities in rough weather, and impracticable to keep 

 a great number of samples. The average number of density determinations on each 

 trip was about 65. Particulars respecting the 37 stations of which the most regular- 

 records have been secured are given below in a compressed summary. The exact 

 position is given by bearings, but as a rule the observations were made in positions 

 which varied slightly, being usually fixed by one bearing on land, and confirmed by the 

 depth. Reference to the plates on which the cross-sections of these stations are shown 

 is added. They are all shown on the accompanying key-map (Plate I.) : — 



Place. Bearing, Magnetic, and Distance in Sea Miles and Tenths. 



Gareloch, (Longitudinal Section 17, Plate 9) — 



Garelochhead. — Mid-channel, \ mile from head of loch, 

 Shandon. — Mid-channel between Shandon and Mamore Farm, 

 Row II. (inward). — Beacon on Spit, S. \ mile, 

 Roio III. (seaward). — Beacon on Spit, N. \ mile, 



Loch Goil, (Longitudinal Section 7, Plate 7) — 

 Lochr/oilhead. — Pierhead, E. by S. \ mile, . 

 Stuckbeg. — Stuckbeg Farm, E.S.E. \ mile, . 



Loch Long, (Longitudinal Section 6, Plate 7) — 

 Arrochar. — Church, S.E. J mile, 

 Thurnbank. — Thornbank House, S.E. \ mile, 



II'i.Y Loch, (Longitudinal Section 5, Plate 7) — 



Kilmun. — Pier, E.N.E. \ mile, . . . . . 10 



* Sixth Annual Report of Fishery Board for Scotland, 1887. 



ction. 



Plate. 



D 



epth. 



17a 



9 



10 fathoms 



17b 



9 



21 



jj 



17c 



9 



23 



» 



... 





12 



tJ 



8a 



7 



27 



Jl 



8b 



7 



40 



») 



16a 



9 



10 

 35 



ft 



