2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 60 



were tightly corked. As a rule samples of the surface water were 

 also taken whenever the ship was stopped for dredging. 



On account of the motion of the ship and because of the great 

 amount of other more pressing work to be done it was not possible to 

 work out the density of the water specimens at sea. They were there- 

 fore allowed to accumulate in the laboratory until a harbor or quiet 

 anchorage was reached when the accumulation was disposed of. 



The variation in the reading of my salinometers resulting from 

 changes in temperature was not known. But any error arising from 

 such variation was eliminated so far as possible by taking the read- 

 ings always in approximately the same temperature, the water 

 samples being stored in a room in which the temperature was at all 

 times fairly constant. 



There are three important sources of error involved in the method of 

 storing water samples in cork stoppered soft glass bottles until such 

 time as determination is possible. The water will dissolve a certain 

 amount of the glass, thus increasing the density ; irregularities from 

 this source, however, are probably negligible so far as my results are 

 concerned, because of the large probable error involved in the method 

 used. Again the date and time of day, or the station number, were 

 written on paper labels and placed in the water ; of course the sizing 

 and certain other of the constituents of the paper and of the pencil lead 

 dissolved out, increasing the salt content of the water ; but the labels 

 used were of uniform size and composition, and the amount of writ- 

 ing was always the same so that approximately the same error is 

 involved in each reading and may be considered as compensated in 

 the general correction applied. The third source of error lies in the 

 evaporation through and around the corks ; occasionally there is 

 an evident discrepancy due to this cause (as for instance in Nos. 9 

 and 12), but ordinarily the error was probably so small as to be, in 

 view of the probable error as calculated, negligible. In the rough 

 northern waters it was often necessary to retain water samples for 

 several days ; but here it was cold and damp and evaporation was 

 reduced to a minimum ; in the south determinations could be made 

 much more frequently, and this source of error was largely 

 eliminated. 



The apparatus used in determining the specific gravity of the 

 water was Hilgard's ocean salinometer, and the specific gravities 

 were all reduced to the standard temperature of 60 ° Fahr. (i5°.56 C). 

 The tables employed in this reduction are given in the Report of 

 the Commissioner of Fish and Fisheries for 1883 (1885), p. 78. 



