KONGL. SV. VET. AKADEMIENS HANDLINGAB. BAND. 24. N:0 ii. 159 



to the lowest temperature it could endure- without freezing, was overlapped by anothei' of 

 fresher water, ice of that kind called 'pancake-ice' by Scoresby and known as 'bottom-ice' 

 in Bohuslän, formed between the layers and rose to the surface, covering the sea for miles 

 with a thick mäss of lumps and cakes of ice. On figs. 2 and 3 plate X the gradual pro- 

 gi-ess of the outlying branches of the Baltic stream from the 17th to the 23rd of January 

 may be folJowed. 



Three days låter the sea off Väderöbod was entirely covered with bottom-ice, which 

 however was 7iot formed at the bottom, where a remnant of salt water of 32 or 33 °/gg 

 still remained with a temperature of -\- 3° or -f- 4°. The ice was in fact given off by 

 some intervening layer of less saltness (27 — 28 "/oo)- There was also another remarkable 

 circumstance in connection with this phenomenon. 



At the begiiining of January 1879 the strong NE. winds diverted the Baltic stream 

 from the coast westward and salt water of 32 & 33 %„ with a temperature of 3° or 4° C. 

 (i. e. the 'bankwater') penetrated from the central part of the Skagerack into the fjords of 

 Bohuslän N. and E. of Väderö. 



With this water the herring approached the coast. Again simultan eously with the 

 invasion of the fresh water, which began on the 13th of January in the southern part of 

 Bohuslän and reached Väderö on the löth, the herring-fishery ceased. 



B. The next layer of inflowing water of from 34 to 35 %„ salt we have called 

 North Sea-water, because it fiUs a great part of the North Sea in winter and in summer as 

 may be seen from the charts in plates IX and X where it is marked light blue. This 

 water is also subject to periodical changes with the seasons. In spring or summer when 

 the Baltic current to a great extent sweeps the 'bank-water' away from the Skagerack the 

 underlying stratum of 34 or North Sea-water swells in volume and reaches a higher level. 

 On plate V Ekman's section from HåUö to Ö. Risör shows that the isosalinic lines slope 

 more gradually in the direction of the coast in July than in February (see section 9). 

 Water from 34 to 35 "/o^ salt then approaches nearer to the Swedish coast and enters 

 channels and fjords from which it is totally excluded in the winter. Thus the periodical 

 influx of the North Sea-water into the Skagerack causes important changes in the deep 

 coast-channels and fjords of Bohuslän. One of those fjords, the Gullmare-fjord, we have 

 studied closely for a period of a year, from February 1890 to February 1891. Plates VI 

 and VII contain a graphic representation of its hydrographic conditions (lOA in February, 

 lOB in June and July, lOC in August 1890, and lOD in February 1891, see p. 88— 99). 



In February 1890 the isosaline of 34 7oo ^'^'^^ intercepted by the submariné ridge 

 at the mouth of the fjord. Consequently the water of 34 7oo ^^ ^^e deep basin of the 

 fjord was cut off from communication with the corresponding water-stratum in the open 

 sea. How long it had remained so is unknown. We found this water singularly deiicient 

 in oxygen, the liter not containing more than about l,s cc. while the carbonic acid reached 

 the unusually great amount of 50,5 to 51,5 cc. We consider this alteration in the com- 

 position of the gaseous constituents in seawater to be due to the influence of organic life. 

 In June the same year the foul water was displaced by a fresh influx of 34 water. In 

 section 10 B the line of 34 7oo reaches över the bar indicating that a new supply of water 

 enters into the fjord. Numerous analyses of the gases contained in this Avater gave about 



