July zi, 1884] 



NA TURE 



on June 23, in which he suggests a connection between the red 

 sunsets and the frequent rains. During the latter half of the 

 past winter the rains were incessant in the Atlantic States of 

 America, and the writer suggested that they were due to the 

 volcanic dust in the atmosphere, in a letter published in the 

 Philadelphia Public Ledger of February 23. In a subsequent 

 issue, March 8, he called attention to Dr. Aitken's researches. 

 Subsequently Prof. Heilprin, of the Philadelphia Academy of 

 Natural Sciences, offered a similar suggestion. 



Philadelphia, July 16 Chas. Morris 



THE SALTNESS AND THE TEMPERATURE 

 OF THE SEA 1 



PROFESSOR DITTMAR'S researches, an account of 

 which forms Part I. of this volume, have finally 

 proved that, so far as the most refined analysis can go, 

 the mixture of salts dissolved in ocean water has attained 

 a state of chemical equilibrium. But, although there is 

 constancy of proportion between the various salts, the 

 ratio of the total salts to the water varies considerably in 

 different parts of the ocean. 



The great evaporation in the dry tropical areas and the 

 removal of water by freezing in the Polar seas tends to 

 increase the salinity in these places, while in the tropical 

 zones of continual rain and in the Polar fringes where the 

 icebergs melt, there is constant dilution going on. The 

 determination of the salinity at different places and depths 

 is of great oceanographic importance, and the problem of 

 finding the salinity has been attacked in various ways. 

 The most simple and straightforward is to evaporate a 

 weighed portion of the water to dryness and weigh the 

 residue, but this cannot be done without chemical change 

 taking place. The magnesium chloride present decomposes 

 with the water into magnesia and hydrochloric acid, and 

 all the carbonic acid of the carbonates is driven off. Gay- 

 Lussac showed long ago how to avoid the error due to the 

 dissociation of magnesium chloride, but no means have yet 

 been suggested for taking account of the carbonates in a 

 total salt determination. Direct weighing being thus 

 found inexpedient, the next best method would appear to 

 be to find the exact amount of any one element present, 

 and by means of a table of complete analysis, taking 

 advantage of the constancy of composition of ocean salts, 

 to convert that into the salinity by multiplication with a 

 constant factor. This is the method which Prof. Dittmar 

 prefers, and for the purpose he estimates the chlorine or 

 rather the total halogen by means of his refinement of 

 Volhard's process. When the salinity of water has to be 

 determined at sea, this delicate method cannot be con- 

 veniently employed, and it has been customary hitherto 

 to measure the specific gravity of the water very carefully 

 and afterwards to reduce the results to salinities. An 

 attempt has been made with considerable success in the 

 United States to substitute the determination of the re- 

 fractive index of water for that of the density, and thence 

 to deduce the salinity by a formula. This method is pre- 

 eminently adapted for use at sea, but it appears not to 

 possess the necessary delicacy. 



The only method by which the specific gravity of a 

 fluid can be ascertained on board ship is by means of 

 hydrometers, and as the extreme values for sea-water are, 

 according to Mr. Buchanan, ro2/So and ro24oo, appa- 

 ratus of great delicacy must be employed. A very delicate 

 glass hydrometer was used on the Challenger, yet in 

 spite of its fragility and the thousands of observations 

 that Mr. Buchanan made with it in all weathers, he suc- 

 ceeded in carrying the one instrument which he had used 

 during the entire voyage back to this country unbroken. 

 His description of the hydrometer is as follows : — 



1 "The Physics and Chemistry' of the Voyage of H.M.S. Challenger." 

 Vol. i. Part ii. " Report on the Specific Gravity of Ocean Water " By J. Y. 

 Buchanan, M.A., F.R.S.E. Part iii. "Report on the Peep-Sea Tempera- 

 ture Observations obtained by the Officers of H.M.S. Challenger during the 

 Years 1873-76." (London : Longmans and Co., 1884.) See Nature, 

 July 24, p. 292. 



" Preliminary calculations showed that convenient 

 dimensions would be about 3 mm. for the diameter of the 

 stem and about 150C.C. for the volume of the body, and 

 from 10 to 12 cm. for the length of the stem. The tube 

 for the stem w : as selected with great care from a large 

 assortment, and no want of uniformity in its outward 

 shape could be detected with the callipers. The tube for 

 the body of the instrument was also selected from a 

 number, in order to secure such a diameter as would give 

 the instrument a suitable length. In order to provide 

 against accidents, I had four instruments made from the 

 two lengths of tubing. The glass work of the instru- 

 ment being finished — except that the top of the stem, 

 instead of being sealed up, was slightly widened out 

 into a funnel — the instrument was loaded with mercury, 

 until the lower end of the stem was just immersed in dis- 

 tilled water of 16° C. A millimetre scale on paper was 

 then fixed in the stem, and the calibration carried on by 

 placing decigramme weights on the funnel-shaped top, 

 and noting the consequent depression on the scale. The 

 whole length of the scale was 10 cm., and this portion of 

 the stem proved to be of perfectly uniform calibre. Several 

 series of observations were made in order to determine 

 accurately the volume of any length of the stem. . . . 

 When this operation of calibration was finished, the end 

 of the stem was carefully closed before the blowpipe." 



The constants necessary for making a specific gravity 

 observation were all determined with the utmost care. 

 They included the exact weight in vacuo of the instru- 

 ment, the volume of the body, the volume of each division 

 of the stem, and the expansion of the whole instrument 

 for a degree Centigrade. These data were fully tabulated, 

 and in addition tables were made of the total weight when 

 each of a set of brass weights was placed on a small table 

 that could be slipped over the top of the stem. These 

 weights were necessary, as, without them, the stem would 

 require to be of great length in order to serve for waters 

 of different density. 



In making an observation Mr. Buchanan always kept 

 the water sample in the laboratory for a night in order 

 that it might have time to attain the temperature of the 

 surrounding air. He then placed about Soo c.c. in a 

 glass jar supported on a swinging table, and immersed the 

 hydrometer in it after ascertaining its temperature exactly. 

 To insure the greatest possible accuracy two readings 

 were frequently made with different weights on the table, 

 the results separately reduced, and the mean taken as the 

 density. The density was calculated in every case by 

 ascertaining the weight of the loaded hydrometer and 

 dividing it by the immersed volume, which is calculated 

 from the temperature and stem-reading. 



Prof. Dittmar examined very particularly into the 

 probable error in reading Buchanan's hydrometer, and 

 after a long series of experiments, described in the chapter 

 on Salinity in Part I., he came to the conclusion that the 

 difference between the salinity as calculated by it and by 

 his direct chlorine determinations [i.e. Xi ~ X "here x 

 stands for the permilleage of chlorine) amounted to 

 - '042 + S, S being a variable the chances of which being 

 greater or less than '06 are equal, and are 4 to 1 in favour 

 of its being less than '12. The value of x is usually 

 between 19 and 20. 



At first Mr. Buchanan reduced his specific gravities to 

 the temperature of IS°"S6 C. by Hubbard's tables, but 

 Prof. Dittmar, in the course of his investigation of " The 

 specific gravity of water as a function of salinity, tempera- 

 ture, and pressure," succeeded in constructing a much 

 better table in which the variation of the coefficient of 

 expansion with the salinity of the water is taken account 

 of, and all Mr. Buchanan's results published in this 

 volume have been calculated by it. A very ingenious 

 graphic method of comparing Hubbard's results with 

 Dittmar's and converting one into the other is given in 

 Plate I. of Part II. 



