6io 



NA TURE 



[April 2%, li 



chlorine, ice forms at - i°"9 C. The following results 

 are from means of closely-agreeing results : — 



Freezing temperature - 2'o 

 Per cent, chlorine I '94 



Difference O'o6 



I '5 

 I '445 

 0-05S 



0-963 

 0-037 



-o-S 



0-475 

 0-025 



Sea-water resembles a chloride of sodium solution, 

 containing the same percentage of chlorine, and the 

 resemblance is closer the greater the dilution. When 

 the beaker was removed from the freezing-bath, the 

 temperature rose during melting as it had fallen 

 during freezing. In these experiments, which had for 

 their object the determination of the temperature at 

 which the crystals melted, as well as that at which they 

 began to form in the water, it was impossible to remove 

 a sample for analysis large enough to enable the sul- 

 phuric acid to be determined in it. 



For this purpose a series of observations were made, 

 using quantities of 300 grammes of sea-water. Freezing 

 was continued usually until the temperature had fallen 

 o°-3 C. below that at which crystals began to form. The 

 mother-liquor was then separated from the crystals by 

 means of a large pipette with fine orifice, before removing 

 the beaker from the freezing bath. The magma of crys- 

 tals was then brought rapidly on a filter and drained by 

 means of the jet pump. The ice, thus drained, was then 

 melted, and the three fractions were analysed. In the 

 following table (I.) the results of four experiments are 

 given. In the one column (W) will be found the weight 

 of the original water taken and of the fractions into which 

 it was split in freezing ; in the other (R) will be found the 

 ratio of SO3 to CI found by analysis, the chlorine being 

 set down as 100 ; thus, in I, the percentage of chlorine 

 found in the crystals, melting at the lowest temperature, 

 was I '497, and that of the SO3, 0-174 ; 'he ratio (R) is 

 therefore 11-62. 



It will be seen that the ratios (R) found for mother-liquor, 

 drainings, and ice agree with one another quite as 

 closely as those found in samples of pure sea-water from 

 different localities. It is to be remembered that in these 

 experiments the water was frozen gently, that is, the rate 

 of abstraction of heat was low, the temperature of the 

 freezing bath being regulated so as to be about 2°_C. below 

 the freezing temperature of the solution. Much of the 

 error and uncertamty about the freezing of saline solutions 

 arises from the violence of the methods employed. 

 Judging then by the constancy of the relation of the per- 

 centage of CI to SO3 we see that in sea- water, frozen at 

 moderate temperatures, the composition of the saline 

 contents of the original water, the mother-liquor and the 

 ice is identical ; and we are justified in concluding that it 

 is probable that the saltness of the ice is due to unfrozen 

 and concentrated sea- water adhering to it. Ice forming 

 in even very weak saline solutions closely resembles snow 

 (which is ice forming in air), and has the same remark- 

 able power of retaining mechanically several times its 

 weight of water or brine. 



A strict account was kept of the heat removed from the 

 sea-water while the freezing was going on. In Table II. 

 will be found the number of heat-units (gramme ° C.) 



removed during the freezing in the case of Nos. III. 

 and IV.; and this number, divided by 79-25, gives the 

 weight of pure ice, which could have been formed at 

 0° C. by the removal of heat. 



Table II. — Calculation oficeformeJ. 



Weight of original water (grammes) 



Per cent. CI in ditto 



Percent. CI in mother-liquor 



Weight of mother-liquor W p= 



Weight of ice W- L = 



Mean freezing temperature (° C) 



Heat abstracted (grammes ° C) 



Equivalent ice formed (grammes) 



- 2.05 

 I4230 

 53-4 



239-3 

 60-7 



Sea-water, like other saline solutions, is easily cooled 

 several degreesbelow itsfreezing-point before crystals begin 

 to form. While cooling down to and below what was known 

 to be its freezing-point, simultaneous obser\-ations of the 

 temperature of the sea-water and the freezing bath were 

 made from half-minute to half-minute. From these obser- 

 vations, the rate of abstraction of heat for different differ- 

 ences of temperature of sea-water and bath was found. 

 At a given moment a minute splinter of ice (weighing 

 much less than a drop of water) was introduced. Crystals 

 immediately began to form, and the temperature rose in 

 from ten to fifteen seconds to the freezing-point. During 

 the freezing the temperatures of bath and sea-water 

 were observed at regular intervals. The heat re- 

 moved is thus made up of that eliminated during the 

 few seconds when freezing began and the temperature 

 rose to the freezing-point, which is found by multiplying 

 the rise of temperature by the weight of liquid, and that 

 removed during the subsequent cooling, which is deduced 

 from the duration of the operation and the rate of loss 

 of heat observed before freezing commenced. The 

 specific heat of the solution is taken as unity. In the 

 table are further given the weight of the sea-water used, 

 the percentage of chlorine in the original water and in 

 the mother-liquor, the weight of the mother-liquor on the 

 assumption that it contains all the salt of the original 

 water, and, by difference, the weight of the ice formed. 

 The agreement between the two quantities of ice formed 

 as calculated by the different methods is as close as could 

 be expected. 



It has thus been shown that the composition of the 

 saline contents of the ice formed as above described is the 

 same as that of the original water, and this of itself is 

 almost conclusive that the salt is contained in adhering 

 brine and not as a solid constituent of the ice. Assuming 

 that this is so the amoimt of ice formed as deduced fron-» 

 the composition of the mother-liquor agrees well with the 

 amount deduced from the thermal exchange taking place 

 during the freezing. 



It has, moreover, been proved by Guthrie, Riidorff, and 

 others, that, in solutions of the salts occurring in sea- 

 water, ice does separate out at first, and continues to- 

 separate out until the concentration has become many 

 times greater than that of sea-water. Assuming that in 

 sea- water all the chlorine is united to sodium, 87 per 

 cent, of the water would have to be removed as ice before 

 a cryohydrate would form, and if it contained nothing 

 but sulphate of soda in the proportion corresponding to 

 the sulphuric acid formed in it, over 90 per cent, of the 

 water would have to go as ice, before the cryohydrate 

 would be formed. 



In my experiments, about 15 per cent, of the weight of 

 the water w-as frozen out as ice, causing a lowering of 

 freezing-point by o°-3 C. In nature it is probable that 

 the ice forming at the actual freezing surface does so at an 

 almost uniform temperature, the local concentration pro- 



