514 



NATURE 



[January g, 19 13 



the berg. There does not seem to be much difference 

 in the conditions in the two cases so far as they are 

 given in the papers ; and as for the sea temperatures, 

 thev are alike, in both cases a little more than 4° C. 



The two cases seem to me to represent two totally 

 different conditions to which we have no clue. The 

 first condition is what we might, according to the old 

 theory, expect a ship would experience when sailing 

 past a berg with a stern wind — the berg surrounded by 

 ice-cooled water on the surface, extending one mile to 

 windward and six miles to leeward, the temperature 

 falling quickly when approaching the berg, and again 

 rising slowly as the distance increased. In the second 

 case this cold surface current is entirely absent, there 

 is no fall of temperature as the berg is approached, 

 but rather a decided rise of temperature, and the 

 observer found no diluted sea water close to the berg. 

 Still, that does not explain the difference. If the fall 

 in temperature in approaching the berg in the first 

 case was' caused by diluted sea water on the surface, 

 W'hy was there none in the second? 



As to the explanation of the rise of temperature 

 when nearing icebergs, found by Prof. Barnes, in 

 his Royal Institution discourse he attributes it to the 

 sun's action on the fresher water of the surface cur- 

 rent. He thinks that as the water of this current 

 tends to keep to the surface and get warmed by the 

 sun, it will have less tendency to mix with the lower 

 water than sun-heated sea water ; but as sun-heated sea 

 water also tends to keep to the surface, any advantage 

 in that way of the fresher water will not be great, and 

 in both cases the mixing will be determined principally 

 by the waves. Prof. Barnes, however, in his letter 

 in Nature of December 12, departs from this theory 

 of surface heating, after finding there was no 

 evidence of any weak sea water near the ice- 

 berg, and he now attributes the high temperature 

 near the ice to a surface current which he thinks 

 flows towards the ice. This current he considers will 

 prevent thp vertical circulation which elsewhere tends 

 to keep the surface of the sea cooler ; but lie does not 

 mention how, in the absence of ice, this vertical 

 circulation is produced. 



The results of Prof. Barnes's investigations are 

 extremely interesting, but they are so much at variance 

 wuth each other, and with the results obtained by 

 others with less delicate methods, that we cannot help 

 hoping he will continue his work under different con- 

 ditions of temperature, &c., so as to help us to under- 

 stand more clearly what is taking place near icebergs. 

 In investigations 6f this kind, more information is 

 required as to the size of the berg observed, since its 

 size will determine the area of disturbance, and the 

 amount of the deviation of the temperature from that 

 of the surrounding area. We also require to know 

 something about the force of the wind and the waves, 

 as they have much to do with the mixing of the hot 

 and cold waters ; and information is also required 

 as to the drift of the berg in relation to the surface 

 water. If bergs surround themselves with indraught 

 and outflow currents, then these currents will tend to 

 have definite boundaries, and rapid changes of tem- 

 perature may take place in small differences of depth. 

 Thermometers at different depths might therefore give 

 valuable information. We also require records of sun- 

 shine, as well as information with regard to the tem- 

 perature of the sea at the depth of the bottom of the 

 iceberg. 



After reading Prof. Barnes's last letter I made 

 some further investigations as to what takes place 

 while ice is melting in sea water. The experiments 

 are all laboratory ones, and therefore somewhat in- 

 conclusive, since, as I have alrcadv said, we cannot 

 reproduce all the conditions in nature. Further, I 



NO. 2254, VOL. 90] 



may mention that owing to the distance to the nearest 

 sea I liad to experiment with common household 

 salt and water of the density of sea water, 

 but this is not likely to produce any important differ- 

 ence in the results. 



The temperature circulation in the sea is profoundly 

 modified by the presence of the salts. In a lake of 

 fresh water attaining its maximum density at a 

 temperature of about 39° F. the downward currents 

 produced by surface cooling stop at about that tem- 

 perature, and further cooling of the water causes it 

 to tend to keep at the surface. In the sea, however, 

 this is not the case, as sea water does not attain its 

 maximum density until it is cooled much below 39° F. 

 If the salinitv be small, then the temperature of maxi- 

 mum density may only be lowered some two or three 

 degrees below that of fresh water. But if the salinity 

 be that of ordinary sea water, then the temperature 

 of maximum density is below the freezing point of 

 fresh water. We see from this that whenever sea 

 water is cooled by the presence of melting ice 

 it is made denser and so caused to sink. It is this 

 cooled sea water which causes Dr. Carpenter's dow-n- 

 ward current. 



One cannot help wondering what was the nature of 

 the oceanic circulation before the salts began to accu- 

 mulate, and how this circulation was gradually 

 modified and the temperature at the depths slowly- 

 lowered to what it is at the present day. 



It seems strange that when ice is melting in sea 

 w^ater we should have Dr. Carpenter's downward 

 current of cold water, while parallel with it all round 

 the ice there should be a rising current of cold but 

 weak sea water ; and one of the questions which 

 suggested itself was : Is there any intermingling of 

 the oppositely flowing streams? Does the cooled 

 weak sea water of the upward current mix with the 

 cooled downward one. In other words, does any of 

 the melted ice go to mix with and cool I he downward 

 current? To test this point some ice was prepared 

 by freezing some water which had been previously well 

 boiled to expel the air, so as to avoid any air bubbles 

 in the ice which might aid the rising current. Before 

 freezing there was added to the boiled water a little 

 aniline blue. The product was a piece of transparent 

 blue ice. This ice was moored in the sea water at 

 about mid-deplh, so that there was water both above 

 and below it. As the ice melted a stream of blue 

 water w'as seen rising from it and spreading itself 

 over the surface of the water, but not a vestige of 

 blue could be detected in the bottom water, showing 

 that the downward current is cooled by radiation. 



Another question that may be asked is : May not 

 the downward current overpow-er the upward one? 

 If the. ice goes deep down in the water the descending 

 current will gain volume and velocity as it descends ; 

 may it not, therefore, overcome and carry down the 

 melted ice current? Of course, one cannot in a labora- 

 tory get .1 satisfactory answer to such a question. The 

 best I could do was a miniature berg of about one 

 foot deep. A rod of blue ice of that length was pre- 

 pared. This was placed vertically in a tall jar of 

 sea water and the currents noted, but no difference 

 in tlie circulation could be seen. A current of blue 

 water flowed up close to the rod, and a downward 

 cold current flowed to the bottom, but no blue went 

 to the bottom of the jar. The rod, instead of stand- 

 ing vertically, was next placed at a considerable 

 angle ; but even then all the blue went to the top, 

 the cold blue water, though it no longer travelled 

 alonsr the rod, but left it and rose over the rod, fight- 

 ing its way to the surface through the clear water. 



In order to study what takes place in the neigh- 

 bourhood of molting ice the following experiment was 



