138 



NATURE 



[December i, 1910 



iceberg is passed. In nearly every record we have there 

 is a small rise of temperature above the surrounding sea 

 temperature before the fall occurs, which seems charac- 

 teristic of an iceberg effect. 



In the light of the microthermograms we have obtained, 

 the usual method of taking temperatures at sea seems 

 decidedly inadequate. Thus, even if temperatures are 

 taken over the side of a moving ship every fifteen minutes, 

 readings are obtained at about two- to three-mile intervals, 

 which obviously cannot be of much value in determining 

 the temperature gradient characteristic of an iceberg. 

 They might easily, as some of our charts show, indicate 

 a rising rather than a falling temperature. The ordinary 

 marine thermometer, with a degree one-eighth of an inch 

 long, would miss entirely temperature effects made per- 

 fectly plain by the microthermometer. The persistence of 

 a temperature gradient in the direction of a ship's^ course 

 is one thing which I think can be relied on to give ice- 

 berg warnings, but when the whole temperature drop is 



temperatures show no diurnal variation, except in the 

 former case when near land. It seems to be well knowi> 

 to biologists that small temperature variations in the sea 

 may be set up by the existence of marine life, and it is 

 difficult to think of any other cause for what we have 

 observed. It is interesting as indicating how important 

 a part marine life probably plays in the conservation of 

 solar energy. H. T. Barnes. 



McGill University, October 27. 



Fig. 2. — Microthermogram of the Temperature of the Sea. 



fewer than two degrees in six miles it is evident that very 

 sensitive thermometers must be used to detect it. 



Besides the disturbing influence of ice, the proximity of 

 land within a few miles produces effects of great magni- 

 tude as compared with the remarkably small variations 

 of temperature in the open sea. In our case this was, no 

 doubt, caused by the cold under-currents being turned up 

 by the shoals and shore line of the Labrador coast. 



For hydrographic work, the determination of current 

 boundaries could be made with great exactness from a 

 comparison of the temperature traces and the determination 

 of ship's position. 



The small inequalities in the temperature of the sea 

 stand out in strong contrast to the uniform temperature 

 of the St. Lawrence River just after the ice has moved 

 out in the spring. These inequalities of temperature 

 suggest at once the possibility of a vertical circulation set 

 up by convection currents, which must be an important 

 factor in the retention of the solar energy absorbed by 

 the sea. It explains why our records of air and sea 



NO. 2144, VOL. 85] 



Dun Coat Colour in the Horse. 



My attention has been directed to a letter in Nature 

 of November 24 over the signature of Prof. J. Wilson. 

 He disputes the accuracy of certain extractions from " The 

 General Stud Book," which originally appeared \n The 

 Veterinary Record, in my paper on the inheritance of dun 

 coat-colour. Prof. Wilson also states that in the Stud 

 Book entries there is a considerable 

 element of doubt. This would appear 

 to be the usual attitude of his mind in 

 relation to data which do not exactly 

 fall in with his own theories. 



Let me first take the case of the 

 mare Silverlocks (foaled 1725). I most 

 emphatically deny that " the Stud 

 Book assumes " that this mare, which 

 is described as a chestnut on p. i, 

 vol. i., is identical with a mythical 

 chestnut mare which Prof. Wilson says 

 was foaled exactly a hundred years 

 later. The animal to which he prob- 

 ably alludes was foaled in 1824, and is 

 entered in the third volume as a " bay 

 colt Silverlock," by Blacklock out of 

 Sheba's Queen. I would direct Prof. 

 Wilson's attention to the fact that the ' 

 chestnut mare Silverlocks (1725) is the 

 only mare of that name in the first 

 four volumes of the Stud Book, and 

 that the Stud Book entry is perfectly 

 authentic, since it was extracted from 

 an early Racing Calendar which de- 

 scribes Silverlocks as a chestnut mare 

 by The Bald Galloway, out of a sister 

 to Chaunter. This mare was raced for 

 some years, and there is no room for 

 doubt that she was the dam of the dun 

 colt Buffcoat, foaled 1742, and of his 

 two dun sisters, foaled in 1738 and 

 1739 respectively, all three being by 

 The Godolphin Arabian (brown or bay). 

 Does Prof. Wilson mean to imply that 

 Lord Godolphin was so dishonourable 

 as to run Buffcoat under a false pedi- 

 gree, for that in effect is what he would 

 have us believe? I have already 

 directed his attention in a private 

 letter to the circumstance that I 

 have a portrait of Silverlocks (1725) 

 which shows her an unmistakable chestnut, and not a dun. 

 In regard to the dun filly Sarah Curran (1892), by 

 Robert Emmett (bay or brown), out of Cellulites (black). 

 Prof. Wilson is certainly misleading, for he fails to 

 disclose the fact that Messrs. Weatherby distinctly state 

 in vol. xviii., p. 727, that "this mare erroneously 

 appeared in the last volume as dead." Now, whether 

 Prof. Wilson likes it or not, the breeder of Sarah Curran, 

 Mr. J. T. Hartigan, returned this mare as a dun. I 

 judge he was in a better position to form an opinion con- 

 cerning her pedigree and colour than my critic, who never 

 saw her. 



Prof. Wilson says that the filly (1886) by Lord Gough 

 (bay) out of Danseuse (brown) is described as a bay. 

 Here again he does not state the whole truth. As a matter 

 of fact, this filly was returned as bay when a foal, but 

 her breeder specially altered the colour to light dun in 

 vol. xix. She had then reached maturity, and was a 

 brood mare. 



I do not wish to take up your space in quibbling as 



