UNDER THE :\IIDNIGnT SUN. 335 



had passed the pnme vertical (about 5.25 p. m.), and 

 as the day closed had tumbled to 1.5°, — a disagreea- 

 ble reminder that one May day does not make a sum- 

 mer any more than one swallow does. 



Mr. Dunbar having in his wanderings come across 

 two sets of bear-tracks about three miles from the ship, 

 along a small lead in the ice, a trap was sent out and 

 set for them. They had evidently caught a seal asleep 

 and eaten it, for blood was on the ice in various jDlaces. 



A calculation of the amount of the leak, or, in other 

 words, the amount of water pumped over board, re- 

 sults in placing it at 300 gallons an hour, — a vast im- 

 provement over February 20th, when 1,647 gallons per 

 hour were pumped out of the ship. The cause of the 

 decrease can only be conjectured, for we may not know 

 it for some time. Either that portion of the forefoot 

 which we assumed to have been broken has been shoved 

 back by ice pressure and closes the leak, or our cement 

 and other material are doing; more toward checkino; the 

 inward flow of the water than we had counted on. I 

 am sufficiently grateful, however, for the saving of fuel 

 thereby resulting, to prevent me from finding fault with 

 either cause. 



May 2d, Sunday. — Since placing the compasses in 

 the binnacles on Friday, I have carefully watched them 

 to get a deviation table made up from my magnetic 

 bearings from the ice. A very curious feature has been 

 observed in connection w^ith them. As the temperature 

 falls each night (sometimes getting down to single fig- 

 ures) the needles are drawn to the right several degrees, 

 and as the temperature increases in the morning they 

 gradually go back again, resuming a normal position 

 when the temperature, generally speaking, is 15° and 

 over, as the needles of the compasses (Ritchie's liquid) 



