THE RETURN TO COLD AXU DARKNESS. 477 



dreaded is our being turned out on the floe by the loss 

 of our ship, and the knowledge that we have to trav- 

 erse two hundred and fifty miles to get to the main- 

 land, which may be safely assumed as a humanly im- 

 possible task. 



October 2od, Saturday. — Our fox is not yet caught, 

 though he visits the place and eats all he can find. 



October 24th, Sunday. — Our friend the fox again 

 visited the trap, and this time got caught in it and 

 escaped. Hair and blood and a closed trap were the 

 proofs of his being caught ; his absence sufficiently 

 proved his escape. 



There is a considerable amount of doubt thrown on 

 all observations taken during such cold weather as we 

 experience in an Arctic winter. Sextants were never 

 designed to be submitted to such contraction as they 

 now undergo in use, and there is no way to allow for 

 or remedy the changes produced in the length of the 

 arc. The greater the cold the greater the contraction 

 of course, but that gives no index error. A sextant 

 very carefully adjusted to-day, and then having an in- 

 dex correction of 30", was found after a short exposure 

 to have an index error of 4' apparently, but how much 

 the arc was shortened it would be impossible to say. 

 The mercury on the index and horizon glasses cracks 

 and splits, and Chipp is kept busy in supplying new 

 backs. For some time, over a month, I have been try- 

 ing to get some satisfactory lunars to check our chro- 

 nometers by, but they have all resulted so ridiculously, 

 and no tw T o alike, that I have despaired of getting any- 

 thing reliable. As another resource I shall break out 

 our zenith telescope, and see if it has power enough to 

 define Jupiter's satellites, by whose eclipses, occulta- 

 tions, or transits I can get chronometer errors. The 



