508 



i:Xl'HDITIOX TO POINT BARROW, ALASKA. 



The monthly means of the bifllar readings appear quite irregular, produced by large dis- 

 turbances and by change in adjustment. The latter became necessary in consequence of 

 the effect of temperature and moisture on the suspension. During the winter the observatory 

 became thickly coated with ice ou its sides and roof, which during thawing weather kept the inte- 

 rior atmosphere in a state of extreme moisture. The observed variations in the length of the 

 suspension fibers and in the torsion of the two declination instruments may be thus accounted for, 

 and the greater or less stiffness of the fibers was probably occasioned by moisture deposited upon 

 it freezing and thawing alternately. The effects on the readings of changes of temperature and 

 gradual loss of magnetism* of the magnet or of such secular change are small compared with the 

 above irregularities from other causes. It would seem desirable to use metallic suspension in the 

 place of silk. 



The September mean (619.5) was corrected to 519.1 by application of a rough correction of 

 318 divisions to the readings of the first six days, found by comparison with the mean of the suc- 

 ceeding six days. 



In August, 1883, the mean reading was higher (639.7) than at ;iny other time, and it was evi- 

 dent that the adjustment of the instrument had from some unknown cause been disturbed. One 

 of the observers (Mr. Maxfield) states that when he took down the instrument on the 27th he found 

 that the adjusting screw which holds the thread and determines the distance between the threads 

 worked rather loosely in its bearings, whereas it was very tight when the instrument was first set 

 up. It is difficult to fix upon a particular time when the rapid increase iu the readings com- 

 menced, but it was most probably between August 7 and 8, and lasted for two or three days before 

 the instrument settled again to a fixed condition. A slow, progressive motion is apparent from 

 the last two days of July. For our present purpose the matter is of little importance, since we 

 shall deal strictly in a ditt'ereutial way, only aiming at roughly comparable absolute readings. In 

 order to reduce the monthly readings during August roughly to a uniform scale a correction of 

 187.0 divisions was applied. 



of monthly mmii rallies (in<-liixirr, of Hlxturhii AWX iiiut inic'in-cc/cd for cJiii<j<s of tem- 

 perature anil rariation* in ncnlc ctt I tie*) of the hourly reailinys of the J-irouke ln'fihir matjnetomt'ti-r at 

 Uglaitmir, Altutka, 188ti-'s:5. 



Mean. 



-.11).! 

 187.8 



100.4 



:i;7. u 



501. t 

 41U.4 



The 



j ure now over thirty years old. They wiv HMM! ;it Washington iu Ib53. 



