160 ANNALS NEW YORK ACADEMY OF SCIENCES 



with the letters C, (D), E and F, expressing that way my view that the 

 variations are identical. 



However, a very important difference must be noted at once ; it is the 

 difference in time of the occurrence of the same pleionian crests and 

 antipleionian depressions. The Pikers Peak maxima occur ten or eleven 

 months later than those of the Port Darwin curve. Xow, the distance 

 between the crests C and E is in both cases 67 months. Some of the 

 details of the crest C of Pikers Peak, as well as of the depression that 

 followed, may be easily observed on the corresponding crest and depres- 

 sion of the Port Darwin curve. If, therefore, the missing pleion of the 

 Pike's Peak curve appears very plainly on the Port Darwin curve, as it 

 does, we are justified in presuming that the interval of more than 5 years 

 separating the crests C and E on the Pike's Peak was indeed abnormal, 

 as well as the depression K, and that this anomaly must be ascribed to 

 the Krakatoa dust veil. 



The crest (D) of Port Darwin is not developed to its normal value. 

 On the diagram the dotted line indicates the portion which must be con- 

 sidered as having been cut away. The anomaly begins with the consecu- 

 tive mean of October, 1882, to September, 1883, that is to say, just one 

 month sooner than on Pike's Peak. This anomaly is not ten or eleven 

 months in advance on the corresponding detail of the Pike's Peak curve, 

 but just one month and occurs one month later than the Krakatoa erup- 

 tion — ^when the dust veil reached Port Dar^dn, after having traveled 

 twice around the world along the equator. The duration of the anomaly 

 extends from the mean of October, 1882,-September, 1883, till the mean 

 of February, 18 84,- January, 1885, or 17 months. The same figure may 

 be adopted in the case of the Pike's Peak curve. 



An important question arises now. Is it possible to estimate the lower- 

 ing of temperature due to the presence of the Krakatoa dust veil? Ac- 

 cording to the dotted lines of the diagrams the lowering of the tempera- 

 ture for the consecutive mean of September, 1883,-August, 1884, may 

 have been 1.9 °F. in the case of Port Darwin and 3.4 °F. in the case of 

 the Pike's Peak observations. But this is an estimate of no scientific 

 \alue. The departures of Tables I and II are of no help. It would be 

 necessary to know what these departures ought to have been. I imagine 

 it would be possible to attempt the calculations by tracing maps and by 

 comparison of the temperature conditions of the dust-affected regions 

 with those over which the dust veil was not spread out. But even in that 

 case comparisons would be most difficult, because we do not know how 

 the dust veil affected the general atmospheric circulation or how the 

 abnormal conditions of one region affected the temperatures of other 

 regions mechanically. 



