184 MARIOX EXPEDITION TO DAVIS STEAIT AXD BAFFIN BAY 



of these weather-map vahies Avith the values of icebergs, shows that 

 an excess of air over the northern Xorth Athmtic has been reflected 

 eitlier in a very light, or at least a moderately' light, iceberg season 

 off Xewfoiindland during the following spring. The opposite type 

 of pressure distribution is unmistakeable in showing a greater num- 

 ber of bergs than normal, but it does not permit of subgrou})ing. 

 The plus type of pressure map, in other words, exhibits a higher 

 correlation with poor iceberg years than does the minus type with 

 correspondingly rich years. This indicates that there are other in- 

 fluences at work, such as variations in the air temperatures and in the 

 water temperatures in the far north; variations in the precijutation, 

 or perhaps sporadic phenomenon, such as an ice jam in the Arctic 

 Archipelago. 



It was found that pressure differences between various points 

 furnished the most useful values for purposes of correlation, because 

 in this method there is no room for the personal bias which may 

 enter when charts are classified according to the types. We corre- 

 lated first the annual variations of icebergs past Newfoundland with 

 the pressure gradient foi- the previous summer over the ice fjords 

 of west Greenland, the actual values employed being the pressure 

 differences, Jacobshavn minus Upernivik, as obtained from the 

 previouslv constructed isobaric maps of the Davis Strait region, 

 1880-1920. 



Tlie months were then examined separately as follows: 



The negative value of the correlation is in the right direction, 

 namely, that with jDressure lower at Jacobshavn than Upernivik, the 

 winds are directed offshore, and therefore tend to drive an abnormal 

 number of icebergs into the current that eventually bear them past 

 Newfoundland. The magnitude of the correlations are. however, 

 insignificant and therefore we mu.'^t conclude contrary to Mecking 

 (1906), that although the offshore summer gradient is real, it is 

 nevertheless of little importance, so far as effecting the dispersal of 

 bero-s .south of Newfoundland. 



