124 
BAUER. 
conform with the rules of official correspondence between 
nations. Many a well and carefully executed magnetic sur- 
vey in the past has had its full importance for world-wide 
investigation destroyed because of the possibility of error in 
the secular variation corrections which must he applied to 
bring its results up to the date of the later data. 
Though I may be judged guilty of defending my own 
policy, I believe the course pursued by the Carnegie Insti- 
tution of Washington in conducting the general magnetic 
survey of the globe is the only way in which this particular 
project, and similar ones to it, could not only he expe- 
ditiously conducted, but so as to realize the chief objects of 
the work. Judging from individual expressions received 
from scientific men everywhere, they appear in agree- 
ment with us. This policy, briefly stated, is: To make, with 
the aid of the friendly and harmonious cooperation of all 
concerned, a rapidly executed magnetic survey of the greater 
part of the globe, so that a general survey, all-sufficient for 
the solution of some of the great and world-wide problems 
of the earth’s magnetism, will be completed within a period 
of ten to fifteen years. At a smaller number of points, se- 
lected in consideration of the prime questions at issue, the 
observations are to he repeated at intervals of five years or 
less, in order to supplement the rather sparsely distributed 
magnetic observatory data. Thus the determination of the 
corrections for reduction of the general work to any specific 
date is continuously provided for. 
Now, had I the time or were this the place, I should like 
to add a paragraph regarding the needful accuracy and the 
prime questions to be considered in the conduct of such a 
piece of work. Suffice it to say that the most evident result 
of all magnetic work in the past is that, for the purposes of 
a general survey, it is far better to make some sacrifice in 
accuracy if thereby it is made possible to secure observations 
at another point. In other words, the errors due to local 
disturbing conditions are far greater than the purely obser- 
vational ones. ITence multiplicity of stations rather than 
extreme accuracy and laborious methods of observation and 
reduction is the prime requisite. 
