Report of the U. S. Coast Survey for 1883. 49 
machine are to be computed beforehand for each station, each 
year of observations by itself according to the author’s plan, 
and that to derive the constants from one year’s observations 
will take an experienced computer several months, we see that 
the efficiency of the machine has been counted from where this 
long preliminary computation left off. Practically, therefore, 
its present efficiency is very materially less than 30 computers. 
When once the tidal observations at a station have been 
completely put through the laborious harmonic analysis and 
the machine constants calculated, then it may be that the ma- 
chine can’ predict for that station in one day as much as it 
would take a computer 80 or 40 days to do. 
A comprehensive understanding of this valuable paper is 
only to be obtained from a careful study of the paper itself 
and of the five sketches with which it is illustrated. 
for the length of the Yolo base line. This base line, it wi 
Yolo base line 
Fee in these reports from year to year show, in spite of good’ 
‘ting, marked differences in this res 
. 
