butlon based on the latest available data and extrapolation, or 

 (c) a predicted distribution based on climatological data for the 

 particular month desired and tempered by the latest information 

 and by experience. It is believed that, if a mass of data as 

 mentioned above was collected and such a technique developed, 

 it would be of great value in delineating one of the main effects 

 upon the thermal structure of the ocean. With this effect evalu- 

 ated, it would then be possible to proceed to the study of other 

 factors. 

 III. Data 



The single feature that has had the greatest effect on the 

 course of this investigation has been the lack of data suitable 

 for conducting it. When the approach to be employed in this in- 

 vestigation was first decided upon, the type of data that would 

 be necessary to employ this approach immediately became clear. 

 It was evident that most of the data that were available from 

 various sources had two serious defects: (a) The data were obtain- 

 ed by the classical technique; that is, both time and distance 

 were allowed to vary in obtaining the serial observations, and 

 (b) none of the data contained measurements of a sufficient num- 

 ber of parameters to evaluate completely all the effects on the 

 thermal structure. 



The first of these defects is by far the more important. 

 It is the opinion of the investigators that, if problems such as 

 this are to be solved eventually, more attention m.ust be given 

 to eliminating the factor either of time or distance from the 

 collection of data. One of the first things that becomes evi- 



68 



