6322 HYDROGRAPHIC MANUAL PaGE 582 



the air, it may cause an error of as much as 5P C. The canvas bucket, used for the 

 surface haul, should be constructed as if it were two buckets, one inside the other, with 

 an air space between for insulation; the outside should be painted so that it sheds 

 water. The haul must be made quickly and the temperature taken with a quick- 

 acting thermometer as soon as the sample is on deck. 



6322. Frequency of Observations 



For use in echo sounding and R.A.R, it must be assumed that the velocity of sound 

 is relatively constant throughout certain areas and during certain periods of time. 

 For economic reasons the values used cannot be changed too frequently. The length 

 of time during which the velocity may be assumed to be constant will vary from an 

 entire season to a 2-week period on the working ground, and the area will likewise 

 vary from that surveyed in an entire season to that surveyed in a 2-week period. This 

 assumes that the temperature and salinity of the water remain relatively constant. 

 A sufficient number of serial temperatures must be observed so that the velocity may 

 be known with the accuracy specified in 6313. The number required depends on the 

 physical characteristics of the water and the physiography of the area. The proximity 

 of an ocean ciu-rent or the discharge from a large river would necessitate more frequent 

 observations, in which particular attention would have to be given to the salinity 

 determinations. 



At least one complete serial temperature should be observed in the deepest part of 

 the area surveyed each month, for use in computing corrections to echo soundings. 



Where little is known about the regional variation in temperature, the first monthly 

 deep-water serial temperature of a season should be accompanied by a second one in 

 deep water at a different locality, down to depths where the temperatures of the two are 

 in approximate agreement. In the limited areas ordinarily sm'veyed during 1 month 

 it will usually be found that this will occur comparatively near the surface (generally 

 within 75 fathoms of it) in the layer where the influence of the seasonal change is 

 greatest. Where this is the case, other serial temperatures taken dm-ing the month 

 may be restricted to the upper layers, possibly to the depth limit of the bathythermo- 

 graph. The hydrogi-apher should endeavor to determine the regional and seasonal 

 variation in an area with a minimum number of serial temperatures at positions 

 distributed throughout the area. 



Physical conditions of sea water are rarely so stable that one serial temperature 

 \vill suffice for the duration and area of a survey. Practicall}^, however, even assuming 

 such a condition, additional observations must be made at intervals to prove the 

 stability. Wliere there is little regional variation and only a slight seasonal change, 

 a minimum of four or five well-distributed serial temperatures each month should 

 suffice. Where the physical characteristics are extremely variable, both regionally and 

 seasonally, as during the summer months along the Atlantic Continental Slope in 

 the vicinity of Nantucket Island, Massachusetts, a minimum of about 15 serial temper- 

 atures should be observed each month. In such conditions it is better to take too 

 many observations rather than too few to ensure that the requirements specified 

 in 6313 will be met. 



For most localities there is insufficient knowledge of the physical conditions to 

 enable the hydrographer to predict the temperature gradient or to extrapolate a tem- 

 perature curve far beyond an observed value, so he must depend on his own observa- 

 tions. But prior surveys in an area or published studies by other organizations may 



