175 



18 20 22 24 26 28 30 32 34 



Temperature °C 



Fig. 4. Variation of apparent conduc- 

 tivity cell as a function of frequency and 

 temperature. Insert is schematic of 

 circuit containing conductivity cell. 

 Sequence of measurements was 16 C, 

 70 to 3000 cps; 20°C, 70 to 3000 cps; 

 etc. to 32°C, and then repeat as system 

 cooled to 16°C. Open circles measure- 

 ment on heating cycle; closed circles 

 while cooling. 



measurements showed a slow decrease 

 conductance was 0.3 millimho below its 

 hos are nearly equivalent to changes in 



ured in field calibrations. This change 

 would require frequency to drift by 40 

 cps. 



In addition to giving a measure 

 of the frequency coefficient of the H-cell, 

 the measurements shown in figure 4 in- 

 dicate what appears to be either a long 

 term drift in the apparent conductance 

 of the cell or a hysteresis produced by 

 heating and cooling. The two measure- 

 ments at 16°C, taken approximately five 

 hours apart, show a small increase in 

 apparent conductance. The magnitude 

 and direction of the long term drift noted 

 above was observed over a period of 

 three days by measuring the apparent 

 conductance of the cell at frequent in- 

 tervals at 70 and 1000 cps. Except dur- 

 ing the 70 cps measurements, the bridge 

 was energized at 1000 cps. During a 

 three hour period the apparent conduc- 

 tance was seen to decrease, pass 

 through a minimum, and then slowly in- 

 crease. During the three days a spread 

 of nearly 0.3 millimho was observed. 

 At the end of the three days' immersion 

 the cell was covered by what appeared to 

 be a bacterial slime. (The test solu- 

 tion was filtered sea water.) Remioval 

 of the slime by passing a pipe cleaner 

 through the cell increased the apparent 

 conductance by 0.6 millimho. Careful 

 cleaning with nitric acid followed by 

 thorough rinsing in distilled water in- 

 creased the apparent conductance an 

 added 1.06 millimhos. After cleaning, 

 for 17 hours at which time the apparent 

 original value. The changes in millim- 

 chlorinity units. 



The temperature coefficient estimated from these data is similar to that 

 predicted from the Thomas et al. (1934) measurements, i.e., 0.01 CaO.Ol 

 CI %o. 



The results of these measurements are given not for their quantitative 

 aspects but rather to show that the condition of a set of electrodes will have a 

 marked influence on the measured conductivity. The results suggest that for 

 precision in situ measurements some standardized method of cleaning and aging 

 the electrodes may be desirable. 



(3) An estimate of precision obtained by a statistical analysis of simul- 

 taneous measures from two CTI instruments was made by Mr. Blair Kinsman, 

 Instructor of Oceanography at the Chesapeake Bay Institute. He reports as fol- 

 lows: 



