214 I'nivcrsitii af Califoniia I'ublkalions in Zoology [Vol. Lj 



hauls were also eoutaminated, and that the correspouding tempera- 

 tures and salinities were erroneous. 



The precise reason for such i-epeated failure is unknown. It is 

 true that the guiding sleeves (text-figs. D and E, s') of the cover of 

 the Kofoid bottle had become worn during years of previous use, and 

 it is po.ssible that the trouble was largely cau-sed by the undue "play" 

 of these sleeves combined with a possible inequality in tension of the 

 thermometer and water-bottle trip-springs, and the usual inclination 

 of the cable. It does not seem likely that the trouble was entirely 

 due to the worn sleeves, since the bottle rarely failed to close properly 

 prior to the addition of the attachments. 



However, whatever the cause of defective operation may have been, 

 the relations observed between the temperatures and corresponding 

 salinities could not have arisen had the apparatus operated properly. 

 On the whole, the hydrographic results can be accounted for only on 

 the assumption that the depths at which the thermometers reversed, 

 those at which the water-sample bottle closed, and those at which the 

 Kofoid bottle closed differed materially and in an erratic way from 

 each other and from the depths indicated by the amount of cable out. 

 In fact, the depths indicated by temperatures and those indicated by 

 the corresponding salinities proved to be so discordant as to preclude 

 estimating, even roughly, the correct depth in any one case. For this 

 reason, while all the subsurface hauls, because of their biological value, 

 are entered in table 3. the corresponding hydrographic data are not 

 tabulated. 



■2. ERROES IN TEMPERATURE, SALINITY, AND DEPTH 

 All hydrographic ob.servati(ms are subject to several important 

 sources of error. Some errors are inherent in the instrument used, 

 and some are due to occasional mistakes, but the most difficult to 

 detect are consequent upon the impossibility of constructing any piece 

 of oeeanographic apparatus that will not occasionally operate defec- 

 tively. The causes of defective operation are exceedingly numerous, 

 and, as stated formerly (1915, p. 21) : "we doubt if anyone who has 

 not been in actual charge of such work could anticipate the extent to 

 which trouble is nearly always encountered." Even though all prac- 

 ticable precautions are taken, important errors unexpectedly arise, 

 making it necessary to examine critically each series of observations 

 in order to detect the erroneous ones and to reject those that may be 

 misleading. 



