22 University of California Puhlications in Zoology [Vol. 15 



anyone, who has not been in actual charge of such work, could antici- 

 pate the extent to which trouble is nearly always encountered. In 

 some instances the source of trouble is apparent, in others it is verj' 

 obscure, while in still others it remains an unsolved mystery. In the 

 first category may be noted that, after a haul has been made with the 

 Kofoid water bottle from depths exceeding three hundred meters, the 

 temperature of the water in the upper part of the bottle dili'ers by 

 two or three degrees from that in the lower part, thus causing con- 

 siderable trouble in determining the correct temperature (see p. 14). 

 In the last category may be mentioned cases in which a Richter ther- 

 mometer suddenly balks and, like a stubborn donkey, absolutely re- 

 fuses to work until it has had its rest. Most of this trouble, however, 

 does not lead to error and has little bearing upon the reliability of 

 the data, but some of it relative to hauls made with the Kofoid and 

 Nansen nets as well as to the preservation of water samples does lead 

 to error and must, therefore, be discussed. 



(a) The Kofoid net. — In using this net one is at times struck by 

 the almo.st total absence of organisms obtained during a particular 

 haul. In some such instances it has been found that the spring of the 

 opening jaw had broken or slipped so that the first messenger failed 

 to open the net and the second reversed both jaws at the same time. 

 In other instances the net when thorough!}^ tested on deck operated 

 perfectly, the presumption therefore being that it failed to capture 

 anything because there were no organisms to be captured. During 

 July, 1912, a method was devised (see p. 13) by which an Ekman 

 current meter could be suspended below the Kofoid net in such a way 

 that it would be started when the net opened and- stopped when it 

 closed. "With such a combination it is clear that, if any time elapsed 

 between the opening and closing of the net, the current meter would 

 show a registration and thus a definite means was at hand for deter- 

 mining whether or not the net could actually be open and filtering 

 water without catching any organisms. 



A large number of hauls were made with this api^aratus, distributed 

 in depth from the surface to 550 meters, with the result: (1) that no 

 instance came to light in which there was an absence of organisms 

 combined with a current-meter registration; and, (2) that in a num- 

 ber of cases the current meter failed to register, thereby showing that 

 no haul had been made. These facts forcibly indicate that the net, 

 when properly opened, may be depended upon to make its capture and. 

 consequently, whenever an "empty haul" is discovered, it may be 



