212 Universitjj of California Puhlicafious in Zoology [Vol. 15 



3. The covers (co.) which project in front of the orifices of the 

 nets when open (fig. B), may modify the pressure-distribution of in- 

 flowing water so as to reduce significantly the rate of filtration. 



How serious these disadvantages are remains to be determined. 

 As stated on page 209, the second can be largely overcome by pressing 

 out instead of casting the metal parts of the apparatus. Moreover, 

 the preliminary collections made during May, 1915 (water samples 

 4918-5019, table 1; hauls 4062L-4205R, table 3) were much more 

 satisfactory than we had reason to anticipate, because the cable then 

 in use did not permit attaching as heavy a w-eight as seemed necessary. 

 In spite of this, only eight cases w^ere imperfect out of seventy-two 

 operations, and these were due to insufficient tension on the closing 

 trip.s. Furthermore, in the twenty-eight trials made on the last day, 

 ilay 16. only one case of imperfect operation was discovered. On the 

 whole, these results are very encouraging, but as further experience 

 will probably suggest certain desirable modifications of the apparatus, 

 a detailed description is not given. ^ 



h. Attachments to the Kofoid water bottle. — To meet the require- 

 ments of a microplankton trip made during September, 1913, two 

 mechanically independent attachments were added to the Kofoid water 

 bottle (Kofoid, 1905) in order to permit the u.se of reversing ther- 

 mometers, and to obtain an additional sample of water for salinity 

 determination (text-figs. D and E). The temperature in situ is always 

 given more accurately by reversing thermometers than by estimates 

 based on the temperature of any sample inclosed in a non-insulated 

 water bottle, and, as pointed out formerly (1915. p. 14), it has been 

 impossible to make any .such estimate for depths exceeding 350 meters. 

 Again, in order to avoid withdrawing any of the organisms contained 

 in the Kofoid bottle itself, the attachment for obtaining an additional 

 water sample was added. 



Frequent testing before and during actual collecting gave no 

 evidence of defective operation. But the nature and degree of cor- 

 relations between temperature and depth, salinity and depth, and 

 temperature and salinity demonstrated that in fully ninety per cent 

 of the 176 subsurface hauls made, the Kofoid bottle closed at some 

 unknown depth significantly less than that indicated by the amount 

 of cable out. Indeed, it seems probable that the remaining nineteen 



1 During the summer of 1916 about two hundred trials of this apparatus were 

 made from the IT. S. S. "Albatross." The very satisfactory results of this 

 severe test proved the apparatus to be fully as reliable and durable as the 

 Ekman water bottle. 



