18 University of California Publications in Zoology [Vol.15 



(oi- water sample), are written iu India ink previous to sailing. How- 

 ever, to provide against loss of tags, a few tag-books are carried in 

 duplicate in which these numbers are not entered. When the par- 

 ticular tag is properly filled out it is placed inside the jar containing 

 the haul or water sample, and the latter is then filed until the 

 "Agassiz" reaches home. 



In addition to the tags a log-book is kept in which is also entered 

 the same data for each haul and water sample, together with the 

 details of operating the apparatus, accidents, weather conditions, and 

 in fact anything that might have the remotest bearing upon the 

 collections. 



Upon completion of the day's work the recorder checks the entries 

 in his log-book against those the captain has kept, adding the sextant 

 bearings and other remarks relative to the day's navigation. As soon 

 as possible after the material reaches the laboratory the hauls are 

 condensed into four-ounce pomade bottles, and filed, the salinities of 

 the water samples are determined (see p. 28), and the data are 

 arranged in final form and typewritten, usually in triplicate. 



C. Errors in Field Observations 

 1. ERRORS DUE TO POSITION 



(o) Latitude and longitude. — Three methods are used for de- 

 termining the latitude and longitude at which collections are made, all 

 of which are sub.ject to errors varying in magnitude consequent on a 

 variety of conditions. The sextant method, being the most accurate, 

 is always used when practicable. Otherwise the position is ascertained 

 b.y a pelorus or, when the use of this instrument is impracticable, 

 recourse is had to the less accurate method of "dead-reckoning." i.e., 

 of calculating the position from course and log. 



Location by the sextant as used by us consists of the selection of 

 three prominent points on land, and the measurement of two angles, 

 one between the horizontal lines connecting the observer with the 

 central and left-hand points ; the second between the lines connecting 

 the observer with the right-hand and central points. Whenever pos- 

 sible, choice is made of prominent and sharply dffined headlands or 

 other salient points which can readily be identified on the United 

 States Coa-st and Geodetic Survey charts. Care is taken to avoid the 

 selection of points lying on or near a circle pa.ssing through the 

 observer, for under such conditions the location is indeterminate. 



