METHODS 



In I9U8, periranent netting stations were selected in Pool 8 at five 

 locations shown in Figure 1. Beginning on May l5, one trap net was set 

 at each of the stations. Each net was lifted daily between 8 and 11 A.M., 

 the catch removed and the net immediately reset. Fishing in this manner 

 was continued for 1^ days after which the nets were removed for a "rest 

 period" of h to 6 days.. During the rest period the nets v;ere repaired 

 and treated with copper napthenate preservative. 



Such 10-day netting periods alternating with h to 6 day idle periods 

 continued through September 25, making a total of nine netting periods. 

 In following portions of this paper these netting periods will be desig- 

 nated by the Roman numerals I through IX. They are dated as follows: 



Period I - May iS through May 2k 



II - May 31 through June 9 



III - June 17 through June 26 



IV - July 1 through July 10 



V - July 18 through July 27 



VI - August 3 through August 12 



VII - August 18 through August 27 



VIII - September 1 through September 10 



IX - September I6 through September 25 



During the nine netting periods 325 successful trap-net lifts were 

 made. The results of 125 other lifts were discarded because of muskrat, 

 beaver or human interference with the nets. One of the five netting 

 stations was abandoned during Period IX because of muskrat activity. 



All fish taken in the trap nets -were counted, weighed and measured. 

 After the middle of June scales were taken from a selected sample of each 

 species of fish. Gross examination was made of the state of gonad develop- 

 ment of many of the fish. 



During the netting a careful record was kept of water temperature and 

 water turbidity at the time each net was raised. Turbidity readings were 

 made with a platinum needle turbidimeter. Water level and rainfall records 

 were obtained from the U. S. Weather Bureau at La Crosse. 



Data for each net lift were recorded separately in code suitable for 

 tabulation on an IBM punch card machine. 



STATISTICAL CONSIDERATIONS 



The use of trap-net catches for evaluation of size and structure of 

 a fish population is essentially a sampling problem. Means calculated 

 from trap net catch dat^ are subject to sampling error and some method 



