maximum fishing. Daily coverage was on 

 an 8-hour basis from 9:00 a.m. to 12:00 

 noon and fromi 2:00 p.m. to 7:00 p. mi. 

 Where feasible, adjacent fishing areas 

 were covered by one man. Catch data 

 were collected daily from the state-con- 

 trolled area. 



At bridge and bank areas, observa- 

 tions and counts were made hourly to 

 determine rate of catch for estimating 

 catch and effort for hours of non-coverage. 

 Catch and effort for the 2 noncensused 

 hours between 12:00 noon and 2:00 p.m. 

 were obtained by averaging the results 

 of the previous and following hour. The 

 catch and effort figures determined from 

 surveys conducted from 6:00 p.m. to 7:00 

 p.m. were doubled to obtain an estimate 

 of number of shad caught and number of 

 fishermen from 6:00 p.m. to end of fishing 

 day (until dark). Few fish are caught 

 after 6:00 p. mi., and therefore any error 

 caused by this approximation was negli- 

 gible. Nunnber of fishermen and nunnber 

 of shad caught and kept were recorded 

 daily. No attempt was made to estimate 

 number of shad caught and released. 



To census the boat fishery, drop-in 

 card boxes were mounted on posts at 

 each boat -launching area with signs giving 

 the purpose of the survey and requesting 

 each fishing party to fill out a census 

 card. Each card contained space to record 

 number of male and female shad kept, 

 number of fishermen in party, and date. 

 Card number and landing site information 

 were completed by the biologist. Enough 

 census cards were attached to the boxes 

 each day to make certain that one was 

 available for each party. 



The boats and boat fishermen were 

 counted each census day by biologists 

 working in the bank and bridge areas. 

 In the early part of the season emphasis 

 was placed on the boat areas, and the 

 majority of the boat fishermen were con- 

 tacted at the beginning and end of the 

 fishing trip to acquaint them with the 

 card system and to help them complete 

 the census cards. Also, during the season 

 "reminder cards" (creel census cards 

 with an additional note) were placed on 

 the windshields of boat -fishermen's cars 

 to remind those acquainted with our sys- 

 tem to fill out the card and to acquaint 

 new fishermen with our program. The 

 card boxes were checked each day, and 



the information obtained from connpleted 

 cards was proportioned to the total boat 

 count to estimate the day's catch. By 

 doing this, niany faults inherent in a creel 

 census by the voluntary -type card system 

 were nullified. 



Prior to the 1958 survey, an effort 

 was nnade to evaluate the method used 

 the previous season. The 1957 6 -day week 

 sampling schedule was applied to the state- 

 controlled area catch data. Results showed 

 less than 1 percent error in estimating 

 the catch. Because funds and personnel 

 were not sufficient for a 6 -day sampling 

 schedule in 1958, a hypothetical 5 -day 

 week census, including all week ends and 

 holidays, was applied to the 1957 state- 

 controlled area catch data. Results showed 

 less than 3 percent error from the known 

 catch, and therefore a 5-day week sannpling 

 schedule was adopted for the 1958 study. 

 Two biologists were assigned to make 

 the census during the 1958 season. In 

 all other respects the 1957 and 1958 creel 

 census miethods were the same. Scale 

 samples and length-weight data by sex 

 were collected from the catch throughout 

 both seasons. 



The Salmon and Scantic Rivers were 

 checked during both years. Most of the 

 shad caught in these rivers were taken 

 by trout fishermen during the early part 

 of the season. Approximate estimates of 

 fishing success and effort were calculated 

 fronn data obtained from weekly angler 

 interviews. Since the catch in these areas 

 was small, little error was introduced 

 into the total sport catch estimate by 

 these approximations. 



ANALYSIS OF RESULTS 



In 1957 the estimated catch for bank 

 and bridge fishing was based on the 6 -day 

 week sampling schedule and the boat catch 

 was estimated from a 37.3 -percent return 

 of census cards. In 1958 the estimated 

 catch for bank and bridge fishing was 

 based on the 5-day week sampling sched- 

 ule and the boat catch on 43.5 -percent 

 return of census cards. Returns of boat 

 fishermen cards by area in 1957 were: 

 Farmington River--42.4 percent; Windsor 

 Locks--27.1 percent; Kings Island--70.3 

 percent; and Willimansett Bridge--29.3 

 percent. Percentage card return during 

 the 1958 season showed a slight increase 



