sexes were combined for the calculation, but could have slightly 

 influenced egg production estimates. No differences were found in the 

 length-frequency distribution of males and females tagged as compared to 

 those recaptured. Thus, recapture probabilities were not bi^i&ed by size 

 selectivity. 



A significant difference was found in the proportions branded and 

 recaptured at the stations sampled. The proportions were significantly 

 different between a group comprised by stations 52 (16% branded and 20% 

 recaptured), 2 (5%, 7%), and 1 (17%, 19%) and that by 53 alone (17%, 

 11%). This was not unexpected as winter flounder concentrated in the 

 channels of tlie lower river before leaving the estuary and were more 

 available for recapture there. They also probably withdrew from the 

 northern portion (station 53) of the upper river arm into the southern 

 portion (52) as well. 



Two sources of Information were available to examine for loss of 

 marks during the survey. Two of the 67 fish that were double-branded 

 and recaptured one or more times had only one brand visible. Of the 

 fish both branded and disc-tagged, only 1 of the 91 fish re-examined had 

 a missing brand. These data indicated that the number of recaptures was 

 perhaps underrepresented by about 2 to 3%. According to Arnason and 

 Mills (1981), the small loss of brands should not have biased the 

 abundance estimate (N) nor its standard error, although it could have 

 affected the number joining (B) and its standard error. They also noted 

 that the precision of the estimates might have been reduced slightly, 

 but in our study corrections were unnecessary since the bias was small 

 and precision (CV=19% for 1983) was relatively good. No bias occurred 

 because of capture-prone fish; only 15 fish were recaptured twice (4% of 

 total) and 1 specimen was taken 3 times (0.2%). 



Any movement of marked fish out of and back into the Niantic River 

 during the abundance survey could have introduced serious bias into the 

 Jolly estimate (NUSCo 1980; Balser 1981). Several significant 

 differences were found in the proportions of fish marked and recaptured 

 during various weeks. As in previous years, the significant differences 

 were thought to have been due to sampling error rather than from an 

 acutal temporary outmigration of branded winter flounder. The 

 differences in 1983 were associated with a decrease in the number of 



16 



