FISHERY BULLETIN: VOL. 85, NO. 4 



1984, trawl samples were obtained each month 

 from July 1983 through January 1985 during 

 tows of 1-3 h duration. Similar observations of 

 trawl catches were obtained from cruises made by 

 the CT DEP's research vessel (a 13 m stern-rigged 

 trawler) from May through November in 1983 

 and 1984 during tows of 0.5-2 h duration. Com- 

 mercial pot fishery sampling was conducted inter- 

 mittently during summer and fall in both years. 

 All samples were obtained in Long Island Sound 

 west of long. 72°52'W. 



American lobsters were examined for evidence 

 of physical damage, denoted as "new" (resulting 

 from the current net tow or pot haul) or "old" 

 (previously sustained). Old damage was inferred 

 by the presence of discolored and healed tissue 

 and by the absence of bleeding. New damage was 

 further categorized as "major" and "minor". 

 Major damage was defined as death, broken or 

 crushed body parts or claw(s), and multiple in- 

 juries, even when each individual wound was 

 minor. Minor injury included walking leg loss or 

 damage, antenna damage, minor breakage to a 

 claw or rostrum tip, and recently autotomized 

 claw(s). Recent autotomy (reflex amputation) was 

 defined as a fresh partition without discoloration 

 of the covering membrane at the "breaking 

 plane" (Herrick 1909). Additional data were 

 recorded on length, sex, shell hardness, presence 

 of eggs, and absence of claws. Soft, newshell, and 

 hardshell condition generally followed stages A, 

 B, and C, respectively, as described by Passano 

 (1960). In this paper, the term "newshell" in- 

 cludes both soft and newshell lobsters (stages A 

 and B). 



In comparing damage rates between commer- 

 cial and research vessels, we found that hardshell 

 American lobsters <81 mm CL were damaged 

 more frequently in research than in commercial 

 samples (x^ = 6.143). Since the research vessel 

 used a smaller mesh (51 mm cod end) than com- 

 mercial vessels (75-114 mm cod end), it caught a 

 proportionately greater number of American lob- 

 sters <81 mm CL (86% research vs. 72% commer- 

 cial). Therefore, only lobsters >81 mm CL were 

 analyzed because the greater proportion of sub- 

 legal American lobsters and other biota in the 

 small mesh sample might have resulted in an 

 overestimate of damage relative to the commer- 

 cial fishery, due both to intraspecific, agonistic 

 behavior and to compacting of net contents. 



The incidence of major damage, compared be- 

 tween pot and trawl gears, large and small trawl 

 vessels, and commercial trawlers and research 



trawler were analyzed by chi-square goodness of 

 fit (Sokal and Rohlf 1969). 



Delayed Mortality 



Damaged and undamaged American lobsters 

 less than the minimum length were taken from 

 the same area of Long Island Sound, on the same 

 date, from commercial pot vessels and from either 

 commercial trawl vessels or research vessel. Sam- 

 ples were obtained in November 1983 and May, 

 July, and August 1984. They were transported in 

 tanks with circulating seawater to a laboratory 

 equipped with an open circulating seawater sys- 

 tem and were observed twice daily for a period of 

 14 days. Lobsters were fed daily with assorted 

 fish species. Dead lobsters were removed upon 

 observation. Lobsters were selected nonrandomly 

 from catches to ensure that damaged lobsters 

 were adequately represented in the tests. Since 

 damaged lobsters often made up only a small per- 

 centage of the total, our sample consisted of pro- 

 portionally more damaged lobsters than their ac- 

 tual proportion in the catch. 



American lobsters with minor damage were 

 combined with undamaged ones for analysis be- 

 cause undamaged lobsters rarely experienced de- 

 layed mortality (2/309) and those with minor 

 damage never did. To estimate total delayed mor- 

 tality from each gear, the observed delayed mor- 

 tality rate in each test category (undamaged 

 hardshell, damaged hardshell, and newshell) was 

 extrapolated to the corresponding trawl catches of 

 undamaged, damaged, and newshell lobsters ob- 

 served during the same months in 1984. The lab- 

 oratory mortality rate for newshell lobsters was 

 applied to all newshell lobsters in commercial 

 catches rather than only damaged newshell lob- 

 sters since internal (unobserved) damage may 

 have occurred. Field data from 1984 commercial 

 catches were used to estimate total trawl-induced 

 mortality because the proportion of newshell lob- 

 sters in commercial catches was highest during 

 1984, and we believe that that year's data most 

 accurately represent the proportion of newshell 

 lobsters taken in the fishery. 



The chi-square goodness of fit (Sokal and Rohlf 

 1969) with Yates' correction for continuity (Zar 

 1974) was utilized to determine whether lobster 

 mortality rates differed between large and small 

 trawl vessels and between commercial and re- 

 search trawl vessels. The effect of season and molt 

 condition was examined using a log-likelihood 

 ratio test (G-test; Sokal and Rohlf 1969). Seasonal 



738 



