RICHARDS ET AL.: BEHAVIORAL INTERACTIONS OF AMERICAN LOBSTER AND CANCER CRABS 



served as a control. This arrangement was neces- 

 sitated by poaching of the 8-lobster treatments 

 when they were at the ends of strings. The stock 

 rates approximated natural catch rates for lob- 

 sters, but were considerably lower than could be 

 achieved for crabs. The use of two stocking densi- 

 ties allowed us to assess the effects of both the 

 species identity and the stocking density upon 

 catch rates. 



Traps were hauled daily, weather permitting, 

 rebaited with about 1 kg of flounder or floun- 

 der carcasses, and experimental traps were re- 

 stocked if necessary. The number, size, sex, and 

 proportion of the catch in each trap compart- 

 ment were recorded for each of the three species. 

 A total of 336 trap hauls were made. 



The sizes of stocked animals were C. borealis, 

 95-115 mm carapace width (CW); C. irroratus, 

 90-115 mm CW; and H. americanus, 75-85 mm 

 carapace length (CL). Carapace width of crabs 

 was measured as the distance between the two 

 most lateral notches on the carapace; carapace 

 length of lobsters was the distance between the 

 posterior edge of the carapace and the postero- 

 dorsal edge of the eye socket, parallel to the lon- 

 gitudinal axis. Stocked animals were assigned to 

 traps unsystematically with respect to size and 

 sex. 



To further assess the effects of lobsters on catch 

 rates of crabs, laboratory studies were under- 

 taken from July through October of 1979. Two 

 rectangular wooden lobster traps (69 X 34 X 51 

 cm) were covered with 2.5 cm mesh wire to simu- 

 late the mesh size of traps used in the field ex- 

 periments. The baited traps containing either 8 

 or (control) lobsters (70-85 mm CL) were placed 

 in the center of two large indoor tanks (3.4 X 

 1.5 X 0.5 m and 3.1 X 1.5 X 0.5 m) supplied with 

 ambient seawater running at about 2 1/min. 

 Inflow and outflow were at opposite ends of the 

 tanks, thus the water flowed through the traps. 

 Each tank was provided with 10 clay pipe shel- 

 ters (10.2 cm in diameter, 31 cm long, with two 

 open ends). For each trial, 15 individuals of C. ir- 

 roratus (80-110 mm CW) or C. borealis (85-115 

 cm C W) were placed in the tank. After about 24 h 

 the catch was counted and removed, and the loca- 

 tion of animals in the trap was recorded. Crabs 

 were used only once; stocked lobsters were used 

 twice, in different traps. Prior to experimenta- 

 tion, each species was held separately in large 

 outdoor tanks supplied with running seawater 

 and fed every third day with a variety of species 

 of fresh fish. Individual traps were alternated as 



experimental and control treatments to avoid 

 bias due to differences between traps and tanks. 

 Ten replicates of each experiment were per- 

 formed. 



Behavior 



Behavioral mechanisms affecting trap effi- 

 ciency were investigated by direct observation in 

 the laboratory. A rectangular wooden lobster 

 trap was modified to improve visibility by re- 

 placing the top with 2.5 cm mesh wire and paint- 

 ing the bottom white. The trap was baited with 

 thawed whole flounder or flounder carcasses, 

 stocked with 5 or (control) lobsters (70-85 mm 

 CL), and placed in a 3.1 X 1.2 X 0.6 m tank pro- 

 vided with 10 clay pipe shelters and ambient sea- 

 water running at about 2 1/min. An hour after 

 the trap was placed in the tank, 20 C. irroratus or 

 C. borealis (80-110 mm CW) were added. Tape- 

 recorded observations began 15 min later and 

 continued during alternate 15-min periods. A 25- 

 watt incandescent red light suspended 1.2 m 

 above the tank provided the only light. Kennedy 

 and Bruno (1961) have shown lobsters to be rela- 

 tively insensitive to these wavelengths. 



Observations were carried out intermittently 

 from July through October 1979. One sunset-to- 

 sunrise observation for each combination of stock 

 treatment (0 or 5 lobsters) and catch species (C. 

 irroratus or C. borealis) revealed that activity 

 peaked between sunset and midnight. Subse- 

 quent observations were made during these 

 hours. Lobster-stocked and control observations 

 for each crab species were done within 2 wk of 

 each other to minimize seasonal effects. A total of 

 11.5 h of observation in three separate periods 

 was made on each combination of stock treat- 

 ment and catch species. 



All animals were held in conditions similar to 

 those described previously for tank experiments, 

 and were in captivity from 2 d to 1 mo before use. 

 No animal was used more than once. 



Data collected included frequency and nature 

 of inter- and intraspecific interactions and trap 

 entry and escapement. Positions of animals in 

 the trap were recorded every 15 min. 



RESULTS 



Trap Efficiency 



We assumed that the relative effect of the ex- 

 perimental treatments would not differ between 



53 



