Water conditions were maintained via an air lift 

 filter. The animals were fed a dry pellet diet ap- 

 proximately every other day (see diet #5, Balazs et 

 al. 1973). A 12-12 photoperiod with one-half hour 

 twilight lighting at "sunrise" and "sunset" was 

 employed. 



Prior RLsiticnt 1 xpcrmicni 



Eai'lier experiments revealed a foi'm of shelter 

 preference or selection operating in M. rosenhcrgii 

 (Peebles 1977). The shelters used in this experi- 

 ment were identical to those most frequently 

 selected by animals in the earlier experiments. 

 One shelter was placed in each experimental tank. 

 A shelter consisted of six concrete bricks arranged 

 into a double open ended square tunnel ( 19.3 x 19 

 X 11.4 cm tall). 



Refuge othei" than the shelter was eliminated by 

 the use of oblong experimental tanks ( 137 x 75 x 

 92 cm deep) and the suspension of the air lift filters 

 just below the water surface ( the usual position for 

 these filters was on the bottom i. Water depth was 

 34 cm. 



Adults from commercial ponds were placed in 

 two separate holding tanks, where they were kept 

 for no longer that 1 wk. Two animals were re- 

 moved, one each from the separate holding tanks. 

 Three body characteristics were measured: stan- 

 dard length (tip of telson to orbit of eye), and 

 lengths of left and right chelae. The animals were 

 tagged by means of a small plastic "bread bag 

 twist-tie" that was color coded and tied around the 

 tail. It took about 15 s to attach. Following tagging 

 the two animals were placed separately in ex- 

 perimental tanks. Three observations were made 

 before the introduction of the "immigrant" and 

 four observations were made after the introduc- 

 tion. The preintroduction observations were made 

 on the second, third, and seventh days after the 

 animals were placed in their separate experimen- 

 tal tanks. There were three observations per ani- 

 mal, each lasting 3 min. After the preintroduction 

 week a coin was flipped to determine which animal 

 would be the immigrant. The immigrant was des- 

 ignated as the introduced specimen and was 

 moved via a dip net from its tank to the resident's 

 tank. The resident was the animal that was not 

 moved from one experimental tank to another. 

 The postintroduction observations were made on 

 the day of introduction, and the second, third, and 

 seventh days after introduction. The observation 

 performed on the day of introduction was 15 min 



and designed to monitor agonistic interactions as- 

 sociated with the initial encounters of the paired 

 animals. The remaining three postintroduction 

 observations were 3 min each and designed to re- 

 cord the animal's position within the tank. All 

 observations were made between 1000 and 1530. 

 Since these animals are nocturnal, movement and 

 behavioral interactions were minimal during the 

 daytime. 



A total of 36 animals (18 immigrants, 18 resi- 

 dents) were used. Paired animals were of the same 

 sex. This controlled for the possible confounding 

 effect heterosexual courtship behavior might have 

 on competition foi- shelter occupancy. 



Simultaneous Introdut tion Fxpcrinicnt 



The treatment of the simultaneous introduction 

 experiment differed from the prior resident exper- 

 iment in four ways: 1 ) only males were used; 2) the 

 animals were simultaneously introduced into the 

 oblong tanks; 3) two additional body characteris- 

 tics were measured (body weight and carapace 

 length); and 4) the animals were not separately 

 observed prior to introduction. 



Fifteen trials were run employing a total of 15 

 pairs or 30 animals. Observations were made on 

 the day of simultaneous introduction, and the sec- 

 ond, third, and seventh days after introduction. 

 The observation performed on the day of simul- 

 taneous introduction was 15 min and designed to 

 monitor agonistic interactions associated with ini- 

 tial encounters of the paired animals. The remain- 

 ing three postintroduction observations were 3 

 min each and designed to record the animal's posi- 

 tion within the tank. 



Control Experiment 



Eleven controls were run to test the effect of 

 handling. Animals were selected, measured, tag- 

 ged, and placed individually in experimental 

 tanks. One week later the control was netted, held 

 in the air, and reintroduced into the same experi- 

 mental tank. Observations were made for the 

 week before and the week after netting (mock im- 

 migration). 



Operation.il Definitions 



Successful: an animal that was in a shelter at 

 the end of the 7-day period following immigration. 

 Unsuccessful: an animal that was not in a shel- 



906 



