FOGARTY ET AL.: MOVEMENT OF AMERICAN LOBSTER OFF RHODE ISLAND 



Figure 5. — Straight linedistance between release and recovery 

 sites for American lobster at inshore locations off Rhode Island. 

 Release locations are composites of several release sites in each 

 area. Total distance travelled is noted for tracks which are trun- 

 cated by borders. 



did undertake extensive movements. Appropriate 

 test statistics and directional components (Saila 

 and Flowers 1968) were computed for each total 

 data set and for data partitioned according to time 

 at large (Table 2). Missing data for some returns 

 prevented the use of all the recoveries for these 

 analyses. Mean vector angles (G) for the three 

 locations ranged from 164.53°(T) to 193.69°(T), 

 and Rayleigh tests indicated a nonuniform dis- 

 tribution of returns at each inshore location (Table 

 2). In general, the north-south vector components 

 were consistently stronger than the east-west 

 components for each location. When partitioned 

 by time at large, the relative magnitude of the 

 north-south and east-west vector components 

 were more nearly equal for the first time period 

 (0-20 d) at the Sakonnet River site, possibly re- 

 flecting an initial random dispersal of released 

 lobsters. The lack of a statistically significant 

 mean vector bearing for this period (Table 2) sup- 

 ports this inference. 



The consistently low estimates of V ' for lobsters 

 tagged at inshore locations were due, in part, to 

 physiographic constraints since east-west move- 

 ments were often limited by the coastline, particu- 

 larly in Narragansett Bay ( Figure 1). The negative 

 north-south vector components were indicative of 



net southerly movement since the cosine of angles 

 ranging from 90° to 270° would be negative. Simi- 

 larly, negative values of V imply a westerly dis- 

 placement since the sine of angles from 180° to 

 360° would be negative. 



The mean square dispersion coefficient (a^) var- 

 ied considerably by location and the time period 

 under consideration (Table 2). The quantity a^ 

 measures the relative degree of undirected move- 

 ment of any individual with respect to the group 

 directional average. Some caution is necessary in 

 interpreting these values since dispersion coeffi- 

 cients are likely to be overestimated when move- 

 ments are nonrandom (Jones 1959). 



To examine the possibility of directed seasonal 

 movements, we pooled data from inshore release 

 locations and regrouped them according to release 

 period and time at large. We compared movement 

 statistics for lobsters released prior to 1 July 1975 

 and recaptured prior to 1 September 1975 with 

 those released after 1 July 1975 and recaptured 

 prior to 1 September 1975 (Table 3). We noted a 

 nonuniform distribution of returns at all levels of 

 analysis (Table 3i. The north-south directional 

 components consistently dominated the east-west 

 components for both groups. The negative V val- 

 ues reflect the strong southerly directionality for 

 both groups while the east-west components ( V") 

 varied considerably when further partitioned by 

 time at large (Table 3 ). For lobsters released after 1 

 July 1975 the north-south vector components were 

 two to three times higher than for late spring- 

 early summer releases, indicating a sharp in- 



TaBLE 2. — Mean vector angle (B) from true north, mean square dispersion coefficient (a^), north-south (V) and east-west (V) 

 directional components, Rayleigh test statistics {R and Z ) and sample size (n)for lobsters released in the Sakonnet River, Narragan- 

 sett Bay, Rhode Island Sound, and Cox Ledge. The mean square dispersion coefficient is a measure of random movements; V and V 

 indicate nonrandom components of movement. Negative values of V and V are indicative of net southerly and westerly movements. 

 The Rayleigh test is a test for uniform concentration of points around a circle of unit radius. 



•p<o.oi. 



775 



