tagged population at a theoretical summed rate of 23% 

 per annum (12% tag loss, 6% natural mortality, 5% 

 recapture). In calendar year 1972, for example, the 

 theoretical number of tagged lobsters outstanding at the 

 beginning of the year was 3,630; the expected number of 

 returns for 1972 based on the average rate of returns 

 (0.049) in years 1968-71 is 179, in sharp contrast to 67 ac- 

 tual returns. 



The distribution of recaptured tagged lobsters is con- 

 sidered representative of the distribution of the lobster 

 population. Fishermen search for commercial quantities 

 of lobsters throughout the year at depths of 10-350 

 fathoms (18-640 m), which is a considerably greater 

 range than the 35-300 fathoms (64-549 m) depth interval 

 from which lobsters for tagging were initially captured. 



DATA REDUCTION, PLOTTING, AND 

 FORMAT 



Data received on individual recaptures varied 

 considerably. Data sought included date and position of 

 recapture (latitude and longitude, or loran A coor- 

 dinates), sex, carapace length, presence or absence of ex- 

 ternal eggs, cheliped configuration, and designation of 

 any missing chelipeds and walking legs. The most 

 critical data were location and date of recapture, and 

 carapace length from which both migration trends and 

 growth could be determined; this was received on 350 of 

 the recaptured lobsters. Recapture location and date 

 only were received on 576 individuals and provide the 

 basis for analysis of movements. 



Data was listed and keypunched in two different for- 

 mats. The data format (see appendix tables) for this 

 study includes growth increments for reader reference, 

 but this element of the study is being treated separately; 

 return data is listed chronologically by sex. The basic 

 data deck provided input for computer calculation for in- 

 dividual recaptures of great circle distance traveled from 

 point of release to point of recapture, days at large, and 

 other standard computations such as mean distance 

 traveled and mean time at large by various groupings of 

 individuals. The same data deck served as input for a 

 Cal-Comp Plotter, Model 663, from which release coor- 

 dinates, recapture coordinates, or combinations of both 

 were plotted in various combinations to reveal and dis- 

 play the overall features of dispersion within and 

 between release groups and to show the overall monthly 

 distribution of recaptures. The Cal-Comp Plotter was 

 simultaneously programmed and fed a series of coastline 

 coordinates, isobath coordinates, and titular information 

 such that the finished plot was a Mercator chart drawing 

 to the nominal scale of 1:1,200,000. 



Among the 945 recaptured lobsters, 584 (61.8%) were 

 reported by specific location, 183 (19.4%) by generalized 

 location — usually by reference to a named submarine can- 

 yon, and 178 (18.8%) without location information of 

 any kind. Those recaptures reported by approximate 

 location are hand-plotted in distinctive fashion within 

 the machine plots of the various subgroups of returns 

 with specific location. 



In order to facilitate interpretation of recovery data, 

 we have treated the 52 original releases as 29 according to ' 

 the constraints footnoted in Table 1. This action 

 minimized the plotter executions and gave more 

 coherence to the individual plots. Test plots of recovery 

 coordinates showed a number of cases where overplotting 

 or tight grouping of recovery points resulted in a confu- 

 sion of points and numbers. In these instances we used a 

 plotting subroutine which plotted all points within or 

 upon the eastern and southern side of a given 6-minute 

 square (0.1° square) as a single point with collective 

 number at the diagonal center of the square; the average 

 displacement of any single point plotted in this manner 

 is well under 3 nautical miles (5.6 km) which we have 

 accepted as within the limits of navigational accuracy, 

 reporting, or both. 



Because the tag releases were effected in greater or 

 lesser increments over a long period of time, they con- 

 stitute a series of repetitive experiments and are treated 

 accordingly; the overall presentation which follows takes 

 the form of an atlas which provides a pictorial analysis of 

 the results of the various releases. Additionally, we have 

 developed a generalized treatment of th,e monthly dis- 

 tribution of offshore lobsters in relation to bottom 

 temperature. 



Original Station Locations 



A total of 52 releases of tagged lobsters were made on 

 the outer continental shelf and slope commencing in 

 March 1968 and ending in May 1971 (Table 1). Cruise 

 numbers and station numbers are not wholly in con- 

 secutive order because interim cruises involving coastal 

 area tagging were also conducted in the same period. 

 Thus, station 66 occupied during Cruise 20 was actually 

 the 52nd and last release during a total of 14 cruises con- 

 cerned with offshore tagging. 



The original release localities (Fig. 1) show their loca- 

 tion relative to major features of the continental shelf 

 and to each other. Most (86%) of the tagging was ac- 

 complished from the vicinity of Block Canyon eastward 

 because of more productive lobster fishing in these areas 

 and because other aspects of cruise objectives required 

 cruise orientation to the east of Block Canyon to max- 

 imize time sharing of the research vessels Delaware I, 

 Delaware II, and Albatross IV. 



Composite Station Locations 



Thirteen (25%) of the 52 original releases are plotted at 

 their original release locality (Fig. 2). The remainder 

 were combined in groups of two or three and assigned 

 location coordinates with averaged latitude and 

 longitude rounded to the nearest whole minute (Table 1). 

 Maximum distance between any two original release 

 sites comprising a composite station was 4 nautical miles 

 (7.4 km). The purpose of this treatment was to effect a 

 logical pooling of release and recapture information that 

 would expedite both plotting and evaluation of the data. 

 Computations of distance traveled and time at large are, 



