Stoner and Davis: Outplanting queen conch, Strombus gigas 



395 



Cumulative mortality was examined by using the 

 same procedure as enclosure experiment I. Mortal- 

 ity was calculated three times throughout the experi- 

 ment, and analyzed statistically at the midpoint (day 

 45) and end (day 90) of the experiment. Growth rates 

 were calculated for two growth periods: 12 April to 

 27 May and 27 May to 11 July 1990. 



Experiment II The second 3-month tethering experi- 

 ment was conducted during the winter ( 7 February- 

 3 May 1991) at site CI. Too few of the original hatch- 

 ery-reared conch remained alive to set up the experi- 

 ment at the second site. Hatchery-reared conch 

 ranged from 100 tol38 mm SL (mean=116, SD=8), 

 and wild conch were 111-133 mm SL (mean=124, 

 SD=5). Tethers were set up as in experiment I with 

 four replicated rows of 20 individuals ( 10 hatchery- 

 reared and 10 wild conch), except the conch them- 

 selves were not tagged. The conch were checked for 

 mortality three times during the experiment and 

 analyzed statistically at days 42 and 84. 



Because cable ties were secured behind long api- 

 cal spines, escape from tether apparatus would be 

 possible only in the event of failure in the cable tie, 

 monofilament line, or connections. Failure appears 

 to be unlikely because nearly all kills observed in 

 this study were found as empty shells attached to 

 the tether apparatus or as crushed shells within 1 m 

 of the original location. 



Data Analysis 



Analysis of variance (ANOVA), following the guide- 

 lines of Day and Quinn (1989), was used extensively 

 in the interpretation of growth and mortality data. 

 The statistical procedures started with full model 

 ANOVA that included all independent effects. When 

 interactions were significant, one- or two-way 

 ANOVAs were performed to examine the effects of 

 site and stock type, the variables most critical in this 

 study. For brevity, non-significant interaction terms 

 in multiple-way ANOVAs are not addressed in the 

 text but are reported in tables. Mortality data were 

 normally examined at the mid-point of individual 

 experiments and at the end. 



Cochran's test was used to test for homogeneity of 

 variances. Log and arcsine transformations of data 

 were used in some cases to remove heteroscedasticity; 

 these are noted in the text. Where repeated measure- 

 ments were made within one experimental enclosure 

 (i.e. growth rates determined for conch in one pen), 

 mean growth rates in the enclosures were used as 

 replicates rather than individual measurements to 

 eliminate pseudoreplication (Hurlburt, 1984). Analy- 

 sis of covariance (ANCOVA) was used to test for dif- 

 ferences in morphological characteristics ( shell weight, 



shell diameter, and tissue weight) between hatchery- 

 reared and wild conch. Shell length was the covariate. 

 After release, tagged, free-ranging conch dispersed 

 from the initial 1-ha study sites. Chi-square analy- 

 sis was used to compare dispersion of the two stock 

 types, where the distribution of tagged wild conch 

 was used for the expected frequency in different sur- 

 vey zones. 



Results 



Conditions at the outplant sites 



During the 15-month study period, bottom-water 

 temperature ranged from 24°C in February 1990 to 

 30°C in late September 1991, then declined rapidly 

 and remained between 24 and 25°C until early May, 

 when temperature rose to 27°C (Fig. 3). 



Density estimates made 1 month prior to the be- 

 ginning of the free-ranging experiment (2 March 

 1990) showed that the density of conch at site CI 

 was 16 times higher than at site C2 (Table 1). Be- 

 tween March and May, conch density at site CI in- 

 creased to over 1.2 conchm -2 , owing to immigration 

 of the natural population. Transplanted conch from 

 the free-ranging experiment made up 7-15% of the 

 conch in the density estimates; however, on 2 May 

 1990, transplanted conch accounted for 88% of esti- 

 mated density. In July, there were nearly equal den- 

 sities of conch at sites CI and C2, but in November 

 1990 and February 1991, densities were close to the 

 original values first observed in March 1990. This 

 may be due to directional changes in movement of 

 conch (towards the northeast) during the winter, 

 which took them away from the transplant sites CI 

 and C2 (see Behavior). 



As expected, shell length measurements taken 

 during the density surveys show an increase in length 

 over time for wild conch tagged at CI (from mean=105 

 ± 14 SD in March 1990 to mean=129 ± 13 SD in Feb- 

 ruary 1991 ((Table 1). This represents an overall 

 growth rate of 0.07 mmday" 1 , similar to that mea- 

 sured in free-ranging tagged conch. No growth rate was 

 calculated for C2, because density surveys yielded low 

 numbers; however, the mean sizes appear to be com- 

 parable to those measured at site CI (Table 1). 



Handling and tag effects 



Transporting hatchery-reared conch appeared to 

 have little adverse effect on their subsequent 

 survivorship in the field. Conch were left out of wa- 

 ter for 7 hours, and all remained alive during the 7 

 days after transport while they were tagged and 

 placed in enclosures for Experiment I. 



