BULLETIN OF MARINE SCIENCE. 52(3): 873-885. 1993

GROWTH OF POST SETTLEMENT JUVENILES OF THE FLORIDA STONE , MENIPPE MERCENARIA (SAY) (: XANTHIDAE), IN THE LABORATORY

Wendy A. Tweedale, Theresa M. Bert and Susan D. Brown

ABSTRACT We combined relevant data from three laboratory studies to define molt increment and intermolt period of post settlement « IO.S-mm carapace width [CW)) juveniles of the , Menippe mercenaria. Regression of post molt CW on premolt CW yielded a linear relationship. Mean growth (proportional increase in CW) per molt was 18%, but variation among individuals was high. Molt increment increased significantly and linearly as size of the crab increased. Intermolt period increased significantly with increasing crab size in a log- linear relationship. We attempted to define biologically meaningful size classes (instars) by using mean size-at-instar of individuals raised from the megalopal stage and size-at-molt of individuals who completed eight or more molts in the laboratory. Three determinants (mean, median, and maximum of intermolt period) and two techniques (one based on observed complete intermolt periods and one based on estimates of intermolt period) were used to predict age-at-size of instar-12 juveniles (> 10.4 mm CW). Within a size class, estimates of mean, median, and maximum intermolt periods generally were not significantly different from corresponding values for the three determinants calculated using the observed intermolt periods. However, summing the size-specific intermolt period values to calculate age resulted in disparities among the six estimates of age at the end of instar II. Nevertheless, our analyses suggest that under normal salinity and temperature conditions in Florida, postsettlement juvenile M. mercenaria may require up to 12 months to complete growth to approximately 10mmCW.

As in all , growth in is a discontinuous function. Although small increases in size may occur during intermolt stages, when the arthrodial membranes extend (Kurata, 1962; Hartnoll, 1983), size increases dramatically only immediately following ecdysis, through uptake of water (Passano, 1960; Hartnoll, 1982, 1983). Because the integument is shed at each molt, no permanent record of growth remains. Thus, methods of age determination typically used for vertebrates are unsuitable for estimating growth in crabs; age is commonly cal- culated from observations of molt increment and intermolt period. Laboratory studies of growth have been conducted on many species of crabs. The degree of success in elucidating growth pattern has varied among these studies, principally because laboratory conditions may affect both molt increment and intermolt period (Kurata, 1962; Hartnoll, 1982). Generally, as the size of the organism increases, molt increment increases and percentage increment (growth factor; Kurata, 1962) decreases (Kurata, 1962; Hartnoll, 1982), but the percentage increment is highly variable both within and among species (Hartnoll, 1982). lntermolt period also generally increases as size of the organism increases (Kurata, 1962; Hartnoll, 1983), and variation in both molt increment and intermolt period increases as the organism ages (Tagatz, 1968; Chittleborough, 1976; Restrepo, 1989). Quantitative descriptions of growth in commercially valuable marine species are particularly important. A thorough understanding of the growth rates of these species in all life stages can be important for projecting yield per recruit, for determining the length of time required to reach legal size, and for estimating

873 874 BULLETIN OF MARINE SCIENCE, VOL. 52, NO.3, 1993 other parameters essential for management of the fishery. Growth studies, in conjunction with recruitment and population studies, provide the basis for fishery management. The stone crab Menippe mercenaria (Say, 1818) is an important commercial species in Florida (Bert et al., 1978; Lindberg and Marshall, 1984), and a number of studies related to growth of this species have been conducted. Larval devel- opment in the laboratory has been described (Ong and Costlow, 1970; Field, 1989; Brown et al., 1992). Growth increments and growth rates of large juveniles and adults have been derived from both laboratory and field studies (Savage and McMahan, 1968; Savage, 1971; Yang, 1972; Yang and Krantz, 1976; Savage and Sullivan, 1978; Sullivan, 1979; Bert et aI., 1986), but sample sizes in these studies were small. Also, holding conditions were highly variable and inconsistent both within and among the laboratory studies, and information from the field studies was largely anecdotal. In addition, many laboratory studies describing aspects of growth have focused on the role of claw regeneration (Savage et aI., 1975; Savage and Sullivan, 1978; Sullivan, 1979; Simonson and Steele, 1981; Simonson, 1985) because in the stone crab fishery, only the claws are harvested. Despite the dif- ferences in methodologies and the emphasis on claw regeneration, laboratory studies of growth in M. mercenaria have elucidated the general qualitative rela- tionships of molt increment and intermolt period for subadults and adults. Al- though substantial inter- and intra-individual variation exists, larger crabs gen- erally have larger molt increments (Savage, 1971), smaller percentage increments (Savage and McMahan, 1968; Bert et al., 1986; Restrepo, 1989), and longer intermolt periods (Savage and McMahan, 1968; Savage and Sullivan, 1978) than do smaller crabs. Quantitative assessments ofthe components of growth of stone crabs in Florida have been made only for larger juvenile and adult males (Restrepo, 1989). Similar analyses are lacking for females of a comparable size range and for postsettlement juveniles. We described the growth of postsettlement (metamorphosis to 10.5-mm car- apace width [CW]) juvenile M. mercenaria utilizing the relevant data gleaned from three independent studies in which postsettlementjuvenile stone crabs were reared in the laboratory. We analyzed both the molt increment and intermolt periOd of post settlement juvenile M. mercenaria, and we estimated age at ap- proximately 10.5 mm CW using three determinants (mean, median, and maxi- mum of intermolt period) and two methods of calculating age-at-size (using ob- served intermolt period, using partial intermolt period).

METHODS AND RESULTS Data Available. -Our values for molt increment and intermolt period were ob- tained from postsettlementjuvenile Menippe mercenaria that had molted at least once in one of three laboratory studies. In two of the experiments (studies A and B), variations in temperature and salinity were controlled and photoperiod was constant (light: dark = 12h: 12h); in the third study (study C), variations in temperature, salinity, and photoperiod were not controlled. In study A (Florida Marine Research Institute, unpubl.), each juvenile was held in water of 27°C, 300/00 (ambient summer conditions) for 1 week. Then, depending upon the experimental treatment, water temperature was abruptly decreased by 3°C, 5°C, or 7°C for either 4 or 8 days. Following the treatment, water temperature was returned to 27°C for the duration of the experiment; total experimental time was 7 weeks. Two groups of juveniles (controls) were held at 27°C throughout the experiment. Study B (Brown et al., 1992) was a factorial experiment in which each juvenile was held in a specific water temperature-salinity combination for 7 weeks. Temperature TWEEDALE ET AL.: JUVENILE STONE CRAB GROWTH 875

Table I. Results of analysis of covariance comparing least-square regressions of postmolt carapace width on premolt carapace width for three data sets used to define growth parameters of post settlement juvenile Menippe mercenaria (see text for definitions of data sets). A. Study A versus Study B. B. Study B versus Study C. C. Study A versus Study C. ns: not significant, *: significant at a :5 0.05 with sequential Bonferroni adjustment (Rice, 1989)

Source df SS MS A. Sum of groups 253 40.487 0.160 Pooled within 254 41.208 0.162 Total 255 41.583 0.163 F"ope = 4.508 ns Feley.';on = 2.309 ns B. Sum of groups 351 44.686 0.127 Pooled within 352 44.946 0.128 Total 353 46.485 0.132 F"ope = 2.041 ns F"ey.tion = 12.053* C. Sum of groups 444 73.378 0.165 Pooled within 445 73.819 0.166 Total 446 75.006 0.168 F"ope = 2.664 ns Feleyat'on = 7.156* and salinity conditions in the factorial array ranged from 5°C to 35°C and 100/00 to 400/00, in 5°C and 10%0 increments. In both studies A and B, each juvenile was held separately in 100-150 ml of artificial seawater (Instant Ocean®), and the water was changed on alternate days. In study C (Florida Marine Research In- stitute, unpubl.), juvenile stone crabs were collected sporadically from October 1965 through September 1969 and held for various lengths of time in large outdoor tanks of natural seawater, permitting the temperature, salinity, and photoperiod to vary with ambient conditions (temperature and salinity ranges: 15°-30°C, 230/00- 350/(0). In each study, ample food (usually molluscan muscle) was provided on altemate days. Because the temperature and salinity conditions in which the crabs were held differed among studies, our initial steps in the analysis were to determine the useful component of the data on molting within each study and to test the validity of combining the useful data from the three experiments into a single data set. First, crabs with missing appendages were eliminated from each data set. In study A, goodness-of-fit G-tests (Sokal and Rohlf, 1981) established that no significant differences in the percentage of crabs surviving or in molt frequencies occurred between controls and treatments; therefore, all data on molting were pooled and used in our analysis (N = 175). In study B, Brown et al. (1992) determined that low temperatures (~l O°C) or salinities (~150/00) affected both survival and molting rates of post settlement juvenile M. mercenaria; therefore, from that study only growth data from juveniles held in temperature-salinity combinations of 15°C, 200/00 or higher (N = 82) were included in our analyses. From study C, growth data were used from all individuals that were within our defined size range (N = 273) because the temperature and salinity conditions recorded for that study were never below the critical levels defined by Brown et al. (1992). To examine the feasibility of combining the three data sets into a single data set, least-squares regressions of post molt CW on premolt CW generated separately for each study were compared to each other in pairwise regressions. The variances of the data sets were homogeneous, allowing us to use analysis of covariance 876 BULLETIN OF MARINE SCIENCE, VOL. 52, NO.3, 1993

14 o

12 "...... o E E 10 •...... •

~ 8 U +- 6 0 E +- en 4 o Study A v Study B 0 o Study C Q.. 2

0 0 Premolt CW (mm)

Figure I. Scatterplot of postmolt carapace width (CW) vs. premolt CW for all data used for analyzing growth of postsettlement juvenile Menippe rnercenaria. Studies are defined in the text.

(ANCOV A; Snedecor and Cochran, 1969) for the analyses. The sequential Bon- ferroni technique (Rice, 1989), a statistical adjustment for determining the sig- nificance of multiple tests of a single hypothesis, was used to determine the tablewide significance of each ANCOVA category. The slopes of all lines were homogeneous, but Study C differed significantly from studies A and B in elevation (Table 1). The homogeneity of slopes indicated that growth increment in the three studies was similar, and a scatterplot of the data confirmed that the data points from the three studies were thoroughly intermingled (Fig. 1). Therefore, all data were combined for further analyses. Molt Increment, - We wished to use each molt as an independent data point for the analysis of growth. However, because some juveniles used from studies A and B had molted as many as three times in the laboratory, and many juveniles used from study C had molted numerous (up to 11) times in the laboratory, it was first necessary to evaluate whether the length of time these crabs were held in the laboratory affected molt increment. We grouped the data by laboratory molt number (1-9; sample sizes were inadequate for molts 10 and 11) and computed the linear regression for growth (postmolt CW on premolt CW) separately for each molt number through the ninth laboratory molt. All regressions had high r\dj values (0.95-0.99; N = 10-375), indicating that growth, as described by each molt number, fit a linear model. This similarity in pattern of growth among molts allowed us to compare the nine molt-specific regression lines in all pairwise com- binations using ANCOV A; table-wide significance of each ANCOV A category TWEEDALE ET AL.: JUVENILE STONE CRAB GROWTH 877

14

,-... 12 E -..-E 10 ~ 8 U

6 -0 E en 4 -0 a.. 2

0 0 2 4 6 8 10 12 Premolt CW (mm) Figure 2. Regression lines for molt increment oflaboratory-reared postsettlementjuvenile Menippe mercenaria. Solid line: calculated using values for all molts recorded (Y = l.l3X + 0.16, N = 530); dashed lines show 95% confidence interval. Dotted line: calculated using only first laboratory molts (Y = l.12X + 0.22, N = 301). CW: carapace width.

(slope and elevation) again was determined using the sequential Bonferroni tech- nique (Rice, 1989). We found no significant differences in the slopes or elevations of any pairwise combination oflines, indicating that the length of time the crabs were held in the laboratory did not significantly affect molt increment. Therefore, we considered each molt in the data set to be an independent data point and combined information from all molts to describe growth. For postsettlement juvenile M. mercenaria, growth can be modeled by the equation Y = 1.13X + 0.16 (1)

2 (r adj = 0.97; P 4:: 0.001), where Y = postmolt CW and X = premolt CW (Fig. 2). To further check for possible effects of laboratory holding time on growth, we used ANCOV A to compare the regression line generated using all molts (Equation 1) to a regression line generated using only the first laboratory molt (the molt presumed to be least affected by laboratory holding time). The lines were virtually identical (Fig. 2), further supporting the idea that laboratory holding time had no discernible effect on molt increment and the validity of using every laboratory molt in our analysis of growth. To define the relationship between molt increment (postmolt CW minus pre- molt CW) and crab size, a regression line was generated, yielding the model Y=0.13X+0.17 (2)

(r2adj = 0.28), where Y = molt increment and X = premolt CW (Fig. 3A). Although 2 the variance was high, as indicated by the relatively low r adj value, molt increment 878 BULLETIN OF MARINE SOENCE, VOL. 52, NO.3, 1993 A. "E 6 E •••••••• 5 +-c 4 0 CD E 3 0 CD 0 0 •••• 0 ...,Sl u 2 c 1 +- 0 0 ::IE 0 2 4 6 8 10 12

8. "'C 0 6 e_ •••• CD 5 a.. - - +- 4 0 E 3 •••• CD +- 2 c o 1 " c 0 ~ 0 246 8 10 12 Premolt CW (mm) Figure 3. Relationship of growth parameters to crab size (premolt carapace width leW]) for labo- ratory-reared postsettlement juvenile Menippe mercenaria. Dashed lines define 95% confidence in- tervals. A. Molt increment. Solid line: Y = 0.13X + 0.17, N = 530. B. Intermolt period (days). Solid line: In Y = 0.14X + 2.64, N = 213. TWEEOALE ET AL.: JUVENILE STONE CRAB GROWTH 879

Table 2. Size (mean [x] and standard deviation [SO] of carapace width [mm CW)) of postsettlement juvenile Menippe mercenaria at each crab instar. Definitions of data sets used for determining size- at-instar and size-at-molt are presented in the text

Size·al-instar Size-at-molt Size classes (N - 5) (N = 10) (N = 15) Instar xCW SO xCW SO CW Range

I 2.2 (0.2) 1.9-2.4 0.8 2 2.8 (0.3) 2.7 (0.4) 2.5-2.9 0.5 3 3.4 (0.5) 3.2 (0.6) 3.0-3.5 0.6 4 3.8 (0.6) 3.8 (0.7) 3.6-4.1 0.6 5 4.3 (0.6) 4.5 (0.7) 4.2-4.8 0.7 6 5.1 (0.9) 5.2 (0.9) 4.9-5.5 0.7 7 5.8 (0.9) 5.8 (0.9) 5.6-6.3 0.8 8 6.7 (1.1) 6.4-7.3 1.0 9 7.7 ( 1.5) 7.4-8.3 1.0 10 8.7 (1.7) 8.4-9.3 1.0 II 9.4-10.4 1.1

increased significantly (P < 0.001) as CW increased. Mean percentage of growth per molt was 18% (SD = 0.08) over the entire size range and was not related to SIze. Intermolt Period.-Intermolt period (number of days between molts) typically increases as the size of the organism increases. We used regression analysis to determine the relationship of the length ofintermolt period to the size of the crab (Fig. 3B). The relationship was best described by the equation In Y = 0.14X + 2.64 (3)

2 (r •dj = 0.18; P < 0.001), where Y = intermolt period and X = premolt CWo Throughout the size range of the crabs in our study, intermolt period increased logarithmically as CW increased. Consequently, we grouped the data into size classes to evaluate size-at-age. Rather than simply divide the data into uniform size-class increments, our objective was to establish size classes that presumably had biological relevance. We attempted to define size classes that reflected the sizes of crab instars by calculating the mean and standard deviation of crab size (mm CW) separately at each molt using two subsets of the data. The first data subset consisted of values for the intermolt periods of crabs raised from megalopae (N = 5); from this subset we calculated mean size-at-instar for each of the first seven instars. The second data set, from which we obtained size-at-molt, was composed of the values for intermolt period of crabs that had been brought into the laboratory at a size of ::::;3.2mm CW (the size of the largest instar-2 crab in the first data subset) and that had molted at least eight times in the laboratory (N = 10). We presumed that these crabs had molted only once prior to collection and were therefore initially in the second crab stage. The means and standard deviations of size-at-instar (calculated using data subset one) and size-at-molt (calculated using data subset two) agreed closely (Table 2). Therefore, we used the combined subset data to construct size classes that presumably reflect the size of post settlement stone crabs at instars 1-10 (Table 2). Estimates for size class 11 were made by averaging the size of all crabs ~9.0 mm CW (N = 8). Within each defined size class, the data consisted of values for the intermolt periods of crabs that had been held in the laboratory for various lengths of time 880 BULLETIN OF MARINE SCIENCE, VOL. 52, NO.3, 1993

Table 3. Results of one-way tests for size-class-specific significant differences in intennolt period among molt numbers in laboratory-reared postsettlement juvenile Menippe rnercenaria. See Table I for definition of size classes. ANOYA: analysis of variance; K-W: Kruskal-Wallis test; Duncan: Dun- can's New Multiple Range test; N: sample size; P: probability level; x MP: mean length of intennolt period (days); *: significant at Ci = 0.05 with Bonferroni adjustment (Rice, 1989). Molt numbers are in parentheses. Underlined groups of mean intennolt periods show statistically homogeneous means

Size class Test N Statistic P xMP

I 2 ANOYA 9 F=12.21* 0.01 Duncan 39.0 (3) 26.0 (4) 12.8 (2) 3 ANOYA 20 F = 1.84 0.18 4 ANOYA 32 F= 4.50 0.01 5 ANOYA 25 F = 1.58 0.32 6 K-W 34 x2 = 9.98 0.13 7 ANOYA 29 F = 1.65 0.18 8 ANOYA 27 F = 0.54 0.77 9 ANOYA 15 F= 3.12 0.07 10 ANOYA 10 F = 3.04 0.13 11 ANOVA 7 F = 1.16 0.51

and that had molted various numbers of times. Therefore, before calculating an estimated intermolt period for each size class, we first determined if the time the crabs were held in the laboratory affected intermolt period. Within each size class, we tested for significant differences in mean length of intermolt period among molt numbers (which represented holding time in the laboratory). When the log-transformed data met parametric assumptions, we used single classification analysis of variance (ANOV A) and Duncan's New Multiple Range test; when parametric assumptions were not met after log transformation, we used the Krus- kal-Wallis test on untransformed data. To determine true significance for multiple tests ofthe same hypothesis, the sequential Bonferroni test was applied to F-values for the parametric analyses. No significant differences in intermolt period among molt numbers within a size class were found except in size class 2, in which sample sizes of outlying intermolt periods (MP) were small (MP-3: N = 1; MP-2: N = 2; Table 3). Therefore, within each class, we pooled all intermolt periods to calculate observed length of the intermolt period for that size class. For each size class, three values for length of intermolt period (maximum, median, mean) were calculated using each of two subsets of the data. One subset was composed of the values for all complete intermolt periods observed in the laboratory; the other was generated using partial intermolt periods (time until first molt in the laboratory; following Restrepo [1989]). For calculating length of in- termolt period within a size class using the partial intermolt periods, the estimate for maximum length ofintermolt period was made using the maximum time until first molt of an individual in the size class and the estimates for median and mean lengths of inter molt period were made using twice the calculated size-class-specific median and mean times, respectively, until first molt. These estimates were based on the assumption that the intermolt periods of individuals collected from the field were uniformly distributed in the intermolt cycle (Restrepo, 1989). As might be expected, the size-specific estimates of maximum intermolt period made using the partial intermolt periods were usually shorter than the corre- sponding values obtained using the observed complete intermolt periods (Table 4). Conversely, size-specific estimates of median and mean intermolt periods calculated using the partial intermolt periods typically were longer than the cor- responding median and mean values calculated from the observed intermolt pe- TWEEDALE ET AL.: JUVENILE STONE CRAB GROWTH 881

Table 4. Size-class-specific observed (Obs) and estimated (Est) maximum, median, and mean (stan- dard deviation [SD] in parentheses) lengths of intermolt period (days) for laboratory-reared postset- tlement Menippe mercenaria. Median and mean of estimated intermolt period are calculated as median (X 2) and mean (X 2) days until first laboratory molt. Asterisks denote size classes (defined in Table I) in which observed and estimated values were significantly different (P :s;0.05)

N Maximum Median Mean (SD) Size class Obs Est Obs Est Obs Est Obs Est

I 5 24 30 29 23 32 15 (8) 22 (7) 2 9 35 43 40 16 26 20 (12) 28 (7) 3 20 44 42 30 18 25 22 (13) 29 (9) 4 32 50 79 49 27 25 31 (18) 30 (10) 5 25 28 63 52 32 22 31 (15) 31 (13) 6 34 40 64 59 37 43* 36 (17) 49 (13)* 7 29 28 96 54 40 54 43 (22) 53 (17) 8 27 12 96 58 36 34 43 (19) 48 (18) 9 15 15 76 62 40 80 46 (18) 66 (19)* 10 10 6 88 61 55 97 56 (21) 84 (16) II 7 5 88 55 20 80* 38 (22) 78 (12)*

riods. Mann-Whitney V-tests performed separately on each size class indicated that the mean of the estimated intermolt period was significantly different from the mean of the observed intermolt period for size class 11 (P = 0.02), and was marginally significant for size classes 6 and 9 (P = 0.05). The medians of the estimated intermolt periods were significantly different (median test, PROC NPARI WAY; SAS Institute, Inc., 1985) from the medians of the observed in- termolt periods for size classes 6 and 11 (P = 0.04 and P = 0.01, respectively). In crabs, age is usually estimated by summing the intermolt periods to generate a growth curve (Hiatt, 1948; Kurata, 1962; Hartnoll, 1982, 1983). We estimated age (number of days after metamorphosis to crab stage) of juvenile M. mercenaria at the end of instar 11 (approximately 9.4-10.4 mm CW) by using the maxima, medians, and means of calculated intermolt periods from Table 4 to construct growth curves for postsettlement juvenile M. mercenaria (Fig. 4). The estimates of time necessary to complete growth through size class 11 varied widely (from 344 days for median of observed intermolt period to 765 days for maximum of observed intermolt period), and the estimated and observed data sets generally produced highly disparate projections of age at the end of this instar for each type of estimate. In addition, because the maximum value for age at the end of instar 11 is based on the cumulative intermolt period values of single crabs in each size class, the observed complete intermolt period deviated markedly from the cor- responding median and mean values, whereas all three predictions of age at this size derived using the partial intermolt period data were more similar. For the maximum intermolt period, the cumulative effect of size-specific differences be- tween the values obtained was that a considerably longer time interval (31%) was projected for growth through size class 11 using observed intermolt period data. More importantly, for the more commonly used median and mean intermolt periods, projections of age through this size class derived from the observed intermolt period data were notably shorter (46% for the median; 39% for the mean) than those obtained from the partial intermolt period data.

DISCUSSION Our characterization of molt increment in postsettlement juvenile Menippe mercenaria is similar to that found in other crustaceans (Kurata, 1962; Hartnoll, 882 BULLETIN OF MARINE SCIENCE, VOL. 52, NO.3, 1993

A. 10 ~ 8 0 Median en 6 T 344 - o c 4 Te 518 2 0

B. ,------10 . ~ 8 Mean 0 en 6 T 381 -c 0 4 Te 518 2 0

c. , 10 .---- .. ~ 8 0 Maximum en 6 -c T 765 4 0 ,rr. T 549 2 e 0 0 200 400 600 800 Time (days)

Figure 4. Growth curves of laboratory-reared Menippe mercenaria constructed from the six different estimates of intermolt period listed in Table 3. A. Median intermolt period. B. Mean intermolt period. C. Maximum intermolt period. Solid lines: observed intermolt periods; dashed lines: estimated in- termolt periods. Values ofT are estimates of age (days after metamorphosis to crab stage) at approx- imately lO-mm carapace width, calculated by summing the mean observed (TJ or estimated (Te) intermolt periods for size classes I-II (defined in Table I).

1982 and references therein): absolute molt increment increases as size of the organism increases. However, contrary to other reports in which the percentage growth increment in M. mercenaria decreased as crab size increased (Savage and McMahan, 1968; Bert et al., 1986; Restrepo, 1989), proportional growth in our TWEEDALE ET AL.: JUVENILE STONE CRAB GROWTH 883 study was not correlated with size, probably because we considered only a limited SIzerange. The growth parameters documented in laboratory studies are frequently pre- sumed to be representative of the growth parameters of in the wild. However, laboratory studies of growth for which there are no comparisons with field observations should be interpreted conservatively (Hartnoll, 1982) because molt increments of individuals reared in the laboratory may be equivalent to (Kurata, 1962; Tagatz, 1968), smaller than (Hiatt, 1948; Savage and McMahan, 1968; Sweat, 1968; Savage, 1971; Hogarth, 1975), or larger than (Chittleborough, 1976) molt increments of individuals in the field. In using laboratory-generated growth parameters to estimate age-at-size, we infer that our laboratory growth rates can be extrapolated to the field. We detected no change in molt increment attributable to laboratory holding time, indicating that laboratory rearing had either no effect or an immediate and consistent effect on molt increment. Two lines of evidence suggest that laboratory effects on molt increment in the indi- viduals we analyzed were minimal or absent: (1) no difference in molt increment among molt numbers was found, despite the high degree of variation in laboratory holding conditions among the three studies used in our analysis; and (2) unlike earlier laboratory studies of growth in M. mercenariajuveniles and adults (Savage and McMahan, 1968; Savage, 1971; Savage and Sullivan, 1978), we observed no decrease in molt increment at laboratory molt 1 compared to the molt increments of subsequent laboratory molts. Thus, unless undetected laboratory effects oc- curred, the growth rates we observed should reflect those that occur in nature. Our relationship of premo It carapace width to postmolt carapace width is com- parable to those generated for larger juvenile and adult M. mercenaria (Savage and McMahan, 1968; Savage, 1971; Restrepo, 1989); however, the slope of our equation is somewhat less than the slopes calculated for growth relationships of crabs in larger size classes. Reported size-specific molt increments of adult M. mercenaria measured in the field (Sullivan, 1979; Bert et al., 1986) are greater than those projected by our equation when the predicted molt increment is ex- trapolated to adult sizes. Our equation may be specific only to the growth pattern of postsettlement M. mercenaria; growth pattern typically changes with age, and separate regressions for different life stages are usually required (Hiatt, 1948; Kurata, 1962; Hartnoll, 1982). The intermolt period of post settlement M. mercenaria followed a pattern typical for crustaceans, increasing log-linearly with size. A similar exponential relation- ship between intermolt period and size was calculated for male M. mercenaria (Restrepo, 1989). Although no effect of laboratory holding time on intermolt period was statistically detectable from our analyses, our intermolt periods were long when compared to those reported for juvenile Florida stone crabs reared in different types of environments (Savage, 1971; Yang and Krantz, 1976). Holding containers and holding conditions used in the three studies we evaluated were diverse but, nevertheless, may have somehow affected intermolt period below a statistically detectable level. Alternatively, an undetected effect may have occurred immediately after capture. To calculate the three determinants of inter molt period, we used the commonly employed method of observed complete intermolt periods and a method, pre- viously applied to male M. mercenaria, that uses statistical estimates (Restrepo, 1989). Within each method, the mean and medians derived from the observed intermolt periods yielded similar age-at-size predictions. However, our predic- tions of age at the end of instar 11 obtained using the mean and median intermolt periods calculated from the partial intermolt periods were longer than those made 884 BULLETIN OF MARINE SCIENCE. VOL. 52. NO.3. 1993 using the means and medians derived from the observed lengths of intermolt period. Error is inherent in each estimate of intermolt period; when intermolt periods for individual size classes are used for projecting long-term growth, sum- ming intermolt periods to predict age-at-size compounds the effect of the error component. Predictions of age-at-size from our equations of molt increment and intermolt period suggest that approximately one year is required to attain 10-mm carapace width. Our growth model is consistent with the minimum rate of growth of postsettlement juveniles inferred from data collected in Tampa Bay, Florida, (Florida Marine Research Institute, unpubl.) but these field studies suggest that rapidly growing juveniles may grow to 10-mm CW in the field within six months. Order-of-magnitude growth differences in juvenile Florida stone crabs have been found in a mariculture system that simulated natural conditions (Yang and Krantz, 1976), and considerable variation could also be expected in nature. Nevertheless, the growth rates we established for postsettlement juvenile M. mercenaria may have important implications for management of the Florida stone crab fishery. Slow-growing individuals may require two years to reach the larger juvenile size classes (e.g., 30-35 mm CW)o Previous studies (Savage, 1971; Yang, 1971, 1972; Savage and Sullivan, 1978) may have underestimated the time required for M. mercenaria to progress through the postsettlement period and consequently may have underestimated the time required to grow to adulthood. Studies of growth for all life-history stages of other commercially important crabs (Tagatz, 1968; Orensanz and Gallucci, 1988) and evidence presented here suggest that a thorough survey of growth in all life stages of M. mercenaria is needed to understand the time required for this species to attain reproductive maturity and legal size.

ACKNOWLEDGMENTS

We thank H. Cruz-Lopez, J. Risser, D. Romero and P. Steele for field and laboratory assistance and D. Camp for statistical assistance. We also thank T. Perkins and G. Bruger for reviewing the manuscript. This research was funded by grants from the Department of Commerce, National Ocean- ographic and Atmospheric Administration (P.L. 99-659 [Project 2-11-2]) and from the Florida Sea Grant College Program (Project R/LR-B-24), and by the State of Florida, Department of Natural Resources, Florida Marine Research Institute.

LITERATURE CITED

Bert, T. M., R. E. Warner and L. D. Kessler. 1978. The biology and Florida fishery of the stone crab Menippe mercenaria (Say), with emphasis on southwest Florida. Fla. Sea Grant Tech. Pap. NO.9. 82 pp. --, J. Tilmant, J. Dodrill and G. E. Davis. 1986. Aspects of the population dynamics and biology of the stone crab (Menippe mercenaria) in Everglades and Biscayne National Parks as determined by trapping. South Fla. Res. Cent. Rep. SFRC-86/04. 77 pp. Brown, S. D., W. A. Tweedale, T. M. Bert and J. J. Torres. 1992. The effects of temperature and salinity on survival and development of early life stage Florida stone crabs (Menippe mercenaria). J. Exp. Mar. BioI. Ecol. 157: 115-\36. Chittleborough, R. G. 1976. Growth of juvenile Panulirus longipes cygnus George on coastal reefs compared with those reared under optimal environmental conditions. Aust. J. Mar. Freshwater Res. 27: 279-295. Field, C. J. 1989. Effects of salinity and temperature on survival and development time of the larvae of stone crabs, Menippe mercenaria (Say, 1818) and Williams and Felder, 1986 (Decapoda: Brachyura: Xanthidae). M.S. Thesis, Louisiana State University, Baton Rouge, Lou- isiana. 157 pp. Hartnoll, R. G. 1982. Growth. Pages 111-196 in L. G. Abele, ed. The biology of Crustacea, Vol. 2. Academic Press, Inc., New York. --. 1983. Strategies of growth. Mem. Aust. Mus. 18: 121-131. Hiatt, R. W. 1948. The biology of the lined shore crab, Pachygrapsus crassipes Randall. Pac. Sci. 2: 135-213. TWEEDALEETAL.:JUVENILESTONECRABGROWTH 885

Hogarth, P. J. 1975. Instar number and growth of juvenile Carcinus maenus (L.) (Decapoda, Brachy- ura). Crustaceana 29(3): 299-300. Kurata, H. 1962. Studies on the age and growth of Crustacea. Bull. Hokkaido Reg. Fish. Res. Lab. 24: I-liS. Lindberg, W. J. and M. J. Marshall. 1984. Species profiles: life histories and environmental require- ments (South F1orida)-stone crab. U.S. Fish Wildl. Servo FWS/OBS-82/11.21. U.S. Army Corps Eng. TR EL-82-4. 17 pp. Ong, K. S. and J. D. Costlow. 1970. The effect of salinity and temperature on the larval development of the stone crab Menippe mercenaria (Say), reared in the laboratory. Chesapeake Sci. 11: 16-29. Orensanz, J. M. and Y. F. Gallucci. 1988. Comparative study ofpostlarvallife-history schedules in four sympatric species of Cancer (Decapoda: Brachyura: Cancridae). J. Crustacean BioI. 8(2): 187-220. Passano, L. M. 1960. Molting and its control. Pages 473-536 in T. H. Waterman, ed. Physiology of Crustacea, Yol. 1. Academic Press, Inc., New York. Restrepo, Y. R. 1989. Growth estimates for male stone crabs along the southwest coast of Florida: a synthesis of available data and methods. Trans. Am. Fish. Soc. 118: 20-29. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43(1): 223-225. SAS Institute, Incorporated. 1985. SAS user's guide: statistics, version 5 ed. SAS Institute, Inc., Cary, North Carolina. 956 pp. Savage, T. 1971. Effect of maintenance parameters on growth of the stone crab, Menippe mercenaria (Say). Fla. Dep. Nat. Resour. Spec. Sci. Rep. No. 28. 19 pp. -- and M. R. McMahan. 1968. Growth of early juvenile stone crabs, Menippe mercenaria (Say, ]819). Fla. State Board Conserv. Spec. Sci. Rep. NO.2!. ]8 pp. --- and J. R. Sullivan. 1978. Growth and claw regeneration of the stone crab, Menippe mercenaria. Fla. Mar. Res. Publ. No. 32. 23 pp. ---, --- and C. E. Kalman. 1975. An analysis of stone crab (Menippe mercenaria) landings on Florida's west coast, with a brief synopsis of the fishery. Fla. Mar. Res. Publ. No. 13. 37 pp. Simonson, J. L. 1985. Reversal of handedness, growth, and claw stridu]atory patterns in the stone crab Menippe mercenaria (Say) (Crustacea: Xanthidae). J. Crustacean Biol. 5(2); 281-293. --- and P. Steele. 1981. Cheliped asymmetry in the stone crab, Menippe mercenaria, with notes on claw reversal and regeneration. Northeast Gulf Sci. 5(1): 21-30. Snedecor, G. W. and W. G. Cochran. 1969. Statistical methods. Iowa State University Press, Ames, Iowa. 593 pp. Sokal, R. R. and F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman and Company, San Francisco, California. 859 pp. Sullivan, J. R. 1979. The stone crab, Menippe mercenaria. in the southwest Florida fishery. Fla. Mar. Res. Pub I. No. 36. 37 pp. Sweat, D. E. 1968. Growth and tagging studies on Panu/irus argus (Latreille) in the Florida Keys. Fla. State Board Conserv. Tech. SeT. No. 57. 30 pp. Tagatz, M. E. 1968. Growth of juvenile blue crabs, Callinectes sapidus Rathbun, in the St. Johns River, Florida. Fish. Bull. (U.S.) 67(2): 281-288. Yang, W. T. 1971. Preliminary report of the culture of the stone crab, Menippe mercenaria (Say). Proc. 2nd Annu. Workshop, World Mariculture Soc. 2: 53-54. Yang, W. T. 1972. Notes on the successful reproduction of stone crabs, Menippe mercenaria (Say) reared from eggs. Proc. 3rd Annu. Workshop, World Mariculture Soc. 3: ]83-184. --- and G. E. Krantz. 1976. "Intensive" culture of the stone crab Menippe mercenaria. Fla. Sea Grant Tech. Pap. No. 35. 15 pp.

DATEACCEPTED: May II, 1992.

ADDRESS: Department of Natural Resources, Florida Marine Research Institute, 100 Eighth Ave. SE, St. Petersburg, Florida 33701 .. PRESENTADDRESS: (W.A. T.) 313 Laurel Street #6. Palatka, Florida 32 J 77. Send request for reprints to T.M.B.