MURAWSKI KT AL.: GROWTH OF OCEAN QUAHOG. ARCTICA ISLANDICA 



marginal increments of shell growth from the 

 last external and internal marks to the shell 

 edge, whereas winter samples had recently 

 formed annuli (Fig. lc; Table 3). Thus, mark for- 

 mation probably occurs during the last half of 

 the calendar year. These observations are consis- 

 tent with data presented by Jones (1980). In a 

 study of the seasonality of incremental shell 

 growth, he noted that internal growth bands in 

 shell cross sections were formed from September 

 to February. The formation of growth bands 

 apparently overlaps the spawning period (Jones 

 1980); however, both events may be related to 

 other physiological or environmental stimuli 

 since specimens that were reproductively imma- 

 ture formed bands concurrently with mature 

 ocean quahogs. 



Back-calculated mean lengths at age varied 

 considerably depending on the subset of data 

 analyzed in Table 3. Mean lengths at age for all 

 year classes (bottom rows in Table 3) were gener- 

 ally smaller than mean lengths at the last com- 

 plete annulus (rightmost diagonal vector), and 

 growth of recent age groups (2-8) appeared more 

 rapid than for older ocean quahogs (Lee's phe- 

 nomenon; see Ricker 1969). However, conclu- 

 sions regarding the growth of older age groups 

 (9-18) are tenuous due to the relatively small 

 numbers of these ages sampled (87% of the sam- 

 ples were <8-yr-old). 



Age analyses were limited to ocean quahogs 

 that exhibited suitable contrast on the shell sur- 

 face to discern external concentric rings. Thus, 

 the oldest aged ocean quahogs (particularly ages 

 14-18) may represent the smallest, slowest grow- 

 ing individuals of their year classes; faster grow- 

 ing individuals may have reached sizes asso- 

 ciated with color changes of the periostracum. 

 Nevertheless, back-calculated mean lengths at 

 age for 14- to 18-yr-old ocean quahogs did not 

 tend to be progressively smaller than means for 

 ages 9-13, perhaps indicating that size selectivity 

 of older individuals was not a significant bias 

 (Table 3). 



The objectives of fitting statistical models to 

 age-length data were to describe growth during 

 the juvenile and early adult phases of life, and 

 more importantly, to predict ages associated 

 with the lengths of the smallest recaptured speci- 

 mens (59-65 mm) thereby linking the age-length 

 data and mark-recapture results into a contin- 

 uous growth function. Recognizing the disparate 

 nature of data subsets in Table 3, a series of ex- 

 ponential and logistic growth equations were 



fitted to: 1) weighted mean back-calculated 

 lengths at age for all quahogs, 2) weighted mean 

 lengths at age for ages 2-8, and 3) mean lengths 

 at the last completed annuli (rightmost diagonal 

 vector) for ages 2-10 and 2-13. For our purposes, 

 the applicability of a particular model fit was 

 judged not only by the total amount of variance 

 between length and age explained by the equa- 

 tion, but by predicted annual growth increments 

 in the 59-65 mm range. An appropriate model 

 would fit as much of the age-sample data as pos- 

 sible and yield calculated annual growth incre- 

 ments consistent with those observed from re- 

 captured specimens. 



Exponential equations utilizing weighted 

 mean back-calculated lengths for ages 2-8, and 

 lengths at the last complete annulus for ages 2-13 

 yielded unacceptable fits by our criteria. The 

 former equation was calculated with informa- 

 tion from the linear portion of the growth curve, 

 predicted lengths beyond age 8 were unrealistic- 

 ally high. The latter equation incorporated one 

 negative growth increment (between ages 11 and 

 12) and thus the calculated asymptote was only 

 62.8 mm; predicted annual growth near the 

 asymptote was considerably less than observed 

 increments for that size (Fig. 2). 



The logistic growth equation fitted to weighted 

 mean lengths at age for all ocean quahogs (SL = 

 52.09/1 + exp(2.4722 - 0.4702(0)) was superior 

 to the respective exponential fit considering the 

 residual sums of squares criterion. The reverse 

 was true for the logistic equation describing 

 mean lengths at the last annulus for ages 2-10 

 (SL = 43.12/1 + exp(2.9361 - 0.8069 (*))). How- 

 ever, asymptotic lengths were, for both logistic 

 equations, well below the range of shell lengths 

 considered in the mark-recapture experiments. 

 Thus, extrapolation of logistic age-length re- 

 lationships, necessary for initializing the Ford- 

 Walford equation, was not feasible. On the 

 contrary, the two exponential equations yielded 

 reasonable asymptotic lengths and adequately 

 described ocean quahog growth relative to that 

 inferred from modal progressions in 1970 and 

 1980 length-frequency distributions (Fig. 6) and 

 observed growth increments (Fig. 2). 



Exponential growth equations computed from 

 weighted mean lengths at age for all ocean qua- 

 hogs and mean lengths at the last annulus for 

 ages 2-10 were: SL = 75.68-81.31 (0.9056)' and 

 SL = 72.70-75.22 (0.8935)', respectively. Mean 

 lengths at age predicted from the two equations 

 generally reflect differences among data sets 



31 



