broken down into legal (>2.54 cm thickness) and il- 

 legal sizes. The actual area raked inside 5 pairs of 

 plots was then systematically sieved by hand through 

 6 mm mesh to a depth of 12 cm to estimate the num- 

 bers of legal- and illegal-sized clams missed during 

 raking. These observations permit a quantitative 

 comparison of the two rakes: 1) Rate of hard clam 

 capture and 2) efficiency of hard clam capture for 

 both size classes of clams in the sand-flat habitat. 



Seagrass Bed Methods 



We chose matched 2 X 2 m plots, which we then 

 marked with 1.6 m stakes. Plots used here were 

 smaller than in the sand flat, because the presence of 

 seagrasses and higher clam densities slowed the rak- 

 ing and reduced the area covered in 6 min. We select- 

 ed 5 groups of 3 plots each, two for application of 

 each rake and a third as a control to estimate initial 

 seagrass biomass. Raking, sieving, and data record- 

 ing were carried out in the same fashion as in the sand 

 flat. In addition to measuring the area covered by 

 each rake in 6 min for each plot and counting the 

 numbers of hard clams (in the two size classes) 

 collected, we also excavated by hand and placed into 

 buckets all fresh seagrass material left behind in each 

 raked area and from aim 2 area within each control 

 plot. We returned all seagrass material to the labo- 

 ratory where we washed away salt and sedi- 

 ments, separated by clipping aboveground com- 

 ponents (blades and shoots) from belowground 

 components (roots and rhizomes), and weighed each 

 separately after drying to constant weight at 105°C. 

 These data permit a quantitative comparison of the 

 two rakes in 1) rate of hard clam capture and 2) ef- 

 ficiency of hard clam capture for both legal- and 

 illegal-sized clams in the seagrass habitat, analogous 

 to the sand-flat contrasts. By subtracting the dry 

 weight of seagrass remaining in raked areas from the 

 dry weight in the matched controls, we were also able 

 to estimate the mass of above- and belowground 

 seagrass removed by each rake. We then used these 

 figures to estimate the relative environmental impact 

 of each rake in the form of estimated dry weight of 

 seagrass removed 1) per unit time, 2) per unit area 

 raked, and 3) per legal-sized clam captured in the 

 seagrass habitat. 



Results 



Sand-Flat Habitat 



habitat, with a mean catch more than 50% higher 

 than that of the bull rake (Table 1). The rate of cap- 

 ture of illegal-sized clams was equally low for both 

 rakes in this habitat. Although both rakes were 100% 

 efficient in their capture of legal-sized clams inside 

 the areas raked in this environment, the pea digger 

 covered significantly more area during a fixed period 

 of time (Table 1) and, therefore, was able to catch 

 more clams than the bull rake. Because of equal cap- 

 ture efficiency, the average numbers of legal-sized 

 and illegal-sized clams caught per unit area raked did 

 not differ significantly between rakes in the sand-flat 

 environment (Table 1). 



Table 1. — Hard clam capture rate per unit time, per unit area raked, 

 and capture efficiency of two clam rakes from 14 paired replicate 

 plots of a sand flat. Complete excavation to estimate capture ef- 

 ficiency was done for only 5 of the 14 pairs. F- tests revealed no 

 significant difference between treatments in variance, except for 

 area raked which required a log transformation prior to performing 

 the (-test. 



The pea digger produced significantly more legal- 

 sized hard clams per unit time of use in the sand-flat 



1 * = P<0.05;" =P<0.01; ns = P >0.05. in a two-tailed paired Mest. 



2 >2.54 cm thick. 



3 Back-tr8nsformed mean of arcsin-transformed percents of clams 

 captured. 



'Insufficient densities of small clams prohibited replicate estimates of capture 

 efficiency, thus these percents are based on pooled totals (4 and 3. respectively) and 

 were tested by Fisher's exact test. 



Seagrass Habitat 



In the seagrass bed, the two rakes again differed 

 significantly in average catch of legal-sized clams per 

 6 min of raking; however, in contrast to the sand-flat 

 results, the bull rake was the more productive imple- 

 ment (Table 2). The bull rake also tended to catch 

 more small clams per unit time, although the num- 

 bers of clams caught in this size class were small and 

 the differences between rakes not statistically signi- 

 ficant (Table 2). The greater return from use of the 

 bull rake was mainly a consequence of the signifi- 

 cantly greater area raked per unit time. The number 

 of clams captured per unit area actually raked and 

 the efficiency of clam capture in areas actually raked 

 did not differ significantly between rakes for either 

 size class of hard clam (Table 2). 



A 6-min application of the bull rake in the seagrass 

 habitat caused an estimated loss of seagrass biomass 



431 



