PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING 



produced by the 2 controls (Fig. 3). In the seagrass, 

 M. mercenaria recruitment may also have been 

 reduced by harvest treatments (Table 4), but the 

 conclusion depends upon the assumption that the 

 shelly control I was an adequate control for recruit- 

 ment data. Given the enhanced survivorship of M. 

 mercenaria recruits in shell (Castagna and Kraeuter 

 1977) and the significant illegal clamming in sea- 

 grass control I, this assumption is questionable. 



It is possible that removal of adult hard clams 

 enhances larval settlement over a larger spatial scale 

 than the 1,225 m- experimental plots because 

 depletion of larvae by feeding from the water col- 

 umn should extend over a larger spatial scale (Peter- 

 son 1982b). Although it is possible that our sampling 

 was on too fine a scale to detect such an effect, our 

 sampling occurred on a far larger spatial scale by 

 3 orders of magnitude than any previous experi- 

 mental test of adult-larval interactions and, thus, 

 should have provided for greater opportunity to 

 detect any positive effect of adult hard clam 

 removal. The failure to demonstrate a response in 

 the sand flat may be a different consequence of 

 scale. Newly recruited hard clams may settle more 

 heavily where adult densities have been reduced but 

 the effect may be diffused away by the physical 

 dispersal of new recruits by tidal currents and 

 waves. As a consequence of such multiple interpre- 

 tations, we can best conclude that on the scale of 

 our experiments no dramatic increase in hard clam 

 recruitment occurs vdth intense mechanical harvest 

 of adult hard clams in seagrass and harvest may 

 even reduce recruitment in both unvegetated and 

 vegetated areas. 



The effect of various clam harvest treatments in 

 the seagrass bed on seagrass biomass (Fig. 4) is the 

 most obvious result of this study. Clam harvest of 

 all types had an immediate impact in reducing the 

 seagrass biomass. Reduction of seagrass increased 

 with harvest intensity, as was demonstrated both 

 by the enhanced effect of the second treatment ap- 

 plication, which was much more intense than the 

 first, and also by the larger effects of intense kick- 

 ing as compared with the other treatments (Fig. 4). 

 Although the seagrass biomass in the raking and 

 light-kicking matrices recovered to levels predicted 

 from the controls within a year's time, the seagrass 

 biomass in the intense-kicking matrices did not even 

 begin to recover for 2 years and had not fully re- 

 turned to predicted, control levels after 4 years. 

 These results imply that if sufficient seagrass is 

 destroyed, recovery is slow. Because our intense- 

 kicking treatment removed only an estimated 50% 

 of available hard clams and because the efficiency 



of hard clam capture per unit time of harvest was 

 greater in the intense treatment than in the light 

 treatment in the seagrass habitat, we suspect that 

 commercial clam kickers would apply even more 

 harvest intensity than we did in the this intense- 

 kicking treatment. Consequently, the effects of com- 

 mercial clam kicking in seagrass beds are probably 

 underestimated by our data (Fig. 4). Furthermore, 

 by using both control matrices (including the shelly 

 one) in estimating the effects of harvest on seagrass 

 biomass, we intentionally provide an additional con- 

 servative bias. Clam kicking at a low level (=15% 

 of available hard clams harvested) does not appear 

 to be any more destructive of seagrass than hand 

 raking that same number of clams, but the lack of 

 replication of these two types of treatment matrices 

 renders this a tentative conclusion. 



The extremely slow recovery of seagrass in the 

 intensely kicked seagrass matrices raises the possi- 

 bility that seagrass beds and unvegetated sand flats 

 may exist as alternative stable states (Sutherland 

 1974; Connell and Sousa 1983; Peterson 1984) on 

 many of the same shallow bottoms of sounds and 

 coastal lagoons. That is, a given shallow bottom may 

 exist as either a seagrass bed or an unvegetated 

 sand flat, but whichever state it occupies it is likely 

 to retain for a relatively long period of time. Trans- 

 formation from one state to another may require 

 some input of external energy. Because great 

 changes in current regime and surface sediment 

 character are associated with the presence and 

 growth of seagrasses (Ginsburg and Lowenstam 

 1958; Orth 1977; Fonseca et al. 1983; Peterson et 

 al. 1984; Eckman in press), it is reasonable to 

 hypothesize that destruction of seagrass may result 

 in sufficiently higher energy at that site that natural 

 reestablishment could be difficult. Certainly, the 

 slow return of seagrass following intense clam kick- 

 ing in our experiments implies that seagrass re- 

 covery even in previously vegetated areas is ten- 

 uous. If seagrass beds and unvegetated bottoms do 

 tend to represent alternative stable states for large 

 areas of the estuarine and sound bottom, then 

 denuding of vegetation would have long-lasting ef- 

 fects, even beyond what we have demonstrated. 

 Furthermore, transplantation of relatively dense 

 seagrass may be necessary to produce rapid rever- 

 sion back into a vegetated system (for reviews of 

 disturbance, recovery, and transplantation of sea- 

 grasses see Zieman 1982; Thayer et al. 1985). 

 Because of the important roles that seagrasses play 

 in promoting estuarine productivity and coastal 

 fisheries (Thayer et al. 1975), intense clam kicking 

 in vegetated areas could have long-lasting and 



295 



