were the result of large predators. The shells were 

 clean and some were mixed within the surface 

 layer on the bottom. In addition, the predators had 

 created pits 50 cm in diameter and 6-10 cm deep in 

 the aggregate and substrate which had been cover- 

 ing the clams. 



To eliminate effects of losses due to predation 

 during the first year's study and concentrate on 

 the effects due to these predators, we have utilized 

 the estimate of the mean number of clams from the 

 July 1977 samples as 100% of those present for 

 further predation. The estimated numbers of 

 clams in each experimental plot for July 1977 are 

 given in Table 2, and the number of clams remain- 

 ing for the corresponding treatments from the Oc- 

 tober to November 1977 sampling are given in the 

 same table. Several important aspects not evident 

 from the ANOVA table are apparent. A combina- 

 tion of baffles, gravel, pens, and traps was essen- 

 tial for high survival. Pens were significant only 

 because of the predation between July and Oc- 

 tober. The percentage survival between these two 

 sampling periods seemed to indicate that gravel 

 somehow negatively interacted with the baffles 

 (compare percent survival B + G and B + NG, 

 Table 2) when pens were absent. This was not the 

 case, but resulted from heavy predation in the 

 baffle + gravel sites with no pens. Since there were 

 more clams in these areas in July, the percent 

 survival was lower, but total survival was better 

 than in the baffle + no gravel sites (Table 2). The 

 higher survival in the baffled sites was due to the 

 protection the baffles offered the clams when the 

 predators entered the area. Almost all clams found 

 in these areas were close to, or beneath, the cross 

 rods supporting the bottom of the baffle. This same 

 shadow effect was the cause of the nonsignificant 



Table 2. — Total number of clams estimated from mean number 

 per sample (July 1977), total counts (October- November 1977), 

 and the percent mortality between the sampling dates. — = not 

 calculated because of estimate in July. P = pens,T = traps, G = 

 gravel, B = baffle. The prefix N = absence; NP = no pen, etc. 



Month or period 



Item 



B + G NB + G B + NG NBNG 



July 1977 



Oct.-Nov. 1977 



% survival 

 July to Oct.-Nov. 



baffle + gravel interaction in the October- 

 November ANOVA table. If more clams had been 

 present in the B + NG sites, the clams would have 

 been in the center of the plot and thus vulnerable 

 to predation. 



The impact of predation to the mariculture of 

 clams can be seen by comparing survival inside 

 and outside the penned sites (Table 2). Estimated 

 survival inside a penned area was always more 

 than 76%, and the average survival for both 

 penned sites was 94% from July to October- 

 November. Average survival for the same period in 

 the unpenned sites was 8.75%. The greatest survi- 

 val in the unpenned areas was 65% but, as ex- 

 plained above, was due to protection provided by 

 baffle frames. These data indicate that at least 

 85-90% of the observed losses in the unpenned 

 sites were due to predation. The importance of 

 these data is amplified when the size of the clams 

 is considered. 



The average size of clams in July 1977 was 3.2 

 cm and by October was 3.9 cm (hinge to lip). The 

 percentage marketable clams (1 in (25.4 mm) 

 thick) was the same for both the penned and un- 

 penned sites (58.5 and 58.6%) in October. This 

 indicates no size selection of clams, but that clams 

 of all sizes were consumed. The loss of such large 

 clams represents 2 yr of work and a product of 

 market size. 



Flounders, known to prey on young Mercenaria 

 mercenaria and to selectively eat the neck of adult 

 clams, have been eliminated as potential pred- 

 ators because they are not capable of crushing the 

 shell of 3 cm high hard clams. Of the seven species 

 of fish capable of forming pits and crushing the 

 shell of 3 cm size hard clams (Table 3), only two are 

 known to be common away from the inlets and 

 near the planted areas (Richards and Castagna 

 1970; Musick 1972). These two species, Z)asya^/s 

 centroura andRhinoptera bonasus, are prime sus- 

 pects for causing the destruction in our unpro- 

 tected plots. The former cannot be eliminated be- 



TABLE 3. — Potential fish predators on 3 cm hard clams in Vir- 

 ginia. Information from Richards and Castagna ( 1970) and 

 Musick (1972). 



Scientific name 



Common name 



Dasyatis americana 

 D. centroura 

 D. sayi 



Myliobatis freminvillei 

 Aetobatus narinari 

 Rhinoptera bonasus 

 Pogonias cromis 



Souttiern stingray 

 Roughtall stingray 

 Bluntnose stingray 

 Bullnose ray 

 Spotted eagle ray 

 Cownose ray 

 Black drum 



540 



