length PVC pipe, sealed at the base, with a removable cap, and 

 fitted with a handle for the submarine's manipulator arm (Figure 

 5). 



The arrays were deployed from the surface after sites were 

 chosen using side-scan sonar imagery. The arrays were attached 

 to 60 m lengths of anchored line, stretched out laterally during 

 deployment (practical in the <60 m depths selected) . After 

 deployment, the "Delta" submersible descended and searched first 

 for the aligned anchor line, and then followed it to the array. 

 Future surveys would be more practical if a pinger were attached 

 to the array. Upon location of the array, the traps were 

 installed, uncapped, and the site was photographed (Figure 6) , 

 paying particular emphasis to bedforms. The Kennebec paleodelta 

 arrays were emplaced in June, 1987, and revisited one month 

 later, at which time the traps were collected and replaced. The 

 Saco Bay arrays were emplaced in July, 1987. All arrays will be 

 revisted in the summer of 1988. 



Sediment samples were analyzed by standard sieve techniques 

 (sand fraction) , and by Micromeritics Sedigraph 5000 ET (mud 

 fraction) . Sediment trap samples were pre-treated with bleach to 

 remove organics. Size statistics were calculated with our 

 laboratory program "LGRAIN" (Belknap, unpub.), which yields Folk- 

 Ward and Method of Moment statistics. 



Finally, direct observations of the sedimentary environments 

 were made through submersible observations. Twenty-nine dives 

 over the past four years have traversed more than thirty 

 kilometers of sea floor (Table 1) . Still photos, annotated 

 videotapes, and logs of personal observations comprise the record 

 of these dives, which are on file at the University of Maine. 

 Most dives were directly along traverses that had previously been 

 studied by seismic reflection profiling and/or side-scan sonar. 



153 



