Only one test (test 31) was performed in a pre- 

 dominately silt seafloor. Site information to 10 feet 

 indicated a reasonably competent material would be 

 encountered at the expected penetration depth of 20 

 to 30 feet; as a result, the small sand fluke was 

 chosen. This trend proved to be false; soil shear 

 strength did not increase with depth at the rate 

 expected as evidenced by the soil recovered from the 

 fluke. The holding capacity was 28,000 pounds, more 

 than design, but less than what should have been 

 obtained from the larger 2 x 4-foot clay fluke, which 

 is more optimum for this intermediate strength soil. 



Tests results in stiff clay also indicate that the 

 nominal design capacity is readily achieved; in partic- 

 ular, note test 23. This stiff clay was found in 12,200 

 feet about 120 miles off Norfolk, Virginia. The exact 

 composition of the seafloor was unknown when the 

 small sand/stiff clay fluke was chosen; it was a lucky 

 choice. Seafloor composition was determined from 

 the soil recovered on the fluke and from a subsequent 

 unsuccessful attempt (later coring cruise) to take a 

 long core. The corer bent after penetrating a few feet 

 of surficial silt. 



Several tests were performed in soft clay, and the 

 results are inconclusive. This is due to the variation in 

 holding capacity for soft clay even though various 

 fluke sizes are used. The short-term holding capacities 

 of both small and large flukes were generally about 

 20,000 pounds, except for tests 21 and 22 where an 

 extremely soft hemipelagic clay was encountered. 

 The small fluke was used for test 21, which resulted 

 in 13 kips, and the 2 x 4-foot fluke was used for test 

 22. The 8-kip load recorded for test 22 is misleading 

 because the fluke penetrated deeper than expected, 

 thereby apparently damaging the cable and fittings 

 when it ran out of cable travel while moving. As the 

 downhaul cable was being brought on deck, it was 

 noticed that it had a badly damaged fitting and that 

 the fluke was not hanging on the line. This fitting 

 parted under only the cable's weight and fell to the 

 seafloor. A longer downhaul cable was used in all 

 subsequent tests in soft clay. The larger clay fluke 

 could have been used for each of these tests. 



Test 27 was performed in a red clay, north of 

 Hawaii. It was difficult to keep the main load line in 

 line with the axis of the dynamometer sheave; there- 

 fore, false readings were recorded during the test. 

 Actual values were determined through post 



calibration. The 17-to-24-kip range indicates the 

 degree of uncertainty evident during post-calibration. 



Rock Tests 



Only three tests were completed in rock and a 

 rock conglomerate — tests 11, 12, and 29. The 

 results, which were scattered, indicate that con- 

 siderably more testing in this general type of seafloor 

 is necessary to more clearly define behavior. One test, 

 test 11, resulted in a spectacular holding capacity of 

 107,000 pounds in direct uplift with only 2-1/2 feet 

 of penetration in basalt. Test 12 in the same area 

 resulted in a capacity of 20,000 pounds with a total 

 penetration of 7 feet. A 20-to-30-foot-diameter 

 plateau with 45-to-60-degree sloping sides was used 

 for both tests. The first fluke was fired direcdy into 

 the center of the plateau, while the second hit the 

 slope; it was surprising that the second one even pene- 

 trated. The basalt on the slope was very weathered 

 and fractured easily. The basalt in the center did not 

 fracture; the fluke just became wedged in the basalt. 

 In the same general vicinity of Anacapa Island, a third 

 test was performed (test 29) in a rock conglomerate. 

 Solid basalt was desired; the conglomerate was 

 unexpected. Nonetheless, the holding capacity was 

 30,000 pounds in what divers described as a rubble. 

 Keying in this material would have been desirable. 



DISCUSSION 



Test Correlation 



An attempt was made to find an empirical param- 

 eter that would suitably describe observed short-term 

 anchor performance. This effort resulted in Figure 

 18, which is a plot of the parameter 17, the ratio 

 between measured holding capacity and kinetic 

 energy, versus general soil type. Three distinct phases 

 were apparent when looking at the parameter rj, and 

 they were best separated by soil type. The reasons for 

 this reasonably good correlation are not yet apparent. 

 In fact it is somewhat contradictory to expected 

 behavior. For example, the plot indicates that holding 

 capacity is controlled more by the available kinetic 

 energy than fluke size, especially for the soft clay 

 where the majority of data exists. It indicates that 

 with equal kinetic energy, a small fluke will hold 



21 



