AbstractThe performance of the novel fish anchor has been assessed through field tests. The tests were carried out at three different locations in the Swan River, Perth. The water depths were 2.91–4.73 m. The riverbed soils consisted of clay (Location 1), silty clay (Location 2) and silt (Location 3). The anchor was installed in the riverbed by dropping first through an air column followed by a water column, and finally striking the riverbed at an impact velocity of 10.4 (Location 1), 11.43 (Location 2), 11.72 (Location 3) m/s. The anchor tip embedment depth in the riverbed was measured at 1–2.5 times the anchor length. For similar impact velocity, the tip embedment depth in clay was 1.56–1.68 and 2–2.1 times of that in silty clay and silt, respectively. The capacity was found to reduce with loading angle at the padeye to the horizontal. The ultimate capacity was 1.5–2.6 times the weight of the anchor submerged in water for loading angle 90°, which increased to 8.2–15.6 times the weight for θ0=∼16°. The fish anchor dove, as opposed to pull out of the riverbed, for loading angles ≤37–47°, resulting in nonelliptical failure envelopes, which have been expressed mathematically.