METIIOIIS OF SAMPLING THE BENTHOS
213
polychactes occurring down to 15 cm, the maximum depth sampled.
This is one of the only examples of direct measurement of the depth
at which offshore animals live, but the results are probably not generally
applicable to the open sea; for the rest we must rely on the relative
catching powers of different samplers.
Johansen (1927) made a careful comparison on several sandy
grounds of the quantity of sediment and weight of invertebrates
taken by the Knudsen sampler as compared with 0.1 and 0.2 m2
Petersen grabs. Judging from the volume of sediment brought up,
the average depth of penetration was 25.6 cm for the Knudsen sampler
and 1.2 and 2.6 cm respectively for the two grabs, while the weight of
invertebrates taken by the Knudsen sampler was of the order of four
times that brought up by the larger Petersen grab. I n addition, the
two grabs failed to take, or took only small specimens, of a number
of species, including Arenicola, Macoma and Pectinaria.
More recently Holme (1949) has compared the sampling ability of
his 0.05 m2 scoop-sampler against the 0.1 m2 Petersen grab. I n these
trials, on three types of ground, the Petersen grab penetrated to average
depths of 2-7, 0.2 and 0.35 cm, while the scoop-sampler penetrated
to 10.7, 4.6 and 4-25 cm respectively on the same grounds. I n samples
totalling 0.5 ni2 the scoop-sampler collected 272 specimens belonging
to a t least thirteen different species, while the Petersen grab took only
thirty-three specimens belonging to five species. A measure of the
sampling depth was shown by some specimens of Ensis, the top 8 cm
of three specimens and two shorter fragments being taken in the scoop
(Fig. 26). The Petersen grab failed to take any, while in a dredge
haul the cut-off tops, each about 2 cm long, of three specimens were
taken.
In another trial, Holme (1953) has compared the numbers of animals
taken on a bottom of fine muddy sand by the 0.1 m2 Petersen grab,
Holme’s 0-05 m2 scoop-sampler and a suction-corer covering 0.02 rn2.
The corer penetrated to an average depth of 16.6 cm, the grab to about
3 cm, and the scoop-sampler to about 6 cm. The results show that the
Petersen grab (230 specimens per m2) failed to sample the fauna
adequately, while the samples taken by the scoop-sampler (976/m2)
and suction-corer (884/m2) were of the same order. This indicates that
the scoop-sampler was on this ground sampling the fauna adequately,
although it should be noted that siphons only of the lamellibranch
Lutraria were taken, the animal being out of reach of all three samplers.
On the Dogger Bank, Ursiii (1954) has compared the sampling
abilities of 0.2 m2 van Veen and Petersen grabs. While the volume
of sediment taken by the van Veen was 8.5 1. as against only 5.8 1. for
213
polychactes occurring down to 15 cm, the maximum depth sampled.
This is one of the only examples of direct measurement of the depth
at which offshore animals live, but the results are probably not generally
applicable to the open sea; for the rest we must rely on the relative
catching powers of different samplers.
Johansen (1927) made a careful comparison on several sandy
grounds of the quantity of sediment and weight of invertebrates
taken by the Knudsen sampler as compared with 0.1 and 0.2 m2
Petersen grabs. Judging from the volume of sediment brought up,
the average depth of penetration was 25.6 cm for the Knudsen sampler
and 1.2 and 2.6 cm respectively for the two grabs, while the weight of
invertebrates taken by the Knudsen sampler was of the order of four
times that brought up by the larger Petersen grab. I n addition, the
two grabs failed to take, or took only small specimens, of a number
of species, including Arenicola, Macoma and Pectinaria.
More recently Holme (1949) has compared the sampling ability of
his 0.05 m2 scoop-sampler against the 0.1 m2 Petersen grab. I n these
trials, on three types of ground, the Petersen grab penetrated to average
depths of 2-7, 0.2 and 0.35 cm, while the scoop-sampler penetrated
to 10.7, 4.6 and 4-25 cm respectively on the same grounds. I n samples
totalling 0.5 ni2 the scoop-sampler collected 272 specimens belonging
to a t least thirteen different species, while the Petersen grab took only
thirty-three specimens belonging to five species. A measure of the
sampling depth was shown by some specimens of Ensis, the top 8 cm
of three specimens and two shorter fragments being taken in the scoop
(Fig. 26). The Petersen grab failed to take any, while in a dredge
haul the cut-off tops, each about 2 cm long, of three specimens were
taken.
In another trial, Holme (1953) has compared the numbers of animals
taken on a bottom of fine muddy sand by the 0.1 m2 Petersen grab,
Holme’s 0-05 m2 scoop-sampler and a suction-corer covering 0.02 rn2.
The corer penetrated to an average depth of 16.6 cm, the grab to about
3 cm, and the scoop-sampler to about 6 cm. The results show that the
Petersen grab (230 specimens per m2) failed to sample the fauna
adequately, while the samples taken by the scoop-sampler (976/m2)
and suction-corer (884/m2) were of the same order. This indicates that
the scoop-sampler was on this ground sampling the fauna adequately,
although it should be noted that siphons only of the lamellibranch
Lutraria were taken, the animal being out of reach of all three samplers.
On the Dogger Bank, Ursiii (1954) has compared the sampling
abilities of 0.2 m2 van Veen and Petersen grabs. While the volume
of sediment taken by the van Veen was 8.5 1. as against only 5.8 1. for