242
N. A. HOLME
preservation had little effect on wet weight, and further determinations
are now needed on other groups of animals.
There is no reason why dry weight values should be calculated
from the weights of either fresh or preserved animals. It may be more
convenient to use a conversion factor based on volume of the animal
(displacement of water, or a calculated value), or on some other
measurement such as length. Thus, M. L. Jones (1961) calculated the
volumes of animals of different size from linear measurements, the
results being finally expressed in terms of volume. Working on penaeid
prawns Hall (1961) established a relationship between carapace length
and weight. Plotting log weight against carapace length, an almost
straight-line relationship was estrtblished for each of twenty-four
species of prawn. Such methods must be far more convenient than
weighing when many individuals of a species are to be determined.
XIII. STATISTICS
It is only in recent years that benthos sampling methods have
been subjected to critical statistical review, and it may be that this
application of statistical techniques represents the major advance in
the field of benthos sampling during the last ten or fifteen years.
Some of the methods employed are common to both plant and
animal ecology, and a good general account of these is given by
Greig-Smith (1967). There are, in addition, special problems associated
with grab sampling of the benthos, and these will be discussed first.
We have already considered the extent to which samplers capture
all the animals below the nominal area sampled. Some animals may
be too deeply buried to be captured at all, but in applying statistical
treatments to the results of a survey the most important question is
how accurately are the shallow-burrowing animals within the reach of
the equipment sampled ? Birkett (1958) has discussed the sampling
ability of different instruments with special reference to the semicirculu profile of the grab sample. In this type of sample the area
sampled decreases with depth, and Birkett expresses efficiency of
sampling in terms of the relationship between the maximum depth
sampled and the average depth, as memured by volume of the sample.
The efficiency of a coring tube wodd be loo%, but that of a van Veen
grab was found to be only 8 little better than 60%. It followed that
d e s s the depths at whicb individuals lived were known and the
penetration of the grab were constant the accurate estimation of densities per square metre would not be possible.
An alternative method, for comparative purposes, is to expresa
N. A. HOLME
preservation had little effect on wet weight, and further determinations
are now needed on other groups of animals.
There is no reason why dry weight values should be calculated
from the weights of either fresh or preserved animals. It may be more
convenient to use a conversion factor based on volume of the animal
(displacement of water, or a calculated value), or on some other
measurement such as length. Thus, M. L. Jones (1961) calculated the
volumes of animals of different size from linear measurements, the
results being finally expressed in terms of volume. Working on penaeid
prawns Hall (1961) established a relationship between carapace length
and weight. Plotting log weight against carapace length, an almost
straight-line relationship was estrtblished for each of twenty-four
species of prawn. Such methods must be far more convenient than
weighing when many individuals of a species are to be determined.
XIII. STATISTICS
It is only in recent years that benthos sampling methods have
been subjected to critical statistical review, and it may be that this
application of statistical techniques represents the major advance in
the field of benthos sampling during the last ten or fifteen years.
Some of the methods employed are common to both plant and
animal ecology, and a good general account of these is given by
Greig-Smith (1967). There are, in addition, special problems associated
with grab sampling of the benthos, and these will be discussed first.
We have already considered the extent to which samplers capture
all the animals below the nominal area sampled. Some animals may
be too deeply buried to be captured at all, but in applying statistical
treatments to the results of a survey the most important question is
how accurately are the shallow-burrowing animals within the reach of
the equipment sampled ? Birkett (1958) has discussed the sampling
ability of different instruments with special reference to the semicirculu profile of the grab sample. In this type of sample the area
sampled decreases with depth, and Birkett expresses efficiency of
sampling in terms of the relationship between the maximum depth
sampled and the average depth, as memured by volume of the sample.
The efficiency of a coring tube wodd be loo%, but that of a van Veen
grab was found to be only 8 little better than 60%. It followed that
d e s s the depths at whicb individuals lived were known and the
penetration of the grab were constant the accurate estimation of densities per square metre would not be possible.
An alternative method, for comparative purposes, is to expresa