J. Exp. Mar. Biol. £col., 1990, Vol. 137, pp.

195-214 195
Elsevier

JEMBE 01413

The use of the size structure of benthic macrofaunal

communities to estimate faunal biomass and secondary
production

Graham J. Edgar
Amakusa Marine Biological Laboratory, Amakusa, Japan; Zoology Department, University of Tasmania,
Hobart, Tasmania, Australia

(Received 11 April 1989; revision received 19 January 1990; accepted 3 February 1990)

Abstract: A general allometric equation (P = 0.0049 * B 0 ·80 T 0 ·89 ) which relates daily macrobenthic produc-
tion P (µg · day- 1 ) to ash-free dry weight B (µg) and water temperature T ( 0 C) was calculated from the
published production rates of 41 invertebrate species. Allometric equations were also calculated from the
same literature set for adult (P = 0.0050 * B 0 ·78 T 0 ·92 ) and juvenile (P = 0.0063 * B 0 ·86 T 0 ·80 ) animals. The
biomass exponent of the general equation (0.80) differed significantly from 0.63, the value predicted by Banse
& Mosher's ( 1980) equation P/B = 0.65Ms- 0 -37 . A new method of estimating the secondary production of
benthic macrofaunal communities based on the use of these equations is outlined. The method requires
relatively little sampling effort and is applicable to marine and estuarine species living in water temperatures
of 5-30 °C. Production is calculated using the mean ash-free dry weight of animals retained by sieves of
differing mesh size, the abundances of animals in different sieve size classes, and the general equations
described above. The mean biomass of animals in the various sieve size classes used in production
calculations can be measured directly, or, providing that the fauna! assemblage does not contain many long
flexible individuals or is dominated by a few very large animals, using the empirically derived equation
log B = - 1.01 + 2.64•logS, where Bis fauna! ash-free dry weight (mg) and Sis sieve size (mm).

Key words: Biomass; Fauna! size structure; Macrofauna; Secondary production

INTRODUCTION

Most investigated macrofaunal communities are described in terms of fauna! abun-

dance or biomass. The production of communities is rarely estimated, not because it
is of less interest than standing stock, but because of methodological and sampling
difficulties. Investigators inevitably sacrifice the estimation of production because suf-
ficient resources are not available or they can then better rationalise time in their studies
by increasing the number of other fauna! parameters to be investigated, the level of
taxonomic identification, the frequency of sampling and the amount of sample replica-
tion. Production also cannot be estimated with any degree of accuracy when diverse
communities with continuously breeding species are studied.
The most practical way to calculate the production of benthic communities has been
Correspondence address: G.J. Edgar, Department of Zoology, University of Melbourne, Parkville,
Victoria 3052, Australia.

0022-0981/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

196 G.J. EDGAR

to determine the production of a few dominant species using cohort analysis ( see Crisp,
1984) and to ignore, or use approximate P/B (annual population production/annual
mean biomass) ratios to estimate, the production of the less important species (e.g.,
Warwick & Price, 1975; Robertson, 1984). The use of P/B ratios requires some informa-
tion about the life history of individual species; Banse & Mosher (1980) described an
empirical relationship between the P/ B ratio of the population and the size of the species
at maturity, while several authors (e.g., Waters, 1969; Robertson, 1979) have found that
the P/B ratio predictably varied with the life span of species. The underlying assumption
in using P/B-life history relationships, that production is influenced by a single major
factor, has been criticised, however, as overly simplistic (see, e.g., McLaren & Corkett,
1984).
Population production is primarily a function of two major independent factors: the
metabolic rate-body size relationship of individuals and the distribution of body sizes
within the population (Dickie et al., 1987). Both these factors can be directly ascertained
for benthic communities, although it is generally impractical to measure the sizes of all
individuals or to determine the body size-production relationship for each species. The
distribution of body sizes within benthic communities can, however, be rapidly and
easily determined by passing fauna! samples through a series of nested sieves. Total
community production should therefore be able to be estimated using data on fauna!
sieve size and generally determined mean daily production rates of animals in different
sieve size classes, providing that there is little difference in weight specific production
between species and between sites. Such a method would be of considerable practical
value because it would allow the production of diverse macrofaunal communities to be
estimated routinely. To facilitate the use of this method, the relationship between body
weight and daily production of benthic invertebrates is examined here. The mean
biomass of animals retained by different-sized sieves has also been determined, allowing
the biomass of macrofaunal samples to be estimated indirectly from fauna! sieve size
distributions. While the biomass ofmacrofaunal samples can be measured directly with
relative ease by drying animals after separation from the environmental matrix, this
process requires the destruction of samples. The sieve method is useful whenever direct
biomass measurements cannot be made, such as when specimens need to be retained
for taxonomic purposes or the material is expected to need further sorting at a later
stage.
METHODS

ESTIMATION OF BIOMASS

The mean ash-free dry weights (AFDW) of different sieve size classes of polychaetes,
molluscs, caprellid amphipods, crustaceans (excluding caprellid amphipods) and platy-
helminths were determined from macrobenthic samples collected in three habitats:
intertidal sand flats and artificial algal habitats, both at Tomioka, West Kyushu, Japan
(32°32'N, 130°2'E; see Tamaki (1985) and Imada & Kikuchi (1984) for descriptions
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 197

of the respective sites), and Amphibolis antarctica seagrass beds at Seven Mile Beach,
Western Australia (29°1l'S, ll4°53'E; see Edgar, 1990). Samples were hosed,
immediately after the addition of 5 % formalin, through a nested series of sieves (8.0,
5.6, 4.0, 2.8, 2.0, 1.4, 1.0, 0.71, 0.50 mm) stacked in descending order of size (Edgar,
1990), and the AFDW calculated as the difference between the dried (80 °C for 48 h)
and ashed (500 °C for 2 h) weights of animals retained by each sieve. Each weight
determination included at least four species and 10 individuals, with no single species
comprising > 35 % of the total animal number.

COMPARISON OF MEASURED AND ESTIMATED SAMPLE BIOMASS

The degree of accuracy with which abundance data can be converted to biomass data
using the general values calculated here was assessed by collecting faunal samples from
Sargassum and Zostera habitats and comparing the directly measured and the estimated
faunal biomasses. The Sargassum-associated fauna was collected from 3 m depth at
Tomioka on 18 March 1988 by enclosing Sargassum patens plants in plastic bags and
then extracting the animals in the laboratory using the methods of Edgar (1983). Zostera
marina samples were obtained using 0.125 x 0.125-m 2 cores (see Edgar, 1990) from 1 m
depth at Takamokujima(32°35'N, 130°25'E) on 7 June 1988. The biomass of samples
was estimated in three ways using the equation B = L n; · b; where B is the estimated
total biomass, n; is the abundance of animals retained by sieve size i and b; is the mean
biomass of animals retained by sieve size i: (1) by grouping all fauna, (2) by summing
separate calculations of biomass for caprellids and other groups, and (3) by summing
separate calculations of biomass for molluscs, polychaetes, crustaceans other than
caprellids, caprellids, platyhelminths and other groups. In these calculations the mean
weights of animals were assumed equal to the mean weights of the noncaprellid groups
(see Table 11) whenever more explicit regression values were not used. Animals
> 5.6 mm sieve size were excluded from the analysis.

ESTIMATION OF PRODUCTION

The relationship between body weight, water temperature and production rate was
determined by estimating, from the figures and tables of published studies, the mean
daily increment in AFDW for cohorts of a variety of marine and estuarine species at
defined temperatures. The species included in this analysis are listed in Table I. A
number of additional species were suitable for use in the analysis but were not included
because body length increments could not be converted to changes in AFDW. Macro-
faunal species which grew much > 1 g AFDW were also excluded from the analysis.
In order to ensure the independence of the data set, whenever the growth of more than
one species or cohort at a site, or the growth of a single species at different sites, has
been described in the literature, only a single cohort was selected for inclusion here. This
was done by first excluding any poorly defined cohorts and then assigning numbers to
cohorts and using a random numbers table.
 TABLE I
Calculated daily production rates of juveniles and adults of 41 macrobenthic species at defined temperatures and biomasses. Life history column lists each species by taxonomic
(C, Crustacea; M, Mollusca; P, Polychaeta), functional (e, epifaunal; i, infaunal) and trophic (c, carnivore; h, herbivore; d, detritivore) grouping.

General Young Adult Life

history
T('C) Biomass P(µg· day-') T('C) Biomass p ( µg . day- I) T('C) Biomass P(µg·day- 1
)

(mg) (mg) (mg)

Undo/a inermis 6 2.02 5.61 7 0.274 2.96 1.55 4.39 Ced Collie (1985)
Gammanu mucronatus 27 0.169 5.2 27 0.022 3.4 27 0.10 8.7 Ced Fredette & Diaz (1986)
Pseudoprotella plasma 12 0.062 0.83 14 0.023 0.96 8 0.12 1.3 Ced Hughes (1978)
Corophium insidiosum 14 0.021 0.9 16 0.014 I.I 16 0.037 1.1 Ced Birklund (1977)
Lembos websteri 12 0.553 12.9 16 0.076 3.1 16 0.50 6.6 Ced Moore (1981)
Parhya/e basrensis 17 5.4 46.2 24 0.20 5.5 14 1.00 8.9 Ced Ali & Salman (1987)
Palaemonetes pugio 16 17.2 395 16 19,8 340 25 64.1 270 Ced Alon & Stancyk ( 1982)
Macrophthalmus latifrons 15 14.3 204 22 3.4 83 16 201 310 Ced Robertson (1984)
Vea pugilator 22 57 465 21 14.3 350 22 57.0 465 Ced Colby & Fonseca ( I 984)
Scopimera g/obosa 16 5.41 26.l 25 0.57 25 23 3.38 73.9 Ced Wada (1981)
Halicarcinus australis 16 45.5 580 25 9.6 310 15 36 380 Ced Lucas & Hodgkin (1970)
Diche/opandalus bonnieri 13 525 1260 12 185 1300 14 788 2500 Cec AI-Adhub & Bowers (l 977)
Crangon crangon 12 1182 1703 18 132 3700 17 525 3300 Cec Henderson & Holmes ( 1987)
Palaemon adspersus 15 38.9 811 16 31 700 18 215 2270 Cec Baden & Pihl ( 1984 J
Ampelisca brevicornis 12 2.25 9.5 16 0.75 19.2 14 3.0 44 Cid Klein et al. (1975)
Ampe/isca macrocepha/a 12 1.32 14.0 8 0.126 3.2 13 1.7 18.1 Cid Kanneworff ( 1965)
Ampe/isca abdita 12 0.58 3.1 20 0.068 1.6 II 0.26 4.0 Cid Mills (1967)
Haustorius canadensis 12 3.0 17 14 2.0 19 10 4.0 12 Cid Donn & Croker ( 1986)
Pontoporeia femaraia 5 2.25 13 16 0.60 9.1 6 LS 12 Cid Wildish & Peer ( 1981)
Amphiporeia lawrentiana 12 0.056 0.86 8 0.082 1.4 5 0.83 7.9 Cid Downer & Steele (1979)
Cyathura can),ata 5 0.52 l.2 18 0.30 3.8 12 0.15 3.8 Cid Olafsson & Persson ( 1986)
Patella vulgata 10 15.1 91.1 16 6.4 65 ll 62 210 Meh Blackmore (1969)
Patella aspera 10 247 887 15 6.4 73 18 129 366 Meh Thompson (1979)
Cel/ana tramoserica 23 86.9 409 24 6.4 78 20 85 460 Meh Underwood (1975)
Montfortula rugosa 24 18.9 74 21 2.0 33 25 19.3 75 Meh Creese (1981)
Donax incarnatus 30 8.1 170 30 12.5 330 30 75 980 Mid Ansell et al. (1972)
Te/Jina fabula 5 0.008 0.082 7 0.019 0.22 10 0.14 0.93 Mid Muus (1973)
Macoma balthica 13 5.0 33.0 9 3.0 5l 9 14.6 143 Mid Chambers & Milne ( 1975)
Mya arenaria 19 285 1450 8 17.8 150 15 160 690 Mid Burke & Mann (1974)
Donax trunculus 20 11 320 ll 6.3 105 8 54 110 Mid Ansell & Lagardere {l 980)
Venus s/riatula 13 31 126 12 6.3 55 5 42 80 Mid Guillou & Sauriau (1985)
Spisula subtruncata 19 14.5 76.4 19 11.3 180 12 56 !80 Mid Ambrogi & Ambrogi (1985)
Cerastoderma edu/e 15 0,034 1.4 Mid Moller & Rosenberg ( 1983)
Cerastoderma edu/e 19 90 1500 18 40 625 Mid Essink & Bos (1985)
Ahra nitida 7 2.3 15.4 18 0,23 3.9 10 3.4 21.6 Mid Josefson (1982)
My1ilus edu/is 6 58.0 181 14 6.3 145 9 3l.4 220 Med Kaut.sky (1982)
My1i/us ga//opmvincialis 24 380 1950 16 300 1950 17 800 2670 Med Ceccherelli & Rossi ( 1984 J
Chlamys opercularis 7 69.2 348 7 42 350 14 390 1740 Med Richardson et al. ( 1982)
Pectinaria californiensis 8 6.15 645 10 1.9 31 8 27 110 Pid Nichols (1975)
Pectinaria koreni 15 13.8 240 15 6,0 220 13 24 200 Pid Nicolaidou (1983)
Nephtys hombergi 15 2.3 22.2 18 1.4 15.5 14 14 160 Pie Warwick & Price (1975)
Nereis diversicolor 12 0.027 2.2 14 0.17 6.4 14 0.88 18 Pie Moller (1985)
200 G.J. EDGAR

Each species' daily production rate was estimated by measuring, for a date arbitrarily
selected using a random numbers table, the tangential slope of the smoothed graph
depicting the growth of the cohort mode through different seasons. The slope measure-
ments ideally showed changes in AFDW with time (e.g., Nephtys hombergi; Warwick
& Price, 1975), however the rate of change of AFDW more frequently needed to be
estimated from length measurements using appropriate transformations. Where authors
have not included the regression relating length to weight, the regressions of Josefson
(1982) for bivalves, Collie (1985) for amphipods of three morphological types, Hughes
( 1971) for limpets, and Pihl & Rosenberg ( 1982) for crabs and carid shrimps were used.
When not given, the AFDW and dry weight of molluscs were considered to be 5 and
6% oflive weight (see Ansell et al., 1972), and the corresponding values for amphipods
were 15 and 22%, respectively (see Collie, 1985). Although the use of these regressions
may have introduced considerable errors into the estimation of the mean weight of a
cohort at a particular time, such errors would be relatively unimportant because produc-
tion rates were related to biomass rather than time or body length. Proportional errors
thus would be incorporated into both the biomass and production estimates, causing
little change to the ratio of biomass to production.
All calculated production rates were related to prevailing mean monthly water tem-
peratures. These ranged from 5 to 30 ° C. Whenever temperature data were not pub-
lished with growth data, temperatures were estimated from studies at comparable sites
in the same geographic region.
The production rates of each species were also separately scaled for adult and young
individuals because it was immediately apparent from the general data set that weight-
specific productivity decreased with increasing animal size much more rapidly within
species than between species.
The sizes of adult and young animals were arbitrarily chosen at ~ 70 and ~ 30% of
the published maximum body length. The production rate of Ampelisca macrocephala,
for example, was calculated from Fig. 1 ofKanneworff(1965) by selecting 10 and 4 mm
body length animals to represent adult and young individuals. By drawing a tangent to
the graph at 10 mm length, the growth rate at this size was estimated to be 2 mm in
1. 7 month (0.0392 mm· day- 1 ). Using the body length-weight relationship determined
by Collie (1985) for Ampelisca agassizi, the increase in length from 10 to 10.0392 mm
during 1 day represents an increase in AFDW from 1.70404 to 1.72213 mg. Individuals
of 1. 704 mg thus had an estimated daily production rate of 18.1 µg · day- 1 in early
October, the time when water temperatures were ~ 13 °C {also from Fig. 1, Kanneworff,
1965). Analogous calculations were made for 4 mm and the randomly selected size
(9.1 mm) of A. macrocephala, and for other species.
The dependence of the daily production rate (P) on animal AFDW (B) and water
temperature (T) was evaluated using the equation P = a* Bb * Tc. The coefficients a,
b and c were calculated from the log-transformed equation log P = log a + b * logB +
c * log Tusing the method ofleast squares. Temperature was expressed in this equation
in logarithmic rather than linear terms because the winter production rates of temperate
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 201

invertebrates are often reported to be insignificant compared to the summer rates (e.g.,
Pihl & Rosenberg, 1982), hence a change in water temperature from 5 to 10 cc was
considered more likely to affect production by a magnitude equivalent to a change from
10 to 20 cc rather than a change from 10 to 15 cc.
The species used in the analysis have been classified in Table I into taxonomic
(Crustacea, Mollusca and Polychaeta), trophic (carnivorous, detritivorous and herbiv-
orous) and functional (epifaunal and infaunal) categories. Because the trophic cate-
gories are broadly overlapping, with few macrofaunal species feeding solely on animal,
plant or detrital material, species were assigned to the trophic group which corre-
sponded with the predominant fraction of their diet. Hence, for example, suspension
and deposit feeding animals were classed as detritivorous although they may also
consume considerable quantities of microalgae and protozoans.

RESULTS AND DISCUSSION

ESTIMATION OF BIOMASS

The mean AFDW values of the various animal groups retained by different sieves are
shown in Fig. 1. The corresponding regression equations are listed in Table II and the
estimated mean weights of different sieve-sized animals, as calculated from the regres-
sions, are shown in Table III.
The slopes of the mollusc, crustacean other than caprellid, platyhelminth and poly-
chaete regression equations did not vary greatly from each other, however the slope of
the regression for caprellid amphipods was markedly lower than for the other animal
groups (all four pairwise homogeneity of slope ANCOVAs, P < 0.001). Although

2~--------------------~
• Crustaceans
a Molluscs
Ol o Polychaetes
E + Caprellids
+
A Platyhelminths
+
+

-2 -+-----,----.---,.--....---.-.......--.-.-----.------.---.---,.--..----.--.......---.--~
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8
-Lo.g. sieve size (mm)
Fig. 1. Mean AFDW values (mg) of animals collected on sieves of differing mesh sizes. Line shows
regression for all faunal groups other than caprellids (log B = - 1.009 + 2.643 * logS; see Table II).
202 G.J. EDGAR

TABLE II
Equations relating mean fauna! AFDW B (mg) and sieve mesh size S (mm). Because each sieve retains
animals ranging from that sieve size to next, Sis expressed as geometric mean for each sieve (i.e., for 0.5 mm
mesh sieve, logS = (log0.5 + log0.71)/2). Coefficients were calculated from logB =a+ b * logS; n, number
of values. Combined equation was calculated using all except caprellid data points.

a SE of a b SE of b r2 n

Crustaceans l.040 0.087 2.672 0.068 0.991 16

Molluscs 1.020 0.086 2.758 0.068 0.991 17
Polychaetes -0.988 0.113 2.494 0.097 0.976 18
Platyhelminths 0.963 0.125 2.641 0.122 0.979 12
Caprellids 1.214 0.099 1.901 0.086 0.976 14
Combined 1.009 0.017 2.643 0.048 0.980 63

caprellids of 0.5 mm sieve size were of similar mean weight to the other fauna! groups,
caprellids of 4 mm sieve size were < 25 % of the weight of other animals.
The slopes of all the regression equations relating sieve size to AFDW were signifi-
cantly < 3 (t tests, P < 0.05). One possible reason for the AFDW of animals not
increasing with the cube of body dimensions is that the larger individuals devote a
proportionately greater amount of their volume to inorganic armour (shell, exoskeleton,
etc.). Elongate animals such as caprellids and polychaetes also become proportionately
thinner and more flexible as they develop, and thus more liable to be entangled over the
grid-wires of sieves. The mean weights of animals retained by large sieves thus tended
to be reduced by a few thin individuals with disproportionately low weight. Elongate
animals nevertheless posses a much greater weight to cross-sectional area than compact
individuals. The fact that the slope of the regression between sieve size and AFDW for
polychaetes was not significantly lower than the slopes of the regressions for molluscs,
crustaceans excluding caprellids, and platyhelminths (three pairwise homogeneity of
slope ANCOVAs, P > 0.05) indicates that the weight of some polychaetes on each sieve
must have been much greater than the mean weights of other major animal groups,
thereby compensating for the small individuals entangled on grid wires. Polychaetes

TABLE III
Mean AFDW values (mg) of animals retained by different sieves as estimated from equations of Table II.

Sieve mesh size (mm)

0.5 0.71 1.0 1.4 2.0 2.8 4.0 5.6

Crustaceans 0.023 0.058 0.143 0.36 0.91 2.30 5.8 14.7

Molluscs 0.023 0.060 0.152 0.40 1.03 2.67 7.0 18.1
Polychaetes 0.028 0.067 0.166 0.37 0.88 2.09 5.0 11.8
Platyhelminths 0.028 0.069 0.170 0.42 1.06 2.65 6.6 16.5
Caprellids 0.023 0.044 0.084 0.16 0.31 0.61 1.2 2.3
Combined 0.025 0.062 0.153 0.38 0.95 2.39 6.0 14.9
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 203

show much greater variance in weight for each sieve size than compact organisms (see
Table 6 in Edgar, 1983).
The relationship between sieve size and AFDW for polychaetes described in Table I
should only be used with caution because it was formulated from samples dominated
by epifaunal nereids, syllids and phyllodocids and infaunal spionids, opheliids and
nephtyids, and included few string-shaped individuals. If, for example, large numbers
of capitellids had been collected, then many animals would be likely to have been
trapped on the grid-wires during sieving and the regression slope depressed below 2.
A number of animal groups not examined here, such as ophiuroids and pycnogonids,
would also show proportionately low AFDW on sieves. When long flexible species or
species with thin rigid appendages are collected in large numbers, their AFDW clearly
needs to be calculated separately from other taxa.
A comparison of the estimated and the directly measured AFDW of the Sargassum
and Zostera samples (Table IV) indicates that the general regression between faunal

TABLE IV
Fauna! AFDW (mg) ofbenthic samples collected from Sargassum and Zostera habitats as estimated using
mean animal weights listed in Table III and abundance data; x, mean sample AFDW. Three methods were
used to estimate total fauna! biomass: Method 1, by grouping all fauna; Method 2, by summing separate
calculations of biomass for caprellids and other groups; and Method 3, by summing separate calculations
of biomass for molluscs, polychaetes, crustaceans other than caprellids, caprellids, platyhelminths and other
groups. Directly measured weights of samples are also shown.

Sargassum: replicate number x ± SD Zostera: replicate number X± SD

2 3 4 2 3 4

Crustaceans 22.0 28.9 34.7 38.2 31.0± 7.1 7.9 4.4 2.5 3.4 4.6± 2.4
Molluscs 20.1 29.5 9.7 12.5 18.l ± 8.9 2.6 30.6 11.0 5.4 12.4 ± 12.6
Polychaetes 6.0 3.7 4.9 9.3 6.0± 2.4 49.5 48.4 77.8 56.2 58.0 ± 13.7
Platyhelminths 2.7 0 0.2 0.9 1.0± 1.2 0 0.4 1.1 0 0.4± 0.5
Caprellids 4.1 5.1 6.8 26.3 10.6± 10.5 0.2 0 0.5 0.1 0.2± 0.2
Other groups 17.9 0.3 6.4 0 6.2± 8.4 0 0.4 0 0.4 0.2± 0.2
Total (Method 3) 72.8 67.5 62.7 87.2 72.6 ± 10.6 60.2 84.2 92.9 65.5 75.7 ± 15.4
Total (Method 1) 81.6 78.3 78.7 130.4 92.3 ± 25.5 59.8 83.3 91.9 64.9 75.0± 15.1
Total (Method 2) 73.5 68.l 66.7 91.5 75.0 ± 11.4 59.6 83.3 91.1 64.8 74.7 ± 14.9
Total (Measured) 64.1 69.6 65.7 76.6 69.0± 5.6 54.6 84.7 75.7 54.1 67.3 ± 15.4

sieve size and fauna! AFDW shown in Table II provides a good estimate of the AFDW
of fauna! samples, except when large numbers of elongate animals with disproportio-
nately low weight are present. The large sieve size classes of the Sargassum fauna at
Tomioka were dominated by caprellid amphipods, consequently the mean estimated
AFDW was much greater than the directly measured AFDW when caprellids were
grouped with other fauna (Method 1). The overestimation of fauna! biomass decreased
from 37 to 10% when caprellids were separated from other groups (Method 2), and was
only 5 % when the fauna was further subdivided (Method 3).
204 G.J. EDGAR

The Zostera samples differed from the Sargassum samples in being predominantly
infaunal rather than epifaunal and lacking a large component of flexible elongate
animals. Consistent estimates of fauna! biomass were consequently calculated using the
three methods, the mean estimates being 11-12% greater than the direct measurement.
Errors in the estimation offaunal AFDW can be greatly reduced by directly weighing
the largest animals. Eight animals > 5.6 mm were collected in the Zostera cores. These
individuals, which were excluded from the above calculations, had a mean
AFDW · sample- 1 (105.3 mg) which was greater than the mean AFDW of the fauna!
component between 0.5 and 5.6 mm sieve sizes (67.3 mg). The difference between the
directly measured and the estimated fauna! AFDW of Zostera samples was < 5 % for
all three methods when AFDW estimates incorporate the directly measured AFDW of
animals > 5.6 mm.

RELATIONSHIP BETWEEN BIOMASS, TEMPERATURE AND PRODUCTION

Body size had an extremely large, and highly significant (ANOVA: df l/38,
F = 510. 7, P < 0.00 I), influence on the daily production rate of invertebrates (Eqn. 1
in Table V, Fig. 2), accounting for 92 % of the total variance in the log production rates

TABLE V

Equations relating daily production P ( µg · day- 1 ) to fauna! AFDW B (µg) and water temperature T ( 0 C)
for different animal groups. Epifauna, Infauna, Crustacea and Mollusca are subsets of general data set.
Coefficients were calculated from logP = a + b * logB + c * log T; n, number of values.

a SE of a b SEofb C SEofc r2 n Eqn.

General -2.31 0.24 0.80 0.04 0.89 0.22 0.94 41 l

Adult -2.30 0.23 0.78 0.03 0.92 0.19 0.95 41 2
Young -2.20 0.24 0.86 0.03 0.80 0.19 0.96 41 3
Epifauna - 1.99 0.43 0.78 0.05 0.68 0.32 0.94 21 4
Infauna -2.46 0.32 0.79 0.07 1.05 0.36 0.94 20 5
Crustacea -2.86 0.38 0.81 0.05 1.32 0.33 0.95 21 6
Mollusca -2.18 0.26 0.87 0.05 0.46 0.24 0.97 16 7

ofmacrofauna. Much of the variance not attributable to body size in Eqn. I (2.5%) can
be explained by the influence of temperature; this parameter also had a highly significant
effect on daily production rates (ANOVA: df = 1/38, F = 16.6, P < 0.001). The log
temperature coefficient (0.89) was slightly < 1, indicating that a doubling in temperature
was accompanied by slightly less than a doubling in production rate.
Eqn. 1 can probably be extrapolated to larger sizes and still provide a reasonable
estimate of macrofaunal production. The investigation was arbitrarily restricted to
animals < 1 g AFDW because I could only determine the production of one mega-sized
invertebrate species, the crab Portunus pelagicus (Potter et al., 1983). Because the
emphasis of this study was on using the sieve method to estimate production, it was
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 205

10000 - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

-1000

-
-0
Cl
..:;;; 100

C:

-
0

0
:::l
-0
10

...
0
c..

.01 .1 10 100 1 000 10000

Weight (mg)
Fig. 2. Relationship between daily production ( µg · day- 1 ) and animal AFDW (mg) at I O C. Production
O

rates of all 41 species as listed in Table I have been adjusted to I O° C using a temperature exponent of0.89.
Solid line shows the regression equation log P = - 1.42 + 0.80 • logB; dashed lines show 50 and 200% of
predicted production.

desirable for the equations to be calculated using animals to which the sieve method
would be applied, and not to have them influenced by an outlying point for a carnivorous
crab. However, the equations do not change greatly even if this data point is included.
Eqn. 1 becomes log P = - 2.37 + 0.81 * logB + 0.90 * log T (r 2 = 0.95) with the addi-
tion of the estimated production of Portunus pelagicus (B = 21.3 g, T = 22 °C,
P= 115mg·day- 1 ).
Perhaps the most surprising aspect of Eqn. 1, and also the other equations listed in
Table V, is the very low proportion of the total variance oflog production which remains
unexplained after the effects of biomass and temperature have been accounted for.
Amongst the factors which contribute to the 6 % of the total variance oflog production
which remains unexplained are: (i) interspecific differences (including taxonomic,
trophic and functional differences), (ii) spatial differences (including food availability
differences), and (iii) ontogenetic differences in production rates, and (iv) the analytical
and statistical errors inadvertently included in my data set. In this latter category are
errors in my estimation of temperature and biomass for taxa where they have not been
explicitly provided, errors in applying a graphical method to calculate production from
figures designed for other purposes, and sampling and statistical inaccuracies in the
modes of the original authors cohorts. Such errors would certainly act to increase the
variance of Eqn. 1, but presumably not greatly affect the coefficients of the equation
because the differences between the "calculated" and "real" values should have a
normal distribution with a mean approaching 0.
Regression equations calculated from the epifaunal and infaunal subsets of the
general data set (Table V: Eqns. 4 and 5) were similar to the overall regression (Eqn. 1),
and to each other. Regression equations calculated from the Crustacea and Mollusca
subsets (Table V: Eqns. 6 and 7) did, however, differ in their temperature coefficients.
206 G.J. EDGAR

This difference was only of marginal significance (homogeneity of regression coefficient

AN COVA, df = 1/31, F = 4.17, P = 0.050) and was probably caused by a Type I error
(sensu Zar, 1974). The temperature coefficient for adult Crustacea (1.02 ± 0.28) did not
differ significantly from that for adult Mollusca (0.70 ± 0.31), and a similar result was
found for young animals. The concordance in the regression equations for the inde-
pendent epifaunal and infaunal, and the Crustacea and Mollusca, data sets indicates
that the influences on daily production of taxonomic and functional groupings of
animals are not great. Furthermore, the regression equations for the total data set should
be reliable.
Insufficient data were available to make the calculation of regressions for different
trophic groups worthwhile. The great majority of the species examined were detritivores;
only five carnivorous shrimps and polychaetes and four herbivorous limpets being
categorised differently. An examination of the residuals of adults of the three trophic
groups (Table VI) nevertheless indicates that trophic grouping probably has a con-

TABLE VI
x values of residuals from Eqn. 2 (Table V) (i.e., differences between logarithms of observed and predicted
daily production) for different trophic groups; n, number of values.

x SD n

Carnivore 0.192 0.107 5

Herbivore -0.235 0.181 4
Detritivore -0.003 0.220 32

siderable influence on production rate. The mean of the residuals for the carnivorous
species was significantly higher (P < 0.01) and for herbivorous species significantly
lower (P < 0.05) than for detritivorous species (SNK test; ANOVA: df = 2/38,
F = 4.70, P = 0.015). However, because the herbivorous species were all limpets with
similar intertidal grazing habits and the carnivorous species comprised only two groups,
shrimps and polychaetes, too much reliance should not be placed on these results. The
low production rates of limpets may have resulted, for example, from exposure to air
rather than a herbivorous diet. Parry (1982) found that the production rate of a
sublittoral limpet was higher than the production rates of three adjacent eulittoral
species. Terrestrial species are also generally reported to have much lower growth and
production rates than marine relatives (see, e.g., Tamura & Koseki, 1974; Tanaka,
1988), so aerially exposed intertidal animals may show similar trends. Further investi-
gations into the influences of diet and aerial exposure on production rate are nevertheless
clearly warranted and may allow production estimates to be more accurately refined.
The coefficient of log biomass in the production equation for adult animals
(0.78 ± 0.03: Eqn. 2, Table V) differed from the corresponding biomass coefficient for
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 207

young animals (0.86 ± 0.03: Eqn. 3, Table V). At small body sizes young animals had
only slightly greater production rates than adult animals, whereas at body sizes > 10 mg
the production rates of juveniles were more than double those of adults (Fig. 3). The
significance of this difference cannot be properly estimated because the two equations
were calculated from the same literature set and hence were not completely independent.

10000 - - - - - - - - - - - - - - - - - - - - - - - - - ,

:;-1000
0)
::i
._, 100

C:
O 10
t;
:::i
-0 0
...
0
a. 0 ,, 9-'
•
o
Adult
Juvenile
, 0

.01 .1 10 1 00 1000
Weight (mg)
Fig. 3. Daily production (µg· day- 1 ) as a function of adult and juvenile animal AFDW (mg) at 10 °c.
Production rates of all 41 species as listed in Table I have been adjusted to 10 °C using temperature
exponents of 0.92 (adults) and 0.80 (juveniles). Solid line shows regression equation for adult animals:
log P = 1.38 + 0.78 * logB; dashed line shows corresponding regression for juvenile animals:
log P = - 1.40 + 0.86 * logB.

A homogeneity of regression coefficient ANCOVA nevertheless results in an F value

of 3.06 (df = 1/76); the probability of this occurring by chance, if the two sets had been
independent, is 0.08. It therefore seems likely that the difference between the biomass
coefficients of the adult and young equations is real because the signal to noise ratio
between related data sets is greater than for independent data sets. If the differences
between the biomass coefficients are, however, due to statistical noise, then the adult
and young equations certainly differ greatly in elevation (ANCOVA: df = 1/78,
F = 21.15, probability assuming independence <0.001).
Much of the difference between the regression equations relating log production to
log AFDW for adult and young animals (Fig. 3) probably resulted from reproductive
losses not being included in calculations of production. As adults often expend con-
siderable energy in reproduction, their somatic growth rates should generally be less
than young individuals of the same size, and hence Eqn. 2 should have a lower elevation
than Eqn. 3. The difference between the coefficients oflog AFDW in Eqns. 2 and 3 (i.e.,
the slopes in Fig. 3) nevertheless was unexpected and its cause, if real, can only be
speculated upon. Perhaps size-selective predation or other factors cause mortality to be
strongly negatively correlated with body size, in which case small species may maximise
208 G.J. EDGAR

TABLE VII
Secondary production calculations for Ampharete acutifrons as based on data of Table 4 of Warwick & Price
( 197 5). Production has been estimated as sum of monthly production increments (see Warwick & Price,
1975) and by using Eqns. I (El), 2 (E2) and 3 (E3) (Table V) to calculate daily production· ind- 1 for
animals at mean weight at cohort. Water temperature data are from Hibbert (1977).

T(OC) Density Mean Production Daily production (µg) Monthly production (g)
(m-2) weight increment
(mg) (g) El E2 E3 El E2 E3

Oct 14 50 2.72
Nov II 18 2.60 -0.0041 22.3 21.0 37.1 0.012 0.011 0.020
Dec 8 10 3.27 0.0136 21.7 20.0 37.9 0.007 0.006 0.011
Jan 7 0 0 0 0 0 0 0 0 0
Feb 7 320 0.11 0.0176 1.2 1.2 1.7 0.012 0.011 0.016
Mar 7 1056 0.20 0.0619 1.9 1.9 2.9 0.061 0.060 0.091
Apr 8 1400 0.56 0.4421 4.9 4.7 7.7 0.207 0.199 0.324
May II 880 1.49 1.0602 14.3 13.6 23.0 0.378 0.358 0.607
Jun 15 656 2.03 0.4147 24.1 22.9 38.3 0.474 0.451 0.754
Jui 18 230 2.19 0.0709 30.2 28.7 47.3 0.208 0.198 0.326
Aug 19 168 3.16 0.1930 42.4 40.1 67.6 0.214 0.202 0.341
Sep 19 98 3.10 -0.0080 41.8 39.5 66.5 0.123 0.116 0.196
Oct 14 80 3.78 0.0605 37.3 34.9 62.0 0.090 0.084 0.149

Total 2.323 1.784 1.697 2.834

reproductive output by concentrating production towards growth and hence survival,

whereas individuals of larger species channel most assimilated energy towards repro-
duction on reaching a certain size because they suffer low levels of mortality and have
little incentive to continue growing. Whatever the cause, a difference between the
coefficients of log AFDW in Eqns. 2 and 3 will have ecological consequences.
The two equations cannot be extrapolated much below the size range examined here
(i.e., < 3 µg AFDW) or the somatic production of mature animals becomes greater than
the production of juvenile animals of the same size. This situation is highly unlikely,
hence either (i) the two equations are approximations for curvilinear relationships which
merge together at lower body sizes than examined here, (ii) the equations are not
relevant to meiofauna but only to a macrofaunal animal grouping, or (iii) the difference
in the coefficients of log AFDW for adults and juveniles is not real. An examination
of the distribution of the data points of Fig. 3 provides no support for the first hypothesis;
the residuals at both large and small body sizes are evenly distributed either side of those
predicted by the regressions.
The increasing difference between the production rates of young and adult animals
as body size increases is consistent with Banse & Mosher's (1980) finding that P/B
declines with body size more rapidly within the macrofauna and meiofauna than
predicted from regressions calculated using animals with a larger range in body size. The
coefficient of log biomass in Banse & Mosher's regression relating log P/B to log
biomass (a parameter hereafter referred to as the biomass exponent) nevertheless differs
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 209

substantially from that predicted by Eqn. 1. Eqn. 1 can be used to calculate population
production ( see below) and should therefore be directly transformable to a regression
relating body mass to population P/B. Such a transformed regression has a biomass
exponent ± SE of 0.20 ± 0.04, significantly higher than the value of - 0.37 ± 0.03
derived by Banse & Mosher.
The source of the discrepancy between the biomass exponent deduced here and that
derived by Banse & Mosher probably lies in their data selection. Banse & Mosher
recognised that a discontinuity existed between the P/B ratios of planktonic and benthic
species in the meiofaunal size range. They consequently decided, probably on the basis
that they had three good data points for calanoid and cyclopoid copepods and another
for a cladoceran but only one such point for a harpacticoid copepod, to exclude the
benthic harpacticoid rather than the planktonic species from their regression. However,
as Banse (1982) has since suggested that cladocerans and perhaps copepods do not
show body mass-production relationships typical of larger animal groups, and as the
macrofaunal species used in their analysis were overwhelmingly benthic or terrestrial
rather than planktonic (a chaetognath being the exception), open-water rather than
benthic species would have been better excluded from their calculations. Open-water
species tend to be continuously active and therefore possibly have greater feeding,
respiration and production rates in general than benthic species which move inter-
mittently and remain inactive for part of the day. If the five planktonic species are
excluded from, and the harpacticoid copepod included in, Banse & Mosher's analysis,
their regression between Ms (body mass at maturity) and P/B becomes
P/B == - 0.185-0.230 log Ms. The biomass exponent in this regression is similar to that
deduced for adult animals in this study ( - 0.20).
Schwinghamer et al. (1986) described the relationship between P/B and body mass
for marine animals ranging in size from macrofauna to microbes. They found an overall
biomass exponent of - 0.21, and also considered that the biomass exponents within
macrofaunal and meiofaunal/microfaunal groupings were much greater ( ~ - 0.30) than
the overall coefficient. This latter contention needs to be reconsidered, however, as it
was made in the light of Banse & Mosher' s biomass exponent for macro fauna of - 0.3 7.
Schwinghamer et al.'s biomass exponents for the macrofaunal and meiofaunal/
microfaunal groupings were each influenced by single outlying points (see their Table I
and Fig. 1) and had wide confidence intervals which included the overall exponent of
- 0.21.
Dickie et al. (1987) calculated an overall biomass exponent of - 0.18 for animal
groups ranging in size from mammals to meiofauna. This value is insignificantly dif-
ferent from the biomass coefficient of Eqn. 3 (young) when the latter is expressed in
terms oflog P/B rather than log P (i.e., as - 0.14). Dickie et al. further deduced, largely
on the basis of Banse & Mosher's macrofaunal biomass exponent of - 0.37 and
Schwinghamer et al.'s meiofaunal biomass exponent of - 0.34, that within various
animal groups ecological factors cause the biomass exponent to decrease to ~ - 0.35.
The results of my analysis agree with Dickie et al.'s major thesis that the primary
210 G.J. EDGAR

determinant of community production, the metabolic rate of individuals, has a biomass

exponent in the order of - 0.18, and that ecological scaling within the macrofauna
depresses the biomass exponent below this value. My results nevertheless also suggest
that amongst macrofaunal species ecological factors do not cause the biomass exponent
to fall below - 0.25, let alone approach - 0.35. Ecological scaling possibly also
becomes insignificant if production is calculated as the sum of somatic and gonadic
production, rather than somatic production alone.

ESTIMATION OF COMMUNITY PRODUCTION

The small proportion of the variance of log production contributed by factors other
than biomass and temperature indicates that a general equation has value in estimating
community production. A general equation has been used in the associated paper
(Edgar, 1990), producing estimates of macrofaunal community production consistent
with the results of other studies in similar habitats.
Calculations of community production should generally be made using Eqn. 1. Eqn. 2
(derived from the production rates of animals at ~ 70 % of maximum length) is, how-
ever, the better equation to use in studies such as Edgar (1990) where animals at ~ 70%
of maximum body size dominate the data sets and are near the mean of the biomass
distribution. Such data sets may be quite common because biomass is almost invariably
concentrated amongst the large individuals of a species even though young individuals
are generally much more numerous in a community than old individuals. It should be
noted that the choice between Eqns. 1 and 2 is not a major one as they are almost
identical, and that the data used to calculate Eqn. 2 were based on the production rates
of animals at ~ 70 % of the size ranges given in the published papers, and not 70 % of
the maximum length for the species. Because these measurements were not at a precise
size, Eqn. 2 should be more generally applicable, but also have a larger proportion of
unexplained variance, than if it had been deduced using production measured at exactly
70% of maximum length. Moreover, Eqn. 2 converges on Eqn. 3 at small body sizes,
hence even though the ratio of young to adult animals decreases with sieve size, the use
of Eqn. 2 (or Eqn. 1) for the smallest sieves will introduce little error into production
estimates.
The accuracy of production estimates based on the general regressions shown in
Table V can be gauged from the distribution of the residuals. The data points for all 41
species were between 35% (Montfortula rugosa) and 350% (Nereis diversicolor) of the
predicted values. The so value of the residuals ofEqn. 1 was 0.26, hence ~ 76% of data
points fall between 50 and 200'.1/0 of predicted values. Based on Eqn. 1, the 95%
confidence interval for the estimated production of a single population is between 30
and 320% of the predicted value.
If the production of a community is calculated, the confidence interval of the esti-
mated production rapidly declines as the number of species increases (by the square root
of the number of species if all species are equally productive). The production con-
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 211

fidence interval extends from 44 to 230% of that predicted for communities with two
equally productive species, from 59 to 170 % for five equally productive species, and
from 69 to 145% for 10 equally productive species. Although no natural communities
possess species with equal productivities, the approximate confidence intervals can be
calculated for real communities because the inverse of Simpson's index ( = 1/'f.p/,
where P; is the proportional productivity of the ith species) provides a measure of the
effective number of equally productive species in a community (see Hill, 1973; Peet,
1974).
The above calculations of confidence intervals provide only very approximate esti-
mates because they incorporate the statistical and analytical errors in my calculation
of the Table I data set, and they also assume that the productivities of different species
within the community are independent. This latter assumption would not be the case
if, for example, exploitative competition is common or a seasonal lack of food resources
limits growth among a guild of species. At present it is difficult to assess the extent of
such interactions amongst benthic species; it would clearly be worthwhile for produc-
tion estimates for a large number of species from one site to be calculated using the sieve
method and compared with those obtained by more direct means to determine whether
consistent biases in secondary production occur within an assemblage.
The allometric relationships of Table V do not need to be used in conjunction with
sieve size data but can be used more accurately by weighing all animals and partitioning
them into weight classes or by using cohort mode data. This latter method is shown in
Table VII where the cohort data of Warwick & Price (1975) for Ampharete acutifrons
(their Table 4) have been recalculated using Eqns. 1-3 to provide production estimates.
The estimate of annual secondary production calculated using Eqn. 1 is 23.2 % less than
the estimate of Warwick & Price based on cohort analysis, while the estimates based
on Eqns. 2 and 3 are 29.6% lower and 22.0% higher, respectively. Calculations for the
other major species in Warwick & Price's study (Nephtys hombergi, their Table 3) give
overestimations of production of 38, 36 and 159% using Eqns. 1-3, respectively. The
large overestimation of secondary production calculated for N. hombergi using Eqn. 3
is indicative of the magnitude of the error introduced using an equation based on young
animals when many adult animals are present.
Estimates of secondary production based on cohort analysis are undoubtedly more
accurate than those calculated using the sieve method. The obvious disadvantage of the
sieve method, that interspecific variability in productivity is ignored, is compounded by
the influence of food availability also being assumed insignificant. Secondary produc-
tion will consequently be underestimated for any populations with, or during any period
of, food surplus and rapid growth, and will be greatly overestimated during seasons of
negligible growth. The daily production rates of laboratory-reared animals fed surplus
food (e.g., Nair & Anger, 1979; Takeuchi & Hirano, 1988) are generally much greater
than the rates predicted using Eqns. 1-3, hence the production rates of field populations
often may be restricted by limited food availability.
The sieve method nevertheless has major advantages over the cohort method in its
212 G.J. EDGAR

ease of use, in being a smooth measure which does not fluctuate greatly between months,
and in not being affected by factors such as size-selective predation which can shift the
cohort modes offield populations. It should be pointed out, however, that the equations
were calculated from the modal shifts in field data, so if size-selective predation is a
widespread phenomenon which has caused many of the data points to be under-
estimated then these errors will have been incorporated into the general equations.
The major benefit of the new method over that of Banse & Mosher (1980) is that
population structure is incorporated into the calculation of production. Banse &
Mosher recognized this deficiency in their model; they excluded exploited and manipu-
lated populations from their data set because the addition of predatory or fishing
pressure "reduces greatly the older year classes and results in an upward shift in PJB".
Banse & Masher's method also cannot be used to estimate rates of production on time
scales of less than a year, and it assumes that water temperature has a negligible
influence on annual production rate when between 5 and 20 °C.

ACKNOWLEDGEMENTS

Constructive criticisms of the manuscript which were provided by S. F. Rainer, A. I.

Robertson and R. M. Warwick were much appreciated. The study was supported by
a Queens Fellowship in Marine Science and a Japan Society for the Promotion of
Science/Australian Academy of Science Postdoctoral Fellowship.

REFERENCES

Al-Adhub, A. H. Y. & A. B. Bowers, 1977. Growth and breeding of Dichelopandalus bonnieri in Isle of Man
waters. J. Mar. Biol. Assoc. U.K., Vol. 57, pp. 229-238.
Ali, M.H. & S.D. Salman, 1987. Growth and production of the amphipod Parhya/e basrensis (Talitridae)
in the Shatt al-Arab region. Mar. Ecol. Prog. Ser., Vol. 40, pp. 231-238.
Alon, N. C. & S. E. Stancyk, 1982. Variation in life-history patterns of the grass shrimp Palaemonetes pugio
in two South Carolina estuarine systems. Mar. Biol., Vol. 68, pp. 265-276.
Ambrogi, R. & A. 0. Ambrogi, 1985. The estimation of secondary production of the marine bivalve Spisula
subtruncata (Da Costa) in the area of the Po river delta. P.S.Z.N.l. Mar. Ecol., Vol. 6, pp. 239-250.
Ansell, A.D., P. Sivadas, B. Narayanan & A. Trevallion, 1972. The ecology of two sandy beaches in
south-west India. III. Observations on the populations of Donax incarnatus and D. spiculum. Mar. Biol.,
Vol. 17, pp. 318-332.
Ansell, A.D. & F. Lagardere, 1980. Observations on the biology of Donax trunculus and D. vittatus at lie
d'Oleron (French Atlantic coast). Mar. Biol., Vol. 57, pp. 287-300.
Baden, S. P. & L. Pihl, 1984. Abundance, biomass and production of mobile epibenthic fauna in Zostera
marina (L.) meadows, western Sweden. Ophelia, Vol. 23, pp. 65-90.
Banse, K., 1982. Mass-scaled rates of respiration and intrinsic growth in very small invertebrates. Mar. Ecol.
Prog. Ser., Vol. 9, pp. 281-297.
Banse, K. & S. Mosher, 1980. Adult body mass and annual production/biomass relationships of field
populations. Ecol. Monogr., Vol. 50, pp. 355-379.
Birklund, J., 1977. Biomass, growth and production of the amphipod Corophium insidiosum Crawford, and
preliminary notes on Corophium volutator (Pallas). Ophelia, Vol. 16, pp. 187-203.
Blackmore, D. T., 1969. Studies of Patella vulgata L. I. Growth, reproduction and zonal distribution.J. Exp.
Mar. Biol. Ecol., Vol. 3, pp. 200-213.
 ESTIMATION OF MACROBENTHIC PRODUCTION AND BIOMASS 213

Burke, M. V. & K. H. Mann, 1974. Productivity and production: biomass ratios of bivalve and gastropod
populations in an eastern Canadian estuary. J. Fish. Res. Board Can., Vol. 31, pp. 167-177.
Ceccherelli, V. U. & R. Rossi, 1984. Settlement, growth and production of the mussel Mytilus gallo-
provincialis. Mar. Ecol. Prog. Ser., Vol. 16, pp. 173-184.
Chambers, M. R. & H. Milne, 1975. The production of Macoma balthica (L.) in the Ythan estuary. Estuarine
Coastal Mar. Sci., Vol. 3, pp. 443-455.
Colby, D.R. & M. S. Fonseca, 1984. Population dynamics, spatial dispersion and somatic growth of the sand
fiddler crab Vea pugilator. Mar. Eco/. Prog. Ser., Vol. 16, pp. 269-279.
Collie, J. S., 1985. Life history and production of three amphipod species on Georges Bank. Mar. Ecol. Prog.
Ser., Vol. 22, pp. 229-238.
Creese, R. G., 1981. Patterns of growth, longevity and recruitment ofintertidal limpets in New South Wales.
J. Exp. Mar. Biol. Ecol., Vol. 51, pp. 145-171.
Crisp, D.J., 1984. Energy flow measurements, In, Methods for the study of marine benthos, edited by N.A.
Holme & A. D. McIntyre, IBP Handbook No. 16, Blackwell, Oxford, pp. 284-372.
Dickie, L. M., S. R. Kerr & P.R. Boudreau, 1987. Size-dependent processes underlying regularities in
ecosystem structure. Ecol. Monogr., Vol. 57, pp. 233-250.
Donn, T. E. & R. A. Croker, 1986. Seasonal patterns of production in the sandy-beach amphipod Haustorius
canadensis. Estuarine Coastal Shelf Sci., Vol. 22, pp. 675-687.
Downer, D.F. & D.H. Steele, 1979. Some aspects of the biology of Amphiporeia lawrenciana Shoemaker
(Crustacea, Amphipoda) in Newfoundland waters. Can. J. Zoo!., Vol. 57, pp. 257-263.
Edgar, G.J., 1983. The ecology of south-east Tasmanian phytal animal communities. I. Spatial organisation
on a local scale. J. Exp. Mar. Biol. Ecol., Vol. 70, pp. 129-157.
Edgar, G.J., 1990. The influence of plant structure on the species richness, biomass and secondary
production ofmacrofaunal assemblages associated with Western Australian seagrass beds. J. Exp. Mar.
Biol. Ecol., Vol. 137, pp. 215-240.
Essink, K. & A. H. Bos, 1985. Growth of three bivalve molluscs transplanted along the axis of the Ems
estuary. Neth. J. Sea Res., Vol. 19, pp. 45-51.
Fredette, T.J. & R.J. Diaz, 1986. Life history of Gammarus mucronatus Say (Amphipoda: Gammaridae) in
warm temperate estuarine habitats, York River, Virginia. J. Crust. Biol., Vol. 6, pp. 57-78.
Guillou, J. & P. G. Sauriau, 1985. Some observations on the ecology of a Venus striatula population in the
Bay of Douarnenez, Brittany. J. Mar. Biol. Assoc. U.K., Vol. 65, pp. 889-900.
Henderson, P.A. & R.H. A. Holmes, 1987. On the population biology of the common shrimp Crangon
crangon (L.) (Crustacea: Caridea) in the Severn estuary and Bristol Channel. J. Mar. Biol. Assoc. U.K.,
Vol. 67, pp. 825-847.
Hibbert, C.J., 1977. Energy relations of the bivalve Mercenaria mercenaria on an intertidal mudflat. Mar.
Biol., Vol. 44, pp. 77-84.
Hill, M.O., 1973. Diversity and evenness: a unifying notation and its consequences. Ecology, Vol. 54,
pp. 427-432.
Hughes, R. G., 1978. Life-histories and abundance of epizoites of the hydroid Nemertesia antennina (L. ). J.
Mar. Biol. Assoc. U.K., Vol. 58, pp. 313-332.
Hughes, R. N., 1971. Ecological energetics of the keyhole limpet Fissure/la barbadensis Gmelin. J. Exp. Mar.
Biol. Eco/., Vol. 6, pp. 167-178.
Imada, K. & T. Kikuchi, 1984. Studies on some reproductive traits of three caprellids (Crustacea:
Amphipoda) and their seasonal fluctuations in the Sargassum bed. Pub/. Amakusa Mar. Biol. Lab., Vol. 7,
pp. 151-172.
Josefson, A. B., 1982. Regulation of population size, growth, and production of a deposit-feeding bivalve:
a long-term field study of three deep-water populations off the Swedish west coast. J. Exp. Mar. Biol.
Ecol., Vol. 59, pp. 125-150.
Kanneworff, E., 1965. Life cycle, food, and growth of the amphipod Ampelisca macrocephala Liljeborg from
the Oresund. Ophelia, Vol. 2, pp. 305-318.
Kautsky, N., 1982. Growth and size structure in a Baltic Mytilus edu/is population. Mar. Biol., Vol. 68,
pp. 117-133.
Klein, G., E. Rachor & S. A. Gerlach, 1975. Dynamics and productivity of two populations of the benthic
tube-dwelling amphipod Ampelisca brevicornis (Costa) in Helgoland Bight. Ophelia, Vol. 14, pp. 139-159.
Lucas, J. S. & E. P. Hodgkin, 1970. Growth and reproduction of Halicarcinus australis (Haswell) (Crustacea,
Brachyura) in the Swan estuary, Western Australia. Aust. J. Mar. Freshwat. Res., Vol. 21, pp. 149-162.
214 G.J. EDGAR

McLaren, LA. & C.J. Corkett, 1984. Singular, mass-specific P/B ratios cannot be used to estimate copepod
production. Can. J. Fish. Aquat. Sci., Vol. 41, pp. 828-830.
Mills, E. L., 1967. The biology of an ampeliscid amphipod crustacean sibling species pair. J. Fish. Res. Board
Can., Vol. 24, pp. 305-335.
Moller, P., 1985. Production and abundance of juvenile Nereis diversicolor, and oogenic cycle of adults in
shallow waters of western Sweden. J. Mar. Biol. Assoc. U.K., Vol. 65, pp. 603-616.
Moller, P. & R. Rosenberg, 1983. Recruitment, abundance and production of Mya arenaria and Cardium
edule in marine shallow waters, western Sweden. Ophelia, Vol. 22, pp. 33-35.
Moore, P. G., 198 I. The life histories of the amphipods Lembos websteri Bate and Corophium bonnellii Milne
Edwards in kelp holdfasts. J. Exp. Mar. Biol. Ecol., Vol. 49, pp. 1-50.
Muus, K., 1973. Settling, growth and mortality of young bivalves in the Oresund. Ophelia, Vol. 12,
pp. 79-116.
Nair, K.K.C. & K. Anger, 1979. Experimental studies on the life cycle of Jassa falcata (Crustacea,
Amphipoda). Helgol. Wiss. Meeresunters, Vol. 32, pp. 444-452.
Nicolaidou, A., 1983. Life history and productivity of Pectinaria koreni Malmgren (Polychaeta). Estuarine
Coastal Shelf Sci., Vol. 17, pp. 31-43.
Nichols, F. H., 1975. Dynamics and energetics of three deposit-feeding benthic invertebrate populations in
Puget Sound, Washington. Ecol. Monogr., Vol. 45, pp. 57-82.
Olafsson, E. B. & L. E. Persson, 1986. Distribution, life cycle and demography in a brackish water population
of the isopod Cyathura carinata (Kroyer) (Crustacea). Estuarine Coastal Shelf Sci., Vol. 23, pp. 673-687.
Parry, G.D., 1982. The evolution of the life histories of four species of intertidal limpets. Ecol. Monogr.,
Vol. 52, pp. 65-91.
Peet, R.K., 1974. The measurement of species diversity. Annu. Rev. Ecol. Syst., Vol. 5, pp. 285-307.
Pihl, L. & R. Rosenberg, 1982. Production, abundance, and biomass of mobile epibenthic fauna in shallow
waters, western Sweden. J. Exp. Mar. Biol. Ecol., Vol. 57, pp. 273-301.
Potter, J.C., P.J. Chrystal & N.R. Loneragan, 1983. The biology of the blue manna crab Portunus pelagicus
in an Australian estuary. Mar. Biol., Vol. 78, pp. 75-85.
Richardson, C.A., A.C. Taylor & T.J. Venn, 1982. Growth of the queen scallop Chlamys opercularis in
suspended cages in the Firth of Clyde. J. Mar. Biol. Assoc. U.K., Vol. 62, pp. 157-169.
Robertson, A. I., 1979. The relationship between annual production: biomass ratios and lifespans for marine
macrobenthos. Oecologia (Berlin), Vol. 38, pp. 193-202.
Robertson, A. I., 1984. Trophic interactions between the fish fauna and macrobenthos of an eelgrass
community in Western Port, Victoria. Aquat. Bot., Vol. 18, pp. 135-153.
Schwinghamer, P., B. Hargrave, D. Peer & C. M. Hawkins, 1986. Partitioning of production and respiration
among size groups of organisms in an intertidal benthic community. Mar. Ecol. Prog. Ser., Vol. 31,
pp. 131-142.
Takeuchi, I. & R. Hirano, 1988. A preliminary report on the life-history of Capre//a danilevskii Czerniavski
(Crustacea, Amphipoda) reared in the laboratory. Bull. Jpn. Assoc. Benthology, Vol. 32, pp. 37-41.
Tamaki, A., 1985. Zonation by size in the Armandia sp. (Polychaeta: Opheliidae) population on an intertidal
sand flat. Mar. Ecol. Prog. Ser., Vol. 27, pp. 123-133.
Tamura, H. & K. Koseki, 1974. Population study on a terrestrial amphipod, Orchestia p/atensis japonica
(Tattersal), in a terrestrial forest. Jpn. J. Ecol., Vol. 24, pp. 123-139.
Tanaka, M., I 988. Ecology of Sesarma (Holometopus) haematochir (de Haan) in Amakusa. II. Growth rates.
Pub/. Amakusa Mar. Biol. Lab., Vol. 9, pp. 77-92.
Thompson, G. B., 1979. Distribution and population dynamics of the limpet Patella aspera (Lamarck) in
Bantry Bay. J. Exp. Mar. Biol. Ecol., Vol. 40, pp. 115-135.
Underwood, A.J., I 975. Comparative studies on the biology of Nerita atramentosa Reeve, Bembicium nanum
(Lamarck) and Cellana tramoserica (Sowerby) (Gastropoda: Prosobranchia) in S.E. Australia. J. Exp.
Mar. Biol. Ecol., Vol. 18, pp. 153-172.
Wada, K., 1981. Growth, breeding, and recruitment in Scopimera globosa and Jlyoplax pusillus (Crustacea:
Ocypodidae) in the estuary of the Waka River, middle Japan. Pub!. Seto Mar. Biol. Lab., Vol. 26,
pp. 243-259.
Warwick, R. M. & R. Price, 197 5. Macro fauna production in an estuarine mud-flat.J. Mar. Biol. Assoc. U.K.,
Vol. 55, pp. 1-18.
Waters, T. F., I 969. The turnover ratio in production ecology of freshwater invertebrates. Am. Nat., Vol. I 03,
pp. 173-185.
Wildish, D.J. & D. Peer, 1981. Methods for estimating secondary production in marine Amphipoda. Can.
J. Fish. Aquat. Sci., Vol. 38, pp. 1019-1026.
Zar, J.H., 1974. Biostatistica/ analysis, Prentice-Hall, Englewood Cliffs, New Jersey, 620 pp.