Comparative productivity in ecosystems-secondary productivity

o. W. Heal and S. F. MacLean Inr.

Introduction
Current emphasis upon the study of ecosystem function has generated much
interest in the patterns, mechanisms and control of energy flow and produc-
tivity. The conventional approach, incorporated in the IBP Biome studies,
involves collection of data on energy flow and productivity in a variety of
ecosystems. These data will be compared-first in related ecosystems, and then
in full variety-in search of patterns. The principles will be derived to explain
the observed patterns.
In the present paper we take a different course in the comparison of secondary
productivity between ecosystems. The energetic properties of consumption (C),
assimilation (A), production (P), and respiration (R) of heterotrophs are derived
as characteristics of the organisms' physiology and energy source. The organiza-
tion of heterotrophs into trophic systems imposes further constraints upon their
energetics. Thus, we examine the biological principles of secondary productivity
(Wiegert & Evans, 1967; Petrusewicz & Macfadyen, 1970) and assemble them
into a simple model to deduce the patterns of secondary productivity within and
between a variety of terrestrial ecosystems.
This deductive approach is adopted partly because of the lack of measurements
of trophic level productivity, rather than of single species productivity, in a
variety of ecosystems. However the results of a few ecosystem studies, including
some reported at this meeting, provide a preliminary evaluation of our deduc-
tions.
The basic carbon and energy source upon which all heterotroph production is
dependent is net primary production (NPP), and this sets obvious constraints
upon heterotroph production. In the balanced or steady state ecosystem total
h~terotroph respiration equals NPP. In an accumulating ecosystem heterotroph
respiration falls below net primary productivity by some amount usually
between 0 and 10 % of NPP. If, then, heterotroph production is related to
respiration in some simple and constant way, secondary productivity would
simply track primary productivity, and the maps presented by Helmut Lieth in
the previous paper could, with adjusted scale, describe secondary productivity
as well. Variation in bioenergetic parameters between groups of organisms, and
variation in the relative abundances of major groups of organisms in ecosystems,

89

W. H. Dobben et al. (eds.), Unifying concepts in ecology

© Dr W. Junk B. V. Publishers, The Hague and Centre for agricultural
publishing and documentation, Wageningen 1975
will lead to deviations between patterns of primary and secondary productivity.
It is a widely accepted concept that new tissue resulting from reorganization
of molecules from food represents production (SCIBP, 1974). Operationally,
production within a population may be determined by summing the growth
increments of individuals within any period of time:

(1)

where LlB j is the change in biomass of an individual in the population within the
time interval, and n is the total population size. Alternatively, production may
be determined by summing change in population biomass (LlB) and elimination
(E: loss of biomass due to death, loss of body material, emigration) (Petrusewicz
& Macfadyen, 1970):

P = LlB +E (2)

Thus, production is estimated at the population level. The sum of the productivity
of heterotroph populations across a trophic sequence, e.g. herbivore-carni-
vore- ..• top carnivore, is often taken as secondary production; this represents
successive reorganization of the same energy-containing molecules. Within the
decomposer cycle the same principle holds, although the reorganization begins
with dead organic matter. Thus, a food chain may be developed, e.g. fungi-
bacteria-protozoa-bacteria, etc., on the death of the organisms. Analogous to
production in the herbivore food chain, the sum of new organisms formed may
be taken as secondary production.
This concept of secondary production differs from the theoretical sequence of
orders of consumption and production, 1°, 2°, 3°, ... , n° recognized in both
the herbivore and saprovore trophic systems (Wiegert & Owen, 1971; Batzli,
1974). We use the term total heterotrophic productivity for the sum of successive
molecular reorganizations by heterotrophic organisms and one purpose of this
paper is to explore the consequences of this concept of productivity.
We use the calorie as our unit of energy flow and production, since it is a
useful common denominator, but it is often used in the sense of chemically
bound energy. We avoid the problem of confusing amounts e.g. production, and
rates e.g. productivity, by considering all values as m- 2 yel.

Assimilation and growth efficiencies

A two-part taxonomic-trophic categorization of heterotrophic organisms
(Table 1) summarizes a wide range of organisms and trophic functions. The
taxonomic distinctions recognize a general gradient in organism size as well as
the energetically important distinction between homeotherms and heterotherms.
Parts of the matrix are considered blank; thus, for simplification, we treat all
microorganisms as saprovores, and vertebrate microbivory is considered nil.

90
Table 1. A simple taxonomic-trophic categorization of heterotrophic organisms. For
each category the characteristic assimilation (A/C) and growth (PIA) efficiencies are
given.
Trophic function
Herbivore Carnivore Microbivore Saprovore
AIC PIA Ale PIA Ale PIA Ale PIA
Microorganisms 0.40
Invertebrates 0.40 0.40 0.80 0.30 0.30 0.40 0.20 0.40
Vertebrate
Homeotherms 0.50 0.02 0.80 0.02
Vertebrate
Heterotherms 0.50 0.10 0.80 0.10

Two measures are of primary importance in the fate of ingested or consumed

energy in a heterotroph population: the proportion of consumed food which is
assimilated (A/C) and the proportion of assimilated food used in production
(PIA). These parameters vary widely in different taxonomic and trophic groups
but only a brief justification of the representative values (Table 1) can be
offered here. Detailed data are referred to in Reichle (1971), McNeill &
Lawton (1970) and were reviewed by Macfadyen at this Conference.

Microorganisms
The concept of consumption efficiency (A/C) used in animal ecology has
little relevance for microorganisms which digest their food externally and
assimilate the products of exo-digestion directly. There are also problems of
comparison in terminology. For example in microbiology, assimilation often
refers to the fraction of substrate which is converted to new cells whereas in
fauna studies, and in the present paper, assimilation includes respiration as well
as production.
The vast literature, cited by Forrest (1969), Payne (1970) and Stouthamer &
Bettenhaussen (1973), on pure culture studies of bacteria, provides many values of
production per unit of substrate utilized (PIA) between 0.30 and 0.60. Continuous
culture techniques have tended to give lower values. Most bacterial studies have
concerned growth with readily available substrates, but yield declines at low
growth rate and with complex substrates. In addition there is a maintenance
requirement during periods of no growth. However fungi show high values of
PIA (0.25-0.50) even when growing on leaves and wood over periods up to six
months (Swift, 1973, personal communication; J. C. Frankland, personal
communication).
In field studies using 14C labelled substrates, growth efficiencies mainly
between 0.35 and 0.60 have been recorded for mixed microbial popUlations in
soil, freshwater, estuarine and marine conditions (Shields et aI., 1973; Behera &
Wagner, 1974; Hobbie & Crawford, 1969; Williams, 1973). These studies used

91
readily degradable substrates but the weight of evidence suggests that mixed
populations under field conditions convert about 0.40 of the substrate utilized
into cell biomass.

Invertebrates
The values of A/C for invertebrates relate to trophic function (diet) in the
general pattern:
saprovores < microbivores ~ herbivores < carnivores
(Reichle, 1971). Plant-sucking insects are often exceptions, passing large
amounts of carbohydrate (energy) while feeding selectively for amino acids
(Dixon, 1971); this gives them a low A/C for energy, but not for nitrogen. The
general pattern is relatable to dietary quality, which might be measured as
per cent long-chain polysaccharides.
The fate of assimilated energy (PIA) varies with life cycle characteristics,
behaviour, growth stage, etc. (McNeill & Lawton, 1970) rather than with diet
(Reichle, 1971). The highest recorded values are for larvae of parasitic insects
(Pteromalidae, Ichneumonidae), with no energy demands other than those
directly associated with growth (Chlodny, 1968). Lowest values are found in
predatory arthropods e.g. chilopods, spiders that actively pursue their prey
(Moulder & Reichle, 1972). Where energy is invested in materials not measurable
as growth e.g. the 'spittle' of spittlebugs, or silk in web-spinning spiders, the
production efficiency calculated from growth data is reduced.
Life cycles with long periods of maintenance of non-growing adult stages, or
periods of growth cessation, lead to low PIA ratios (McNeill & Lawton, 1970).
The PIA ratio is higher when calculated over the actively growing phases of the
life cycle than when periods of no or negative growth are included e.g. pupae or
adults indicating the growth efficiency of the organisms and of the population
respectively. The range of PIA seems to be lower for terrestrial than for aquatic
invertebrates (Welch, 1968).

Vertebrates
The assimilation efficiency of ungulates varies widely, with season and diet
(Short et aI., 1974). Grodzinski & Wunder 1975 found that assimilation
efficiency of ruminants averaged around 50 %, while monogastric vertebrate
herbivores tended to have higher values. Non-ruminants include graminivores
with an easily digested food item that is not included in the ungulate diets. On
comparable diets, the ungulate mode of digestion is probably no less, and
possibly more, efficient than the monogastric mode. The latter occurs in arctic
Alaska, where caribou (Rangifer tarandus) extract a greater fraction of dry
matter and energy from both monocot. and dicot. forage than do brown lem-
mings (Lemmus trimucronatus) (Batzii et aI., in press).
As in invertebrates, vertebrate carnivores show high assimilation efficiency.

92
Carnivores feeding upon vertebrates apparently achieve higher assimilation
efficiency than do carnivores upon invertebrates. This retlects the low digesti-
bility of ingested invertebrate exoskeletons while vertebrate skeletons are often
rejected prior to consumption. The arachnid and phalangid mode of ingestion
results in rejection of prey exoskeleton and a high A/C, comparable to vertebrate
predators upon vertebrates. Limited data on lizards (Pough, 1973) indicate
assimilation efficiencies similar to homeotherms.
PIA for vertebrate homeotherms is very low compared with invertebrates,
reflecting the high energy cost of thermoregulation. Maximum values generally
approach 0.03 (Turner, 1970) but values for microtine rodents with high produc-
tivity and rapid population turnover may exceed 0.04 (Petrusewicz & Hansson,
1975; Batzli et aI., in press). The few data on production efficiency of poikilother-
mic vertebrates indicate that values are intermediate between vertebrate homeo-
therms and invertebrates, for example, Mann (1965) quotes values between
0.07 and 0.10 for five fish species.
The values given in Table I and discussed above lead to some obvious predic-
tions regarding the efficiency with which a particular population converts the
energy consumed into biomass. This is the product of the two ratios, AIC and
PIA, and microorganisms show high efficiency, invertebrates and heterothermic
vertebrates are intermediate, and homeothermic vertebrate herbivores have very
low efficiency.

Organization of trophic systems

A next level of regulation of production comes from the organization of
species populations into food webs or trophic systems. Two more or less distinct
trophic systems exist in virtualIy all ecosystems-a herbivore system based upon
living autotroph tissue, and a saprovore system based on dead organic matter
(Fig. 1). The two trophic systems correspond broadly to an above and below
ground division, but merge to varying extents e.g. root and rhizome feeding
invertebrates, which together with saprovores and microbivores, support
populations of soil carnivores. Saprovorous insect larvae emerge from the soil as
adults and fall prey to above ground predators.
Several important conceptual differences distinguish the two trophic systems.
The herbivore system is biophagic at its base; events occurring within the system
can interact with and modify the rate of input of energy into the system, as in
over-grazed pastures. The saprovore system is saprophagic at its base with no
direct effect upon the rate at which energy enters the system although there can
be indirect effects upon primary productivity.
A second major conceptual difference lies in the pathways of energy transfer
within and through the two trophic systems. In the herbivore system, the energy
contained in organic molecules has two possible fates: it may be respired, with
release of CO 2 and radiant energy, or it may pass to the saprovore system. The
latter occurs when food is cropped but not consumed, consumed but not

93
 3 °Consumers

2° Consumers

1° Consumers

Input
~____p_p____l~1------~~____D_O_M____~
HERBIVORE SYSTEM SAPROVORE SYSTEM
Figure 1. A generalized trophic structure for terrestrial ecosystems and the consumption
efficiencies (CnIPn_l) used in the calculation of heterotrophic productivity in grass-
lands. PP = primary production. DOM = dead organic matter. For further abbrevia-
tions see Table 2.

assimilated, or when organisms suffer non-predatory mortality. Thus even at

the first level of consumption 50 %of the energy consumed passes to the sapro-
vore system since herbivores assimilate only about 50 %of the consumed food.
Once in the saprovore system, at least in a steady ecosystem, a unit of energy
has but one ultimate fate: release through respiration. Energy that is 'wasted' at
any trophic exchange, or energy in non-predatory mortality, returns to the pool
of dead organic matter (DaM) where it is available, once more, for assimilation
by microorganisms or saprovores. Thus, the saprovore system is conservative of
energy through the characteristic of recycling through the pool of dead organic
matter. This feature, we believe, is critical in allowing the greater length,
complexity, and biomass of saprovore over herbivore food chains.
Animals of the herbivore system tend to be larger than in the saprovore
system; this may reflect physical constraints on body size of soil-dwelling
organisms. Vertebrates are few in the saprovore system, and function primarily
at the carnivore level. Vertebrate saprovores-scavengers such as vultures-use
only recently dead animals, which are only a minor part of the input to DOM,
and are consequently limited in abundance in much the same way as are preda-
tors. Vertebrates are common at both the herbivore and carnivore levels of the
herbivore system. Thus with the high metabolic rate of homeothermy another
restriction on the flow of energy through the herbivore system is added.

94
Consumption efficiency
In the flow of energy through the two systems, a key parameter is Cn/Pn-1;
consumption at any trophic level, n, to production at the previous trophic level,
n - 1. The difference (P n-1 - Cn) represents energy passing to the dead organic
matter pool. Generalized values are given in Figure 1.
Vertebrate herbivores, especially ungulates and rodents, reach their greatest
abundance in grasslands and up to 50 % of available primary production may
be consumed (Wiegert & Evans, 1967). This is probably an overestimate, since
it is based upon above-ground plant parts. Thus, 25 %of the annual net primary
production is probably a more realistic maximum for terrestrial ecosystems. For
most ecosystems, e.g. forests, probably less than 5 % of the annual net primary
production is consumed by vertebrate herbivores.
Invertebrate herbivores usually consume less than 5 % of above ground
primary production (Reichle & Crossley, 1967). Exceptionally, during popula-
tion peaks of locusts or larvae of lepidoptera, the annual leaf production may
be consumed. This still represents less than 50 %of total primary production in
grasslands and less than 20 %in forests, although the effect on primary produc-
tion may be drastic (Rafes, 1970).
Consumption of prey by carnivores is usually related to prey abundance but
not prey production (Pearson, 1966, 1971). In studies of single-species popUlation
dynamics the causes of mortality can often not be quantified. The task of
generalizing is further confounded by density-dependent numeric and/or
functional response relationships between predator and prey populations.
In Ngorongoro in East Africa nearly all mortality of wildebeest resulted from
predation by the spotted hyaena. Since the prey population was fairly stable,
production equals mortality, and 100 %of all production was harvested by pred-
ators. In Serengeti there were many observations of non-predatory mortality
representing unharvested prey production passing to the saprovore trophic
system. About 60 %of the mortality of hyaenas resulted from predation (Kruuk,
1970). For Mustela nivalis feeding on Microtus in an old field, Cn fPn-1 was
estimated at 0.37 (Golley, 1960). Thus, consumption efficiency for vertebrates
preying on vertebrates may exceptionally reach 1.0, but is probably below 0.5 in
most cases.
Predation by vertebrates upon invertebrates usually accounts for less than
25 % of prey mortality. For example, about 20 % of the mortality of spruce
budworm (Choristoneurafumiferana) results from predation (Morris & Miller,
1954). Annual energy utilization by titmice (Parus spp.) and shrews (Sorex spp.)
in an oak wood was estimated at 5 %of the production of the main invertebrate
prey (Varley, 1970).
Little is known of the consumption efficiency by soil fauna, especially micro-
bivores for which it is necessary to estimate microbial production. Most
estimates are indirect, but radioisotopes allow direct estimation. Coleman &
McGinnis (1970) found a very small transfer from labelled fungi to soil mites

95
'I
McBrayer & Reichle (1971), found that consumption by fungivores averaged 7 %
of dry body weight day-l at field temperatures and total consumption was 24 mg
fungus m- 2 day-I. Healey (1967) reported that the fungivorous collembole
Onychiurus procampatus consumes II to 17 % of its body weight daily. Assuming
10 % and a 250-day season, annual consumption is 25 x the mean population
biomass. This may account for a significant part of microbial production.
Data on consumption by invertebrate saprovores are similarly scarce.
Consumption by Lumbricidae estimated from laboratory feeding rates, reaches
large proportions (Satchell, 1967). In areas without Lumbricidae, consumption
by litter-layer saprovores was I % of dry body weight per day, compared with
7% for fungivores (McBrayer & Reichle 1971). Consumption by 'total sapro-
vores' in a pine forest was 4.6 % of body weight daily giving a total annual
consumption of only I % of the annual litter fall (Kowal & Crossley, 1971).
Reichle (1968) found that food consumption of forest floor arthropods was
related to body weight: Consumption = 0.063 wtO. 68 • This gives values between
those quoted above.
McBrayer & Reichle (1971) estimated consumption by soil predators in a
deciduous forest to be 2.0-2.5% of body weight and 17 % of the consumption
by saprovores. Using the values for invertebrate saprovore AIC and PIA (Table
1) saprovore production is estimated from consumption. The resulting C pr~datorl
PSRPro •.-ore of 0.68, is high; however, the predators also feed upon microbivores,
so that consumption of saprovores is less than is indicated. Based upon biomass
and respiration of predators v. saprovores and microbivores, CpIPS+M is
probably around 0.25.
Vertebrate carnivores may feed upon invertebrates in the saprovore system;
however. their effect is small. Thus, less than 1 %of the production of tipuJids is
consumed by major predators in moorlands (Anthus pratensis, meadow pipit)
and in pasture (Sturn us vulgaris, starling) (Coulson & Whittaker. in press;
Kluijver, 1933). In northern Alaska conditions are optimal for the consumption
of invertebrate production by avian insectivores (MacLean, 1973) but they
account for only about 10% of the production of the major prey (MacLean,
in press).

Estimation of heterotrophic productivity

These values may now be applied to the general model of ecosystem organiza-
tion (Fig. I) to deduce productivity and energy flow in an ecosystem. The values
for AIC and PIA (Table I) are used for all calculations. since these are characteris-
tics of the organisms and are not likely to change systematically between
ecosystems and are based on extensive research. Values for CfI/NPfI - 1 were
varied. simulating differences in abundances and ecological organization in
different ecosystem types and reflecting the lack of data on consumption
efficiencies. Our first calculations used a value of 0.25 which approaches maxi-
mum for consumption of net primary production by vertebrate herbivores

96
Table 2. Calculated ingestion, production, respiration and egestion by heterotrophs
(kcal m-2 yc1) per 100 kcal m- 2 net annual primary production in a grassland ecosystem.
Efficiencies of consumption, assimilation and production in Figure 1 and Table 1 were
used in the calculation.
Ingestion Production Respiration Egestion
Herbivore system
Herbivores, vertebrate (HJ 25.000 0.250 12.250 12.500
invertebrate (Hi) 4.000 0.640 0.960 2.400
Carnivores, vertebrate (C y ) 0.160 0.003 0.123 0.031
invertebrate (C i) 0.170 0.040 0.095 0.034
Saprovore system
Saprovores, invertebrate (Si) 15.153 1.212 1.818 12.122
microbial (Sm) 136.377 54.551 81.826
Microbivores, invertebrate (Mi) 10.910 1.309 1.964 7.637
Carnivores, vertebrate 0.041 0.001 0.032 0.008
invertebrate 0.648 0.155 0.363 0.130
Total 192 58 99 35

% passing through
Herbivore system 15.2 1.6 13.5 42.9
Saprovore system 84.8 98.4 86.5 57.1

(Fig. 1). Energy flow per 100 kca1 m- 2 yr- 1 input from net primary production
was calculated from the energetic parameters in Table 1 and Figure 1, and is
given in Table 2.
We reiterate that we use the concept of production in the sense of organic
matter added to the biomass of a particular heterotroph species population.
Through predation, biomass may be consumed, and part assimilated and
either respired or added to predator biomass (production), while the remainder
is lost as faeces to the pool of dead organic matter. Non-predatory mortality
represents return of energy to the pool of dead organic matter from which it is,
again, assimilated by saprovores; that is, it is recycled. Thus a unit of chemical
potential energy, originally fixed in primary production, can be recorded in the
consumption, assimilation, and production of a number of different heterotroph
populations. The only constraint in the steady state (non-accumulating)
ecosystem is that heterotroph respiration equals annual net primary production.
Since the saprovore system is conservative of energy a unit of energy reappears
in production until it is, ultimately, respired.
In the example (Table 2) 86.5 % of the energy passed to the dead organic
matter pool and the saprovore system: 71.0 %as unconsumed primary produc-
tion, 14.9 % as faeces and 0.6 % as non-predatory mortality of heterotrophic
organisms. The herbivore system respires only 13.5 %of the energy entering the
system, even when herbivore consumption is set at its highest for terrestrial
ecosystems. Thus, the two trophic systems of Figure I are clearly of two different
sizes. It follows that differing levels of vertebrate herbivore ingestion, e.g. forests

97
(0.05) versus grasslands (0.25), have little effect upon patterns of total hetero-
trophic productivity.
It comes as no surprise that production and respiration of microorganisms
exceeds that of animals. This high microbial production allows a production of
microbivores which is greater than that of invertebrate saprovores.
Recycling within the saprovore food chain was considered by taking the
energy entering the pool of dead organic matter, initially 86.6 kcal/IOO kcal
NPP, following this through the saprovore system, and summing egestion and
non-predatory mortality to give energy re-entering dead organic matter for
another pass through the system. Thus, we treat recycling conceptually as a
series of discrete passes through the saprovore system.
For the system considered here, 56.7 %of the energy entering DOM is lost in
respiration at each pass; 43.3 %is recycled in organic matter. Six passes through
the saprovore system were necessary before the amount recycling fell below 1 %
of the NPP (0.433 6 X 86.6 = 0.57). Within the decomposer subsystem theoreti-
cal distinctions of trophic levels have been made. For example, Batzli (1974)
recognized 10 ,2 0 , • • • ,n° detritus, representing new and recycled input into
the DOM pool while saprovores (saprophages) were identified as 10 ,2 0 , • • • ,n°
depending upon the category of detritus used. Such distinctions are operation-
ally impossible under field conditions, since (a) the DOM forms a spatially,
structurally and chemically complex pool that is not easily subdivided, and (b)
the same organisms often function simultaneously at a number of levels. Their
activities cannot readily be apportioned to levels, as for example, with above-
ground omnivores using a variety of discrete food types. Thus, when microbial
or saprovore production is estimated from the sum of increases in population or
organism size, the value includes production based upon recycling of energy
containing molecules. While production cannot operationally be separated into
primary and recycled components, the approach described here allows estimation
of energy recycling through the DOM pool.

Evaluation of the model

Conclusions derived from such an exercise must be evaluated. We have only
the most preliminary picture of total ecosystem energetics, and the lack of
measurements of total net primary production, of different trophic levels and in
a variety of ecosystems, reduces the possibility of evaluating adequately the
accuracy of the predictions. Nevertheless, a first attempt has been made using a
limited set of data for production and respiration which, to a large extent, did
not contribute to the definition of the parameters used in the calculation. To
calculate heterotrophic production for a site, the predicted values for each
trophic level per 100 kcal m- 2 ye 1 were multiplied by the measured net annual
primary production for the site. For grasslands the predicted values in Table 2
were used; for non-grassland sites revised values were based on a consumption
efficiency by vertebrate herbivores of 0.05.

98
 4
10 /
/ /0 9

-
/
/
/
103 / /
/ /
l>... //6 6 6 5
/
/
/

/
102 6' / t1 6:' /
'"I ~
/ ( 10 L3 8 - /
E 6' J /,
/
r 7/
r
/
/
, lOr "6.4
/ ;2/
iii 10
/sr 1
/
/
U
.::.: \.J' 0 7 2 / 'e2 /
'-"
.-() 3 tJ
/ ~ /
c
0 Ai 7· /
/

·zU /
/
/
/

:J / /

..
"0 / 1 J4 /
0 / /
-1 02 / 10 /
10
c.. / 7- /

.4
/ /

"...
/ /
<II / /
-2 / /
U 10 /

.
/
/ /
"0 /
G.I /
/ /
Q. / /
10. 3 /
/
/
/
/
10- 4
-4 10- 3 10- 2 10- 1 10 3 10 4
10 10 10 2

Observed production (kcal m-1 yr- f )

.Figure 2. Heterotrophic production in a variety of ecosystems-observed and pre-
dicted values in kcal m- 2 yr-I, for vertebrate herbivores (.), invertebrate herbivores
(0), vertebrate carnivores (_), invertebrate carnivores (D), invertebrate saprovores
and microbivores (M, microbial saprovores (0). Observed values are from-Grass-
lands: 1, Pawnee, U.S.A. (Andrews et aI., 1974).2, Moor House, U.K. (Coulson &
Whittaker, in press). 3, Point Barrow, Alaska (Brown et aI., in press) including a range
of annual values for vertebrate herbivores. 4, Devon Island, Canada (Bliss et aI., 1973).
Non-grasslands: 5, Meathop Wood, U.K. (Satchell, in press) including a range of
values for invertebrate herbivores. 6, Oak Ridge, U.S.A. (Moulder & Reichle , 1972;
Re ichle et aI., 1973). 7, Moor House, U.K. (Coulson & Whittaker, in press). 8, Virelles,
Belgium (Froment et aI., 1971).9, Meerdink, Netherlands (Nagel-de Boois, 1971). 10,
Wytham Wood, U.K. (Varley, 1970).

The sites used for evaluation range from tundra to tropical rain forest. The
results (Fig. 2) indicate that observed values are generally within an order of
magnitude of predictions. Differences between actual and predicted values may
result from (a) large year to year variations some of which are indicated, (b)
errors in estimation of production and respiration for field populations, as well
as (c) errors in estimation made by the model.
In most cases predicted values are greater than observed values (Fig. 2). Field

99
.....
10 4
--.
>.
..... 10 3
I
E

~
"'v 10 2
'-"
C
0 10
....
..."'
c...
<II

...
41

"'0
....C1IV 10- 1
"tl
...
41
Il.. 10- 2

Observed respiration (kcal m-2 yr-f

Figure 3. Heterotrophic respiration in a variety of ecosystems-observed and predicted
values, in kcal m- 2 yr- 1 , for trophic levels as in Figure 2. Observed values are as listed in
Figure 2, plus-Grasslands: II, Armidale, Australia (K. J. Hutchinson, personal
communication). Non-grasslands: 12, Alpine shrub to tropical forest, Japan and
Thailand, (Kitazawa, 1970).

estimates of production rarely include all of the organisms functioning at a

particular trophic level, and thus tend to be underestimates. The only observed
value for annual microbial production (Nagel-de Boois, 1971) is a preliminary
estimate of production of fungal mycelium . This included only the L, F and H
horizons of the woodland site and was considered to be an underestimate for
methodological reasons. Estimates of bacterial production per gram of soil in
various temperate and tundra sites of the USSR (Parinkina, 1974) indicate
annual production of the order of 102-103 kcal m- 2 yr- 1 •
Annual respiration predicted by the calculations is also compared with some
observed values in the literature (Fig. 3). The result is similar to that for produc-
tivity. Although many of the data are for the same sites and organisms, the range
of values for respiration is less than the range of production (Fig. 2). This
reflects the low production but high respiratory cost of homeothermy. In both
Figures 2 and 3 predicted values exceed observed values, especially in carnivores,
indicating that the balance between production and respiration, determined by
PIA, is reasonable. As assimilation efficiencies are based on good data and it is
unlikely that site primary production is overestimated, then the errors probably
result from overestimation of consumption efficiency and underestimation of

100
observed values. Without further data, modification of model parameter values
is unreasonable but it emphasises areas requiring further research. With the
exception of the differences in vertebrate herbivore production in grassland and
non-grassland ecosystems major variations in the pattern of heterotrophic
production in different ecosystems are not recognisable, the predicted pattern of
heterotrophic productivity simply follows primary productivity.
A model of this sort is a very soft one, being based upon many parameters, and
variation of anyone may alter the results throughout. Thus, by trial and error
variation of parameters, the model output could be adjusted to fit most field
data; however, this violates the deductive nature of the exercise and would
provide little insight. The degrees of freedom of the model (the number of sets
of input parameters that produce reasonable output) are limited by (1) the
requirement that a set of output parameters, rather than one, show reasonable
values simultaneously, and (2) the fact that efficiencies of assimilation (A/C)
and production (PIA) are fairly well documented for most fauna and well
approximated by the present values. The main areas of ignorance which affect
production by soil organisms and may introduce serious errors into our
predictions are (I) the proportion of dead organic matter consumed by micro-
organisms and invertebrate saprovores (Fig. 2), (2) the consumption efficiency
(CnIP n-l) of fauna especially microbivores, and (3) the production efficiency of
mixed populations of microftora under natural conditions. Since all organisms
in the saprovore subsystem are linked through the phenomenon of recycling,
variation in any of these parameters has effects throughout the system. The
effects may be explored with the model to indicate the range of values which
produce reasonable results.

The meanings of production

Comparison of productivity within or between populations or trophic levels
indicates the extent to which different organisms are able to harvest and re-
organize the organic molecules of their food. Such comparisons provide insight
into the functioning and ecological limitation of different populations or trophic
levels, particularly in the two trophic subsystems. The apparent anomaly of
heterotrophic consumption exceeding autotrophic production results from
summation between trophic levels and recirculation of organic matter in the
saprovore system. Indeed, there is nothing that, theoretically, prevents hetero-
trophic production from exceeding primary production. Given an average PIA
value of greater than 0.5, this would occur. The value of secondary production
summed over a sequence of heterotroph trophic levels is not strictly comparable
to the value of primary production, which represents the single trophic level.
We should not compare heterotrophic assimilation with autotroph production
in search of the balanced ecosystem (Andrews et aI., 1974).
Because this concept of production includes recycled energy in the saprovore
subsystem it is not comparable to the concept of yield. Increasing the fraction of

101
Table 3. Estimated annual throughput (g m-2) of dry matter, nitrogen and phosphorus
by microflora (M) and invertebrate (I) saprovore populations per 100 kcal m- 2 input
from primary production (NPP). Concentrations of Nand P are given as %dry weight.
Throughput relative to input is shown in italics.
NPP Microflora Invertebrates
gm- 2 % gm- 2 % M/NPP gm- 2 % I/NPP
Dry matter 22.22 12.44 0.56 0.24 0.01
Nitrogen 0.33 1.5 0.75 6.0 2.27 0.02 10.0 0.06
Phosphorus 0.014 0.065 0.037 0.300 2.64 0.002 1.000 0.14

herbivore production harvested by carnivores will not necessarily influence

herbivore productivity, unless the total harvest exceeds production and thus
reduces herbivore biomass. In contrast, increasing the harvest of microbial
production by microbivores results in increased loss of energy through micro-
bivore and carnivore respiration, thus less energy recycling to support further
microbial production.
Because of differences in efficiencies the apparent importance of different
organisms varies according to the measures of consumption, production, and
respiration (Table 2). For example, in these calculations, vertebrate herbivores
contributed 13 %of total consumption, 12 %of respiration and 0.4 %of produc-
tion. (Clearly, the food-producing potential of vertebrate herbivores is limited.)
The importance of invertebrate saprovores is highest when judged by consump-
tion-the act by which they influence microbial activity through comminution
of litter and other processes.
The accumulated production estimate for a population or trophic level, when
multiplied by nutrient concentration gives the amount of nutrient cycled
annually through the population. Since nutrient requirements for tissue main-
tenance are ignored, it is a minimum estimate of mobilization. The nutrient
concentration of primary production is usually lower than that of microorgan-
isms or fauna (Reichle, 1971); thus, nutrient turnover relative to input is greater
than the relative turnover of fixed energy, or dry matter. An example, Table 3,
is based on the calculated annual production of microflora and of invertebrate
saprovores of 55 and 1.2 kcal per 100 kcal input from net primary production
(Table 2).
The biomass turnover rate or relative production (PI B) is a useful measure of
the rate of functioning of a population and is of practical value. From the ratio,
based on biological theory, the productivity of a population can be estimated
from easily collected biomass data (Phillipson, 1973). Annual production and
biomass of microorganisms are very difficult to determine. However, accumula-
tion of data may reveal a trend that is of predictive value. The relationship is
likely to be temperature dependent and data in Parinkina (1974) and Nagel-de
Boois (1971), and estimates of generation time, suggests values for P/B of 5-50.

102
 At the opposite extreme of size and biological complexity, vertebrate homeo-
therms regulate their productive environment and rate of acquisition of energy.
Thus the P/B ratio will vary more with size and life history strategy than with
external variables such as temperature. Large, slow growing, long lived animals
have a relatively low P/B, e.g. 0.2 for caribou (Rangifer tarandus), with
seasonal breeding. Small, rapidly growing and short-lived animals have a high
rate of biomass turnover e.g. about 6.0 for the brown lemming (Lemmus
trimucronatus) with year-round breeding and a high reproductive potential
(Batzli et aI., in press). Phillipson (1973) suggests 2.0 and 0.2 for small and large
mammals respectively. A population that is heavily harvested by predators or by
man will have an age distribution favouring younger, more productive
individuals and a higher P/B ratio. These characteristics correspond to "r_" and
"K" life history strategies (Pianka, 1970), thus we might expect a pattern of
decreasing P/B values approaching the tropics, or somewhat greater productivity
per unit of vertebrate homeotherm biomass in temperate and high-latitude
ecosystems. The major determinant, however, is probably the relative abundance
of large (e.g. ungulate) versus small (e.g. rodent) homeotherms.
Productivity of heterotherms is strongly affected by environmental tempera-
ture, e.g. the life cycles of tundra invertebrates often extend over a number of
years, in contrast to related species in more temperate areas (MacLean, 1973,
1975). Thus annual production is low. Because of the extended life cycles, a
number of separate generations or year classes coexist resulting in a large
abundance and biomass of individuals, each contributing a small amount to
annual production; P/B is low, with decreasing values as life cycles increase. The
pattern of change with latitude is opposite to that we predicted for homeotherms.
High invertebrate biomass with relatively small seasonal variation allows the
development of an abundant and diverse predator fauna and a large harvest of
prey production by carnivores (CnIP n- 1 ). Thus, this parameter may change
latitudinally with changing P/B relationship.
For annual species, which are more prevalent in temperate ecosystems, the
population biomass replaces itself each year and P/B ~ 1. In yet warmer regions
several generations may be completed in one year, the biomass present at any
time represents only one generation and P/B exceeds 1, increasing as the number
of generations completed in a year increases.
The production to biomass ratio in invertebrates, then, may be very sensitive
to temperature, length of the growing season, and other factors affecting life
history characteristics. It may be very useful for predicting rate of functioning
of animal populations, and possibly for microorganisms as well, from biomass
data. Variations in this parameter, as well as in heterotrophic composition of
different ecosystems, produce patterns of heterotrophic production which
probably characterize the world's ecosystems-but which remain undefined.
This discussion has concerned the ecological organization and function of
heterotrophic organisms, as formalized in a simple model. Time and space

103
limitations have denied us the opportunity to explore the use of the model in
predicting the amount and allocation among trophic types of heterotroph
production in different environments or ecosystem types. Given data on
abundance and rate of functioning of various taxonomic and trophic types, the
patterns of productivity can be predicted.

Summary and conclusions

Currently there are few measurements of heterotrophic production for trophic
levels, as opposed to species, from which patterns of variation can be examined.
We have adopted an alternative approach, using results from the many single
species studies on energetic efficiencies to derive a general model to predict
patterns of heterotrophic production. This assumes that annual heterotrophic
production is a function of

1. the input from primary production,

2. the consumption (CnfPn-l), assimilation (A/C) and growth (PIA) efficiencies

of the populations which are characteristic of their taxonomic and trophic
positions,

3. the organization of heterotrophs into herbivore and saprovore trophic

systems.

Heterotrophic production in an ecosystem is therefore the logical consequence of

the combination of these factors.
The results emphasise some fundamental distinctions between herbivore and
saprovore trophic systems, in particular the conservative nature of the latter.
Organic matter which is not respired by organisms within the saprovore system
returns to the pool of dead organic matter where it is again available for
assimilation. This recycling, in conjunction with high growth efficiency in
microorganisms, results in very high levels of production within the saprovore
trophic system. The estimates of accumulated production times the nutrient con-
centration of the organisms provides minimal estimates of nutrient throughput
in the saprovore system. Nutrient incorporation in biomass is often quite high
relative to nutrient input from primary production, indicating rapid nutrient
recycling.
A preliminary evaluation of the model against observed values for a number of
ecosystem studies shows general agreement, with heterotrophic production
directly related to primary production. Predicted values tend to be higher than
observed values and emphasise the need for a better understanding of consump-
tion efficiency at all trophic levels and for a range of good observed measurements
of production. Thcse limitations prevent us from detecting variations in the
pattern of production between different terrestrial ecosystems, and the present
results suggest that, despite major differences in species composition, most

104
ecosystems are similar in their organization and efficiency of heterotrophic
production.
At this stage the model is very generalized and resolution is limited. It provides
a means of predicting the amount of production of various taxonomic-trophic
categories in ecosystems, and of exploring the effects of variations in composition
and efficiencies of different groups of organisms. By adjusting the input from
primary production to correspond with regional variations (Lieth, this volume)
approximate estimates of heterotrophic production can be obtained on a global
scale.

Acknowledgements
We are very grateful to many research workers, especially those at Merlewood
Research Station and in the US Tundra Biome, for stimulating discussion and
constructive criticism. The preparation of this paper was supported by the
US National Science Foundation (Grant GV 29342 to the University of Alaska),
under joint N.S.F. sponsorship of the International Biological Programme
Tundra Biome and the Office of Polar Programs.

References
Andrews, R., D. C. Coleman, J. E. Ellis and J. S. Singh. 1974. Energy flow relationships
in a short-grass prairie ecosystem. Proc. 1st Int. Congr. Ecology. Pudoc, Wageningen.
22-28.
Batzli, G. O. 1974. Production, assimilation and accumulation of organic matter in
ecosystems. J. Theor. BioI. 45: 205-217.
Batzli, G. 0., R. G. White, S. F. Maclean Jm. and B. D. Collier. In press. The herbi-
vore-based food chain. In: The structure and function of the Tundra ecosystem:
synthesis volume. J. Brown, F. L. Bunnell, S. F. Maclean Jm., P. Miller and
L. L. Tieszen. eds. Dowding, Hutchinson and Ross, Philadelphia.
Behera, B. and G. H. Wagner. 1974. Microbial growth rate in glucose amended soil.
Soil Sci. Soc. Amer. Proc. 38: 591-594.
Bliss, L. C, G. M. Courtin, D. L. Pattie, R. R. Riewe, D. W. A. Whitfield and P.
Widden. 1973. Arctic Tundra Ecosystems. Ann. Rev. Ecol. Syst. 4: 359-399.
Brown, J., F. L. Bunnell, S. F. Maclean, P. Miller and L. L. Tieszen (Eds.) In press.
The Structure and Function of the Tundra ecosystem: synthesis volume. Dowding,
Hutchinson and Ross, Philadelphia.
Chlodny, J. 1968. Evaluation of some parameters of the individual energy budget of
the larvae of Pteromalus puparum (L.) (Pteromalidae) and Pimpla instigator (Fabr.)
(Ichneumonidae). Ekol. pol. Ser. A. 24: 505-513.
Coleman, D. C, and J. T. McGinnis. 1970. Quantification of fungus-small arthropod
food chains in the soil. Oikos. 21: 134-137.
Coulson, J. C and J. B. Whittaker. In press. Fauna of moorland soils. In: The ecology
of some British moors and montane grasslands. O. W. Heal & D. F. Perkins. eds.
Springer-Verlag, Berlin.
Dixon, A. G. F. 1971. Aphids. In: Methods of study in quantitative soil ecology:
population, production and energy flow. J. Phillipson. ed. IBP Handbook 18. Black-
well, Oxford. 233-246.
Forrest, W. W. 1969. Energetic aspects of microbial growth. In: Microbial growth.
P. Meadow & S. J. Pirt. ed. Cambridge University Press. 65-86.

105
Froment, A., M. Tanghe, P. Duvigneaud, A. Galoux, S. Denaeyer-de Smet, G. Schnock,
J. Goulois, F. Mommaerts-Billiet and J. P. Vanseveren. 1971. La chenaie melangee
calcicole de Virelles-Blaimont en haute Belgique. In: Productivity of forest eco-
systems. Proc. Brussels Symp. 1969. P. Duvigneaud. ed. Unesco, Paris. 635-665.
Golley, F. B. 1960. Energy dynamics of a food chain of an old-field community. Eco!.
Monogr. 30: 187-206.
Grodzinski, W. and J. Wunder. 1975. Ecological energetics of small mammals. In:
Small mammals: their productivity and population dynamics. F. B. Golley, K.
Petrusewicz & L. Ryszkowski. eds. Cambridge University Press. 173-204.
Healey, I. N. 1967. The energy flow through a population of soil collembola. In:
Secondary productivity of terrestrial ecosystems. K. Petrusewicz. ed. Panstwowe
Wydawnictwo Naukowe, Warsaw. 695-704.
Hobbie,1. E. and C. C. Crawford. 1969. Respiration corrections for bacterial uptake
of dissolved organic compounds in natural waters. Limno!. Oceanogr. 14: 528-532.
Kitazawa, Y. 1971. Biological regionality of the soil fauna and its function in forest
ecosystem types. In: Productivity of forest ecosystems. Proc. Brussels Symp. 1969.
P. Duvigneaud. ed. Unesco, Paris. 485-514.
Kluijver, H. N. 1933. [Contribution to the biology and the ecology of the starling
(Sturnus vulgaris vulgaris L.) during its reproductive period]. Vers!. Meded. Plziek-
tenk. Dinst Wageningen, 69.
Kowal, N. E. and D. A. Crossley. 1971. The ingestion rates of micro-arthropods in
pine mor estimated from radioactive calcium. Ecology, 52: 444-452.
Kruuk, H. 1970. Interactions between populations of spotted hyaenas (Crocuta crocuta
Erxleben) and their prey species. In: Animal populations in relation to their food
resources. A. Watson. ed. Blackwell, Oxford. 359-374.
MacLean, S. F. 1973. The life cycle and growth energetics of Pedicia hannai atenatta
Alex. (Diptera: Tipulidae), an arctic crane fly. Oikos, 24: 436-443.
MacLean, S. F. 1975. Ecological adaptations of tundra invertebrates. In: Physiologi-
cal Adaptation to the Environment. J. Vern berg. ed. Intext, New York. In press.
Maclean, S. F. Jm. In press. The saprovore-based trophic system. In: The structure
and function of the Tundra ecosystem: synthesis volume. J. Brown, F. L. Bunnell,
S. F. Maclean Jnr., P. Miller and L. L. Tieszen. eds. Dowding, Hutchinson and
Ross, Philadelphia.
McBrayer, J. F. and D. E. Reichle. 1971. Trophic structure and feeding rates of forest
soil invertebrate populations. Oikos, 22: 381-388.
McNeill, S. and J. H. lawton. 1970. Annual production and respiration in animal
populations. Nature, Lond. 225: 472-474.
Mann, K. H. 1965. Energy transformations by a population offish in the River Thames.
J. Anim. Eco!. 34: 253-276.
Morris, R. F. and C. A. Miller, 1954. The development of life tables for the spruce
budworm. Can. J. Zoo!. 32: 283-301.
Moulder, B. C. and D. E. Reichle. 1972. Significance of spider predation in the energy
dynamics of forest floor arthropod communities. Eco!. Monogr. 42: 473-498.
Nagel-de Boois, H. M. 1971. Preliminary estimate of production of fungal mycelium
in forest soil layers. In: Organismes du sol et production primaire. iv. Colloquium
Pedologie. Inst. Nat. Res. Agric. Paris. 447-454.
Parinkina, o. M. 1974. Bacterial production in tundra soils. In: Soil organisms and
decomposition in tundra. A. J. Holding, O. W. Heal, S. F. MacLean & P. W. Flanagan
eds. Tundra Biome Steering Committee, Stockholm. 65-77.
Payne, W. J. 1970. Energy yields and growth of heterotrophs. Ann. Rev. Microbio!.
24: 17-51.

106
Pearson, O. P. 1966. The prey of carnivores during one cycle of mouse abundance.
J. Anim. Ecol. 35: 217-233.
Pearson, O. P. 1971. Additional measurements of the impact of carnivores on California
voles (Microtus califamicus). J. Mammal. 52: 41-49.
Petrusewicz, K. and L. Hansson. 1975. Biological productivity in small mammal
populations. In: Small mammals: their productivity and population dynamics.
F. B. Golley, K. Petrusewicz & L. Ryszkowski. eds. Cambridge University Press.
153-172.
Petrusewicz, K. and A. Macfadyen. 1970. Productivity of terrestrial animals, principles
and methods. IBP Handbook 13. Blackwell, Oxford.
Phillipson, J. 1973. The biological efficiency of protein production by grazing and other
land-based systems. In: The biological efficiency of protein production. J. G. W.
Jones. ed. Cambridge University Press. 217-235.
Pianka, E. R. 1970. On r- and K- selection. Amer. Nat. 104: 592-597.
Pough, F. H. 1973. Lizard energetics and diet. Ecology 54: 837-844.
Rafes, P. M. 1970. Estimation of the effects of phytophagous insects on forest produc-
tion. In: Analysis of temperate forest ecosystems. D. Reichle. ed. Springer-Verlag,
Berlin. 100-106.
Reichle, D. E. 1968. Relation of body size to food intake, oxygen consumption and
trace element metabolism in forest floor arthropods. Ecology, 49: 538-542.
Reichle, D. E. 1971. Energy and nutrient metabolism of soil and litter invertebrates.
In: Productivity of forest ecosystems. Proc. Brussels Symp. 1969. P. Duvigneaud.
ed. Unesco, Paris. 465-477.
Reichle, D. E. and D. A. Crossley. 1967. Investigation on heterotrophic productivity
in forest insect communities. In: Secondary productivity of terrestrial ecosystems.
K. Petrusewicz. ed. Panstosowe Wydawnictwo Naukowe, Warsaw. 563-587.
Reichle, D. E., R. V. O'Neill, S. V. Kaye, P. Soli ins, and R. S. Booth. 1973. Systems
analysis as applied to modeling ecological processes. Oikos, 24: 337-343.
Satchell, J. E. 1967. Lumbricidae. In: Soil Biology. A. Burges & F. Raw. eds. Academic
Press, London, 259-318.
Satchell, J. E. (ed.) in press. The ecology of an English oakwood. Springer-Verlag,
Berlin.
SCIBP (Special Committee for the International Biological Programme). 1974.
Quantities, Units and Symbols. Publ. Int. Council Sci. Unions. Lond.
Shields, J. A, E. A Paul, W. E. Lowe and D. Parkinson. 1973. Turnover of microbial
tissue in soil under field conditions. Soil BioI. Biochem. 5: 753-764.
Short, H. L., R. M. Blair and C. A Segelquist. 1974. Fiber composition and forage
digestibility by small ruminants. J. Wildl. Mgmt. 38: 197-209.
Stouthamer, A H. and C. Bettenhaussen. 1973. Utilization of energy for growth and
maintenance in continuous and batch cultures of microorganisms. Biochim. Biophys.
Acta. 301: 53-70.
Swift, M. J. 1973. The estimation of mycelial biomass by determination of the hexosa-
mine content of wood tissue decayed by fungi. Soil BioI. Biochem. 5: 321-332.
Turner, F. B. 1970. The ecological efficiency of consumer populations. Ecology. 51:
741-742.
Varley, G. C. 1970. The concept of energy flow applied to a woodland community. In:
Animal populations in relation to their food resources. A. Watson. ed. Blackwell,
Oxford. 389-404.
WeIch, H. E. 1968. Relationships between assimilation efficiencies and gro\\th
efficiencies for aquatic consumers. Ecology. 49: 755-759.
Wiegert, R. G. and F. C. Evans. 1967. Investigations of secondary productivity in

107
 grasslands. In: Secondary productivity of terrestrial ecosystems. K. Petrusewicz. ed
2: 499-518.
Wiegert, R. G. and D. F. Owen. 1971. Trophic structure, available resources and
population density in terrestrial versus aquatic ecosystems. J. Theor. BioI. 30: 69-81.
Williams, P. J. Le B. 1973. On the question of growth yields of natural heterotrophic
populations. In: Modern methods in the study of microbial ecology. T. Rosswall
ed. Bull. Ecol. Res. Comm. Stockholm. 17: 400-401.

Author's addresses:
O. W. Heal
Institute of Terrestrial Ecology
Merlewood Research Station
Grange-over-Sands
Cumbria
England

S. F. MacLean Jnr.
Institute of Arctic Biology
University of Alaska
Fairbanks, Alaska
U.S.A.

108