Download Just proportions in food webs

I
' NATURE . VOL 341 . 14 SEPTEMBER 1989
community structure ignores evidencc
that herbivorcs not only dcpend on, but
sti'ongly influence, the plants they eat. In
The Origin of Species, Darwin rcported
that, on a small cleared plot, he "marked
all the seedlings of our native weeds as
they came up, and out of 357 no less than
295 were destroyed, chiefly by slugs and
insects." More recent studies show clearly
that the fcedback effects of herbivores on
plant populations are both negative"' and
positive". In the light of these studies, it is
far from an expected finding that the
complex interplays between plants and
herbivores sllould yield a simple proportionality between NAP and NSP. Moreover, the constant of proportionality, onethousandth, is far lcss than might have
been guessed from widely quoted ecological efficiencies of the order of 10-20
per cent. Nor would the trickle-down
theory explain why consumption by herbivores rises as the square of net foliage
production, as McNaughton ctal. find.
If studies of food-web structure demonstrate a rough proportionality between the
numbers of basal and intermediate species
(trophic species, really, since biological
species with identical predator species and
identical prey species are counted as a
single unit), and if future studies of foodweb function confirm the finding of
McNaughton et al. that NSP and NAP are
proportional, the next problem, and
IVEWSANDVIEWS
p
-
I
ECOLOGY
Just proportions in food webs
Joel E. Cohen
I N FOOD webs of natural communities
with vcrying numbcrs of species, the
number of green plants and other species
that consume no prey in the wcb (basal
species) is ncarly proportional, on ilvcrage. to the number of species that have
both predators and prey in the wcb (intermediate species), accordi~lg to studies
rcported over the past dozen yc;lrsl'.
Empirical rules"' of that typc describe
structural or i~natomic;~lproperties of
food wcbs. An import;int qucstion in
community ecology is wllcthcr parallel
regularities exist in the way in which food
webs function"", On page 142 of this
issuev, McNaughton el ul. report remarkable regularities in measures of food web
function. Their findings indicate thc existence of a physiology of food webs that
gcncralizcs ocross natural cornn~unities.
Their report opens up tllc distant. but
highly attractive, prospect of linking the
previously known structural regularities
with new functional patterns.
McNaughton cl 111. show that thc biomass, consutnption and nct secondary
productivity (NSP) of herbivores ;Ire all
rel;~tcdto the net abovc-ground primary
productivity (NAP)of a natural habitat by
simplc powcr laws. Particularly striking is
that the exponent of the power law that
relalcs NSP to NAP is 1101 sig~~ificantly
dilfcrcnt from 1: to a first al~proximation.
NSP and NAP arc dircctly proportional.
In cl'fcct, the cncrgy convcrlcd inlo herbivorcs pcr square mctrc of sulfi~ceurea
pcr year is, on average, iibout onethousandth of the e n e r g th:it is converted
into plants per square metre of surface
area pcr year. The ecological energy
pyran~id has a much largcr base than
second storey.
This proportionality would b e a triviality if scco~ldaryproductivity wcrc passively controlled and limitcd by primary
productivity. But this tricklc-down view of
NATURE . VOL 3 4 1 . 14 SEPTEMBER 1989
opportunity, is to relate the two proportionalities. In the 35 studies where NSP
and NAP were both measured directly,
are the numbers of basal trophic species
proportional to the numbers of herbivore
trophic species? What is the relation
between primary or secondary productivity and the diversity of biological and
trophic species in these studies? Stay
tuned!
But as Mark Twain advised a young
reporter: first get your facts straight, then
you can distort 'em. McNaughton et al. I very young science. A backward glance to
acknowledge that their findings would be the early stages of other, now maturer,
considerably strengthened by more data sciences shows the tremendous power of
from tropical forests and other structur- empirical proportionalities. Joseph Louis
ally complex systems. At present, their Proust proved that elements combine in
results rely heavily on structurally simple definite proportions by weight when they
form chemical compounds (the law of
vegetation such as grassland and tundra.
Moreover, the wide scatter of the assem- definite composition), laying a corner-,
bled data with respect to the log-log linear stone of the empirical support for the
regression proposed to describe them atomic theory". Gregor Mendel found
brings to mind the statistical maxim that. fixed proportions of different phenotypes
on log-log coordinates, even an elephant in, offspring in his experiments on plant
looks like a straight line. The report of breeding; his explanation in terms of
McNaughton et al. does not have all the genes was phenomenological. Erwin
answers, but it hassome and, betterstill, it
1. Cohen. J. NBlure 270.165-166 (1977).
Briand. F. 8Cohen;J. Nature307.264-267 (1984).
raises important new questions.
, 2.3. Jeffrles.
MJ. & Lawton.
J.H. Freshwater Biolo~v
~.
- 15.
.
Both the structural and the functional
105-112 (1985).
proportionalities found in food webs are
4. Suglhars. G.. Schoenly. K. 1Trombla. A. Science 245.
48-52 (1989).
purely descriptive. Though a model exists
5. Cokn. J.E.. Briand. F. 4 N m s n . C.M. Communily Food
that explains much of food-web strucWebs: Data and meory Blornslhernslicr Vol. 20
(Sprlnger, Heldelberg. 1989).
t u r ~ " ,it too is phenomenological; and
6. Lawton. J.H. 8 Warren. P.H. Trends Ecol. Ewl. 3. 242there is no hint of a mechanistic explana245 (1888).
7. Plmm. S. Foodwebs (ChapmanandHall. London. 1982).
tion for the observed patterns in the report
8.
Cole.
JJ.. Flndlay. S.E.G. PLovell. G.M. (eds) Cornparaof McNaughton et a/.. But ecology is a
NEWS AND VIEWS
Chargaffs discovery that .the concentration of adenine equals that of thymine
and that the concentration of guanine
equals that o f cytosine in DNA from
various sources was one of the chief clues
to Watson and Crick's solution of DNA
structure. Time will reveal the scientific
consequences of the new proportionalities
in food-web structure and function. The
possibilities are exciting.
0
Joel E. Cohen is Professor of Populations at
The Rockefeller University, New York, New
York, 10021-6399,USA.
9.
lo.
11.
12.
13.
tlw Analyses OlEcosystems: Patterns. ~echanismsand
Theories (Springer. New York. 1990).
McNaughton. SJ.. Oeslerheld. M.. Frank. D.A. 8 Willisrns.
K.J. Nature 341. 142-144 (1989).
Dirzo. R. In Perspectives on Plant Populalion Ecolo.~
(eda Dirzo. R. 4 Ssrukhln. J.) 141-165 (Sinsuer. Sunderland. Maasachusetu. 1984).
Blanchl. 1,s..Jones. C.G. & Shschok. M. Trends Ecol.
EWl. 4.234-238 11989).
Cohen. J.E. (L
C.M. ~ r m R.
. soc. 0224.421448 (1885).
Jafk. 0. Crucibles: me ~ t o ~ o l ~ h e m r s r r y ( ~ a w New
cett.
York. 1957).
c man.