Proc. Nat. Acad. Sci. USAVol. 70, No. 2, pp. 585-590, February 1973

Competition Among Plants

F. W. WENT

Laboratory of Desert Biology, Desert Research Institute, University of Nevada, Reno, Nev. 89507

Competition is a word of various meanings. In biology, itoriginally was introduced to account for the low survival rateof the potential offspring of all creatures. The number ofseeds formed by a pea plant may be a dozen; most annualplants produce hundreds or thousands of seeds; cottonwoodtrees and orchids seeds run to the millions; and, in the case offern and mushroom spores, there are billions formed by asingle individual. Since in a state of equilibrium each plantcan be replaced by only a single other one, processes wereconsidered that eliminated the excess offspring (such as theactivity of predators).With Darwin's evolution theory, competition took on addi-

tional meaning in relation to survival of the fittest. Competi-tion was not anymore a struggle between equals, but a mech-anism to award superiority. Competition became a contest,and considerations of combat, struggle, territorial exclusion,and even war entered in the wake of Darwin's ideas. As Warm-ing (1) states, competition is "a consideration of the means bywhich plants oust each other from habitats." But, it is hardto conceive of any mechanisms by which stationary plantscan combat each other to result in an ousting.

In an important experimental investigation, Clements,Weaver, and Hanson (2) studied competition. They con-cluded that "Competition is purely a physical process. Withfew exceptions, such as the crowding up of tuberous plantswhen grown too closely, an actual struggle between competingplants never occurs.. In the exact sense, two plants, no mat-ter how close, do not compete with each other as long as thewater-content, the nutrient material, the light and heat are inlexcess of the needs of both. When the immediate supply of asingle necessary factor falls below the combined demands ofthe plants, competition begins."When growing sunflower, wheat, and other plants at differ-

ent distances of each other, Clements et al. (2) found that thecloser the plants were spaced to one another, the more theyinhibited each other. But, it appeared from their data (seeTable 1) that: (i) all plants in a competition plot were equallyreduced in growth, and (ii) with increasing density of theplanting, the production of the plants per unit area tended toreach a maximum value, which was not changed with furtherdecreases in spacing. This is a common experience in all agri-cultural spacing tests, a result that shows that this form ofcompetition does not provide a mechanism for selection, sinceall individuals are equally inhibited. The same experience wasgained from observations in the field.

In the center of Death Valley near the headquarters of theNational Monument, with an average yearly rainfall of about40 mm, the vegetation is exceedingly poor. Only very fewshrubs-such as Larrea, Atriplex hymenelytra, and Tidestroe-mia-grow in that area, and the number of seedlings of annualplants appearing after rain is small (see Table 2). With the

585

exception of 1969, almost as many ripe seeds were produced asseeds germinated, which means that a rather precarious dy-namic equilibrium exists for the few annual plants growing inthe center of the valley (Geraea canescens, Chorizanthe rigida,and Chaenactis carphoclinia). Exceptional seed years like1935, 1947, and 1969 are needed to replenish the seed popula-tion in the driest parts of Death Valley, where normally noseed surpluses are produced to feed harvester ants (Veromes-sor) and seed-eating rodents (Dipodomys). Consequently,these seed eaters are mostly absent in the driest areas of thedesert.

In the less dry parts of Death Valley 12 plots, each about0.2 i2, were surveyed from January through June. Table 3shows the germination and survival data for these plots in theyears 1968 and 1969. A total of 2893 seedlings were marked inthese plots, of which 1178 (41%) survived, flowered, andset seed, and 33,951 mature seeds were formed. This is anaverage of 29 seeds per plant, or 12 new seeds per germinatedseed. In other years and other deserts a 10- to 20-fold increasein the number of seeds produced per seed germinated was alsofound. The survival of 41% of all seedlings was slightly lowerthan in other deserts [Southern California (3), over 50%;Southern Nevada (4), 56%; Southern Nevada (Went 1972,unpublished data), 53% ]. Some seedlings disappeared becausethey were eaten by rodents or insect larvae (Oenothera clavae-formis by Altica torquata), but most died in the early stages ofgermination when their roots did not penetrate properly in thesoil. But, once established, the seedlings survived for practi-cally 100% to flowering. Only in a few of the taller plants(Mfalacothrix californica, Atrichoseris platyphylla, and Chaen-actis carphoclinia) was reproduction poor, because their flower-buds were grazed off. The general conclusion to be drawn fromthese observations, therefore, is that the selection of the sur-vivors in the population of annuals in the desert is not a resultof competition among themselves. Since there is a 10- to 20-fold increase in seeds with each germination event, what is theselection process that keeps the desert seed population fromincreasing exponentially? Anywhere from 90 to 95% of allseeds produced have to disappear. This disappearance is notdue to decomposition of seeds by microorganisms. In the firstplace, we do not find partially digested seeds in these desertsoils. Besides, we know that the seeds of annuals under drydesert conditions remain fully viable for at least 20 years (5).Removal of seeds by seed-eating animals must be considered

seriously. Tevis (6) found in the desert that "the estimatedamount of seeds taken by the ants (Veromessor pergandei)from an acre in 12 months compared with the estimated num-ber of seeds -produced in a poor year showed that the insectsdo not seriously affect the total seed supply." But Tevis' antshad gone through a long drought period, and were very muchreduced in numbers. An ant nest (also of Veromessor per-

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Proc. Nat. Acad. Sci. USA 70 (1973)

TABLE 1. Effects of spacing on Helianthus annuus 80 daysafter planting (2)

Plants Height, Leaf area, cm2 Dry weight, gramsper 10 cm perm2 plant per plant per m2 per plant per m2

1 220 27000 2700 491.4 494 235 9800 3900 279.2 7016 207 2600 4200 85.5 13564 150 580 3800 20.8 130

250 115 64 1600 4.6 1151000 100 41 4100 2.1 210

gandei) observed in Death Valley was at least 10 times aspopulous as those observed by Tevis, and its population col-lected over 10 times as many seeds. Therefore, it seems likelythat the harvester ants adjust rapidly to the total amount ofseeds available, and, together with rodents [with a comparablebiomass in the desert (7)], actually remove most of the over-production of seeds, resulting in an "ever-normal granary."Therefore, it is not competition among themselves, but preda-tion of their seeds by ants and rodents, that keeps the popula-tion of desert annuals on an even keel.

This lack of competition among seedlings and mature plantswas also found by Kooper (8) among weeds in fallow fields inJava, where the number of seedlings was counted soon afterplowing. This contrasts with an observation of Darwin, whomentions in The Origin of Species (9): ". on a piece of ground3 feet long and two wide, dug and cleared, and where therecould be no choking from other plants, I marked all the seed-lings of our native weeds as they came up, and out of 357 noless than 295 were destroyed, chiefly by slugs and insects."In Darwin's case, only 18% of the weeds survived, althoughnot by competition with each other.Whereas the previous observations show that close spacing

of seedlings does not result in differential survival and elimina-tion of-presumably-the weaker plants, the individuals re-main smaller the closer the spacing. The mechanism of thismutual inhibition has not been established, but even in themost extreme case it does not result in elimination of plants.

In the desert, kangaroo rats (Dipodomys) and pocket mice(Perognathus) collect seeds that they carry in cheek pouches

TABLE 2. Germination of annual plants and seedproduction in a 1_M2 plot in the center of Death Valley

Number of seeds Number of newYear germinated seeds produced

1966 17 111967 30 311968 37 91969 32 2741970 0 01971 19 351972 1 0

to their nests. In years with abundant seed production, theymay bury their cheek pouch contents in spots all over thedesert and, after an appropriate rain, these superficially buriedseed caches will germinate. Then, anywhere from 100 to 200seedlings will sprout in an area of less than 1 cm2; always these

green tufts contain only a single species of seedlings, and al-ways over 95%, often 100%, of all the seeds germinate. Allof the seedlings survive in these most extreme cases of com-

petition, and most of them manage to flower and fruit. In one

of those tufts of Plantago insularis, I counted 126 plants andthree ungerminated seeds, of which two plants had producedtwo flowers, and 74 had one flower each. Since per flower twoseeds are produced, this patch of plants of 1 cm2 produced 152seeds per 129 seeds in the original cache. A well-developedPlantago insularis, growing on 1 dM2 of ground, probablywould have produced 20 ears with 30 flowers each, or 1200seeds, or only 10% of what the closely spaced plants of theseed cache managed to produce per unit space. Similar ob-

servations were made on seed caches of Pectocarya penicillataand other Boraginaceae; in each case at least as many seedswere produced as had been present in the original cache.Summarizing all my desert observations, I found that (i)

competition under most extreme conditions results in a 1:1replacement of all germinated seeds, (ii) under the usual desertconditions there is a 10- to 20-fold increase in seed populationafter each germination event, (iii) there is no differentialsurvival during growth of these desert annuals, and (iv) thenumber of seeds available for germination in successive sea-

TABLE 3. Number of seedlings germinated after early winter rains in 0.1- to 1-M2 plots in Death Valley, the numberthat survived to the flowering and fruiting stage, and the number of ripe seeds produced per plot

1968 1969

Altitude in m Germinating Surviving Seed Germinating Surviving Seed

Valley Bottom -80 37 5 9 32 19 274Road to Beatty -40 303 42 270 63 50 897Bennett's Camp -80 148 69 3093 28 1086Bennett's Well -80 33 11 210Jubilee Pass 100 149 74 140 - 48 2778Titus Canyon 200 35 12 372 63 33 1553Weir 600 186 51 375 80 72 150Grapevine SW 800 222 47 612 170 62 2830Grapevine NW 800 145 13 82 124 61 736Grapevine SE 800 102 16 686 61 45 500Grapevine NE 800 144 45 1005 104 33 600Entrance 1000 194 168 6000 151 83 6258Scotties 1100 272 80 3335 75 11 100All plots - 1970 633 16,189 923+ 545 17,762

586 Went

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Proc. Nat. Acad. Sci. USA 70 (1973)

sons is controlled by removal of the excess seeds by ants androdents, and not by differential production related to differ-ential survival of mother plants. Therefore, these annuals donot exert any measurable selection pressure on each other;their survival is conditioned by (a) the germination condi-tions, (b) the physical conditions immediately after germina-tion that determine whether the seedling roots will penetrateproperly into the soil, (c) the degree to which they are eatenby animals (a minor factor for these diminutive desert an-nuals), and (d) the degree to which their seeds have not beenremoved by ants, rodents, birds, and, to a small extent, liz-ards. These findings perhaps account for the fact that exceptfor germination controls these desert annuals do not show anyobvious adaptations to desert conditions.

In the previously discussed cases much of the interplantinhibition was strong or very strong, when a 100- or even 1000-fold size difference occurred between solitary and closely-spaced plants. But, there was no elimination of individuals;when germination occurred at the same time, the mutual in-fluence was evenly distributed and quantitative.The situation is very different when we consider the es-

tablishment of new plants in an existing vegetation. In aclosed vegetation essentially no seeds germinate. Exceptionsare seeds of parasites, such as Striga and Euphrasia, whichneed the secretions of their host plants for germination. Also,bulbils and other vegetative propagules (e.g., of Polygonumvwiparum) can develop in a closed vegetation. But even inopen spaces of an established vegetation, germination andgrowth of plants is severely inhibited; these examples offer thebest cases of plant competition. Typical examples of suchapparent competition are the open unvegetated spaces aroundSalvia leucophylla and other chaparral shrubs, the lack ofundergrowth in walnut orchards and Eucalyptus forests, andthe absence of desert annuals under Encelia farinosa shrubs.Hundreds of other cases have been described in the literature.They are generally attributed to the effects of allelopathicsubstances, produced by the dominant plant, which eitherprevent germination or inhibit growth of other plants. Anextensive literature has sprung up about Allelopathy, sinceMolisch (10) coined this term 35 years ago. Much informa-tion can be found in Grummer (11), Evenari (12), and Whit-aker (13).

Allelopathic substances may be volatile, for example, theymay be terpenes (14) and operate via the air, or they may dif-fuse through the soil; they may be produced by living roots,or may be derived from decaying above-ground parts; theymay inhibit other plants indiscriminately or be most effectiveagainst younger plants of the same species; they may persistfor a long time in the soil or lose their effectiveness as soon asthe inhibiting plant disappears; some are present in the seeds,but most are produced during the growth of the plant. Along list of chemicals has been established with inhibitoryeffects on other plants but, curiously, there are very few allel-opathic substances known that have growth stimulating ef-fects.

Yet, there are many positive interactions between plants,where the presence of one plant enhances rather than inhibitsgrowth of another, where cooperation rather than competitionreigns, where inclusion rather than exclusion is involved incommunity development. This relationship was pointed outby Kropotkin (15) for animal communities, but the same

correlations between epiphytic orchids, mosses and lichens,and the host tree (e.g., ref. 16). In the desert, several plants(Rafinesquia neomexicana) grow only near shrubs [Franseriadumosa, (17)1. Many mushrooms are always growing in associ-ation with certain trees, and grasses and legumes, like clover,are mutually stimulatory. This effect is so important thatmost of the pasture development in Australia and New Zea-land is based on the coexistence of grasses and subterraneanclover.Thus far, we have looked only into interplant competition.

An entirely different aspect is intraplant competition. Thiseffect has, of course, no evolutionary significance, for it willnot change the species composition of the vegetation.Each plant should be considered not as an individual, but

as a colony, very much like in animals a coral, a bryozoon, or asyphonophore is a colony of similar or of differentiated in-dividuals. There is a strict control of growth that keeps theindividual parts of a plant-stems, roots, buds, flowers-inequilibrium, and that insures a harmonious development.Through correlative inhibition only a limited number of budsdevelops into shoots, and this control is part of the mechanismby which plants competing for space keep their proportions.The more space a wheat plant has, the more basal buds de-velop into tillers and produce ears. Since the expansion growthof new leaves is controlled by existing leaves, the leaf size ofclosely spaced sunflowers, already reduced by strong competi-tion, will remain small (see Table 1). In this way, the inter-plant competition causes a more severe intraplant competi-tion, through hormonal control. In the case of the sunflowers ofTable 1, the intraplant competition almost completely bal-ances the interplant competition, a balance that results in aconstant amount of plant mass per unit growing surface.This effect makes the planting density uncritical for yield peracre for a farmer, at least beyond a minimal density. At higherdensities the farmer may waste seed in planting, and he mayhave trouble in controlling weeds, but he does not significantlychange his total yield per acre.

In a montane Sierran forest, intratree competition turnedout to be more severe than intertree competition. In LittleValley, a valley located 40 km south of Reno between LakeTahoe and Washoe Lake at 1900 m altitude, the meadow inthe center of the valley is bordered by dense forests of Pinusmurrayana, where the ground water stays within 10-100 cmof the surface the year around. In the densest stands the lightintensity is only a few percent of that outside the forest, andin the darkest areas there is no undergrowth, except for somepine and fir seedlings and saprophytic phanerogams such asSarcodes sanguinea, Pterospora andromedea, Corallorhizamaculata, and Pyrola secunda.In one of the darker parts of this forest, with a pure stand of

Pinus murrayana, the following observations were made. InFig. 1 the percent of living trees in any one of five size clas-sifications was entered. The age of the trees is fairly well cor-related with the diameter of the stem, with the 24- to 40-cmdiameter trunks probably 90 years old, when Little Valley waslogged to provide timbers for the mines of Virginia City, 15km east of Little Valley. It is interesting that about half ofthe trees were alive, even among the smallest ones. Thesesmall ones had grown up in the deep shade of the large trees,with 2-6% of the outside light reaching their needles, a lightintensity at which all branches of the older trees were dead.

effect is important in plants as well. There are strong positive

Competition Among Plants 587

Fig. 2 shows the distribution of living branches with green

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Proc. Nat. Acad. Sci. USA 70 (1973)

100

50

0 3 6 12 24 40

FIG. 1. Percentage of living Pinus murrayana trees in a mon-

tane forest (ordinate) as a function of their size (abscissa: diameterat base in centimeters).

needles along the height of three different-sized trees growingwithin 1 In of each other. In the largest and tallest tree, allside branches over 6.5 m from the ground were alive; therewas one live branch at 5.3 m, while all other branches below6.5 m, in deep shade, were dead. Of a 9.5-cm diameter tree,all branches below 4 m were dead, from 4 to 5 m 70% of thebranches were alive, and in its upper 2 m all branches were

alive with green needles, even though they were stronglyshaded most of the time by the surrounding taller trees. Atree with a 5-cm diameter trunk was only 3.9 m tall, and en-

tirely shaded by all surrounding trees. Yet, in its upper 0.4-mtop all branches were alive, with living branches to within 1m from the ground. Trees growing outside the forest in themeadow had only live branches, down to ground level. Thiswas the typical behavior of all trees in this forest: the youngestand smallest ones, of which almost half were alive, had livingbranches far below the level where the branches of older,taller trees were dead, not on account of age (because theselower branches were all alive in free-standing trees) but be-cause of intratree competition. And, where through intratree

14

12

10

8

6

2

A M°° ° B100

C ° D

FIG. 2. Percentage of living primary lateral branches of Pinusmurrayana (abscissa) as a function of height of insertion on tree(ordinate, in meters). Trees A, B, and C were growing within 1 m

of each other, tree D was growing in an open meadow. Stemdiameters: 5.0 cm in A, 9.5 cm in B, 30 cm in C, and 40 cm in D.

competition all branches below 4- to 5-m height were killed,the branches of young trees were alive. Therefore, intertreecompetition is weaker or less effective than intratree com-petition. It was also interesting to note that all dead treesover 5 m in height had produced cones, still attached to theirdead branches. So, even those trees that had succumbed at afairly early age had still had a chance to reproduce.

In this case the number of seedlings that develops in theestablished forest is small, and only after good seed years afew Pinus murrayana seeds germinate. Their seedlings havethe ability to grow at a very low light intensity, perhaps be-cause they are partially fed by mycorrhiza, which are able tosupply not only minerals but also organic food to their roots.This feeding they accomplish in a manner similar to that bywhich the mycorrhizal fungi feed phanerogamic saprophytes.This food is derived from the decomposition of forest litter,which is available in abundance in the Pinus murrayanaforest and which greatly stimulates pine seedling growth(18). A fair number of these seedlings grow into young trees,which are able to develop even in very deep shade, as de-scribed earlier, and where the similar death rate of trees of allsizes (Fig. 1) suggests that little intertree competition occurs.

In several tropical forests, where according to commonconcensus the fiercest interplant competition occurs, the seed-ling situation is similar to that in the Pinus murrayana forest.In rain forests, seedlings grow very slowly in the deep shade;after 5-10 years they may be still only a few decimeters tallwith very few leaves, far too small to compete with each other.In a dark forest in the Central Amazon basin near Manaus,with hardly any herbs as undergrowth (0.5/M2), there were10.2 5- to 10-year-old tree seedlings per M2, equally distrib-uted over classes with 1, 2, 3, 4, 5, 6, 7, and 8 leaves, withonly 1.0 dead seedling per M2. A similar situation was foundin a forest near the mouth of the Amazon near Belem (Table4), with on the average only five seedlings of the 5- to 10-year-old class per M2. Similar numbers were found near Borbaalong the Madeira River and below Tingo Maria along theHuallaga River, both tributaries to the Amazon.With my experience in a mountain rain forest in Java (16)

and observations in forests in Trinidad and Costa Rica, I canstate positively that in tropical rain forests normally only afew tree seedlings germinate per year per M2, and that manyof these seedlings survive for a long time, with hardly anyfoliage, in deep shade, probably partially fed by mycorrhiza,for all of them have an extensive superficial root system in theforest litter. The only two exceptions to this rule I have seenwere (i) in a Javanese mountain forest, where thousands of

TABLE 4. Number of trees and tree seedlings in PirelliForest near Belem*

Light areas Dark areasTrees (4 plots) (2 plots)

Seedlings with cotyledons only 8.7 3.0Seedlings less than 20 cm 51.7 37.5Seedlings 20-80 cm 30.5 16.0Seedlings 100-200 cm 7.2 5.0Trees over 200-cm tall 10.5 9.5Dead plants (% of live) 3.2 (3.0%) 6.0 (8.3%)

* Each plot is 10 M2.

588 Went

41

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Proc. Nat. Acad. Sci. USA 70 (1973)

Cestrum aurantiacum seedlings covered the forest floor (butCestrum was the only introduced tree in that forest) and (ii)a group of 12 closely-spaced unidentified seedlings of the samespecies in the Mocambo forest near Belem, of which three haddied (through competition?). These were the only cases inwhich I found seedlings of the same species adjacent to eachother, and the only cases of extensive death. Otherwise, Ifound in the above-mentioned Mocambo forest only one totwo dead plants per 100 for those less than 30 cm in diameter,and six dead per 100 trees of over 50-cm diameter. When onetakes into account that big dead tree stumps last much longerthan small ones, this finding would suggest a similar deathrate for old as for young trees.

Table 4 suggests that in the Pirelli forest near Belem thereis a higher death rate of seedlings in the early years, but still amajority survives to provide a pool from which a few grow upto tall trees when an old shading tree dies and provides achance for fast growth of the suppressed undergrowth. Thispattern was confirmed in the Tjibodas forest in Java, wheremany trees 10 cm or less in diameter had not grown at all in40 years, whereas the big trees of 50 cm and more in diametereither had died or had grown much in the 40 years since Koor-ders had measured and numbered those trees.Summarizing these observations in tropical rain forests, we

can say:(i) Germination of most tropical tree seeds is a rare phe-

nomenon. Most of these seeds have a very short life span, andthe very rapid fungal decomposition of organic matter intropical soils prevents the build-up of seed reserves in the soil.

(ii) The few seedlings that come up (a few per m2 per year)grow very slowly, even those with very large seeds.

(iii) In the deep shadow of the forest, where so little photo-synthesis is possible that very few herbs will grow there, mosttree seedlings have only a few leaves, although measurementshave shown that they photosynthesize slowly. Probably theirprincipal food source comes from mycorrhiza.

(iv) Survival of these tree seedlings is remarkably high, witha half-life of 10-20 years. Their death is not attributable tocompetition, but to attrition by predators, diseases, and me-chanical injury (dropping branches, e.g.).

(v) Suppression by their physical environment, rather thancompetition by the surrounding vegetation, is the term to beused in connection with the development of these tree seed-lings. Thus, none has an advantage over its neighbors and nodifferential survival exists.

(vi) The same situation continues during the later growth ofthe trees. They still are suppressed by the forest giants shad-ing them, and all 5- to 10-m tall treelets have an even chanceto take over the role of principal tree when a shading tree dies.

(vii) The natural attrition, resulting in a few percent deadtrees of all ages, apparently does not favor or disfavor anyspecies; this lack of competition explains the great abundanceof tree species in the rain forest. The main characteristic allthese rain forest trees have in common is the ability to growvery slowly under suppressed conditions. This slow growth inyouth makes most of them unsuitable for forestry practice.

(viii) We find completely opposite behavior in trees of thesecondary forest, such as Cecropia, Triplaris, and Ochroma,which germinate in abundance in a clearing and which growvery rapidly, 5 to 10 m/year.We have now considered several cases in which competition

occurs among plants. In one set of cases, where one plant

excretes or exudes inhibitory substances that reduce or pre-vent growth of other plants in their neighborhood-so-calledallelopathic substances-the term competition, in the sense ofstrife against one another, ultimately resulting in the ousterof one, seems appropriate. It is remarkable, however, thatsuch inhibition can be quite specific, and that in several cases,specifically among desert plants such as Larrea, the strongestinhibition is exerted on individuals of its own species. In thosecases it seems appropriate to consider the allelopathic sub-stances not as means to cut down competition, but to preventexcessive growth under temporarily favorable conditions, suchas after a heavy rain. If desert shrubs immediately respondedto such rain with fast growth, they might very well overgrowtheir capacity to survive a subsequent drought period, forwhich they should be prepared with a low shoot to root ratio.A second set of cases was discussed in which dense planting

causes mutual inhibition of adjoining individuals. In this casewe can use the term competition in the sense of a scramble forthe same commodity, such as light, nutrients, or water. If theplants grow so close together that they shade each other, theycompete for light and can only grow to the extent that light isavailable. This type of competition has no evolutionary sig-nificance, since all competing individuals are equally inhibited,while the total plant mass and seed produced per unit surfacetends to reach a maximum as the individuals grow closertogether. The mechanism of this type of competition betweenindividuals of the same species is probably hormonal, with alimiting amount of, for example, leaf growth hormone pro-duced per unit leaf or ground surface. It is likely that growthfactors produced in the root system play an important part inthis growth control.A third set of cases involves the development of desert an-

nuals. Here, no competition is found at all in the early stagesof germination and growth. Under extreme conditions, such asin the center of Death Valley, later growth of the annualsseems to be limited by the small amount of precipitation, yetseed production offsets the attrition of the seedlings and youngplants due to drought, and a seed germination to seed produc-tion ratio close to 1:1 occurs. Everywhere else there is a 10-to 20-fold increase in seed production over the number ofseeds that germinates per germination event. The excess seedis removed by ants and rodents. Thus, the equilibrium thatwe find in the desert between the seeds produced and seedsused in propagation is controlled by predation of animals, andno competition in the sense of differential survival or survivalof the fittest exists. The vegetation that develops after a rainis strictly the result of germination of the seeds left by rodentsand ants.

In a fourth set of cases, the development of trees in a coniferforest and in tropical rain forests was considered. It turns outthat intraplant competition is an important factor in the de-velopment of conifer trees, whereas interplant competition isof secondary importance. In tropical rainforests, there doesnot seem to be any competition between plants, at least be-tween seedlings of the forest trees. They are suppressed bythe lack of light due to the dense canopy, but there is no obvi-ous struggle between seedlings or between young trees.

It would be possible to discuss dozens of other cases whereeither there is or is not interaction within or between plantsthat might come under the heading of competition. The gen-eral conclusion about plant competition would still remain

Competition Among Plants 589

the same.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Proc. Nat. Acad. Sci. USA 70 (1973)

There definitely are cases where competition in the senseof strife exists among plants, and where this competition isa factor in survival, and where allelopathy is a factor in evolu-tion. We do get elimination of such forms that are less effec-tive in their struggle for existence. Yet, the successful survivalof so many primitive plants such as algae or conifers makes uswonder about the general effectiveness of competition inevolution. Besides, only competition between organisms thathave not reached the 100% level of adaptedness could resultin differential survival, and it is remarkable in how few genera-tions the 100% adaptedness can be reached in breeding experi-ments [examples: sugar level in beets and sugar cane, growthrates in tomato plants (19) ]. If selection through competitionwere the all-important factor in evolution, we could expectevolution to stop at some time in the future.But we have also seen that competition plays no, or only a

minor, role in environments that are still in full evolution,such as in a tropical rain forest. Therefore, my general con-clusion is that competition is an overrated factor in the plantworld. It exists, but does not have the overriding importancethat is imputed to it in the animal kingdom, where activecompetition and elimination of competitors is possible. Evenamong zoologists the importance of competition in evolutionhas been questioned (15).

Much of the observational work reported was supported byGrants GB 6681 and GB 17731X of the National Science Founda-tion.

1. Warming, E. (1896) Lehrbuch der okologischen Pflanzengeo-graphie (Fischer, Jena).

2. Clements, F. E., Weaver, J. E. & Hanson, H. C. (1929)"Plant competition an analysis of community functions,"Carnegie Inst. Wash. Publ. 398.

3. Juhren, M., Went, F. W. & Phillips, E. (1956) "Ecology ofdesert plants. IV. Combined field and laboratory work on

germination of annuals in the Joshua Tree National Monu-ment, California," Ecology 37, 318-330.

4. Beatley, J. C. (1967) "Survival of winter annuals in theNorthern Mojave desert," Ecology 48, 745-750.

5. Went, F. W. (1969) "A long term test of seed longevity,"Aliso 7, 1-12.

6. Tevis, Jr., L. (1958) "Interrelations between the harvesterant Veromessor pergandei (Mayr) and some desert ephemer-als," Ecology 39, 695-704.

7. Went, F. W., Wheeler, J. & Wheeler, G. C. (1972) "Feedingand digestion in some ants," BioScience 22, 82-88.

8. Kooper, W. J. C. (1927) "Sociological & ecological studies onthe tropical weed vegetation of Pasuruan (the island ofJava)," Rec. Trav. Bot. Neer. 24, 1-255.

9. Darwin, C. (1859) The Origin of Species, from Reprint of6th ed.

10. Molisch, H. (1937) Der Eingfluss einer Pflanze auf die andere-Allelopathie (Fischer, Jena).

11. Grummer, G. (1955) Die gegenseitige Beeinflussung hohererPflanzen- Allelopathie (Fischer, Jena).

12. Evenari, M. (1961) "Chemical influences of other plants(allelopathy)," Encycl. Plant Physiol. XVI, 691-736.

13. Whittaker, R. H. (1970) "The biochemical ecology of higherplants," in Sondheimer and Simeone eds., Chemical Ecology(Academic Press, New York), pp. 43-70.

14. Muller, C. H. (1966) "The role of chemical inhibition(allelopathy) in vegetational composition," Bull. TorreyBot. Club 93, 332-351.

15. Kropotkin, P. A. (1917) Mutual Aid; A Factor of Evolution,(Knopf, New York).

16. Went, F. W. (1940) "Soziologie der Epiphyten eines tropis-chen Urwaldes," Ann. Jard. Bot. Buitenzorg 50, 1-98

17. Went, F. W. (1942) "The dependence of certain annualplants on shrubs in southern California deserts," Bull.Torrey Bot. Club 69, 100-114.

18. Went, F. W. (1973) Effect of soil litter and Mycorrhiza ongrowth of Pinus murrayana, in press.

19. Went, F. W. (1957) "The experimental control of plantgrowth," Chron. Bot. 17, XVII + 343.

590 Went

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1