Plant responses induced by herbivores

TF’EE vol. 3, no. 2, February

1988

the unsatisfactory maximum parsim3ny criterion will have to suffice. The techniques of cladistic analysir; of morphological data and molecular phylogeny reconstruction have now been applied to the M,ammalia for over ten years and the results have not lived up to the expectations of either group of ar alysts. However, broad assessments of the few areas of agreement and those of disagreement are necessary9f18 before any resolution can be achieved. It would be foolish to assume that one method is more likely to achieve success than the other.

Acknowledgements thank Michael Novacek, Malcolm McKenna and Andre Wyss for permission to use

Physiological and chemical traits of many plblnt species change in response to real or sinzulated herbivory. These changes often have significant impacts on behavior, growth, survivorship, feeding and oviposition of insects. However, evidence that plernts gain direct or indirect protection from insect enemies thereby is equivocal at pn!sent. Evidenceis lacking for an impact of induced defenses on insect population dynamics, 6ut few studies have sought it. M’,re detailed studies of plant physiology, biochemistry, genetics and net benefit to individual plants ure needed to identify the adaptive significance of induced defenses. Plants have long been known to respond to pathogen attack by changing cell or tissue structure or b;/ producing chemical deterrents’. The observation that similar changes might occur in response to herbivore attack is more recent2. Siudy of these changes, called ‘induced defenses’ because of their presumed protective effects’, has greatly accelerated in the past ten years. At least 50 examples of changes in physiological or morphological traits, or in bioassay results related to real or simulated herbivory have been published since 1974, involving well over 40

lack C. Schultz is at the Pesticide Research Labora-

tory,Dept

of Entomology,

Pennsylvania State Uni-

versity. University Park, PA 16802, USA.

0

988

Elsevter

P~b,‘cat.oni

Limbridge

0169-5347

88’$02

data

from an unpublished

paper

of theirs

(Ref. 91.

References

I Gregory, W.K. (1910) Bull. Am. Mus. Nat. Hist. 27, l-524 2 Matthew, W.D. 119451 Trans. Am. Philos. sot. 30, I-510 3 Simpson, C.G. ( 19451 Bull. Am. Mus. Nat. Hist. 85, I-350 4 Simpson, G.G. ( 1978) Proc. Am. Philos. Sot. 122,318-328 5 Gingerich, P.D. ( 19861 MO/. Bio/. Evol. 3, 205-22 I 6 McKenna, M.C. ( 1975) in Phylogeny of the Primates ILuckett, W.R. and Szalay, F.S., edsl, pp. 21-46, Plenum 7 Novacek, M.I. ( 19821 in Macromolecular Sequences in Systematic and Evolutionary Biology (Goodman, M.,ed.), pp. 59-8 I, Plenum 8 Shoshani, I.119861 Mol. Biol. Evol. 3, 222-242 9 Novacek, M.1.. Wyss, A.R. and McKenna, M.C. I 1988) in Phylogeny of the Tetrapods

(Vol. 2) (Benton, M.I., ed.), Oxford University Press (in press) IO Wood, A.E. (1957) Evolution I I, 417-425 I I Fischer, MS. ( 1986) Cour. Forschungsinst. Senckenb. 84, l-l 29 I2 Prothero. D.R., Manning, E. and Fischer, MS. in Phylogeny of the Tetrapods (Vol. 21 (Benton, M.I., ed.l, Oxford University Press (in press) I3 Romero-Herrera. A.E., Lehmann, H., loysey, K.A. and Friday, A.E. I I9731 Nature 246,389-395 I4 Goodman, M., Czelusniak, I. and Beeber, I.E.ll9851Cladistics I, 171-185 I5 Shoshani, I., Goodman, M., Czelusniak, 1. and Braunitzer, G. II9851 in Evolutionary Relationships among Rodents (Luckett, W.P. and Hartenberger, I-L., edsl, pp. I9 l-2 IO, Plenum lb Miyamoto, M.M. and Goodman, M. ( 19861 Syst. 2001.35, 230-240 17 McKenna. M.C. II9871 in Moleculesand Morphologyin Evolution (Patterson, C., ed.1. pp. 53-93, Cambridge University Press I8 Wyss, A.R., Novacek, M.I. and McKenna, M.C. 11987) MO/. Bio/. Evol. 4.99-l Ih

PlantResponsesInducedby Herbivores Jack C, Schultz plant species. Various aspects of the subject have been reviewed in the last four years’-4. Because plants possess many passive characters which may protect them against herbivores, responses to herbivore stimuli must be distinguished from ‘constitutive’ traitslJf5 which may change in response to other external or internal influences (e.g. weather, tissue age) but are otherwise stable. For example, leaf phenolic composition and concentrations in Acer saccharum vary seasonally, with leaf age and position, among individuals, among sites, and through timeb, but also change in response to pathogen infection7 and artificial defoliatior?. Only the last variation can be said properly to be ‘induced’ by damage. Induced responses are therefore a form of phenotypic plasticity expressed in response to stimuli from herbivores. Responses are very diverse. Changes in both current growth and regrowth following damage have been called ‘induced responses”-4. In some cases, regrowth tissues differ from those actually damaged by 00

herbivores only because they are youngerg. In these cases, it is doubtful that the plant ‘responds’ to herbivore stimuli by altering its traits; damage merely alters the frequencies of tissues that ordinarily differ in constitutive traits. If the herbivore can only use older tissues, such a plant could become more ‘resistant“? however, the same plant may become more susceptible to a herbivore preferring younger tissues. Similarly, the production of younger tissues out of their normal phenological schedules (e.g. young tree leaves late in the growing season) could reduce or increase plant susceptibility to some herbivores within a single generation’,‘. Herbivore attack may also shift phenology of perennial plants in subsequent years’,2. For herbivores restricted to feeding on tissues of a specific age, this may desynchronize herbivore and plant life histories, making the plants more resistantlO. This is one way that feeding by one herbivore generation could have an impact on performance of the next. Responses like these, which act over more than

TREE vol. 3, no. 2, February

one herbivore generation or plant growing season, have been called ‘long-term’r,2. Those occurring in a single season and decaying more rapidly are called ‘rapid’ or ‘shortterm’rf2. In addition to phenological shifts, inducible plant traits can include production of spines, early leaf abscission, increased resin exudation, and a decrease in nutrients such as nitrogeniJ. The most frequently studied and best documented class of responses is including inphysico-chemical, creased concentrations of fiber and secondary chemical compounds phenolics, alkaloids, such as specific proteinase terpenes,

m

16

c t:

_

;

E 8

.

P=OOO5

j 20

10

1

30

I

40

50

60

!I$ :& 0” 3

< .*. 70-

.

B

$

y= 75 3 +O 42X r- to43 P= 0 004

8

??

.

60 10

20

30

TREE -WIDE

40

50

DEFOLIATION

60

inhibitorsi-4 and silicates’ I. Limited investigations indicate that increases in some secondary compounds (e.g. coumarins’21 cannot be stimulated by herbivory. The spatial and temporal extent of chemical changes arising from damage can vary widely. Significant changes in tissue quality have been detected chemically or with bioassays within leaves13, within branchesI or throughout whole plants and tree canopiesr-4. These changes are detectable within a few hours, days, or weeksr-4,r3,r4, and last a few hours, days, weeks or yearsr-4 (Fig. I). Maintaining some responses requires continuous or repeated damager,3,5p7,8.Young tissues and/or plants are often found to be more responsive to damage’,*, perhaps because of their greater photosynthetic or metabolic activities. Whatever the mechanisms involved, the ability to respond to attack in some way appears to be widespread among plant taxa. Although there is a lamentable tendency to avoid publishing ‘failures’, I am aware of only three experiments in which investigators failed to find some sort of damageinduced change in chemistryr5-r7. In two of these studiesi5,i7 bioassays used branches removed from the host tree (Pinus sy/vestrid5; Ahus rubra’71; this may have blocked plant responses. In the third, using five arctic shrub species, experimenters cut and removed whole branches to simulate herbivory. This, too, may prevent plant responses. Although most positive examples involve woody plants, especially trees, this may be a sampling artifact; herbaceous species in the Labiatae, Solanaceae, Cucurbitaceae and Graminae among others exhibit increases in secondary compounds when leaves are damaged’-4 and species in several other families exhibit damage-induced resistance’,2,4.18.

(%I

Fig. I. Three phenolic measures as functions of manual defoliation of 45 red oak (Ouercus rubral trees. Regression equations, data points and solid lines are for data from foliage inside bags with gypsy moth (Lymantria disparl larvae; dashed lines are for data from leaves randomly sampled from the open canopy. Differences between the two regression lines suggest that larval presence or immediacy of damage or both are needed for maximum response. % TAE = % tannic acid Istandard equivalents). From Ref 5. with permission.

Are plant responses‘incidental’? So little is known (especially by ecologists) of the regulation of plant development and biosynthesis of secondary compounds, that it is reasonable to ask whether ‘induced’ responses represent anything other than an inevitable and incidental consequence of tissue

1988

lossr9. This issue is important because it is a variant of the question, ‘Are these plant responses evolved (or coevolved) adaptive traits?‘. Plant responses to pathogen invasion often involve the de novo synthesis and translocation of antibiotic chemicals (‘phytoalexins’) whose actions can be specific to pathogen genotype and have a plant genotypic basis’. Evidence of this specificity, which suggests that phytoalexin production is an adaptive trait arising primarily due to selection by pathogens, is so far lacking for responses by induced herbivory. However, several reports indicate that leaf damage produced by herbivores is more effective in stimulating plant responses than is artificial (e.g. tearing) damage5r19-21 (Fig. I). The possibility that pathogens introduced during damage may actually provide such cues4fr9 has not been investigated rigorously. Since herbivory may remove metabolic substrates needed to produce constitutive defenses, ‘induced’ responses could be driven primarily by shifts in resource availability22. As an example, the production of secondary chemicals whose molecular structure is based primarily on carbon could depend on the existence of a metabolic ‘surplus’ of carbon in the plant’s biosynthetic pathwayG2. Such a ‘surplus’ might develop in a defoliated plant, such as a tree, which has large stored carbon reserves Carbon and some mineral nutrients (e.g. nitrogen) are lost when leaves are removed; however, carbon can be mobilized from storage while nutrients cannot be obtained readily until the canopy is replaced22. A resulting carbon surplus might be directed into the synthesis of carbon-demanding secondary chemicals. However, some plants possess specific damage-recognition mechanisms and the ability to regulate the synthesis of secondary metabolites directly on the basis of such stimulils2,23. Undamaged individuals of some species appear capable of responding to volatile substances produced by damaged neighbors, increasing or initiating synthesis of secondary compounds”J4, and may thus become more resistant to some insects’.

TF’EE vol. 3, no. 2, February

1988

Many plant species exhibit systemic responses to removal of minute amounts of tissueI-3.8.18J3 and ca-bon-based defenses can be induced in annual plants which do not have large stored reservesi,2. These observations are inconsistent with the nutrient allocation hypothesis22. To the extent that we understand them, the biosynthetic pathways of many plant secondary compounds appear to involve both substrate availability and specific cue regulation. Hence the activities of many key regulatory enzymes in the synthesis of a wide variety of phenolic and other secondary chemicals frequently shown to be induced, art: known to be sensitive to both resource availability and specific cuSes. A recent study25 showed that genes coding for the biosynthetic enzymes central to production of many commonly induced phenolics are regulated by wound-produced ethylene. In the only direct test of the ‘defensive-incidental’ alternatives fertilization did not prevent an induction response in birch trees (8etula pubescensl, as predicted resource availability the by theorylg. Indeed, fertilizing this same plant species can apparently cause induction, especially when insect frass is used as fertilizer’~*. Some ‘induced’ responses are merely normal physiological phenumena stimulated by damage and clearly are best described as ‘incidental’ consequences of herbiinclude vore attack. Examples herbivore impacts on plant architecture26, osmotically regulated leaf abscission27, compensatoly growthz8, and increased photosynthesis as the plant canopy is thinned29. Whether these changes should even be called ‘induced defenses’ is questionable. The adaptive nature of induced re$;ponses cannot be determined without understanding in greater detail how biosynthesis of induced traits is regulated. In addition, studies of induction need to be better with plant integrated basic physiology. Carbon:nutrient ratios are! likely to vary as a function of photosynthetic rates, which can be influenced by nutrient availability and tissue damage. These relationships are probably not identi-

cal for all plant species, and must be ascertained for more cases before valid generalizations can emerge. The heritability of responses has not yet been studied. Of course, these hypotheses&e not mutually exclusive. There is no reason why natural selection cannot preserve an incidental by-product of plant organization which confers superior fitness in some secondary way. Are inducedresponses defensive? ‘Defensive’ can be defined in at least two different ways. In the broad sense, any negative impact on a plant’s enemies could be described as defensive. More strictly, a plant can really only be ‘defended’ by a trait if that trait actually reduces net losses in terms of tissues or fitness. By the first of these definitions, increased defense has been demonstrated frequently, mainly as increased host plant ‘resistance’ in bioassays. In these cases, ‘resistance’ has been quantified as reduced herbivore performance (growth rates, final weights, population densities), not plant gains (Fig. 2). As described above, the mechanisms by which performance is altered are rarely elucidated.

A

B

. I

I

I

IO

20

1

30

TREE-WIDE

7 1 ~.$23-70.007x

.

.

P= 1

0.016 I

I

40

60

DEFOLIATION

60 1%)

Fig. 2. Mean female gypsy moth (Lymantria dispar) pupal weight as a function of defoliation on each of 45 red oak (Ouercos rubral trees. Mean pupal weights for each tree were weighted by the number of pupae obtained on that tree (n = 451. Defoliation depresses pupal weight and fecundity (which are linearly related51 approximately 30% below the insect population mean. Residual variation is due in part to genetic variation in the insects, maternal food quality, and constitutive leaf quality variation (Ref. 5; Rossiter and From Ref. 5, with permission Schultz, unpublishedl.

Correlative and experimental results identify leaf and wood phenolics, tannins and terpenoids as widespread resistance mechanisms. There is considerable debate about the impact of induced defenses on population dynamics of herbivores, perhaps because some authors have emphasized the superficial similarity between postulated plant responses and models of density-dependent population regulationl~2,30. It has been particularly tempting to suggest that cyclic

C

D

A

B

PLANT DEFENSE

DAMAGE TIME Fig. 3. Hypothetical interactions between induced plant defenses and insect population dynamics. At ‘A’, increasing attack (an ‘outbreak’) stimulates plant responses in density-dependent fashion. Peak responses begin to have a negative impact at ‘B’: growth slows, fecundity declines, and the insect population peaks and begins to decline. At ‘C’, insect populations have declined to the point of possible control by natural enemies, and plant responses wane. At ‘D’, improved host plant quality results in insect population growth beyond control of natural enemies, and a second outbreak begins.

47

TREE vol. 3, no. 2, February

The gypsy moth, Lymantria dispar, is one of few defoliating insects shown to induce a chemical change in its host plant (mainly red oak, Ouercus rubral which can have a negative impact on its performance and possibly on its population dynamics.

or irruptive herbivore population dynamics could be driven by plant responses to attackrJ*30 (Fig. 3). The germane evidence is of two types. First, many aspects of herbivore performance influenced by induced responses (e.g. growth rate, insect pupal weight, fecundity, insect egg weight, mortality, susceptibility to enemies) clearly can influence population dynamics. However, there has been no direct attempt to compare these effects with other factors that could be just as influentia14. Making a series of assumptions that are largely unverified experimentally, Fischlin and Baltensweiler30 have provided a model that simulates larch budmoth population dynamics solely as a function of host plant responses. However, the magnitude of induction’s impact on fecundity of most insect species studied thus farr-5 may not be sufficient to drive population dynamics without a concomitant impact on mortality. Secondly, population sizes of some herbivores can be reduced by host plant resistance acquired via damage. All of these positive examples involve insectsr,r8, and none identifies the resistance mechanism. Other studies have failed to find significant population level impacts of induced plant traits in vertebratessre32. It is important to remember that selection ought to act most directly on individual plants. As a result, adaptive advantage and evolution of

48

induced responses may occur without noticeable effects on herbivore population dynamics, as long as plants benefit in evolutionary time. By the second definition of ‘defensive’, there is little evidence that plant losses are actually prevented by induced responses. The best exception is provided by terpene deinsect-stimulated fenses in conifers; tree protection clearly results from this response33. If we include in the loss calculation the potential costs of producing the it is safe to say that response, there is as yet no evidence. In the one published field study of the of induction for consequences plant growthr8, induced resistance reduced herbivore (shown via populations) produced no gain or preservation of plant yield. Obtaining evidence of net gain is complicated by the difficulty of measuring potential negative costs of producing induced responses (e.g. secondary chemicals), a lack of data on the impact of defoliation on plant fitness, and the necessity of integrating losses and benefits over many years for long-lived plant species. Herbivore attack can also reduce host plant resistance. In some cases the herbivores may manipulate host quality biochemically’, while in others the intensity of damage appears to overwhelm the plant’s ability to respond14*3?*34. It would be naive to expect host plant re-

1988

sponses to be omnipotent or inexhaustible (e.g. Ref. 151, to expecl every herbivore to be susceptible to them4, or to ignore likely intraspecific variation in responsiveness4. For example, any plant response that depends on photosynthesis may become impossible beyond critical defoliation levels may vary geographically and with soil quality, and could differ among genotypes. Because many induced plant responses (e.g. phenolics) may have negative impacts on organisms other than herbivores, these effects must be considered in determining net benefit to the plant. Tannins which are among the most frequently studied inducible plant traits, have significant antibiotic insect and plant impacts on pathogens35pj6. It is conceivable that induced increases in these ot other compounds could ( 1) benefit herbivores by reducing their mortality; (21 benefit the plant by reducing its pathogen load; and/or (31 merely regulate regrowth37. These interactions are virtually unstudied yet they could be very important; for example, perhaps induced defenses contribute to forest insect outbreaks by releasing insects from regulation by pathogens, We may err in seeking induction’s major influence in ecological time. The greater significance oi these variations in apparent defense levels may be evolutionary They resemble in many ways the tactics used to slow the evolution 01 pesticide resistance in agricultural pests. in managed systems, withholding pesticides until some ‘economic threshold’ is reached slows the evolution of resistance. Inducibility may represent analogous means of slowing herbivore adaptation and prolonging the utility of plant defenses. However, it is premature to conclude that induced responses have no net positive benefit to the plants producing them (e.g. Ref. 4) Until studies are carried out to measure physiological costs of defense and the negative impact of defoliation, and integrate plant responses with other factors that influence plant susceptibility (e.g enemies of the herbivores, plant stress), we will be unable to reach a conclusion about the net benefit of induced responses.

TRPE vol. 3, no. 2, February

1988

Ecological and experimental implications There are several important consequences arising from the above observations. It is clear that plants display a wide variety of physical and chemical responses to herbivo*e attack, and that these responses can have biologically meaningful impacts on herbivores and other organisms. It is much less clear whether these responses protect plants in either the short or long term, and whether they can influence herbivore population dynamics. The first of these uncertainties can be resolved by careful experimentation. Resolving the second, however, will be problematic because of the complexity, difficulty and cost of carrying out studies which are simultaneously controlled to identify the contribution of induced responses and to account for other influences on herbivore populations. Because the proposed population interactions demand a significant plant defense ‘relaxation’ period’JJ0, a population study in forests would require a decade of continuous study (and funding). ‘JIhether or not herbivore population dynamics are influen~:ed by induced responses, it is very likely that many interactions among taxa as disparate as vertebrates and fungi are mediated by them. Early-feeding herbivores may influence food quality or availability for later-feeding species, during a single season or over a period of yearsl.2.21. Densities and spacing of conspecific individuals on a host plant could depend on the frequency and occurrence of uninduced tissues, in turn influencing density-dependent factors (e.g. predation and parasitism). We are only beginning to explore these complex interactions. If benefit to the induced plant can be demonstrated, induced resistance offers potential for biological control of pests. There is some evidence of hormonal regulation of biochemical induction (LT. Ba,‘dwin, pers. commun.). Perhaps plaint hormones, herbivore produds, volatile plant cues, or mechanical damage could be used to ‘immunize’ crop plants. Inducible changes in plant tissue quality must be accounted for in all studies of plant chemistry, herbivore feeding, plant physiology, etc.

Plant tissues must be sampled in such a way as to avoid causing these changes, and the possibility of pheromonal interactions makes it important to separate experimental treatments spatially. The quality of a herbivore’s food can be judged only by sampling tissues at the time of consumption; constitutive plant traits and postdamage samples provide an incomplete characterization of host food quality5. Finally, it is clear that the interaction between plants and their herbivores is more complex, dynamic and biochemically sophisticated than previously suspected. We sorely need a better integration of plant and animal physiology, animal behavior and biochemistry if we are to answer any of the significant and fascinating questions that arise from this realization. References

I Rhoades, D.F. ( 1983) in Variable Plants and Herbivores in Natural and Managed Systems (Denno. R.F.and McClure, M.S., edsl. pp. 155-220, Academic Press 2 Haukioja, E. and Neuvonen, S. in insect Outbreaks: Ecological and Evolutionary Perspectives (Barbosa, P. and Schultz, l.c., eds), Academic Press (in press) 3 Edwards, P.1. and Wratten, S.D. ( 19851 Oikos 4470-74 4 Fowler, S.V. and Lawton, I.H. ( 1985) Am. Nat 126. 181-195 5 Rossiter, M.C., Schultz, J.C. and Baldwin, I.T. Ecology (in press) 6 Schultz, l.c., Baldwin, I.T. and Nothnagle, PJ. ( I98 I I Am. 1. Bat. 69, 753-759 7 Baldwin, I.T. and Schultz, 1.C. ( 1987) 1. Chem. Ecol. I 3, 1069-l 078 8 Baldwin, I.T. and Schultz, j.C. ( 1983) Science 22 I, 277-302 9 Bryant, I.P. II981 1Science 213,889-890 IO Haukioja. E. and Niemela, P. (1977) Ann. Zoo/. Fenn. 1448-52

II McNaughton, 478-486

S.J. (1985) Oecologia

65,

I2 Berenbaum, M.R., Zangerl, AR. and Nitao, I.K. ( 1986) Evo/uGon 40, I2 15-l 228 I3 Bergelson, I., Fowler, S. and Hartley, S. (1986) Ecol. Entomol. I I, 241-250 14 Raupp, M.J. and Denno, R.F. I I9841 Ecol. Entomol. 9,443-448 I5 Niemela, P., Tuomi, I., Mannila, R. and Oiala, P. (1984) Z. Angew. Entomol. 98, 33-43 I6 Chapin, F.S., Bryant, I.P. and Fox, I.F. ( 19851 Oecologia 67,457-459 I7 Myers, I.H. and Williams. K.S. ( 19871 Oikos 48, 73-78 I8 Karban, R. (I9861 Entomol. Exp. Appl. 42, 239-242 19 Haukioja, E. and Neuvonen, S. ( 19851 Ecology 66, 1303-l 308 20 Capinera, I.L. and Roltsch, W.]. f 19801 1. Econ. En tomol. 7 I, 366-368 21 McNaughton, S.l.and Tarrants, J.L. II9831 Proc. Nat/ Acad. Sci. USA 80, 790-79 I 22 Tuomi, I., Niemela, P., Haukioja, E., Siren, S. and Neuvonen, S. (1984) Oecologia 61, 208-2 IO 23 Ryan, CA. ( 19791 in Herbivores: Their Interaction with SecondaT Plant (Rosenthal, G.A. and Janzen, Metabolites D.H., edsl, pp. 599-618. Academic Press 24 Kimmerer, T. Phytochemistry lin press) 25 Ecker, J.R. and Davis, R.W. (I9871 Proc. NatlAcad. Sci. USA 85,5202-5206 26 Craig, T.P., Price, P.W. and Itami. 1.K. II9861 Ecology 67,4 19-425 27 Cockfield, S.D. and Potter, D.A. I I9861 Oecologia 7 I, 4 l-46 28 McNaughton, S.I. ( 1985) Oikos 40,329-336 29 Heichel. G.H. and Turner, N.C. ( 19831 Oecologia 57, 14-19 30 Fischlin, A. and Baltensweiler, W. i 1979) Mitt. Schweiz. Entomol. Ges. 52, 273-289 31 Lindroth, R.L. and Batzli, G.O. 11986) I. Anim. Ecol. 55,43 l-449 32 lonasson, S., Bryant, I.P., Chapin, F.S. and Andersson, M. (1986) Am. Nat. I28,394-408 33 Raffa, K.F. and Berryman. A.A. 119831 Ecol. Monogr. 53,27-49 34 Danell. K.and Huss-Danell, K. (1985) Oikos 44,75-8 I 35 Taper, M.L. and Case, T.J. ( 19871 Oecologia 7 I, 254-26 I 36 Keating, S.T. and Yendol, W.G. Environ. Entomol. (in press) 37 Green, F.B.and Corcoran. M.R. (I9751 Plant Physiol. 56,80 l-806