Ecology, 69(4), 1988, pp. 1118-1127© 1988 by the Ecological Society of America

HERBIVORE-CAUSED GREENFALL IN THE SOUTHERN APPALACHIANS1

LANCE S. RisLEY2 AND D. A. CROSSLEY, JR.^Department of Entomology, University of Georgia, Athens, Georgia 30602 USA

Abstract. Freshly fallen green leaves (greenfall) were collected from plots on fourforested watersheds differing in aspect and treatment history at the Coweeta HydrologicLaboratory in southwestern North Carolina. Greenfall was categorized as "herbivore-caused" or "unexplained." In almost all cases herbivore-caused greenfall biomass wassignificantly greater than was unexplained greenfall. Herbivore-caused greenfall was sub-divided into "clipped" (caused by petiole-clipping caterpillars), "petiole-damaged" (causedprimarily by petiole borers), "mined" (primarily by microlepidoptera on hickories), and"orts" (discarded leaf fragments). Leaves from 25 species or species groups of deciduoustrees and woody vines (« 35 species) were represented in foliar litterfall collected from littertraps; all of these contributed to herbivore-caused greenfall.

Despite differences in aspect and stand age (10 yr, 23 yr, mature), patterns of greenfallwere similar among watersheds. Seasonal inputs of greenfall to the forest floor were <5%of total foliar production but were greater than combined inputs of insect fecal pellets andinsect body parts. Greenfall occurred continuously during the growing season and resultedfrom the activities of numerous species of insect herbivores. We suggest that greenfall is awidespread phenomenon that may be important as a resource for decomposers and thatit should be included in measures of herbivory.

Key words: ecosystems; feeding behavior; forests; greenfall; herbivory; insects; leaf miners; litterfall;nutrient cycling; role of herbivores.

INTRODUCTION

Insect herbivores are integral components of forestecosystems that establish an important link betweenthe forest canopy and the forest floor. Occasionally,during outbreaks, the herbivore-associated link be-comes the principal factor regulating primary produc-tion (Mattson and Addy 1975). More often this link ismaintained by chronic, nominal rates of herbivory andis characterized by low energy flow (Haukioja 1979).Even nominal herbivory may be important in shapingplant species composition by influencing competitiveinteractions between plants (Schowalter 1981, Lamb1985) and by subtly regulating ecosystem structure(Mattson and Addy 1975). The materials that link in-sect herbivores and decomposer communities includeinsect feces and body parts, nutrients (e.g., potassium)leached from damaged leaves, and damaged foliar ma-terial (leaves, orts). These materials are attacked readi-ly by decomposer organisms, which in turn acceleraterates of decomposition, thus speeding the rate of nu-trient cycling (Springett 1978, Kitchell et al. 1979, Sea-stedt and Crossley 1984).

Production of insect herbivore fecal pellets has beenmeasured in a number of studies (Carlisle et al. 1966,Gosz et al. 1972, Nielsen 1978, Ohmart et al. 1983,Risley 1986), and where herbivory was typically low,fecal pellet production represented <5% of total foliar

1 Manuscript received 27 July 1987; revised 6 November1987; accepted 17 December 1987.

2 Present address: Division of Pinelands Research, RutgersUniversity, Pinelands Field Station, P.O. Box 206, New Lis-bon, New Jersey 08064 USA.

production. Measurements of insect body parts thatoccur as litterfall are almost nonexistent. Ohmart et al.(1983) measured insect body parts as 0.2 to 0.3% offoliar litterfall in several Australian Eucalyptus forests.

Mechanical damage to leaves enhances leaching ofmobile elements from them (Tukey and Morgan 1963).Leaf injury caused by insect herbivores may thereforeincrease foliar susceptibility to leaching (Tukey 1970).Neodiprion sertifer (Geoff.) (Hymenoptera: Tenthridin-idae) was associated with losses of 134Cs in leachatefrom damaged red pine (Pinus resinosa) foliage in Con-necticut (Kimmins 1972). Seastedt et al. (1983) foundthat throughfall from untreated black locust (Robiniapseudoacacia) trees contained significantly greater con-centrations of K than that from locust trees sprayedwith an insecticide to reduce herbivore feeding. In con-trast, an outbreak of the California oakworm (Phry-ganidia californica Packard) (Lepidoptera: Notodon-tidae) did not result in increased leaching losses of Nand P from oak (Quercus spp.) foliage (Hollinger 1986).The study by Hollinger (1986) illustrated that timingof precipitation in relation to herbivory was an im-portant factor in leaching losses associated with chew-ing herbivores.

The influence of insect herbivory on seasonal leaffall is most often described for a single herbivore specieson its host plant. The literature is replete with examplesof sucking insects, gall formers, leaf miners, and ex-ternal leaf chewers that damage foliage sufficiently tocause premature leaf fall. The influence of chronic her-bivory by multiple species of herbivores on single ormultiple plant species has also been described (Reichle

August 1988 HERBIVORE-CAUSED GREENFALL 1119

TABLE 1. Selected study areas at the Coweeta Hydrologic Laboratory, United States Forest Service, Otto, North Carolina.Greenfall and litterfall were collected from middle elevations on each watershed.

Water-shed

27

13

18

AspectSouth-facingSouth-facing

East-facing

North-facing

Slope<%)6057

49

52

Area(ha)1259

16

13

Elevation(m)

709-1004722-1077

725-912

726-993

Treatment historyUndisturbed since 1923.Cattle grazing in lower cove 1941-1952.

and cable-logged in 1977.All woody vegetation cut in 1939. Recut

removed in either treatment.Undisturbed since 1923.

Commercial clearcut

in 1962; no products

and Crossley 1967, Thomas 1969, Chabot and Hicks1982, Fox and Morrow 1983, Ohmart et al. 1983), butrarely measured. Larsson and Tenow (1980) measuredneedle fall in a Scots pine (Pinus sylvestris) forest andreported a 1.5 kg/ha increase due to the feeding activ-ities of "endemic" (defined by Wallner 1987) insectherbivores. Under conditions of typicaly low herbiv-ory, premature leaf fall did not differ significantly be-tween untreated black locust trees and those treatedwith an insecticide (Seastedt et al. 1983). Generally,abiotic effects (e.g., wind, drought) have been consid-ered more important than herbivores as determinantsof early leaf fall (Gosz et al. 1972, Pook 1984, Ring etal. 1985, Hollinger 1986).

Previous work at the Coweeta Hydrologic Labora-tory revealed that percent leaf area removed fromchestnut oak and species of hickory decreased follow-ing relatively intense (up to 50% leaf area removed)herbivory early in the growing season (Risley 1983). Itwas hypothesized that damaged leaves were abscisedselectively. In 1982, L. S. Risley (personalobservations)found that there was a light but steady flux of green,herbivore-damaged leaves from the canopy to the for-est floor. Green litterfall (greenfall) was subsequentlymeasured for six deciduous tree species in an undis-turbed hardwood forest (Risley 1986). Inputs of green-fall to the forest floor were 1.3% of total foliar pro-duction for the sampled tree species and comparableto insect fecal pellet litterfall (2.3 g-m-^yr-'). Greenfallalso contained more K and P than did autumn leaves(L. S. Risley, personal observations). In addition, thehypothesis of premature abscission could not explainsatisfactorily the damaged condition of many leavesthat fell green. Insect herbivores seemed to play anactive rather than passive role in causing greenfall (e.g.,clipping leaves free). Because of the potential impor-tance of herbivore-caused greenfall to decomposers onthe forest floor, and to measurements of herbivory basedon canopy samples, a much broader study was initiatedin 1985 that addressed greenfall on several forestedwatersheds.

In this paper we attempt to answer several questionsregarding greenfall. What mechanisms are responsiblefor greenfall? What proportion is herbivore-caused andwhat insect species are responsible? Is greenfall con-fined to particular plant species? Is greenfall a wide-spread phenomenon that occurs in forests differing in

age and aspect? What are the seasonal patterns andannual production of herbivore-caused greenfall, andhow do these compare with total foliar litterfall andother components of herbivore-associated litterfall?

STUDY SITE

Study areas were located within the 1626-ha Co-weeta basin at the Coweeta Hydrologic Laboratory,United States Forest Service, Otto, North Carolina (35°N, 83° 25' W). Coweeta ranges in elevation from 675to 1592 m; the topography of the area is steep, andslopes average 50% at low to mid elevations (Swankand Crossley 1987). Natural vegetation is forest, pre-dominantly a mixture of deciduous oaks with under-growth consisting of evergreen rhododendron (Rho-dodendron maximum) and mountain laurel (Kalmialatifolia) (Day et al. 1987). Dominant tree species onlow elevation mesic slopes are chestnut oak, red maple,red oak, tulip poplar, and hickories, with floweringdogwood in the understory (Day and Monk 1974).

The four watersheds selected for this study differedin stand age and in aspect (Table 1). Watershed selec-tion was based on previous records of herbivory bychewing insects, topographic similarities, and differingtreatment histories. Selected watersheds had similarslopes, ranges of elevation, and vegetation.

METHODS

Greenfall

A logging road that crossed the middle of each wa-tershed was divided into 5-m sections, 10 of whichwere selected randomly for the establishment of green-fall plots. Each plot measured 5 x 5 m and was locatedalong the road several metres beyond obvious soil dis-turbance or vegetation boundary. Corners of plots weremarked by wooden stakes and sides were marked byplastic flagging at litter level.

At Coweeta, most greenfall occurs from May throughSeptember and collections of greenfall were made every2 wk from 13 May to 20 September 1985. On each ofnine dates all greenfallen leaves and discernible frag-ments of green leaves (orts) were removed from plots,placed into plastic bags, and refrigerated. Periodically,loss of color from freshly fallen green leaves was mon-itored near plots to determine how long these leavesremained green. Greenfall remained identifiable for 2

1120 LANCE S. RISLEY AND D. A. CROSSLEY, JR. Ecology, Vol. 69, No. 4

d on the forest floor (Risley 1986, this study), thusgreenfall from each collection date represented a 48-haccumulation in plots.

In the laboratory, collected greenfall was placed inplant presses and air dried. Dried material was weighedand then examined under a stereomicroscope; a de-scription of each leaf, including petiole and midrib,and any leaf fragments in each sample, was recorded.

Using means (n = 10) for greenfall on each wa-tershed, curves were constructed that described the re-lationship between sampling date and biomass. Mid-April and mid-October were designated as the first andlast dates of greenfall based on the phenology of de-ciduous woody species at Coweeta. Areas under thecurves were calculated using the trapezoidal methodof integration to estimate annual production of green-fall.

Litterfall

Total foliar litterfall, insect fecal pellets, insect bodyparts, and other nonfoliar litterfall were collected usinglitter traps constructed according to a standard designemployed at Coweeta. The interior of each trap mea-sured 62.2 x 66.0 cm (sample area of 0.41 m2), andtop edges of the wood sides were beveled to 45°. Tothe bottom of each trap, 1-mm mesh aluminum win-dow screen was attached. Traps were mounted level= 30-60 cm above the forest floor. Ten litter traps wereused on each watershed; all were located near greenfallplots.

Litter traps were emptied at monthly intervals fromJune through December 1985; one additional collec-tion was made 7 February 1986. Late autumn andwinter collections were necessary for deciduous treespecies (e.g., oaks) that, despite foliage senescence inOctober, abscised dead leaves over a period of severalmonths. The contents of each trap were emptied intoa paper bag and returned to the laboratory. Litter wasair dried; sorted into plant species, insect fecal pellets,insect body parts, and miscellaneous debris; andweighed. Estimates of the annual production of litterbiomass in each category were calculated.

RESULTS AND DISCUSSION

Seasonal dynamics of greenfall

Greenfall resulted from biotic and abiotic factors andoccurred in measurable quantities on all four wa-tersheds during the 1985 growing season. The undis-turbed, south-facing watershed (WS2) produced thegreatest absolute quantities of greenfall biomass; theleast was produced on the 10-yr-old clearcut watershed(WS7). The more mesic 23-yr-old clearcut (WS13) andundisturbed, north-facing (WS18) watersheds pro-duced intermediate amounts of greenfall. Greenfallbiomass was characterized by three distinct seasonalpeaks: in May, June, and August. The peak of greenfall

on 13 May reflected a dramatic premature fall of hick-ory leaflets. A severe thunderstorm increased greenfallon 18 June. Petiole-clipping caterpillars were respon-sible for the August peak.

Greenfall was categorized initially as either "herbi-vore-caused" or "unexplained." Leaves placed in theherbivore-caused category were characterized only bygross petiole damage, because damage to leaf bladesalone did not result in greenfall (Risley 1987). For eachcollection date-watershed combination (n = 36), meanbiomass of herbivore-caused greenfall and of unex-plained greenfall were compared (t test, alpha = .05,n= 10). In most instances herbivore-caused greenfallcontributed significantly greater inputs to the forestfloor than unexplained greenfall. In only 6 of the 36paired comparisons, there was more greenfall in theunexplained category. Three of the six comparisonswere associated with the thunderstorm described above;one of these resulted in the only case where unexplainedgreenfall contributed significantly more (P < .05) tototal greenfall than herbivore-caused. Two of the re-maining three comparisons not associated with thun-derstorms were related to the premature drop of per-simmon and sourwood leaves in moist coves. Theseleaves fell during August and contained large necroticblotches resembling damage caused by Phytophthora,a pathogen responsible for premature leaf abscission(Roberts and Boothroyd 1972).

Twenty-five "species" (we were unable to distinguishspecies within Betula, Carya, Quercus rubra group,Smilax, and Vitis) of deciduous trees and woody vineswere represented in herbivore-caused greenfall (Table2). Although it was observed in plots, greenfall was notquantified for shrubs (e.g., Rhododendron spp., Kalmialatifolia L.) or forbs. During 1985 all plant "species"contributing to total greenfall also contributed to her-bivore-caused greenfall. The correlation of greenfallproduction with foliar litter production on each wa-tershed was calculated (n = number of represented plant"species"). Associations were positive and highly sig-nificant (P < .01) on WS2 (r2 = 0.53), WS7 (r2 = 0.58),WS13 (r2 = 0.78), and WS18 (r2 = 0.64). No correlationwas found between greenfall/foliar litterfall and foliarlitterfall indicating no disproportionate increases ingreenfall biomass with increases in foliar production.Some plant "species," however, contributed propor-tionately more to greenfall than was predicted by theirfoliar production alone. Exceptionally high ratios ofherbivore-caused greenfall to foliar production char-acterized grapes (6.9% on WS18), tulip poplar (7.4%on WS2), and white basswood (11.1% on WS13). Thehigh value for chestnut (9.8% on WS2) was unusualand not representative for chestnut in general.

Plant species richness of greenfall was greater on thetwo watersheds (WS7 and WS13) regenerating fromclearcuts and may have been a reflection of treatmenthistory. For example, Fraser magnolia and cucumber-trees, collected as greenfall on WS13 only, did not ap-

August 1988 HERBIVORE-CAUSED GREENFALL 1121

TABLE 2. Contributions of tree and vine "species" (species or species groups) to total annual herbivore-caused greenfall onfour forested watersheds.*

Plant "species"Acer rubrum (Red ma-

ple)Amelanchier arborea

(Downy serviceberry)Betula spp. (Birch)Carya spp. (Hickory)Castanea dentata

(American chestnut)Castanea pumila (Alle-

gheny chinkapin)Cornusflorida (Flower-

ing dogwood)Diospyros virginiana

(Common persim-mon)

Fagus grandifolia(American beech)

Hamamelis virginiana(Witch-hazel)

Liriodendron tulipifera(Tulip poplar)

Magnolia acuminata(Cucumbertree)

Magnolia fraseri (Fra-ser magnolia)

Nyssa sylvatica (Blackgum)

Oxydendrum arboreum(Sourwood)

Parthenocissus quinque-folia (Virginia creep-er)

Prunus serotina (Blackcherry)

Quercus alba (Whiteoak)

Quercus prinus (Chest-nut oak)

Quercus rubra group(Red oak group)

Robinia pseudoacacia(Black locust)

Sassafras albidum (Sas-safras)

Smilax spp. (Green-briar)

Tilia heterophylla(White basswood)

Vitis spp. (Wild grape)

WS2

Biomassf

0.36 ± 0.21

<0.010.0

3.64 ± 1.26

0.02 ± 0.01

0.0

0.29 ± 0.05

0.0

0.0

<0.01

1.18 ± 0.65

0.0

0.0

0.06 ± 0.03

0.07 ± 0.03

0.0

0.0

0.26 ± 0.20

1.19 ±0.35

1.00 ± 0.39

0.04 ± 0.02

0.0

<0.01

0.00.11 ± 0.06

%|

1.1

0.30.05.6

9.8

0.0

1.7

0.0

0.0

-§

7.4

0.0

0.0

0.3

1.0

0.0

0.0

1.9

1.9

1.4

5.4

0.0

1.0

0.03.5

WS7Biomass

0.35 ± 0.12

0.04 ± 0.020.30 ± 0.270.15 ± 0.09

0.0

0.01 ± <0.01

0.30 ± 0.08

0.01 ± 0.01

<0.01

0.03 ± 0.02

0.12 ± 0.09

0.0

0.0

0.0

0.05 ± 0.03

0.0

0.058

0.0

0.61 ± 0.29

0.13 ± 0.06

0.79 ± 0.29

0.06 ± 0.02

<0.01

0.00.28 ±0.12

%

1.0

5.33.42.3

0.0

0.3

2.8

0.2

0.4

0.4

1.4

0.0

0.0

0.0

1.1

0.0

-§

0.0

1.0

0.3

2.3

0.5

0.2

0.00.9

WS13Biomass

0.12 ± 0.03

0.01 ± 0.010.33 ± 0.190.43 ± 0.14

0.0

0.0

0.30 ± 0.07

0.01 ± 0.01

0.00

0.01 ± 0.01

1.14 ±0.35

0.04 ± 0.04

0.01 ± 0.01

0.01 ± <0.01

0.06 ± 0.03

0.0

0.01 ± 0.01

0.03 ± 0.03

0.85 ± 0.45

0.53 ± 0.21

0.03 ± 0.01

0.01 ± 0.01

<0.01

0.10 ± 0.100.32 ± 0.09

%

0.5

0.62.93.5

0.0

0.0

1.7

0.4

0.0

0.7

0.9

1.0

0.3

0.3

0.7

0.0

1.4

2.7

2.5

1.5

0.6

0.5

0.6

11.11.2

WS18Biomass

0.13 ± 0.04

<0.010.07 ± 0.051.49 ± 0.78

0.01 ± 0.01

0.0

0.17 ± 0.07

0.0

0.0

0.0

0.31 ± 0.17

0.0

0.0

0.02 ± 0.02

0.12 ± 0.05

<0.01

0.0

0.0

0.80 ± 0.29

1.25 ± 0.41

0.05 ± 0.04

0.01 ± <0.01

<0.01

0.00.06 ± 0.04

%

0.3

0.20.72.5

0.9

0.0

1.8

0.0

0.0

0.0

2.8

0.0

0.0

0.9

1.1

<0.1

0.0

0.0

0.8

1.8

3.7

1.4

0.7

0.06.9

* WS2 = south-facing, undisturbed; WS7 = south-facing, 10-yr-old clearcut; WS13 = east-facing, 23-yr-old clearcut; WS18 =north-facing, undisturbed.

t Biomass (g-m-2-yr-') is expressed as mean ± SE (n = 10).^ Percentages represent 100 x the ratio of herbivore-caused greenfall to total foliar litterfall for the designated plant "species."§ Not collected in litter traps.

pear on that watershed until after the first clearcut (Leo-pold et al. 1985).

"Clipped" greenfall

Herbivore-caused greenfall was separated into fourcategories based on patterns of damage, each of whichtypified one kind of herbivore feeding behavior (Fig.1). "Clipped" referred to leaves with abbreviated pet-ioles and considerable (20-50%) leaf area removed,

typical of the effects of petiole-clipping caterpillars(Risley 1986). Inputs of clipped leaves, in terms of leafnumber (not shown) and biomass (Fig. 1), were heavi-est in August. A number of insect species were iden-tified as petiole clippers (Table 3).

Clipped greenfall was represented by 92% (23"species") of the plant "species" that contributed toherbivore-caused greenfall. Two species not repre-sented were Virginia creeper and Fraser magnolia. The

1122 LANCE S. RISLEY AND D. A. CROSSLEY, JR. Ecology, Vol. 69, No. 4

WS2 0 MinedD Petiole-damagedd Orts

Clipped

0.00

0.08

0.04

13 5 18 7 21 4 20 7MAY JUN JUN JUL JUL AUG AUG SEP

WS7

20SEP

0.00 13 5 18 7 21 4 20 7MAY JUN JUN JUL JUL AUG AUG SEP

20SEP

0.00 13 5 18 7 21 4 20 7MAY JUN JUN JUL JUL AUG AUG SEP

20SEP

0.08-

0.04

°-00 13 5 ' 18 7 21 4 20 7 20MAY JUN JUN JUL JUL AUG AUG SEP SEP

SAMPLE DATESFIG. 1. The composition of herbivore-caused greenfall on

four forested watersheds. WS2 = south-facing, undisturbed;WS7 = south-facing, 10-yr-old clearcut; WS13 = east-facing,23-yr-old clearcut; WS18 = north-facing, undisturbed. Heightsof bars represent mean biomass (n = 10) in each category.Mined = primarily by an undetermined microlepidopteran;petiole-damaged = primarily by petiole-boring larvae; orts =discarded leaf fragments; clipped = resulting from activitiesof petiole-clipping caterpillars.

number of plant "species" represented by clippedgreenfall among watersheds varied from 18 (WS7 andWS13) to 14 (WS18) to 13 (WS2). Biomass of clippedgreenfall, as a proportion of herbivore-caused greenfall,was 50.6% on WS2, 42.0% on WS7, 68.3% on WS13,and 55.0% on WS18. Clipped greenfall contributed be-tween 1.3 and 0.5% to foliar litterfall (greenfall"species") on the study areas.

Heinrich (1979) hypothesized that caterpillars pal-atable to insectivorous birds avoid detection by clip-ping petioles of partially consumed leaves, thereby re-moving evidence of their feeding, which might be usedas a cue by the birds. This hypothesis was tested under"seminatural" conditions using Black-capped Chick-adees, mealworms, and several species of palatable cat-erpillars (Heinrich and Collins 1983). The chickadeeswere able to distinguish among tree species and be-tween damaged and undamaged leaves in search ofcaterpillar prey. We believe that bird avoidance is oneof several possible explanations for petiole clipping bycaterpillars.

Rapid induction of plant secondary metabolites inresponse to herbivore damage may also favor petioleclipping. T. Kimmerer (personal communication) foundthat induction of secondary compounds occurred onundamaged eastern cottonwood (Populus deltoides)leaves adjacent to leaves damaged mechanically. Kim-merer also reported that when leaves were damaged,then cut off at the petiole within hours of damage treat-ment, no induction occurred. Therefore, if a caterpillareats a portion of a leaf and clips it free, the caterpillarmight avoid lowering the quality of adjacent leaves.

A third explanation for petiole clipping resulted fromobservations of an undetermined caterpillar species(Lepidoptera: Tortricidae) (C. Smith, personal com-munication) that feeds on sweet birch (Betula lento) atCoweeta. Individual larvae fold leaves longitudinallyand seal the edges with silk, leaving a small openingat the petiole/midrib junction. Each caterpillar thenextends through the opening of its shelter and clips thepetiole, causing the leaf to fall to the forest floor. Re-maining larval development and pupation occurs with-in the leaf purse. A number of these caterpillars were

TALE 3. A partial list of insect herbivores responsible for petiole clipping at the Coweeta Hydrologic Laboratory. All arelepidopteran larvae.

HerbivoresAcronicta americana (Harris)Pangrapta decoralis Hubner*Undetermined species*Undetermined species*Heterocampa guttivitta (Walker)Nadata gibbosa (J. E. Smith)*Ellida caniplaga (Walker)*Papilio glaucus glaucus L.*Callosamia angulifera (Walker)Sphecodina abbottii (Swainson)Undetermined species*

FamilyNoctuidaeNoctuidaeNoctuidaeNoctuidaeNotodontidaeNotodontidaeNotodontidaePapilionidaeSaturniidaeSphingidaeTortricidae

Plants attackedAcer rubrumOxydendrum arborewnCornus ftoridaCastanea dentataCarya spp., Quercus spp.Quercus spp.Tilia heterophyllaLiriodendron tulipiferaLiriodendron tulipiferaVitis spp.Betula spp.

* Additions to list of Heinrich and Collins (1983).

1-

August 1988 HERBIVORE-CAUSED GREENFALL 1123

reared to adults from leaf purses collected from theforest floor and deposited subsequently in the Univer-sity of Georgia insect collection.

"Petiole-damaged" green/all

The second category of herbivore-caused greenfall,"petiole-damaged," was represented primarily by leavesexhibiting evidence of petiole-boring activity. The flux-es of petiole-damaged greenfall were highest in lateJune-early July and in August (Fig. 1). Many of theleaves assigned to this category had abbreviated peti-oles that had broken at the site of most intensive bor-ing. Of secondary importance were leaves with petioleand petiole/midrib damage caused by herbivores feed-ing externally. In our study, some of the petiole-dam-aged greenfall resulted from unsuccessful petiole clip-ping and inefficient feeding (Heterocampa described inCarroll and Hoffman 1980, Lowman 1982) on petioleseither before or after feeding on the leaf blade.

The petiole-damaged category was represented by20 plant "species," slightly less than in the clippedcategory. Oaks, hickories, and red maple were the mostcommon "species" in this category. Petiole-damagedgreenfall contributed 23,41,23, and 22% to herbivore-caused greenfall on watersheds 2, 7, 13, and 18, re-spectively. The proportion of total foliar litter (greenfall"species") in this category was between 0.6 and 0.3%on the same watersheds.

Greenfall in the petiole-damaged category was rep-resented in May primarily by birch and resulted froma petiole borer (probably Apagodiplosis papyriferae[Gagne]) (Diptera: Cecidomyiidae). Much of the green-fall from June and July in this category was due toactivity of the maple petiole borer (Caulocampus ac-ericaulis [MacGillivray]) (Hymenoptera: Tenthridini-dae) on red maple and the hickory shoot borer (Con-otrachelus probably aratus [Germ.]) (Coleoptera:Curculionidae) on hickories. Occasionally, greenfallenhickory leaves were collected with numerous medium-sized phylloxeran galls (probably Phylloxera caryae-caulis [Fitch]) (Homoptera: Phylloxeridae).

Petiole-damaged greenfall was partially the result ofhost plant responses to herbivore damage. Leaf bladesmay contain sufficient auxin to cause retention of allbut the most heavily damaged leaves (Addicott 1982),so blade damage is a poor predictor of premature ab-scission until late in the growing season (Risley 1987).Petiole damage, including damage to the petiole/mid-rib junction, may, however, result in premature fall ofgreen leaves (Brooks 1922, Baker 1972, Johnson andLyon 1976, Martineau 1984).

"Mined" greenfall

The third category of greenfall, "mined," was quitespecific: an unidentified microlepidopteran that at-tacked hickories, and to a lesser extent oaks, in springand midsummer (Fig. 1) was responsible. The fall ofmined hickory leaflets was striking and tracked tree

phenology from low to high elevation watersheds atthe beginning of the growing season. This phenomenonhad been observed previously at Coweeta (W. T. Swank,personal communication) but had not been explained.The presence of a leaf miner in leaflets was not rec-ognized until greenfall samples were processed. Whilein the plant presses, tenuous linear mines developedin 99% of the hickory leaflets collected in May andearly June. Single (rarely, two) mines began at the baseof each leaflet midrib and continued distally for 5-10mm. Blotch mines developed in «50% of the leafletswith linear mines, presumably caused by the samespecies of leaf miner. This was the only leaf minerimplicated in greenfall at Coweeta in 1985.

The mined category was represented by four plant"species" and was dominated by hickories. Oaks wereoccasional contributors and rarely, sourwood. Green-fall categorized as mined contributed 23.8, 3.7, 1.6,and 18.6% to herbivore-caused greenfall on watersheds2, 7, 13, and 18, respectively. Mined greenfall, as aproportion of total foliar litterfall (greenfall "species")was 0.6, 0.04, 0.02, and 0.2% on the same watersheds.

Previous studies of leaf miners have documentedpremature abscission of mined leaves (Nuorteva 1966,Mazanec 1974, Owen 1978, Faeth et al. 1981, Hilemanand Lieto 1981, Proctor et al. 1982, Maier 1983, Prit-chard and James 1984a, b, Potter 1985, Bultman andFaeth 1986). In most of these studies, abscission tookplace late in the growing season. We suggest that thepremature drop of hickory leaflets observed in our studywas not a simple abscission response to damage. Early-season fall of numerous hickory leaflets had been ob-served for at least three seasons previous to this studyand was observed in 1986 and 1987. If leaf minermortality is high in abscised leaflets, it seems unlikelythat populations would remain at observed levels yearafter year. It is possible that the newly hatched leafminer larva girdles the midrib at the point of attach-ment to the rachis, causing the leaflet to fall green tothe forest floor where larval development proceeds topupation (see Kahn and Cornell 1983). Most hickoryleaflet greenfall occurred in moist coves, where fallenleaflets might have remained fresh long enough to sup-port leaf miner larvae. Indeed, a number of leaf minerindividuals completed larval development and pupat-ed in a sample of hickory leaflets collected as greenfallin the spring of 1987.

"Orts" greenfall

The fourth category of herbivore-caused greenfall,"orts," consisted of fragments of leaves large enoughto be found amid other leaf litter in plots. Orts includedmiscellaneous leaf fragments discarded by herbivorouscaterpillars and leaf rolls constructed by female leaf-rolling weevils (Coleoptera: Attelabidae). The seasonaldistribution of orts greenfall (Fig. 1) was similar amongwatersheds and sample dates. Increases were noted inmonths of heaviest caterpillar activity.

1124 LANCE S. RISLEY AND D. A. CROSSLEY, JR. Ecology, Vol. 69, No. 4

250

200-

<M 150'V.01

CO< 100-

Om

50-

1

3 JUN 6 JUL

1

3 AUG 7 SEP

H WS20 WS711 WS13B WS18

II OCT 8 NOV 10 DEC 8 FEB

SAMPLE DATESFIG. 2. Foliar litter production from May 1985 to February 1986 on four forested watersheds. WS2 = south-facing,

undisturbed; WS7 = south-facing, 10-yr-old clearcut; WS13 = east-facing, 23-yr-old clearcut; WS18 = north-facing, undis-turbed. Data are means ± SE.

Orts included 20 "species" of deciduous trees andwoody vines. Orts contributed 2.6, 13.2, 6.8, and 4.7%of herbivore-caused greenfall on watersheds 2, 7, 13,and 18, respectively. The proportion of orts greenfallin total annual foliar litterfall including fall abscission(of only those "species" that contributed to greenfall)was between 0.1 and 0.2% on the same watersheds.

Greenfall in relation to litterfall

Our litterfall measurements were similar to thosereported by Kovner (1955) and Cromack and Monk(1975) for Coweeta deciduous forests (Table 4). Foliarlitterfall from "species" represented in greenfall peakedin autumn (Fig. 2). During the summer (May throughAugust) foliar litterfall contributed small amounts (5.0%onWS2,8.1%onWS7, 5.5%onWS13, 3.3%onWS18)to total foliar litter production.

Insect fecal pellet production was slightly higher thanvalues recorded by Risley in 1983 (2.4 g-m-2-yr-') onWS18 (Risley 1986) and by Carlisle et al. (1966) (2.6-2.9 g-m~2-yr~' postoutbreak) in a British woodland.Fecal pellet production, as a proportion of total foliarlitterfall, was similar on the four watersheds (1.33,1.08,1.23, and 1.09% for WS2, 7, 13, and 18, respectively).Even though herbivory was typically low at Coweetain 1985 (L. S. Risley and D. A. Crossley, personalobservation), our measurements of insect fecal pelletswere lower than expected and probably resulted fromsmall fecal pellets passing through mesh openings.

Insect body parts contributed very little to total lit-terfall (Table 4). Items in this litter category includedwhole insects, head capsules of caterpillars, heterop-teran exuviae, and assorted pieces of exoskeletons. Asa proportion of foliar litterfall, the biomass of insectbody parts ranged from 0.01 to 0.08%.

Total annual inputs of greenfall were small relativeto total foliar production, but were greater than fecalpellet litter and insect body part litter combined (Table4). The proportion of total greenfall biomass to totalfoliar litter biomass (greenfall "species") ranged from2.5% on the south-facing undisturbed watershed (WS2)to 1.2% on the adjacent 10-yr-old clearcut watershed(WS7). Total herbivore-associated litter biomass (fecalpellets, insect body parts, greenfall) ranged from 1.2 to3.0% of total (foliar + nonfoliar) annual litterfall.

IMPLICATIONS OF HERBIVORE-CAUSED GREENFALL

As a resource for decomposers

We have shown that greenfall contributes to the linkbetween herbivores in the canopy and decomposers onthe forest floor. Greenfall contributed <5% to totalfoliar litterfall and was greater than total fecal pelletproduction by herbivorous insects. Yet, the nutrientcontent, especially nitrogen, and timing of herbivore-caused greenfall may multiply its importance com-pared with that calculated by quantity alone.

Major insect herbivore-associated transfers of nitro-

August 1988 HERBIVORE-CAUSED GREENFALL 1125

TABLE 4 . Litter production and herbivore-caused greenfall production on four forested watersheds* at the Coweeta HydrologicLaboratory. Values are means ± SE; n = 10 in all cases except WS13 litter-trap data where n = 9.

Litter categoryProduction (g-

WS2 WS7 WS13 WS18Litter traps

Foliar (deciduous)fFoliar (evergreen)!Insect fecal pelletsInsect body partsMiscellaneous§

Total litterPlots

Total greenfall

320.74 + 15.4910.26 ± 6.834.40 ± 0.560.12 ± 0.03

94.18 ± 18.94429.70

8.14 ± 1.32

279.92 ± 26.373.40 + 2.663.06 ± 0.490.22 ± 0.10

24.25 ± 6.51310.85

3.34 ± 0.37

319.46 ± 18.172.85 ± 2.803.96 ± 0.370.11 ±0.03

45.41 ± 5.01371.79

4.32 ± 0.55

334.85 ± 16.3028.45 ± 13.39

3.94 ± 0.530.02 ± 0.01

69.79 ± 11.16436.23

4.90 ± 1.11* WS2 = south-facing, undisturbed; WS7 = south-facing, 10-yr-old clearcut; WS 13 = east-facing, 23-yr-old clearcut; WS 18 ••

north-facing, undisturbed.t Deciduous plant "species" contributing to greenfall (see Table 2).i Evergreen trees and shrubs.§ Woody Utter, mast, debris.

gen and other nutrients (that are not leached easily fromleaves) from the canopy to the forest floor are in theform of fecal pellets, insect body parts, and foliar lit-terfall. Nitrogen concentrations in fecal pellets are equalto or less than those in consumed foliage (Juritz 1920,Carlisle et al. 1966, Gosz et al. 1972, Hollinger 1986,Schroeder 1986a, b), but because N can be leachedeasily from feces, there is potential for rapid utilizationby decomposers (Fox and Macauley 1977, Silander etal. 1983, Haukioja et al. 1985). After N is lost fromfecal pellets, little else remains except lignin and cel-lulose; both are difficult to decompose. Insect bodiescontain, on average, 7-9% N by mass (Hollinger 1986,Schroeder 1986a, b), but bodies are resistant to de-composition (Seastedt and Tate 1981) and inputs ofbody parts to the forest floor are usually very small.

The third major herbivore-associated transfer of Nfrom the canopy to the forest floor is greenfall. Greenleaves in midseason contain at least 40% more N (bymass) than senesced autumn leaves (Cromack andMonk 1975), which contain typically =»!% N in tem-perate hardwood forests (Carlisle et al. 1966, Gosz etal. 1972, Cromack and Monk 1975, Cotrufo 1977,Muller and Martin 1983). Because N is not leachedeasily from damaged leaves (Tukey and Morgan 1963,Seastedt et al. 1983) and processes leading to greenfall(e.g., petiole clipping) prevent the occurrence of muchtranslocation from leaves, greenfall represents a rela-tively rich resource for decomposers.

Green leaves may be sought and attacked selectivelyby forest-floor macroinvertebrates (Jennings and Bark-ham 1975; C. Blanton, personal communication). Fur-ther, greenfall is decomposed more quickly than au-tumn leaves (C. Blanton and L. S, Risley, personalobservation). While nutrients from greenfall are likelyto become available to nearby root systems and maystimulate new growth (see Owen and Wiegert 1976),we believe that greenfall is of primary importance asa resource for decomposers on the forest floor. At Co-

weeta, greenfall is widespread spatially and temporally,and its occurrence is correlated positively with theabundance (=foliar production) of represented plant"species." These characteristics give greenfall somepredictability, and thus reliability as a resource for de-composers.

As a corrective for estimates ofherbivory

Greenfall may be important as a corrective for es-timates of leaf area removed by herbivores especiallyin plant "species" that lose a considerable proportionof their leaves to herbivores (e.g., hickories; this study).Two components can be considered: petioles that re-main in situ after blades have been removed, and lossof entire leaves from the canopy. We have found thatpetioles remain attached to twigs long after leaf bladeshave been removed if the removal was associated withpetiole clipping or petiole boring at the petiole/midribjunction. Estimates of herbivory based on canopy sam-ples should include these petioles to represent leaveswith 100% blade area removed. Relationships betweenpetiole allometry and leaf area must be determined inorder to calculate absolute quantities of leaf area miss-ing from petioles. For example, Linit et al. (1986) re-ported a significant positive relationship (r2 = 0.90)between petiole diameter and leaf area in northern redoak (Quercus rubra).

Failure to consider herbivore-caused loss of entireleaves may result in underestimates ofherbivory mea-sured during the growing season (Nuorteva 1966, Jour-net 1981, Fox and Morrow 1983, Lam and Dudgeon1985, Risley 1986). Regardless of the extent to whicha leaf is damaged, if it falls prematurely the result is100% leaf area removed by herbivores. Onuf et al.(1977) associated Ecdytolopha sp. (Lepidoptera: Ole-threutidae) with leaf loss in red mangroves (Rhizophoramangle) and incorporated that loss into estimates ofherbivory. If entire leaves are lost as a result of her-bivory, a canopy-based sampling regime should correct

1126 LANCE S. RISLEY AND D. A. CROSSLEY, JR. Ecology, Vol. 69, No. 4

underestimates of leaf area removed by herbivores byincorporating measurements of herbivore-caused leaffall.

ACKNOWLEDGMENTSWe thank Dr. W. T. Swank, Project Leader of the Coweeta

Hydrologic Laboratory, and Coweeta staff for providing ac-cess to study areas and assistance with research design. Effortsby Donna Wilier and John Blair in litter trap construction,placement, and litter collections were greatly appreciated. Dr.Cecil Smith, curator of the University of Georgia insect col-lection, was very helpful with species determinations. ChantalBlanton kindly shared unpublished data from a study of greenleaf decomposition. Earlier versions of this manuscript werereviewed by Drs. J. B. Wallace, W. Berisford, P. Hunter, W.Swank, and Donna Wilier. We also thank F. Slansky and ananonymous reviewer for very useful comments regardingmanuscript revisions. Research was made possible by NSFgrants DEB-8012093 and BSR-8409532 to Dr. D. A. Cross-ley, Jr. and a University of Georgia Graduate School Fellow-ship to L. S. Risley.

LITERATURE CITEDAddicott, F. T. 1982. Abscission. University of California

Press, Berkeley, California, USA.Baker, W. L. 1972. Eastern forest insects. United States

Forest Service Miscellaneous Publication 1175.Brooks, F. E. 1922. Curculios that attack the young fruits

and shoots of walnut and hickory. United States Depart-ment of Agriculture Bulletin 1066.

Bultman, T. L., and S. H. Faeth. 1986. Selective ovipositionby a leaf miner in response to temporal variation in ab-scission. Oecologia (Berlin) 69:117-120.

Carlisle, A., A. H. F. Brown, and E. J. White. 1966. Litterfall, leaf production and the effects of defoliation by Tortrixviridana in a sessile oak (Quercus petraea) woodland. Jour-nal of Ecology 54:65-86.

Carroll, C. R., and C. A. Hoffman. 1980. Chemical feedingdeterrent mobilized in response to insect herbivory andcounteradaptation by Epilachna tredecimnotata. Science209:414-416.

Chabot, B. F., and D. J. Hicks. 1982. The ecology of leaflife spans. Annual Review of Ecology and Systematics 13:229-259.

Cotrufo, C. 1977. Nutrient content in litterfall of an Ap-palachian hardwood stand. Journal of Elisha Mitchell Sci-entific Society 93:27-33.

Cromack, K., Jr., and C. D. Monk. 1975. Litter production,decomposition, and nutrient cycling in a mixed hardwoodwatershed and a white pine watershed. Pages 609-624 inF. G. Howell, J. B. Gentry, and M. H. Smith, editors. Min-eral cycling in southeastern ecosystems. ERDA SymposiumSeries CONF-740513. National Technical Information Ser-vice, Springfield, Virginia, USA.

Day, F. P., and C. D. Monk. 1974. Vegetation patterns ona southern Appalachian watershed. Ecology 55:1064-1074.

Day, F. P., Jr., D. L. Phillips, and C. D. Monk. 1987. Forestcommunities and patterns. Pages 141-149 in W. T. Swankand D. A. Crossley, Jr., editors. Forest hydrology and ecol-ogy at Coweeta. Springer-Verlag, New York, New York,USA.

Faeth, S. H., E. F. Connor, and D. Simberloff. 1981. Earlyleaf abscission: a neglected source of mortality for folivores.American Naturalist 117:409-415.

Fox, L. R., and B. J. Macauley. 1977. Insect grazing onEucalyptus in response to variation in leaf tannins and ni-trogen. Oecologia (Berlin) 29:145-162.

Fox, L. R., and P. A. Morrow. 1983. Estimates of damageby herbivorous insects on Eucalyptus trees. AustralianJournal of Ecology 8:139-147.

Gosz, J. R., G. E. Likens, and F. H. Bormann. 1972. Nu-trient content of litterfall on the Hubbard Brook Experi-mental Forest, New Hampshire. Ecology 53:769-784.

Haukioja, E. 1979. Anti-herbivore strategies in mountainbirch at the treeline. Holarctic Ecology 2:272-274.

Haukioja, E., J. Suomela, and S. Neuvonen. 1985. Long-term inducible resistance in birch foliage: triggering cuesand efficacy on a defoliator. Oecologia (Berlin) 65:363-369.

Heinrich, B. 1979. Foraging strategies of caterpillars: leafdamage and possible predator avoidance strategies. Oeco-logia (Berlin) 42:325-337.

Heinrich, B., and S. L. Collins. 1983. Caterpillar leaf dam-age, and the game of hide-and-seek with birds. Ecology 64:592-602.

Hileman, D. R., and L. F. Lieto. 1981. Mortality and areareduction in leaves of the bog shrub Chamaedaphne caly-culata (Ericaceae) caused by the leaf miner Coptodisca kal-miella (Lepidoptera: Heliozelidae). American MidlandNaturalist 106:180-188.

Hollinger, D. Y. 1986. Herbivory, and the cycling of nitro-gen and phosphorus in isolated California oak trees. Oeco-logia (Berlin) 70:291-297.

Jennings, T. J., and J. P. Barkham. 1975. Food of slugs inmixed deciduous woodland. Oikos 26:211-221.

Johnson, W. T., and H. H. Lyon. 1976. Insects that feed ontrees and shrubs. Cornell University Press, Ithaca, NewYork, USA.

Journet, A. R. P. 1981. Insect herbivory on the Australianwoodland eucalypt, Eucalyptus blakelyi M. AustralianJournal of Ecology 6:135-138.

Juritz, C. F. 1920. Analysis of droppings of caterpillar (An-theraea cytherea). Chemical News 121:181.

Kahn, D. M., and H. V. Cornell. 1983. Early leaf abscissionand folivores: comments and considerations. AmericanNaturalist 122:428^32.

Kimmins, J. P. 1972. Relative contribution of leaching,litter-fall and defoliation by Neodiprion sertifer (Hymenop-tera) to the removal of cesium-134 from red pine. Oikos23:226-234.

KitcheH, J. F., R. V. O'Neill, D. Webb, G. W. Gallepp, S. M.Bartell, J. F. Koonce, and B. S. Ausmus. 1979. Consumerregulation of nutrient cycling. Bioscience 29:28-34.

Kovner, J. L. 1955. Changes in streamflow and vegetationcharacteristics of a southern Appalachian mountain wa-tershed brought about by forest cutting and subsequent nat-ural regrowth. Dissertation. State University of New York,College of Forestry, Syracuse, New York, USA.

Lam, P. K. S., and D. Dudgeon. 1985. Fitness implicationsof plant-herbivore "mutualism." Oikos 44:360-361.

Lamb, D. 1985. The influence of insects on nutrient cyclingin eucalypt forests: a beneficcial role? Australian Journal ofEcology 10:1-5.

Larsson, S., and O. Tenow. 1980. Needle-eating insects andgrazing dynamics in a mature Scots pine forest in centralSweden. In T. Persson, editor. Structure and function ofnorthern coniferous forests—an ecosystem study. Ecologi-cal Bulletins-NFR 32:269-306.

Leopold, D.J.,G.R. Parker, and W. T. Swank. 1985. Forestdevelopment after successive clearcuts in the southern Ap-palachians. Forest Ecology and Management 13:83-120.

Linit, M. J., P. S. Johnson, R. A. McKinney, and W. H.Kearby. 1986. Insects and leaf area losses of plantednorthern red oak seedlings in an Ozark forest. Forest Sci-ence 32:11-20.

Lowman, M. D. 1982. Effects of different rates and methodsof leaf area removal on rain forest seedlings of coachwood(Ceratopetalum capetalum). Australian Journal of Ecology30:477-483.

Maier, C. T. 1983. Effect of the apple blotch leafminer (Lep-idoptera:Gracillariidae) on apple leaf abscission. Journal ofEconomic Entomology 76:1265-1268.

August 1988 HERBIVORE-CAUSED GREENFALL 1127

Martineau, R. 1984. Insects harmful to forest trees. Mul-tiscience Publications, Montreal, Quebec, Canada.

Mattson, W. J., and N. D. Addy. 1975. Phytophagous in-sects as regulators of forest primary production. Science190:515-522.

Mazanec, Z. 1974. Influence of jarrah leaf miner on thegrowth of jarrah. Australian Forestry 37:32-42.

Muller, R. N., and W. H. Martin. 1983. Autumn leaf falland nutrient return in an old-growth and a second-growthforest in eastern Kentucky. Botanical Gazette 144:552-558.

Nielsen, O. 1978. Food resource partition in the beech leaf-feeding guild. Ecological Entomology 3:193-201.

Nuorteva, P. 1966. Leaf development and leaf pathogenson subarctic birches after a calamity ofOporinia autumnata(Bkh.) (Lep., Geometridae). Annales Zoologici Fennici 3:270-286.

Ohmart, C. P., L. P. Stewart, and J. R. Thomas. 1983. Leafconsumption by insects in three Eucalyptus forest types insoutheastern Australia and their role in short-term nutrientcycling. Oecologia (Berlin) 59:322-330.

Onuf, C. P., J. M. Teal, and I. Valiela. 1977. Interactionsof nutrients, plant growth and herbivory in a mangroveecosystem. Ecology 58:514-526.

Owen, D. F. 1978. The effect of a consumer, Phytomyzailicis, on seasonal leaf-fall in the holly, Ilex aquifolium.Oikos 31:268-271.

Owen, D. F., and R. G. Wiegert. 1976. Do consumers max-imize plant fitness? Oikos 27:488-492.

Pook, E. W. 1984. Canopy dynamics of Eucalyptus macu-lata Hook. I. Distribution and dynamics of leaf populations.Australian Journal of Botany 32:387-403.

Potter, D. A. 1985. Population regulation of the native hollyleafminer, Phytomyza ilicicola Loew (Diptera: Agromyzi-dae), on American holly. Oecologia (Berlin) 66:499-505.

Pritchard, I. M., and R. James. 1984a. Leaf mines: theireffect on leaf longevity. Oecologia (Berlin) 64:132-139.

Pritchard, I. M., and R. James. 19846. Leaf fall as a sourceof leaf miner mortality. Oecologia (Berlin) 64:140-141.

Proctor, J. T. A., J. M. Bodner, W. J. Blackburn, and R. L.Watson. 1982. Analysis of the effects of the spotted tenti-form leafminer (Phyllonorycter blancardella) on the pho-tosynthetic characteristics of apple leaves. Canadian Jour-nal of Botany 60:2734-2740.

Reichle, D. E., and D. A. Crossley, Jr. 1967. Investigationon heterotrophic productivity in forest insect communities.Pages 563-587 in K. Petrusewicz, editor. Secondary pro-ductivity of terrestrial ecosystems. Panstwowe Wydaw-nictwo Naukowe, Warsaw, Poland.

Ring, D. R., M. K. Harris, and R. Olszak. 1985. Life tablesfor pecan leaves in Texas. Journal of Economic Entomology78:888-894.

Risley, L. S. 1983. Analysis of aboveground arthropod com-

munities and their influence on nutrient dynamics in a ma-ture and successional forest. Thesis. University of Tennes-see, Knoxville, Tennessee, USA.

1986. The influence of herbivores on seasonal leaf-fall: premature leaf abscission and petiole clipping. Journalof Agricultural Entomology 3:152-162.

-. 1987. Acceleration of seasonal leaf fall by herbivoresin the southern Appalachians. Dissertation. University ofGeorgia, Athens, Georgia, USA.

Roberts, D. A., and C. W. Boothroyd. 1972. Fundamentalsof plant pathology. W. H. Freeman, San Francisco, Cali-fornia, USA.

Schowalter, T. D. 1981. Insect herbivore relationship to thestate of the host plant: biotic regulation of ecosystem nu-trient cycling through ecological succession. Oikos 37:126-130.

Schroeder, L. A. 19860. Protein limitation of a tree leaffeeding lepidopteran. Entomologia Experimentalis et Ap-plicata 41:115-120.

. 19866. Changes in tree leaf quality and growth per-formance of lepidopteran larvae. Ecology 67:1628-1636.

Seastedt, T. R., and D. A. Crossley, Jr. 1984. The influenceof arthropods on ecosystems. Bioscience 34:157-161.

Seastedt, T. R., D. A. Crossley, Jr., and W. W. Hargrove.1983. The effects of low-level consumption by canopy ar-thropods on the growth and nutrient dynamics of blacklocust and red maple trees in the southern Appalachians.Ecology 64:1040-1048.

Seastedt, T. R., and C. M. Tate. 1981. Decomposition ratesand nutrient contents of arthropod remains in forest litter.Ecology 62:13-19.

Silander, J. A., B. R. Trenbath, and L. R. Fox. 1983. Chem-ical interference among plants mediated by grazing insects.Oecologia (Berlin) 58:415^117.

Springett, B. P. 1978. On the ecological role of insects inAustralian eucalypt forests. Australian Journal of Ecology3:129-139.

Swank, W. T., and D. A. Crossley, Jr. 1987. Introductionand site description. Pages 3-16 in W. T. Swank and D. A.Crossley, Jr., editors. Forest hydrology and ecology at Co-weeta. Springer-Verlag, New York, New York, USA.

Thomas, W. A. 1969. Accumulation and cycling of calciumby dogwood trees. Ecological Monographs 39:101-120.

Tukey, H. B., Jr. 1970. The leaching of substances fromplants. Annual Review of Plant Physiology 21:305-324.

Tukey, H. B., Jr., and J. V. Morgan. 1963. Injury to foliageand its effect upon the leaching of nutrients from above-ground plant parts. Physiologia Plantarum 16:557-564.

Wallner, W. E. 1987. Factors affecting insect populationdynamics: differences between outbreak and non-outbreakspecies. Annual Review of Entomology 32:317-340.