http://www.jstor.org

Top-Down vs Bottom-Up Control Influenced by Productivity in a North Derbyshire, UK, DaleAuthor(s): Lauchlan H. FraserSource: Oikos, Vol. 81, No. 1, (Feb., 1998), pp. 99-108Published by: Blackwell Publishing on behalf of Nordic Society OikosStable URL: http://www.jstor.org/stable/3546472Accessed: 16/06/2008 18:29

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=black.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

OIKOS 81: 99-108. Copenhagen 1998

Top-down vs bottom-up control influenced by productivity in a North Derbyshire, UK, dale

Lauchlan H. Fraser

Fraser, L. H. 1998. Top-down vs bottom-up control influenced by productivity in a North Derbyshire, UK, dale. - Oikos 81: 99-108.

Fretwell and Oksanen's theory of trophic dynamics was tested in two plant commu- nities located in a North Derbyshire dale, including: (1) a low productivity calcareous grassland; and, (2) a highly productive Urtica dioica (nettle) patch. Two methods (herbivore removal through pesticide application, and transplanting established, intact turves (0.5 m2) between the two community types) were employed, and analysed in a two-way ANOVA, to test the hypothesis that highly productive communities are controlled by 'top-down' forces and low productivity communities are controlled by 'bottom-up' forces. The Fretwell-Oksanen theory proposes that herbivores limit growth in low productivity communities, not highly productive communities. Therefore, removal of herbivores will result in an increase in plant biomass only in the low productivity community. The results presented in this paper support the Fretwell-Oksanen hypothesis. Furthermore, when small turves were transplanted from the highly productive community to the low productivity commu- nity the removal of herbivores through pesticide application greatly increased the above-ground plant biomass. This result suggests, firstly, that the vegetation grown in a highly productive environment is generally very palatable, and secondly, it strength- ens the evidence that herbivores are limiting plant growth in low productivity communities but not highly productive communities. Individual plant species re- sponse to herbivore removal was related to known relative growth rate values using linear regression and was found to be significant in one case: nettle turves trans- planted into the grassland. In this case, relative growth rate accounted for 68.3% of the variation in the response of the plants to herbivore removal. This suggests that fast-growing plants from a highly productive environment are most likely to respond to the release of a limiting factor, in this case herbivory, within a community.

L. H. Fraser, NERC, Unit of Comparative Plant Ecology, Univ. of Sheffield, Sheffield, UK SO1 2TN (present address: Dept of Biology, Univ. of Ottawa, Box 450, Stn A, Ottawa, ON, Canada KIN 6N5 flfraser@oreo.uottawa.ca]).

There is an increasing volume of theory and evidence

suggesting that primary productivity directly influences the trophic interactions of an ecosystem (Fretwell 1977, 1987, Oksanen et al. 1981, Oksanen 1990a, b, Van de

Koppel et al. 1996, Moen and Collins 1996, Fraser and Grime 1997). Fretwell and Oksanen have proposed that herbivore communities inhabiting a highly productive habitat are primarily controlled by carnivores ('top- down' control), whereas herbivore communities in a low productivity habitat are thought to be primarily

controlled by the supply of available resources ('bot- tom-up' control) (Fretwell 1977, Oksanen et al. 1981). The rationale for a 'bottom-up' controlled community at a low productivity level is that there are not enough resources in the community to support and sustain the third trophic level (carnivores). Productivity also ap- pears to determine to a large extent the plant functional

types likely to be successful in a plant community (Grime 1974, 1977). Fast-growing plants are associated with high productivity communities, and under labora-

Accepted 14 July 1997

Copyright ? OIKOS 1998 ISSN 0030-1299 Printed in Ireland - all rights reserved

OIKOS 81:1 (1998) 99

tory conditions are generally more palatable and vul- nerable to herbivory than slow-growing plants adapted to low productivity communities (Grime et al. 1968, 1997, Feeny 1976, Rhoades and Cates 1976, Grime 1979, Coley et al. 1985, Southwood et al. 1986). How- ever, despite the fact that a certain plant species may be quite palatable, if the herbivore numbers and activity are severely limited by carnivores, that plant species will flourish. Therefore, the extent to which herbivores influ- ence the structure and dynamics of plant communities only becomes understandable and predictable when trophic interactions are known.

Transplanting of established plants into different communities and environments is a commonly used test employed by population ecologists to study fine-scale interactions, specifically competition (Silander and An- tonovics 1982, Turkington 1989). Ecologists have also used turves containing mixtures of established plants to test the communities' responses to abiotic perturbations (MacGillivray et al. 1995) and herbivory (Moen et al. 1993). The methodology employed in the experiment described in this paper is unique because it involves transplanting turves of established plants into very dif- ferent plant communities (in terms of species composi- tion and site productivity) in order to test the varying roles of herbivory.

The purpose of the experiment was to test Fretwell and Oksanen's theory of trophic interactions (Fretwell 1977, Oksanen et al. 1981) by employing two interact- ing methodologies. First, it was intended to study the influence of pesticides on two communities of con- trasted productivity levels. The second purpose was to measure the impact of herbivores from one community on the vegetation of the other (and vice versa) by transplanting turves between the two. A complication which arises when turves are transplanted between communities is the possibility that herbivory will not depend exclusively upon the activity of the invertebrate herbivores originating from the 'host' community. To an unknown extent there may be impacts of inverte- brate herbivores or carnivores transplanted with the turves. The main vertebrate herbivore (sheep) have been absent from the dale for over 5 years. Other vertebrate herbivores, such as rabbits, mice and voles undoubtedly play important roles in the dale ecosystem, but this experiment was designed to test the importance of invertebrate herbivory and its relation to trophic dynamics.

Methods and materials

Study area

The experiment was conducted in a dale located within the Health and Safety Laboratory property in Der- byshire, near the town of Buxton. This dale is charac-

teristic of the shallow-soiled wet calcareous ancient

pastures found in the dales of the high limestone plateau (Balme 1953, Lloyd et al. 1971, Lloyd 1972, Pearce 1987). It is the dale used in studies by Burke and Grime (1996), MacGillivray et al. (1995), and Buckland (1994) and is the site of current climate change experi- ments conducted by the Unit of Comparative Plant Ecology, Sheffield University. The dale has a slope of approximately 30?, and an altitude of 370 m (National Grid Reference SK 057708). The experiment was lo- cated on the north-west facing side of the dale. There is a high annual rainfall evenly spread throughout the year and a short growing season between early May and late September. Virtually no growth occurs during the winter.

Two sites were selected within the dale. One of the sites was located on continuously moist infertile soil

supporting a relatively species-rich calcareous grassland community of slow-growing, mainly evergreen species. The second site was situated on an area where mush- room compost had been dumped ten years previously (MacGillivray 1993). This site consisted of a relatively high productivity patch of fast-growing perennials dominated by nettles (Urtica dioica). Table 1 lists the species that were found in both sites and their mean abundance.

Experimental design

Within each of the two sites selected, 48 0.5-m2 plots were staked out in April 1994. Half of the plots re- ceived a pesticide treatment and two-thirds of the plots were transplanted, either between sites or in situ, for a 2 x 3 factorial combination of six treatments. Each treatment was replicated eight times. The in situ trans- planted plots (turves) were simply removed and re- planted back into their home site, while the other transplanted plots were transferred into the opposite plant community (i.e. a reciprocal transplant). Each of the transplanted plots was excavated to a depth of approximately 200 mm. The control plots were undis- turbed. See Fig. 1 for a map of the study site.

Every 3 weeks, beginning April 1994, 5 g of mol- luscicide (Bio Slug pellets: active ingredient metalde- hyde), 5 g of systemic below-ground insecticide (Bio Chlorophos: active ingredients chloropyrifos and diazi- non), and 100 ml of a systemic above-ground insecti- cide (Bio Long-Last: active ingredients dimethoate and permethrin) were applied to half of the 0.5-m2 plots within each site until the final harvest in October 1994. In order to minimize spray drift, the pesticides were applied between 0400 and 0600 when winds were low. The pesticides were tested independently for possible stimulatory or phytotoxic effects on five of the domi- nant plants occurring within the study sites (Fraser 1996). It was found that the pesticides were lightly

OIKOS 81:1 (1998) 100

Table 1. Lists of species in each site and their mean relative abundance in eight 0.2-m2 control plots. Maximum relative growth rate (Rmax) from Grime and Hunt (1975). Canopy height (can ht) and specific leaf area (sla) from FIBS (Functional Interpretation of Botanical Surveys).

Grassland Nettle patch

% abundance Rmax can ht sla % abundance Rmax can ht sla

Festuca ovina 36.5 1.00 2 10.4 Urtica dioica 60.0 2.35 5 21.3 Carexflacca 19.3 1.38 2 15.6 Arrhenatherum elatius 27.3 1.30 5 31.3 Helictotrichon pratense 9.3 0.75 2 9.8 Elytrigia repens 3.2 1.21 3 23.4 Calluna vulgaris 7.2 0.35 4 11.0 Holcus lanatus 2.7 2.01 3 33.8 Potentilla erecta 7.1 0.83 2 22.7 Deschampsia flexuosa 2.1 0.81 2 16.1 Agrostis capillaris 6.8 1.48 2 28.7 Epilobium montanum 1.3 - 3 28.2 Galium saxatile 4.8 1.16 1 20.5 Rumex obtusifolius 1.1 1.49 2 31.1 Hypericum pulchrum 1.8 - 3 14.9 Poa trivialis 0.9 1.40 1 28.1 Anthoxanthum odoratum 1.7 0.94 2 23.4 Dactylis glomerata 0.6 1.31 3 24.0 Dactylis glomerata 1.3 1.31 3 24.0 Stellaria media 0.4 2.43 2 47.5 Sanguisorba minor 1.1 - 2 11.3 Senecio jacobaea 0.3 1.24 4 25.0 Agrostis vinealis 0.9 1.41 2 18.5 Agrostis stolonifera 0.1 1.48 2 35.5 Senecio jacobaea 0.7 1.24 4 25.0 Viola riviniana 0.4 0.65 2 Bromopsis erecta 0.3 - 2 19.1 Campanula rotundifolia 0.2 0.81 2 20.2 Helianthemum nummularium 0.2 0.70 2 5.0 Trifolium pratense 0.1 - 2 20.9 Pimpinella saxifraga 0.1 - 1 15.4 Holcus lanatus 0.1 2.01 3 33.8 Lotus corniculatus 0.1 1.05 2 19.7 Lathyrus montanus 0.1 0.49 2 19.7 Koeleria macrantha 0.1 0.94 1 16.2 Leontodon hispidus 0.1 0.89 2 23.9 Scabiosa columbaria 0.1 1.26 1 23.8

phytotoxic (the pesticides caused a reduction in biomass), but this result was only significant for one plant species (Urtica dioica).

The final harvest was in early October 1994, before the first frosts. A 200-mm2 quadrat was placed in the centre of each plot and all the above-ground parts of the living plants were harvested within the quadrat. The

plants were then separated into species, oven-dried at 80?C and weighed.

Analysis

A two-way ANOVA was applied to determine the effect of site (which included the transplanting treat- ment) and the pesticide treatment on the total above- ground biomass. Tukey's HSD test was applied to

separate the treatment means. Also, the 95% confidence limits were calculated for each treatment mean.

The effect of the pesticide treatment was calculated for each plant species at its site of origin and at the transplanted site and the species ranked according to their relative abundance. By doing this it was deter- mined which species benefited by the removal of herbi- vores and which species suffered. This procedure was not carried out on the turves which had been excavated then replaced into their home site because it was deter- mined through the ANOVA described above (and pre- sented in Tables 2 and 5 and Figs 1 and 5) that transplanting had no effect on plant biomass.

Linear regressions were carried out to examine the relationship between maximum relative growth rates (Rmax) available from laboratory screening tests (Grime and Hunt 1975) and the change in above-ground biomass due to the pesticide treatment. The Rmax was not known for all the species found in this study, so by necessity the analysis was limited to those species with available values. The Rmax values used in these analyses are presented in Table 1. This analysis was carried out on four subsets of the experimental data: nettle turves in situ; nettle turves transplanted to grassland; grass- land turves in situ; and grassland turves transplanted to the nettle patch.

Results

Nettle community

The nettle patch turves transplanted to the grassland suffered major herbivore attack compared with those remaining in the nettle patch (Fig. 2). Furthermore, the nettle turves that were transplanted to the grassland community responded dramatically to the pesticide treatment. Visible herbivore damage, including damage by leaf chewers and gall-forming insects, specifically the nettle midge (Dasineura urticae), was apparent on the plants not treated with pesticides, while the plants which were treated flourished. The mean dry weight of the total above-ground biomass of the nettle commu-

OIKOS 81:1 (1998) 101

0 m l >e * IN control

control + pesticide

transplanted in situ

transplanted in situ + pesticide

transplanted btw. sites

transplanted btw. sites + pesticide

Fig. 1. Layout of experimental design. Squares represent a 0.5-m2 area.

nity transplanted to the grassland community was 4.34

g without pesticides, and more than three times that, 14.34 g, with pesticides. In fact, this latter treatment was associated with the highest recorded mean biomass. There was no significant difference in the above-ground biomass between the different treatments within the nettle patch. The two-way ANOVA confirmed that the

pesticide treatment had a significant effect on the total

above-ground biomass, while the site treatment did not

(Table 2). The individual species response to the removal of

herbivores was calculated and ranked for both the plots in situ and the transplanted plots (Fig. 3). There was very little difference in the way the species responded to herbivore removal in both sites; the main differences were in the magnitude of change. Urtica dioica had an increase of 0.85 g in mean above-ground biomass with the addition of pesticides in the turves remaining in situ. In the transplanted turves, U. dioica had an in- crease of 4.57 g and Arrhenatherum elatius responded to a similar extent.

The linear regression model comparing the relation- ship between maximum relative growth rate (Rmax) and the change in total above-ground biomass mediated by the pesticide treatment was significant for the nettle turves transplanted into the grassland, but not for the nettle turves in situ (Table 3). The relationship between

Rmax and the change in biomass was positive and the model accounted for 68.3% of the variation within the population (Fig. 4).

Grassland community

The application of pesticides to the grassland turves resulted in effects which were quite different from those observed in the nettle turves. The effects of site and pesticide treatment on the total above-ground biomass were statistically significant (Table 4), but this did not apply to the combined effect of site and pesticides. There was an approximate 50% increase in mean dry weight of the grassland turves in situ (regardless of transplant treatment) when pesticides were applied (Fig. 5). Although this difference was not found to be significant using Tukey's HSD, it was certainly substan- tial. However, since it was not significant sampling error cannot be ruled out. The grassland turves trans- planted into the nettle community had very low biomasses, irrespective of pesticide application, com- pared to the grassland plots in situ.

The individual species response to the removal of herbivores in the grassland community also differed from the response of species in the nettle community. The biomass of most species in the grassland plots remaining in situ was greater with the application of the

OIKOS 81:1 (1998)

grassland

El * N& m

0 x C N

x E l E Il0

0 N s E

0

0

U

N

ttJ Ecl

0

m0 U

Nfieel

N m 0

Elx 1 0 m

102

Table 2. Results of two-way ANOVA examining the effects of site (SITE) and the pesticide treatment (PEST) on the mean above-ground biomass of the nettle turves. This table presents the sum of squares (SS), degrees of freedom (d.f.), mean squares (MS), F-ratios, and p-values (p).

Source of variation SS d.f. MS F-ratio p

SITE 23.288 2 11.644 0.840 0.439 PEST 123.649 1 123.649 8.920 0.005 SITE x PEST 277.298 2 138.649 10.003 0.000 ERROR 582.173 42 13.861

pesticides (Fig. 6a), particularly Festuca ovina. How- ever, the species response in the transplanted plots to the application of pesticides was almost completely opposite to the response of species in the grassland plots in situ (Fig. 6b). Instead of being the species most

benefiting from the removal of herbivores (Fig. 6a), F. ovina suffered the greatest decrease. Other species which showed a similar pattern were Carex flacca and An- thoxanthum odoratum. Some of the species that seemed to benefit from the application of pesticides in the transplanted grassland turves were species native to the nettle community (e.g. Urtica dioica, Poa trivialis, and Arrhenatherum elatius), and were not present in the

grassland turves in situ. The linear regression models between Rmax and the

change in total above-ground biomass caused by the

pesticide treatment were not significant (Table 5), even though the slope tended towards a positive relationship (Fig. 7).

Discussion On the basis of the responses to pesticide treatment Urtica dioica did not appear to suffer much loss

through herbivory within the nettle community and yet

it is a fast-growing herb that has highly nutritious, nitrogen-rich foliage (Davis 1975, 1991). When U. dioica was transplanted into the low-nutrient grassland community though, it was subject to herbivory. In contrast to the nettle community, the application of

pesticides to the grassland community resulted in a substantial increase in mean dry weight of total above-

ground vegetation. In other words, invertebrate herbi- vores seemed to be limiting the biomass of the

vegetation. When the grassland turves were trans-

planted into the nettle community there was no increase in mean dry weight with the addition of pesticides. The effect of transplanting (disturbance of the turves) can be discounted as a factor in this investigation because

a)

0.8-

0.6 -

.c 0.4-

0.2-

Urtica dioica turves in situ o

?e

.a .

m x^ T

?

- F-

25-

20 -

._ S 15-

S

C 10-

E 5-

0-

.2 -

D-P E+P b)6-

5-

4-

M

!. g T

Il

C CT T

Fig. 2. Effect of transplanting and pesticide treatment on the mean total shoot biomass (g) of the nettle turves. In this diagram P = pesticide treatment (+ present, - absent); C = control turves; CT = transplanted turves in situ; T = trans- planted turves. Error bars represent 95% confidence limits.

c o E

OuI E O Ui acE

Urtica dioica turves in grassland

.: B

3-

2-

1-

I- I I I

-1 ai . e.

0

c r m.0 m - .2 Z .

a . E 0 I i...FIlE II I 4 ) LiFi-I 1F- F

I

Fig. 3. Ranked mean abundance measured as the change in dry shoot biomass (g) mediated by the application of pesticides for each species occurring in turves (a) in situ, and (b) in the grassland site. See Table 1 for a key to species.

OIKOS 81:1 (1998)

u.U u ~ - - . .. . .. . .. 5 5-" , . I I n n

Uo-. L y I .. I

. . I

. . I

. I I

A O,- I . .. ',/// e,

I I I I I

I

103

Table 3. Results of linear regression examining the relationship between the maximum relative growth rate (Rmax) and the change in the above-ground biomass of the nettle turves due to the pesticide treatment. This table presents the sum of squares (SS), degrees of freedom (d.f.), mean squares (MS), F-ratios, and p-values (p).

Source of Dependent variable variation

Nettle turves at home Nettle turves in grassland

SS d.f. MS F-ratio p SS d.f. MS F-ratio p

Rmax 0.08 1 0.08 0.75 0.406 10.02 1 10.02 7.52 0.025 residual 1.01 10 0.10 10.66 8 1.33

the enm

pla

Ok (Fr 19c Ok et nit,

results show that there were no significant differ- the Fretwell-Oksanen theory, a community that is :es between the control turves and the turves trans- highly productive can support at least three trophic inted back into their site of origin (Figs 2 and 5). levels. A highly productive community with three The results from this study support Fretwell and trophic levels will be controlled by predators, i.e. it is :sanen's top-down hypothesis of community control 'top-down' controlled. This means that herbivores retwell 1977, 1987, Oksanen et al. 1981, Oksanen should have little influence on the vegetation because 90a, b). This is despite the fact that the Fretwell- their numbers are limited. In the case of the nettle :sanen model explicitly excludes ectotherms (Oksanen community herbivores did not have any significant al. 1981, 1987, Oksanen 1990a). The nettle commu- influence on the mean dry weight of the total above- y was a highly productive community. According to ground biomass. That does not mean invertebrate her-

bivores were entirely absent; pitfall traps confirmed that they were indeed present (Fraser 1996). In fact, herbi-

a). nettle turves in situ vores still seemed to exert some influence on the alloca- tion of growth of Urtica dioica. When herbivores were

0.8- R1R0.264 0 removed, a higher proportion of U. dioica flowered (Fraser 1996). But, overall, herbivores appeared to have

0.6- ? a negligible influence on the biomass of the nettle S~m~~~~~~ ~~~community. Of course, the primary predators within

0.4?- the nettle patch may well be endotherms (e.g. passerine pS^ ao~ ^ birds) but the results of this study show that the

0.2 - Fretwell-Oksanen model can certainly be extended to include ectotherms (see also Fraser and Grime 1997).

??0.0 -0

o The grassland community was at low productivity and

-02 o apparently could not support or sustain a large or

0.5 1.0 1.5 2.0 2.5 effective third trophic level. The evidence for this is that

Rmax when herbivores were removed, the mean dry weight of the total above-ground biomass increased substantially

b) nettle turves in grassland indicating that herbivores were limiting plant growth. If 5.0- a herbivores were limiting plant growth, they in turn

4.0 - R-0.696 must not have been limited greatly by carnivores. There could be three factors that contribute to limiting the

3.0 - abundance, and therefore control, of carnivores in the 0 ? / grassland community. The first factor has already been

c 2.0- stated, which is that there were not enough resources to ;@?~~~ /^^~~~ ~sustain an effective third trophic level. The second

1.0 - factor may be that the patches of high quality food > ? a ? (nettle) were readily detected by herbivores (see Davis

?0. - i/ D ? 1975), but the resulting patches of high herbivore den- sity were so small (in the spatial scale relevant for

0.5 1o 0 1.5 20 26 carnivores) that they were missed by carnivores. Fi- 0.5 1.0 1.6 2.0 2.5

nally, the third possibility may be that there were many ~R~~max ~refuges within the grassland community that protected 4. The relationship between the change in the mean total herbivores from predators. The architecture of the .. ... ^Therbivores from predators. The architecture of the ot biomass (g) due to the pesticide treatment and the

ximum relative growth rate (Rmax) for species present in the grassland community was certainly complex and multi- tle turves (a) in situ, and (b) in the grassland site. layered, more so than the nettle community, and it has

104 4IO

11(98

I

Fig sho ma) nett

C

C

I

104 OIKOS 81:1 (1998)

Table 4. Results of two-way ANOVA examining the effects of site (SITE) and the pesticide treatment (PEST) on the mean above-ground biomass of the grassland turves. This table presents the sum of squares (SS), degrees of freedom (d.f.), mean squares (MS), F-ratios, and p-values (p).

Source of variation SS d.f. MS F-ratio p

SITE 494.36 2 247.18 9.237 0.000 PEST 140.77 1 140.77 5.261 0.027 SITE x PEST 90.00 2 45.00 1.682 0.198 ERROR 1123.868 42 26.76

been shown in predator-prey experiments that refuges for herbivores can be critical for their survival (Huf- faker 1958, Kareiva 1990, Hixon and Beets 1993). Another factor to account for is that the complex architecture of the grassland may simply afford a more suitable environment, regardless of predators, for the major invertebrate herbivores (e.g. molluscs) (see Van de Koppel et al. 1996 for a comparative case).

When turves of the nettle community were trans- planted into the grassland community, top-down con- trol of herbivores was effectively removed. It was assumed that these transplanted turves were too small for viable animal populations from the nettle commu-

nity to persist and establish in the grassland. The effect of transplanting the nettle turves to the opposing site was to increase the influence of herbivores from the

grassland community on the transplants, and the effect was dramatic. Firstly, it proved that Urtica dioica was

very palatable, certainly more palatable than the low nutrient grassland species, otherwise the loss of biomass of the nettle community would have been similar to the losses experienced by the grassland community. It also confirmed that herbivores were not very effective in the nettle community because U. dioica experienced little loss of biomass there.

25-

20 -

. 15-

00% - '

10-

E -

5-

n_

D -P

+P-

/p-

/ C CT T

Fig. 5. Effect of transplanting and pesticide treatment on the mean total shoot biomass (g) of the grassland turves. In this diagram P = pesticide treatment (+ present, - absent); C = control turves; CT = transplanted turves in situ; T = trans- planted turves. Error bars represent 95% confidence limits.

Turves from the grassland community transplanted into the nettle community did not respond as expected. According to the principles already outlined, the trans-

plants should have been released from herbivore pres- sure because of top-down control of herbivores. Yet, there was no increase in mean dry weight of total

above-ground vegetation. On the contrary, there was a decrease, even when pesticides were applied. In retro-

spect, this can be explained by competitive suppression of the transplants through shading and (less likely) mineral nutrient capture from the fast-growing com- petitors surrounding the turves. In fact, some of the small turves were actually invaded by competitors, such as U. dioica and Poa trivialis, from the surrounding nettle community. This result actually corresponds to the predictions of the Fretwell-Oksanen model (Ok- sanen et al. 1981, Oksanen 1990a). Perhaps, though, if larger turves had been used, and shading minimized by cutting surrounding vegetation, an increase in mean dry weight would have occurred. However, an interesting outcome was revealed as a result of the competitive pressure on the transplants. If there was no competition on the transplants from the surrounding vegetation in the nettle community, we would have expected both the transplants with and without pesticides to behave simi-

larly to the undisturbed plots in the grassland commu-

nity that had been applied with pesticides. However, there was a difference in the performance

of species between the transplants applied with pesti- cides to those without pesticides. Even though herbi- vores were controlled in the nettle community, there were herbivores present. It seems that those few herbi- vores were benefiting slow-growing, relatively unpalat- able plants, e.g. Festuca ovina, Carex flacca and Anthoxanthum odoratum. When herbivores were re- moved through the application of pesticides, the fast-

growing competitors, U. dioica, P. trivialis, and Arrhenatherum elatius benefited. A similar result with the same plant species was also found in an outdoor microcosm study (Fraser and Grime 1998). As well, Ostfeld and Canham (1993) showed that fast-growing tree seedlings were released from intense herbivory from voles following vole removal in an old field.

The relationship between the maximum relative growth rate (Rmax) and the change in total biomass due to the pesticide treatment was observed fairly consis-

OIKOS 81:1 (1998)

I I

PI

/~

/-- /0

/1

/1

/l /0

I 00 10, 00 00, 00, 0 00, 00, 000

010

ooo 11 ooo

105

grassland turves In situ Fig. 6. Ranked mean abundance measured as the change in dry shoot biomass (g) mediated by the application of pesticides for each species occurring in the grassland turves (a) in situ, and (b) in the grassland site. See Table 1 for key to species.

?-o. ~ ~ ~ ~ ~ . x

Li i .i [ , : F FF - - - - I

- - . . . . i * ,ij t -

grassland turves in nettle patch

i i

5 t xsa - 0 4

tently within each of the four different subsets of data to which the analysis was applied, i.e. nettle turves in situ; nettle turves transplanted into grassland; grass- land turves in situ; grassland turves transplanted into the nettle patch. In each case, the slope of the regres- sion was positive, meaning that those species with a

higher Rmax had a greater biomass when pesticides were applied. In other words, they responded more quickly than slower growing species to a release of herbivore pressure. However, this relationship was only significant for the model applied to nettle turves trans- planted into the grassland site. This is understandable assuming that the herbivore pressure within the grass- land was much higher than within the nettle patch. Plants with a high Rmax are more palatable than those with lower Rmax's, therefore the herbivores within the grassland would prefer those species with a higher Rmax. In addition, the nettle turves that were trans- planted into the grassland would have received much more light than if they remained within their home site

because the grassland species were much shorter. Plants with a higher Rmax should be able to capitalize on this added resource more rapidly than other plants.

Why was not Rmax a reliable indicator of species response to the release of herbivore pressure in the low

productivity turf environment? Perhaps the answer is that stress-tolerant plants with low relative maximum

growth rates were the dominant species. Unless com-

petitor-type species with high Rmax's were well estab- lished in the community, the plant biomass of the

community would be less likely to quickly respond to the pesticide treatment. A similar response occurs when fertilizers are applied to an established community of slow-growing plants. Rather than an increase in growth rate, slow-growing stress-tolerators accumulate the extra nutrients in their tissue (Chapin 1980). Com- petitor species with high Rmax's were dominant in the nettle turves transplanted into the grassland, and there- fore were able to respond quickly to the pesticide treatment.

OIKOS 81:1 (1998)

a) 3-

2-

'.,-. 0O

cri - Q)

1-

0 'I I , I

_dx o . i

-1 - u

b) .4

0.2

-0.2

.4 B-0

-0.6

-0.8

I I -I I I I I I I I I I I I I II

I

I I I I I I I I I I I I I I

106

Table 5. Results of linear regression examining the relationship between the maximum relative growth rate (Rmax) and the change in the above-ground biomass of the grassland turves due to the pesticide treatment. This table presents the sum of squares (SS), degrees of freedom (d.f.), mean squares (MS), F-ratios, and p-values (p).

Source of Dependent variable variation

Grassland turves at home Grassland turves in nettle patch

SS d.f. MS F-ratio p SS d.f. MS F-ratio p

Rmax 0.26 1 0.26 0.68 0.419 0.09 1 0.09 1.34 0.266 residual 6.89 18 0.38 1.06 15 0.07

a) grassland turves in situ References

*2-- 0 Balme, O. E. 1953. Edaphic and vegetational zoning on the 2.0 o Ro0.191 carboniferous limestone of the Derbyshire Dales. - J. Ecol.

41: 331-344. ~~~~~~1.~6 ~~~- ~Buckland, S. M. 1994. An investigation into the responses of limestone grassland to climatic manipulation and species

CDa5~~~~~~~ 1.~0 - ~introduction. - Ph.D. thesis, Univ. of Sheffield. -.I Burke, M. J. and Grime, J. P. 1996. An experimental study of o 06 - o o plant community invasibility. - Ecology 77: 776-790.

$o ~

,--^Chapin III, F. S. 1980. The mineral nutrition of wild plants. -

??.-0B ?0 o Annu. Rev. Ecol. Syst. 11: 233-260. *.0-P a? D Coley, P. D., Bryant, J. P. and Chapin III, F. S. 1985.

40._- Resource availability and plant antiherbivore defense. - * 0"~~~ zo~~~ ~Science 230: 895-899.

Davis, B. N. K. 1975. The colonization of isolated patches of ..0 05 1. 20 26 nettles (Urtica dioica L.) by insects. - J. Appl. Ecol. 12:

1-14. Rmax - 1991. Insects on nettles. - Naturalists' handbooks, Slough,

Richmond. b) Feeny, P. 1976. Plant apparency and chemical defense. - Rec.

grassland turves in nettle patch Adv. Phytochem. 10: 1-40. 1.0- Fraser, L. H. 1996. Plant-animal interactions: the effects of

R-0.286 invertebrate herbivory and productivity on plant commu- nity dynamics examined in outdoor microcosms and two

0.5 - North Derbyshire dales. - Ph.D. thesis, Univ. of Sheffield. - and Grime, J. P. 1997. Primary productivity and trophic

-,a>~~~~~~~~ ~~~~dynamics investigated in a North Derbyshire, UK, dale. -

o.o- o z P Oikos 80: 499-508. 0 o.o - ? _ P o_o - and Grime, J. P. 1998. Top-down control and its effect on

0 0 the biomass and composition of three grasses at high and low soil fertility in outdoor microcosms. - Oecologia (in

-06- 0o press). Fretwell, S. D. 1977. The regulation of plant communities by

food chains exploiting them. - Perspect. Biol. Med. 20: 0 169-185.

-1.0 - 1987. Food chain dynamics: the central theory of ecology? 0 0.6 1.0 1.6 2.0 2.6 - Oikos 50: 291-301.

Rmax Grime, J. P. 1974. Vegetation classification by reference to strategies. - Nature 250: 26-31.

. 7. The relationship between the change in the mean total - 1977. Evidence for the existence of three primary strategies ot biomass (g) due to the pesticide treatment and the in plants and its relevance to ecological and evolutionary imum relative growth rate (R,mx) for the grassland turves theory.- Am. Nat. 11: 1169-1194. in situ, and (b) in the grassland site. - 1979. Plant strategies and vegetation processes. - Wiley,

New York. - and Hunt, R. 1975. Relative growth rate: its range and

adaptive significance in a local flora. - J. Ecol. 63: 393- 422.

-, MacPherson-Stewart, S. F. and Dearman, R. S. 1968. An investigation of leaf palatability using the snail Cepaea

nowledgements - I wish to thank J.P. Grime for his nemoralis L. - J. Ecol. 56: 405-420. ouragement, support, and advice throughout this study. I - , Cornelissen, J. H. C., Thompson, K. and Hodgson, J. G. grateful to L. Oksanen, S. E. Hartley, and J. H. C. 1997. Evidence of a causal connection between anti-herbi-

nelissen for insightful, useful comments, to J. G. Hodgson, vore defence and the decomposition rate of leaves. - Oikos Thompson and V. K. Brown for species identification, and 77: 489-494. . A. Wells for her tireless help in the field. This work was Hixon, M. A. and Beets, J. P. 1993. Predation, prey refuges, ported by a post-graduate Univ. of Sheffield fellowship and and the structure of coral-reef fish assemblages. - Ecol. NERC Unit of Comparative Plant Ecology. Monogr. 63: 77-101.

OIOf11(98 0

Fig. sho( max (a)

Ack ence am Cor K. to J sup] the

I

.9

I II

I

107 OIKOS 81:1 (1998)

Huffaker, C. B. 1958. Experimental studies on predation: dispersion factors and predator-prey oscillations. - Hilgar- dia 27: 343-383.

Kareiva, P. 1990. Population dynamics in spatially complex environments. Theory and data. - Philos. Trans. R. Soc. Lond. B 330: 175-190.

Lloyd, P. S. 1972. The grassland vegetation of the Sheffield region II. Classification of grassland types. - J. Ecol. 60: 739-776.

-, Grime, J. P. and Rorison, I. H. 1971. The grassland vegetation of the Sheffield region. - J. Ecol. 59: 863-886.

MacGillivray, C. W. 1993. Extreme climatic events and plant community structure. - Ph.D. thesis, Univ. of Sheffield.

-, Grime, J. P., Band, S. R., Booth, R. E., Hendry, G. A. F., Hillier, S. H., Hodgson, J. G., Hunt, R., Jalili, A., Mackey, J. M. L., Mowforth, M. A., Neal, A. M., Ror- ison, I. H., Spencer, R. E., Thompson, K. and Thorpe, P. C. 1995. Testing predictions of the resistance and resilience of vegetation subjected to extreme events. - Funct. Ecol. 9: 640-649.

Moen, J. and Collins, S. L. 1996. Trophic interactions and plant species richness along a productivity gradient. - Oikos 76: 603-607.

- , Gardfjell, H., Oksanen, L., Ericson, L. and Ekerholm, P. 1993. Grazing by food-limited microtine rodents on a productive experimental plant community: does the 'green desert' exist? - Oikos 68: 401-413.

Oksanen, L. 1990a. Predation, herbivory and plant strategies along gradients of primary productivity. - In: Tilman, D. and Grace, J. (eds), Perspectives on plant competition. Academic Press, New York, pp. 445-469.

- 1990b. Exploitation ecosystems in seasonal environments. - Oikos 57: 14-24.

-, Fretwell, S. D., Arruda, J. and Niemela, P. 1981. Ex- ploitation ecosystems in gradients of primary productivity. - Am. Nat. 118: 240-261.

-, Aunapuu, M., Oksanen, T., Schneider, M., Ekerholm, P., Lundberg, P. A., Armulik, T., Aruoja, V. and Bondestad, L. 1997. Outlines of food webs in a low arctic tundra landscape in relation to three theories on trophic dynamics. - In: Gange, A. C. and Brown, V. K. (eds), Multitrophic interactions in terrestrial systems. Blackwell, Oxford, pp. 351-373.

Ostfeld, R. S. and Canham, C. D. 1993. Effects of meadow vole population density on tree seedling survival in old fields. - Ecology 74: 1792-1801.

Pearce, A. K. 1987. An investigation of phenotypic plasticity in coexisting populations of four grasses in a calcareous pasture in North Derbyshire. - Ph.D. thesis, Univ. of Sheffield.

Rhoades, D. F. and Cates, R. G. 1976. Towards a general theory of plant antiherbivore chemistry. - Rec. Adv. Phy- tochem. 10: 168-213.

Silander, J. A. and Antonovic, J. 1982. Analysis of interspe- cific interactions in a coastal plant community - a pertur- bation approach. - Nature 298: 557-560.

Southwood, T. R. E., Brown, V. K. and Reader, P. M. 1986. Leaf palatability, life expectancy and herbivore damage. - Oecologia 70: 544-548.

Turkington, R. 1989. The growth, distribution and neighbour relationships of Trifolium repens in a permanent pasture. V. The coevolution of competitors. - J. Ecol. 77: 717-733.

Van de Koppel, J., Huisman, J., Van der Wal, R. and Olff, H. 1996. Patterns of herbivory along a productivity gradient: an empirical and theoretical investigation. - Ecology 77: 736-745.

OIKOS 81:1 (1998) 108