SEARCH

SEARCH BY CITATION

Keywords:

  • co-existence;
  • dispersal;
  • migration;
  • persistence;
  • Rhopalomyia californica;
  • stability

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

1. Theoretical studies predict that limited amounts of dispersal of individuals in host-parasitoid systems can both enhance the stability of the subpopulations and promote the co-existence of competing parasitoid species. We investigated the effects of dispersal on the population dynamics and parasitoid community structure of a natural host-multi-parasitoid system consisting of the midge Rhopalomyia californica that forms galls on the shrub Baccharis pilularis and the parasitoids that attack the midge.

2. An experiment involving the release of midges into a field with a low background density of galls demonstrated that the midges, on average, travelled approximately 1·7m in their lifetime. This suggests that the appropriate spatial scale at which to look at the effects of dispersal is relatively small.

3. Dispersal of midges and parasitoids between individual bushes was experimentally eliminated in a caging experiment. The midge populations in all of the uncaged replicates displayed dynamics that were similar to each other, while the dynamics of the midge populations in the caged replicates diverged. The midge dynamics on the uncaged bushes were not significantly more stable than those on the caged bushes.

4. Dispersal among bushes was found to play a major role in co-existence of the competing parasitoid species. There was a dramatic drop in the parasitoid species diversity on the caged bushes, with only a single parasitoid species, Platygaster californica, persisting at high numbers in the caged populations. In accordance with theoretical models, P. californica is the parasitoid species in the community that has the highest attack rate and is most effective at searching for hosts in a restricted area. Alternative explanations for this pattern are discussed.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

A great deal of recent theoretical work has been devoted to the role that dispersal may play in the population dynamics of insect–parasitoid and predator–prey systems (Gilpin & Hanski 1991; Hanski & Gilpin 1997). This intense theoretical interest stems from the fact that dispersal potentially offers a simple solution to two fundamental problems in ecology. First, it suggests a possible explanation for the persistence and stability of coupled predator–prey and parasitoid–host systems. Simple consumer-resource models are inherently unstable, yet real populations frequently persist and are often remarkably stable (Murdoch 1994). Models predict that consumer and resource populations that are divided into spatially isolated patches (or subpopulations) can be stabilized by a limited degree of movement between the patches (e.g. Reeve 1988, 1990; Taylor 1988, 1991; Hassell, Comins & May 1991; Sabelis, Diekmann & Jansen 1991). Analogous results have been obtained from models that follow individuals in a lattice in which there are limited movement rates of the consumers and resource, but in which there are no explicit patch boundaries (De Roos, McCauley & Wilson 1991; McCauley, Wilson & De Roos, 1993, 1996; Wilson, De Roos & McCauley 1993). In multi-species models, the term ‘metapopulation’ has been applied both to the situation in which subpopulations in patches frequently go extinct and are recolonized through immigration, and to the situation where dispersal has the effect of stabilizing otherwise unstable dynamics within the patches.

Secondly, dispersal can offer a mechanism by which multiple competing species can co-exist on a single limiting resource. The question of how competitors co-exist is particularly acute in natural insect–parasitoid systems in which there is often a very diverse assemblage of parasitoid species all attacking a shared host species. In the gall-forming midge system that will be discussed in the current paper, the single species of host is attacked by six common species of parasitoids, most of which are specialists on that host. Models suggest that limited dispersal rates can allow for co-existence of competitors in general (Slatkin 1974; Hanski 1983; Tilman 1994; Tilman, Lehman & Yin 1997) and specifically competing parasitoid species (Hassell, Comins & May 1994; Comins & Hassell 1996). In these models, co-existence is most likely if there is a trade-off between dispersal ability and the ability of the parasitoids to compete within a subpopulation.

Given the great deal of recent theoretical work on the effect of dispersal, there is a notable scarcity of supporting empirical evidence on the effects of dispersal from real predator–prey or parasitoid–host systems, and even less from experiments conducted in natural systems in the field. A few, often cited, laboratory experiments provide some indication that spatial subdivision and migration may enhance persistence (Huffaker 1958; Huffaker, Shea & Herman 1963; Pimental, Nagel & Madden 1963), and see the recent examples by Holyoak & Lawler (1996a,b). However, Taylor (1991) searched the literature for potential examples of metapopulation dynamics in predator–prey or parasitoid–host systems in the field, and found no cases that met all of the assumptions of models and that convincingly showed that movement among subpopulations was important to the persistence of the system. One reason for the lack of evidence is that few studies have been designed specifically to test for such an effect. Of the 13 possible examples that Taylor examined, only two were from experimental studies directly testing for the effect of dispersal (Kareiva 1984, 1987; Walde 1991). A more recent survey (Harrison & Taylor 1997) reveals that since Taylor's (1991) review and ‘call to arms’, there have been only two further published studies that experimentally test the effects of spatial subdivision and dispersal on population dynamics in host–parasitoid or predator–prey systems in the field (see Walde 1991; Murdoch et al. 1996). Another consequence of limited host mobility was investigated in Maron & Harrison's recent study (1997) of spatial pattern formation in the tussock moth, Orgyia vetusta. They illustrated that localized outbreaks of the moth could be maintained by the action of more mobile parasitoids. However, we know of no studies that have experimentally demonstrated the role of dispersal in promoting co-existence of competing parasitoid species in the field.

The paucity of experimental studies is partially due to the difficulty in determining what is the appropriate spatial scale at which to look at the effect of dispersal. In models with explicit patches, the appropriate spatial scale is the subpopulation, which is defined as the scale within which most of the processes such as foraging and reproduction occur, but between which movement is relatively rare. Models that follow discrete individuals (e.g. De Roos et al. 1991) define the ‘characteristic spatial scale’ imposed by the individual movement rates, as the scale below which the interaction resembles a well-mixed system. Determination of this scale requires more detailed information about the dispersal distances of hosts and parasitoids than is available for most insect systems (but see Stein et al. 1994; Corbett et al. 1996; Jones, Godfray & Hassell 1996). In addition, few natural systems appear to fit neatly into the scenario envisioned by theoreticians. In many systems the boundaries between subpopulations cannot be clearly defined and parasitoids are often more mobile than their hosts.

In this paper we respond to this scarcity of experimental studies by investigating the role of dispersal in the population dynamics of a gall-forming midge Rhopalomyia californica Felt (Diptera: Cecidomyiidae) and its parasitoids. One obvious first choice for the spatial scale of the investigation in this system is the scale of the individual bush. R. californica forms galls only on the shrub Baccharis pilularis De Candolle (Asteraceae). The habitat of the midge, therefore, naturally consists of patches with B. pilularis shrubs that are suitable for oviposition separated by unsuitable patches of open space or in which other plant species are present. The juvenile stages of the midge and all of its parasitoids occur within galls, so only the adults are mobile. Female midges generally live 1 day or less as adults and appear to be fairly weak flyers, so may not disperse far in most cases. Observations suggest that midges start ovipositing as soon as they come into contact with B. pilularis tips. A single female midge visits a number of tips on a bush, laying eggs in clusters on between 1 and 14 plant tips (mean = 4·5 ± 0·2, at the Big Creek survey location; Latto & Briggs 1995). There is, however, a possibility of longer distance wind dispersal. The various parasitoids that attack the midge are longer lived [with adult lifespans ranging from 1 week to almost a month (Force 1970)], so probably disperse farther.

We first present data from a release experiment showing that most midges disperse only very short distances during their lifetime. This suggests that an individual Baccharis bush may be a reasonable spatial scale at which to examine the dynamical role of dispersal. We then give the results of a field enclosure experiment in which the populations of midges and parasitoids on individual Baccharis bushes were isolated from the effects of immigration and emigration. The elimination of dispersal can potentially affect the populations in a number of ways, and we test a number of hypotheses. (1) High rates of dispersal can synchronize the dynamics among bushes, so subpopulations on isolated bushes may become less synchronous than those that remain connected by migration. (2) If the dynamics of midges on individual bushes are being stabilized by dispersal, then subpopulations should become less stable when isolated in cages (Murdoch et al. 1996). (3) If movement of parasitoids between bushes is necessary for the co-existence of competing parasitoid species, then isolated subpopulations in cages may lose parasitoid species. If this is the case, then only the parasitoid species that is most efficient at searching for hosts at a small spatial scale may persist (Hassell et al. 1994; Comins & Hassell 1996). In this paper, we examine all of these potential effects of dispersal on the midge and parasitoid population dynamics and community structure.

In this paper we are concentrating on the host–multiparasitoid interaction and the effects of dispersal on this interaction. However, the host insect in this system is a gall-forming insect, and in a number of other studies of gall-formers ‘bottom-up’ plant effects have been shown to override the ‘top-down’ effects of parasitoids on the insect abundance and population dynamics (Price et al. 1990; Hunter & Price 1992; Stiling & Rossi 1997). Elsewhere, we have shown that the parasitoids in the Baccharis system impose a very high mortality rate on the midge, and that parasitoids depress the midge population abundance by several orders of magnitude below the parasitoid-free levels (Briggs 1993; J. Latto & C. J. Briggs, unpublished). In the absence of the parasitoids, midge galls can reach extremely high densities and can potentially kill their host plant, however, these extreme densities are rarely observed in natural field conditions in which the parasitoids are always present. Furthermore, a number of authors have discussed the strong competitive interactions between the different parasitoid species that attack the midge (Force 1970, 1974; Ehler 1982, 1992; J. Latto & C. J. Briggs, unpublished). Thus, in this system there are strong interactions, both between the parasitoids and the host, and between the competing parasitoid species, which theory would predict might lead to unstable dynamics or extinctions.

In many other gall systems, the gall-formers can attack only very specific portions of their host plant, and for only a restricted period of time during the year, which leads to an increase in the importance of bottom-up control. For example in their studies of Euura sawflies on willows, Price and colleagues (1990) found strong intra-specific competition between female sawflies for oviposition sites on long, young growing shoots of the plant (Price et al. 1990; Hunter & Price 1992). This resulted in the populations being constrained by the availability of these shoots, with parasitoids having little effect. However, in the Baccharis gall-midge system, the plant puts on new growth throughout the year, the midge will readily lay eggs on any plant tip and galls can be produced on any apical meristem (terminal or axillary buds, including leaf or flower buds). Tests carried out on seedlings in the laboratory (Doutt 1961) and on mature plants in the field (Latto & Briggs 1995) indicate that all plants, and the vast majority of growing tips on the plants, are susceptible to attack by the midge. For these reasons we think that bottom-up effects may be of less importance in this system. In a gall-forming system that is similar in many ways to the Baccharis gall midge system, Stiling and Rossi (1997) found a large top-down effect of parasitoids on the abundance of Asphondylia borriciae in their ‘high nitrogen’ treatments. In their system, parasitoids had no effect only in treatments in which productivity was so low that virtually no galls were present.

Natural history

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

B. pilularis is a perennial evergreen shrub, common in recently disturbed areas and coastal bluffs throughout central and southern California. Adult females of R. californica lay clusters of eggs on terminal buds and growing tips of B. pilularis. The eggs remain on the exterior of the plant and, when the larvae hatch, they work their way between the bud scales and the plant is stimulated to form a gall around a cluster of larvae. Multi-chambered galls can contain from one to over a hundred developing midge larvae (mean = 8·6 ± 0·4 chambers/gall at the Big Creek Reserve site; Latto & Briggs 1995). The entire midge development time can be as short as 30 days on rapidly growing plants under favourable conditions (Force 1970; Hopper 1984), but can take over 70 days in the field (Briggs & Latto 1996). The midge eggs and larvae are subject to parasitism by a suite of six common parasitoid species with total parasitism rate usually reaching over 80% in the field (Force 1970, 1974; Ehler 1982; Hopper 1984; Briggs 1993). There are continuous, overlapping generations of the midge and its parasitoids, with all developmental stages present throughout the year.

One commonly occurring parasitoid species, Platygaster californica Ashmead (Platygasteridae), attacks the midge eggs while they are still on the outside of the plant. Parasitized eggs develop into parasitized larvae that produce galls along with the unparasitized larvae. P. californica has the highest attack rate of any of the common parasitoid species (Force 1970; Hopper 1984; Briggs & Latto 1996), but it is inferior in larval competition to all of the other parasitoid species (Force 1970). All of the other common parasitoid species attack the midge larvae within galls. These include Torymus koebelei Huber (Torymidae), Torymus baccharidis Huber (Torymidae), Zatropis capitis Burks (Pteromalidae), Eupelmus inyoensis Girault (Eupelmidae) and Mesopolobus sp. (Pteromalidae).

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Midge release experiment

In order to determine the typical distance that midges disperse within their lifetime, we conducted a release experiment in a field of B. pilularis bushes on the western edge of the University of California, Santa Barbara, campus (34°20′N, 119°52′W). Prior to the strong rains in 1994, the midge, R. californica, was virtually absent from the Santa Barbara region and was found primarily north of Point Conception, CA. Since the period of uncharacteristically wet weather in 1993–94, R. californica has been present in the Santa Barbara region at low densities.

In a field that had a very low background density of galls, we released approximately 2000 adult female midges onto a central bush between 2 and 10 April 1996. These midges were reared from a large field collection of galls and were released by simply allowing the midges to fly out of a container placed in the central branches of the bush. At this time, the spatial arrangement of B. pilularis bushes and the background density of galls at the site were mapped out.

On 11 and 12 June, 1996 (62 days after the last of the midges were releases) we recorded the number and position of all galls on all bushes in the field. Gall number per bush was converted to galls/m2 of plant surface area by dividing gall counts by the surface area of the bush [(4/3)π(diameter/2)2× height]. At this point, the galls were fully developed, but the next generation of adults had not yet emerged. We removed all of the galls from all of the bushes except the central release bush. The central bush had a large number of galls (2434 galls), which allowed us to monitor the dispersal of the midges for a further generation. In this later generation, the adult midges emerged naturally from the galls, rather than from a container. On August 5 and 6 1996 (after an additional 54 days), we again recorded the number and position of all galls on all bushes in the field. Thus, two measurements were made of the spatial distribution of galls created by midges dispersing from the central bush: first, the adult midges emerged from a plastic box placed in the bush and, secondly, the midges emerged naturally from the galls on the bush.

Field enclosure experiment

At the University of California's Big Creek Reserve (36°04′N, 121°36′W) near Big Sur, CA, a caging experiment was conducted to determine the effect of isolating subpopulations of midges and parasitoids on individual bushes on their synchrony, temporal variability and parasitoid community structure. This site was a coastal hilltop with an elevation of approximately 200 m. All vegetation was cleared from the site 5 years previously, so equal aged, resprouted B. pilularis shrubs dominated the vegetation. Ten B. pilularis bushes and the midges and parasitoids naturally present on them were used in the experiment. Five of the bushes were enclosed in 1 m×1 m×1 m cages composed of fine mesh (180 µ Tetko polyester multi-filiment) over wooden frames. The bases of the cages were buried in the ground, and two Velcro-sealed openings allowed for censusing of the populations. The five other bushes were tagged and left uncaged.

At approximately monthly intervals from June 1991 to October 1992, the midge galls on each bush were counted. The number of midge larval chambers per gall, and the percentage parasitism were measured each census date by collecting and dissecting all old galls from which the insects had emerged. Old galls were censused to avoid any effect of repeated destructive sampling on the population dynamics. The pupal case of the midge is distinguishable from parasitoid pupal cases, so dissection of the galls could reveal if a midge or a parasitoid emerged from each chamber of a gall. This method was used to accurately measure only total percentage parasitism, not parasitism by each species. Throughout the experiment, the species identity of any adult parasitoid observed in the cages or on the uncaged bushes was recorded. At the end of the experiment, all galls remaining on the bushes were collected and the insects reared to determine which parasitoid species were still present.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Midge release experiment

The adult female midges showed very limited dispersal (Fig. 1). A total of 4613 galls were recorded in the first generation and 3137 galls in the second. Most of these galls were formed on the central bush onto which the adults were released. This was true both in the first generation when the adults were released from a plastic box (53% of galls were on the central bush), and in the second generation when the adults emerged naturally from galls (58% of galls were on the central bush). At distances greater than approximately 7 m, any increase in density is indistinguishable from the background density in the field.

image

Figure 1. Distribution of galls at different distances from the central bush in the midge release experiment, after the first (solid circles) and second (open circles) generations. Solid and dashed lines show for the first and second generations, respectively, the fits of an exponential distribution to the densities on bushes near the central bush (at distances < 7 m) and the average background density of galls (distances ≥ 7 m).

Download figure to PowerPoint

On nearby bushes (those < 7 m the release bush), gall density decreased exponentially with distance from the release bush (first generation: gall density = 124 exp (– 0·59×distance), r2 = 0·76; second generation: gall density = 85 exp (– 0·57×distance), r2 = 0·60). The slopes for the loge (density) vs. distance relationships were not significantly different between the first and second generations (t-test, P > 0·5), suggesting that the release of adult midges from plastic boxes did not significantly alter midge dispersal. Thus, assuming that gall densities are a reflection of the distance travelled by the female midges, then this implies that the average distance travelled from the release bush was only about 1·7 m.

Field enclosure experiment

The dynamics in the caged and uncaged treatments were virtually identical for the first 250 days of the experiment (up to March 1992; Fig. 2). The abundance of the midge on all of the uncaged bushes decreased during the second half of the time series, coinciding with the general decrease in midge abundance recorded throughout the region in the field survey. One of the caged replicates (caged C5) was destroyed in a storm after 180 days. After about 250 days the abundance of the midge in the four remaining caged treatments drifted apart. There was an outbreak of the midge in one of the caged replicates (C1), in which almost every available growing tip of B. pilularis was attacked by a gall. By the end of the experiment, parasitism had also increased in this cage and virtually all midge larvae were parasitized (Fig. 3). Midge abundance steadily declined in another replicate (C4), and no midges or new galls were found during the last census. In the other two cages (C2 and C3), the abundance remained fairly constant.

image

Figure 2. . Density of midge larvae (+ 1) on the uncaged and caged bushes in the field enclosure experiment. Cage numbers given are those referred to in the text.

Download figure to PowerPoint

image

Figure 3. . Fraction of hosts parasitized by all parasitoid species on the uncaged and caged bushes in the field enclosure experiment. Cage numbers given are those referred to in the text.

Download figure to PowerPoint

By the end of the experiment, the variance in log-transformed counts of midge abundance among the four caged replicates was significantly greater than the variance among the five uncaged replicates [F-test to compare differences between two variances of final log (n + 1) counts of midge larvae: F3,4 = 16·9, P < 0·05]. The variability in midge abundance among the caged and uncaged replicates started out similar to each other, and remained so for about five generations (300 days) before the caged replicates started to diverge (Fig. 4). The midge populations on uncaged bushes that remained connected by migration stayed similar to each other throughout the duration of the experiment. Thus, the first hypothesis on the effects of dispersal presented in the Introduction was supported: migration had a synchronizing effect on the dynamics of midge subpopulations on individual bushes.

image

Figure 4. . Variance between replicates of log(midge larva counts + 1) on caged and uncaged bushes.

Download figure to PowerPoint

There was no significant difference between the average cumulative variability of the midge populations on caged and uncaged bushes (average cumulative variance of log(n + 1) counts of midge larvae: caged = 3·31, uncaged = 2·85, t-test, P > 0·5). Therefore, the second hypothesis on the effects of dispersal, and a major prediction of metapopulation models was not met in this experiment: subpopulations that were isolated from the effects of immigration and emigration did not show an increase in temporal variability as measured by the cumulative variance.

The fraction parasitized on all of the uncaged bushes remained relatively high throughout the experiment (mean = 0·80, SD = 0·18), and in the range normally found for R. californica in the field (Fig. 3). The fraction parasitized was initially relatively high on all of the caged bushes (except cage C1), but became more variable through time. In the cage where the midges appeared to go extinct (C4) parasitism rates were high throughout, even when midge densities had fallen to low levels. By contrast, the steady increase in abundance of the midge in C1 may have been due to the marked decrease in parasitism rate in that cage after day 260. As the number of midge larvae in that caged increased to over 44,000, the fraction parasitized dipped to less than 20%. The parasitoids did slowly respond to this increase in host abundance, and parasitism rates returned to very high levels (> 95%) by the final sample date. The fraction parasitized also declined through time in cage C2, possibly resulting in the steady rise in midge abundance in that cage.

The parasitoid species diversity in the caged replicates decreased markedly during the experiment (Fig. 5). Therefore, the third hypothesis on the effects of dispersal presented in the Introduction was supported: dispersal promoted the co-existence of multiple parasitoid species at the scale of the individual B. pilularis bush. Because we collected and dissected only old galls (to avoid the problems of repeated destructive sampling of galls), the only information that we have on parasitoid species diversity are the records of parasitoids observed on the uncaged and caged bushes throughout the experiment, plus an exhaustive census of all parasitoids present on all bushes during the last sampling date. Any parasitoid species observed on the caged bushes during the course of the experiment were necessarily present in the galls or on the surface of the plant when the cages were initially erected, but the parasitoids observed on the uncaged bushes may have arrived at any time during the experiment.

image

Figure 5. Number of parasitoid species recorded on the uncaged bushes during the experiment, recorded in the caged bushes during the experiment, and still present in the cages at the end of the experiment. Shown are means over the five or four replicates ± 1 standard error.

Download figure to PowerPoint

Figure 5 shows that on average 5·2 species of parasitoids were observed on the uncaged bushes at any time during the experiment. However, only 3·75 parasitoid species on average were observed on the caged bushes at any time during the experiment. Since no new parasitoid species can immigrate, these 3·75 species must have been present at the start of the experiment. If the moment at which the cages were dropped over the bushes can be assumed be a representative snapshot of time, then only 72% of the parasitoid species that would potentially visit the bushes during the experiment were present on a particular sample date. Figure 5 shows also that on average only one parasitoid species was still present in the caged populations at the end of the experiment. In cage C4 the galls had gone extinct by the final sampling date, so no parasitoid species were present. In cages C1 and C3, the only parasitoid species remaining was Platygaster californica, and in cage C2, 97% of the remaining parasitoid individuals were P. californica, with 3% Mesopolobus sp.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Theoretical studies have shown that limited dispersal abilities of individuals have the potential to both stabilize host–parasitoid interactions and promote the co-existence of competing parasitoid species. In this study, we test for these effects of dispersal in a natural host-multi-parasitoid system. We chose to investigate the effects of dispersal at the spatial scale of the individual Baccharis bush. The midge-release experiment demonstrated that most adult female midges remained and formed galls on the bush onto which they were released. This, in part, defended our choice of the spatial scale of investigation. In release experiments such as this, there is always a possibility of longer distance dispersal completely out of the local field (e.g. by wind). In this study we believe that we have not lost a large proportion of the individuals to long-distance dispersal. In the first generation, we released approximately 2000 adult female midges. Data from an earlier study conducted approximately 2 km from the site of the midge release experiment (Latto & Briggs 1995) suggest that these 2000 midges should produce between 2317 and 7775 galls in the next generation (depending on the specific value of egg and early larval mortality assumed). The actual value of 4613 galls produced is well within this range. Undoubtedly, some degree of longer distance midge dispersal does occur, because midge galls are found on isolated B. pilularis bushes large distances from source populations. Few individuals carried long distances by the wind are probably very important for colonizing these isolated bushes. However, these rare colonization events probably have little effect in stabilizing the population dynamics in established populations.

An important missing piece of information is the dispersal abilities of the parasitoid species, which we are currently trying to quantify. Dispersal ability of parasitoids is notoriously difficult to measure in the field, although there have been a number of recent successes using a variety of techniques (Stein et al. 1994; Corbett et al. 1996; Jones et al. 1996). Existing evidence, both from the field enclosure experiment and from subsequent experiments (C. J. Briggs & J. Latto, unpublished) suggest that individuals of most of the parasitoid species move between a number of Baccharis bushes within their life. In the field enclosure experiment, we found that only 72% of the parasitoid species that were to visit a bush during the course of the experiment were present on a given sampling date. This suggests that parasitoids regularly ‘recolonize’ individual bushes. Observation of parasitoids in the field reveals that most of the species frequently fly between galls, while searching for suitable hosts. The one exception to this is the parasitoid, Platygaster californica that attacks the egg stage. P. californica is frequently observed in the field walking between plant tips, meticulously searching every tip for midge eggs. It is likely that this species has very limited dispersal in the field and may spend its entire life on a single Baccharis bush.

To look at the effects of the elimination of the potential for dispersal at the spatial scale of the individual Baccharis bush, we followed the protocol of Murdoch et al. (1996), and compared the population dynamics on caged and uncaged bushes. The rationale behind this approach is that, if dispersal of individuals between ‘subpopulations’ is having the stabilizing effect predicted by models (e.g. Reeve 1988; Taylor 1988; Murdoch et al. 1992), then the population dynamics on bushes that are isolated from the effects of dispersal (in cages) should become less stable than those that remain connected by migration (uncaged bushes). One measure of stability, that we adopted from Murdoch et al. (1996) is the cumulative temporal variability of the population counts. We found that this did not differ significantly between the caged and uncaged treatments. Although the replication was relatively low in this experiment, it is unlikely that any realistic degree of additional replication would have helped to detect a significant difference in this particular statistic. The reason for this is that, even though there was an outbreak in one of the caged subpopulations, and a steady decline to extinction in another (both highly suggestive of the lack of regulation in the caged subpopulations), the temporal variability in these subpopulations is relatively low (see Fig. 2). Thus, we propose that future experiments should include a more direct assessment of the stability of the subpopulations. The definition of population stability in models is that the population trajectory converges to an equilibrium following a perturbation (although in real systems, the steady-state may be some seasonal pattern or probability distribution rather than a point equilibrium; Turchin 1995). Comparing the rate of convergence after a perturbation to the densities (such as an increase in midge density) would allow for a direct test of the relative stability of the caged vs. uncaged subpopulations.

Between plant movement was found to be very important in maintaining the parasitoid species diversity at the scale of the individual bush. In the caged subpopulations without migration, effectively the only parasitoid species to survive was Platygaster californica. This is in agreement with models of metapopulations of competing parasitoid species (Hassell et al. 1994; Tilman 1994; Comins & Hassell 1996) that predict that, in the absence of dispersal, the species that is most effective at exploiting the local host populations should prevail. In addition to being the only species that searches primarily by walking between tips, P. californica has a higher attack rate than any of the other common parasitoid species (Force 1970, 1974; Hopper 1984; Briggs & Latto 1996). An alternative explanation is that the conditions in the cages were more suitable for P. californica than any of the other species.

When using field enclosure experiments, such as the one presented here, there is always a danger of introducing confounding caging effects. The cages in our experiment, which were made of a very fine mesh fabric, necessarily cut down the light and wind speeds on the caged plants, and probably also altered the humidity. However, any cage control consisting of a partial cage also would have had the undesirable side effect of partially restricting the movement of individuals. An effect of the cages in our experiment, that is possibly more serious than the alteration of environmental conditions, is that the cages may have interfered with the normal foraging behaviour of the parasitoids, either for hosts or for food for adults. Subsequent experiments suggest that movement of parasitoids between bushes is important for suppressing local outbreaks of the midge (C. J. Briggs & J. Latto, unpublished). For most of the parasitoid species (with the exception of P. californica), it appears that we have isolated an area that is within the normal foraging range of the parasitoids, interfering with their aggregative response. This may, in part, explain why an outbreak of the host was able to occur in one of the caged subpopulations. Furthermore, nothing is known about the nutritional requirements of the adult parasitoids in this system. Some parasitoid species need to feed as adults on sources, such as nectar, pollen and honeydew to gain nutrients for survival, and for the production of eggs (Kidd & Jervis 1989, 1991; Jervis & Kidd 1991). If these sources were not available for some of the parasitoid species on the Baccharis bushes within the cages, then the caging had the undesirable effect of discriminating against these species.

One alternative to caging experiments is to use the degree of spatial isolation of bushes to alter the rate of dispersal. Experiments using different spatial arrangements of apple trees have been performed by Walde (1991, 1994) to study metapopulation dynamics in mites. These types of experiments look at the effect of patch number or spatial arrangement on population dynamics, and address different specific theoretical predictions of models than those addressed in our study. We feel that these two types of experiment can complement each other for a fuller understanding of the effects of dispersal and this is an avenue of research that we are currently pursuing.

In conclusion, although our study reveals patterns that are suggestive of a stabilizing effect of dispersal on population dynamics at the spatial scale of the individual Baccharis bush, the evidence is not conclusive. Our study does represent one of the longest (seven to nine generations of the host and parasitoids) experiments looking at the effects of dispersal of hosts and parasitoids in a natural field situation, but further experiments using stronger tests for stability (such as treatments involving density perturbations) combined with greater replication, are necessary.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

This research was supported by a Mildred E. Mathias U. C. Natural Reserve System Student Research grant and a Sigma Xi Grant in Aid of Research. The authors would like to thank Bill Murdoch, Roger Nisbet, Allen Stewart-Oaten, Sue Swarbrick, Sally Holbrook and Mike Bonsall, for useful comments on earlier drafts of this manuscript, and John Smiley and Feyner Arias at the Big Creek Reserve for providing logistical support.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Natural history
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References
  • Briggs, C.J. (1993) The effect of multiple parasitoid species on the gall-forming midge, Rhopalomyia californica. PhD Thesis, University of California, Santa Barbara.
  • Briggs, C.J. & Latto, J. (1996) The window of vulnerability and its effect on relative parasitoid abundance. Ecological Entomology, 21, 128 140.
  • Comins, H.N. & Hassell, M.P. (1996) Persistence of multispecies host–parasitoid interactions in spatially distributed models with local dispersal. Journal of Theoretical Biology, 183, 19 28.
  • Corbett, A., Murphy, B.C., Rosenheim, J.A., Bruins, P. (1996) Labeling an egg parasitoid, Anagrus epos (Hymenoptera: Mymaridae), with rubidium within an overwintering refuge. Environmental Entomology, 25, 29 38.
  • De Roos, A.M., McCauley, E., Wilson, W.G. (1991) Mobility versus density-limited predator–prey dynamics on different spatial scales. Proceedings of the Royal Society of London Series B Biological Sciences, 246, 117 122.
  • Doutt, R.L. (1961) The dimensions of endemism. Annals of the Entomological Society of America, 54, 46 53.
  • Ehler, L.E. (1982) Ecology of Rhopalomyia californica Felt and its parasites in an urban environment. Hilgardia, 50, 1 32.
  • Ehler, L.E. (1992) Guild analysis in biological control. Environmental Entomology, 21, 26 40.
  • Force, D.C. (1970) Competition among four hymenopterous parasites of an endemic insect host. Annals of the Entomological Society of America, 63, 1675 1688.
  • Force, D.C. (1974) Ecology of insect–parasitoid communities. Science, 184, 624 632.
  • Gilpin, M. & Hanski, I. (1991) Metapopulation Dynamics: Empirical and Theoretical Investigations. London: Academic Press.
  • Hanski, I. (1983) Coexistence of competitors in patchy environments. Ecology, 64, 493 500.
  • Hanski, I. & Gilpin, M. (1997) Metapopulation Biology: Ecology, Genetics, and Evolution. San Diego: Academic Press.
  • Harrison, S. & Taylor, A.D. (1997) Empirical evidence for metapopulation dynamics. Metapopulation Biology: Ecology, Genetics, and Evolution (eds I. A.Hanski & M. E.Gilpin), pp. 27 42. San Diego, CA: Academic Press.
  • Hassell, M.P., Comins, H.N., May, R.M. (1991) Spatial structure and chaos in insect population dynamics. Nature (London), 353, 255 258.
  • Hassell, M.P., Comins, H.N., May, R.M. (1994) Species coexistence and self-organizing spatial dynamics. Nature (London), 370, 290 292.
  • Holyoak, M. & Lawler, S.P. (1996a) Persistence of an extinction-prone predator–prey interaction through metapopulation dynamics. Ecology, 77, 1867 1879.
  • Holyoak, M. & Lawler, S.P. (1996b) The role of dispersal in predator–prey metapopulation dynamics. Journal of Animal Ecology, 65, 640 652.
  • Hopper, K.R. (1984) The effects of host-finding and colonization rates on abundances of parasitoids of a gall midge. Ecology, 65, 20 27.
  • Huffaker, C.B. (1958) Experimental studies on predation: dispersion factors and predator–prey oscillations. Hilgardia, 27, 343 383.
  • Huffaker, C.B., Shea, K.P., Herman, S.G. (1963) Experimental studies on predation: complex dispersion and levels of food in an acarine predator–prey interaction. Hilgardia, 34, 305 330.
  • Hunter, M.D. & Price, P.W. (1992) Playing chutes and ladders heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology, 73, 724 732.
  • Jervis, M.A. & Kidd, N.A.C. (1991) The dynamic significance of host-feeding by insect parasitoids: what modellers ought to consider. Oikos, 62, 97 99.
  • Jones, T.H., Godfray, H.C.J., Hassell, M.P. (1996) Relative movement patterns of a tephritid fly and its parasitoid wasps. Oecologia, 106, 317 324.
  • Kareiva, P. (1984) Predator–prey dynamics in spatially structured populations: manipulating dispersal in a coccinellid–aphid interaction. Lecture Notes in Biomathematics, 54, 368 389.
  • Kareiva, P. (1987) Habitat fragmentation and the stability of predator–prey interactions. Nature, 326, 388 90.
  • Kidd, N.A.C. & Jervis, M.A. (1989) The effects of host-feeding behavior on the dynamics of parasitoid–host interactions, and the implications for biological control. Researches on Population Ecology, 31, 235 274.
  • Kidd, N.A.C. & Jervis, M.A. (1991) Host-feeding and oviposition strategies of parasitoids in relation to host stage. Researches on Population Ecology, 33, 13 28.
  • Latto, J. & Briggs, C.J. (1995) Factors affecting distribution of the gall forming midge Rhopalomyia californica (Diptera: Cecidomyiidae). Environmental Entomology, 24, 679 686.
  • Maron, J.L. & Harrison, S. (1997) Spatial pattern formation in an insect host-parasitoid system. Nature, 278, 1619 1621.
  • McCauley, E., Wilson, W.G., De Roos, A.M. (1993) Dynamics of age-structured and spatially structured predator–prey interactions individual-based models and population-level formulations. American Naturalist, 142, 412 442.
  • McCauley, E., Wilson, W.G., De Roos, A.M. (1996) Dynamics of age-structured predator–prey populations in space: asymmetrical effects of mobility in juvenile and adult predators. Oikos, 76, 485 497.
  • Murdoch, W.W. (1994) Population regulation in theory and practice. Ecology, 75, 271 287.
  • Murdoch, W.W., Briggs, C.J., Nisbet, R.M., Gurney, W.S.C., Stewart-Oaten, A. (1992) Aggregation and stability in metapopulation models. American Naturalist, 140, 41 58.
  • Murdoch, W.W., Swarbrick, S.L., Luck, R.F., Walde, S., Yu, D.S. (1996) Refuge dynamics and metapopulation dynamics: an experimental test. American Naturalist, 147, 424 444.
  • Pimentel, D., Nagel, W.P., Madden, J.L. (1963) Space-time structure of the environment and the survival of parasite-host systems. American Naturalist, 97, 141 166.
  • Price, P.W., Cobb, N., Craig, T.P., Fernandes, G.W., Itami, J.K., Mopper, S., Preszler, R.W. (1990) Insect herbivore population dynamics on trees and shrubs new approaches relevant to latent and eruptive species and life table development. Insect–Plant Interactions (ed. E. A.Bernays), Vol. Ii, pp. 1 38. CRC Press Inc., Boca Raton, Florida, USA.
  • Reeve, J.D. (1988) Environmental variability, migration, and persistence in host-parasitoid systems. American Naturalist, 132, 810 836.
  • Reeve, J.D. (1990) Stability variability and persistence in host–parasitoid systems. Ecology, 71, 422 425.
  • Sabelis, M.W., Diekmann, O., Jansen, V.A.A. (1991) Metapopulation persistence despite local extinction: predator-prey patch models of the Lotka–Volterra type. Biological Journal of the Linnean Society, 42, 267 284.
  • Slatkin, M. (1974) Competition and regional coexistence. Ecology, 55, 128 134.
  • Stein, S.J., Price, P.W., Craig, T.P., Itami, J.K. (1994) Dispersal of a galling sawfly: implications for studies of insect population dynamics. Journal of Animal Ecology, 63, 666 676.
  • Stiling, P. & Rossi, A.M. (1997) Experimental manipulations of top-down and bottom-up factors in a tri-trophic system. Ecology, 78, 1602 1606.
  • Taylor, A.D. (1988) Large-scale spatial structure and population dynamics in arthropod predator-prey systems. Annales Zoologici Fennici, 25, 63 74.
  • Taylor, A.D. (1991) Studying metapopulation effects in predator–prey systems. Biological Journal of the Linnean Society, 42, 305 323.
  • Tilman, D. (1994) Competition and biodiversity in spatially structured habitats. Ecology, 75, 2 16.
  • Tilman, D., Lehman, C.L., Yin, C. (1997) Habitat destruction, dispersal, and deterministic extinction in competitive communities. American Naturalist, 149, 407 435.
  • Turchin, P. (1995) Population Regulation: Old arguments and a new synthesis. Population Dynamics: New Approaches and Synthesis (eds N.Cappuccino & P.W.Price), pp. 19 40. Academic Press, San Diego.
  • Walde, S.J. (1991) Patch dynamics of a phytophagous mite population: effect of number of subpopulations. Ecology, 72, 1591 1598.
  • Walde, S.J. (1994) Immigration and the dynamics of a predator–prey interaction in biological control. Journal of Animal Ecology, 63, 337 346.
  • Wilson, W.G., De Roos, A.M., McCauley, E. (1993) Spatial instabilities within the diffusive Lotka–Volterra system individual-based simulation results. Theoretical Population Biology, 43, 91 127.

Received 26 January 1998;revisionreceived 5 April 1999