Drosophila species ( Diptera : Drosophilidae ) oviposition patterns on fungi : The effect of allospecifics , substrate toughness , ovipositor structure and degree of specialisation

To determine if small-scale resource partitioning aids coexistence in fungal breeding Diptera we investigated the avoid­ ance of interspecific larval competition by micro-resource partitioning of the host substrate in several species of Drosophila ovipo­ siting on fungal fruiting bodies. No avoidance was detected and oviposition patterns were unaffected by the presence of allospecifics. There was a general tendency for all species to oviposit preferentially on the cap but no significant differences between species. Egg distribution was not strongly related to the sporophore’s resistance to penetration measured by penetrometer and was also uncorrelated with ovipositor structure even though species had concordant suites of ovipositor characteristics with apparent spe­ cialists typified by hard and sharp ovipositors. We discuss the significance to species coexistence of the absence of micro-resource partitioning of the host.


INTRODUCTION
Insect species that occur on the same host plant might well be expected to show evidence of resource parti tioning (e.g.Jermy, 1984) reducing interspecific competi tion by feeding in different parts of the individual host plant (Price, 1975;Crawley, 1983).Such a phenomenon might particularly be expected among Diptera breeding in fungi since this guild is particularly diverse.In the British Isles twelve of the 34 species of Drosophila recorded develop in fungi and most of these twelve occur sympatrically (Shorrocks & Charlesworth, 1980, 1982;Bingley & Shorrocks, 1995;Law et al., 1997).In consequence, eggs of more than one of these species may well occur in a single fungal sporophore resulting in interspecific compe tition between the developing larvae as has been demon strated in nature (Grimaldi & Jaenike, 1984).There is little evidence that any of the fungal-breeding Drosophila specialise completely on particular species of fungus (Shorrocks, 1975;Grimaldi, 1985;Jaenike & James, 1991).In consequence, the high degree of species coexis tence might be promoted by resource partitioning on a smaller scale, within sporophores, since these are hetero geneous in structure and in toughness.Different species could oviposit in different parts of the fungal body.We therefore tested whether species oviposition patterns were the same when they were ovipositing alone or with allo specifics and we examined whether there was a relation ship between oviposition pattern and the toughness of different parts of the sporophore.We also examined whether there were differences between species in the degree to which eggs were inserted into the substrate rather than laid on the surface since insertion might indi cate better adaptation to oviposition on fungi.
If there were species specific oviposition patterns on sporophores, each species would have an ovipositor selected to suit the particular character of the sporophore region in which the eggs were laid.Each species would be expected to have ovipositors of different shape, hard ness and associated characteristics (cf Kambysalis & Heed, 1971).We therefore characterised ovipositors and examined the relationship of these characteristics to ovi position patterns.

MATERIALS AND METHODS Fly strains and rearing
The flies used were all of wild type and came from strains that had been laboratory reared for only 6-12 months.They were maintained at 20°C and 16L : 8D on a standard laboratory medium (Shorrocks, 1972) with added minced mushroom.Female flies were removed from stock cages under CO2 anaes thesia and kept on medium for 24h at 5/500mm2 thereafter to recover before being transferred directly to the test cages.

Oviposition
Mushrooms (Agaricus bisporus v albida) were used as the oviposition substrate because sporophores of the same size and shape could be obtained easily and consistently.Fungal breeding Drosophila species accept this fungus as a breeding site in the wild and their larvae develop on it easily (Shorrocks & Wood, 1973;Bingley & Shorrocks, 1995).The experiments were carried out with single, open, mushrooms of about 14g placed vertically in small transparent plastic cages (100 x 100 x 130mm) with the stipe supported on a pin held in a sand-filled petri dish.The sand covered the base of the stipe and was kept damp so that the mushroom did not dry out.
We tested the oviposition patterns of D. busckii Coquillet 1901, D. funebris (Fabricius, 1787) or D. phalerata Meigen 1830, three species that are common members of the fungal breeding guild in the British Isles (Shorrocks & Wood, 1973;Shorrocks & Charlesworth, 1982;Law et al., 1997) and elsewhere (Kimura et al., 1977;Burla et al., 1991).Their larvae potentially compete interspecifically as the different species can be present in the same sporophores at the same time.This is indicated by the emergence of adults of more than one species from sporophores or bait fUngi exposed to oviposition for less time than is required for larval development (e.g.Bingley & Shorrocks, 1995).
In the first series of experiments the oviposition patterns of conspecific flies were determined by releasing two females of the same species into each cage and allowing them to forage and oviposit together in the absence of allospecifics.In the second series the effect of the simultaneous presence of allospecifics was determined by forcing a single D. phalerata to forage with either a single D. busckii or a single D. funebris.
After 48 h the eggs on each of the mushrooms were counted and their insertion sites classified as being either on the lower, middle or upper third of the stipe, or on the top, side, fringe or gills of the cap (Fig. 1).The eggs of the species can be distin guished from each other by their different numbers of filaments.To determine whether species do penetrate the fungal surface to oviposit we recorded the depth to which the eggs had been inserted into the mushroom by D. busckii, D. funebris, D. phalerata and also, in a similar separate experiment, by D. subobscura Collin 1936, another common member of the fungal guild (Shorrocks & Charlesworth, 1982;Driessen et al., 1990).

Substrate toughness
The toughness of the substrate was measured using a needle penetrometer (tip diameter 0.19 mm) attached to a force trans ducer which gave output in millivolts.Five measurements were taken on the cap, and on the stipe, of five different mushrooms.The analyses were carried out on the raw data because this was normally distributed.The mean toughnesses are, however, expressed as the force applied (in newtons) and the stress it re presents (newtons m-2).

Ovipositor structure and specialisation
We characterised the ovipositors of five species frequently reared from fungi in Europe on the basis of four features poten tially important in determining the ovipositors' ability to pene trate substrates.Fly species adapted to oviposit in tough substrates would be expected to show concordant suites of ovi positor characterisitics; e.g. to have both harder and more strongly toothed ovipositors.
The more sclerotised, and thus presumably the harder, the ovipositor plates the darker they are.Ovipositor hardness of the four species D. busckii, D. funebris, D. phalerata and D. subob scura, and for the fungal-breeding species D. kuntzei Duda 1924 (Shorrocks & Wood, 1973;Driessen et al., 1990) was therefore categorised by ranking ovipositor colour on a five point scale from dark brown to transparent off-white.The length of the ovi positor, the number of apical teeth and sharpness (the angle between the ovipositor's dorsal and ventral edges) were meas ured on 10 individuals of each species.The means of these measures for each species were also ranked.
The degree to which each species specialises on fungal breeding was derived from rearing records in the general litera ture and in our own field sampling of the species used.

Oviposition
Drosophila busckii and D. phalerata both laid a larger mean number of eggs on the A. bisporus than did D. fune bris (Table 1).To accommodate these differences in spe- cific fecundity the subsequent analyses were carried out on the proportion of eggs laid on each region or, to con trol statistically for the different areas of each region, on the region specific density (proportion/area of region).Eggs were found on the gills of only one mushroom and the gills were therefore omitted from the analyses.
There were no significant differences between species in the region specific density of eggs (Friedman Fr = 8.71, df = 5, 0.10 < p < 0.20).However, in all three species the distribution of eggs between the six regions was signifi cantly different from that expected if each region received eggs in proportion to its area (Table 1).For D. busckii and D. funebris the maximum deviation from expected was in the upper region of the stipe but for D. phalerata it was in the side of the cap.There were also significant dif-TABLE 1. Mean proportions of all eggs laid on different regions of Agaricus bisporus sporophores by three Drosophila species with the number of sporophores used and the mean number of eggs per female.For all three species the propor tions of eggs laid in each of the six regions of the sporophore are significantly different from those expected if numbers of eggs laid was a simple function of area since Dmax for Kolmogorov Smirnov 1-sample tests are significant at p < 0.01.The proportion of eggs laid in the cap as a whole was also sig nificantly greater than expected on the basis of area since all y2 values are significant at df = 1, p << 0.001.ferences in the region-specific density of eggs laid on the cap and stipe with more eggs laid on the cap than would be expected from its area (Table 1).These differences overlay and obscured the differences between regions.There were no significant differences between the distri bution pattern of eggs laid by a species on its own or when it oviposited simultaneously with a second (Table 2).There were, however, significant differences between species in the distribution of eggs between insertion cate gories compared to the proportions expected if deviations between species were due only to chance (Table 3).In particular, over 70% of D. funebris and of D. phalerata eggs were fully inserted.

Substrate toughness
The force transducer readings were significantly dif ferent between the cap and the stipe but not between mushrooms (nested anova: mushrooms F4 = 0.83, p = 0.516; regions within mushrooms Fs = 6.87, p < 0.001).The force, and the stress it represents, required to break the surface was higher for the stipe (0.306 N, 950 kNm~2) than for the cap (0.218 N, 677 kNm-2).There were also differences between the separately determined toughness readings for the cap subregions (824, 606 and 477 kNm 2 for the top, side and edge respectively).Nevertheless, there were no significant correlations (all p > 0.01) between substrate toughness and either the number, den sity or region specific density of eggs except for D. busckii.In this species density was positively, but region-specific density negatively, correlated with tough ness to the same degree (Kendal t = 0.745, z = 2.100, p = 0.0357) as the stipe carried more than the expected number of all eggs, but fewer than the expected number per unit area.

Ovipositor structure and specialisation
Drosophila kuntzei ranked highest for ovipositor length, sharpness and number of apical teeth and also ranked second for degree of melanisation.Drosophila phalerata ranked highest for melanisation and also ranked highly for the other measures.The lowest ranking species was D. subobscura, and D. busckii and D. fune bris had intermediate rankings on all scales (Table 4).All four measures of ovipositor structure for the five species examined were highly concordant (Kendal W45 = 4.175, p <<< 0.01).Mean ovipositor rankings agreed with rank ings produced on the basis of other specific characteris tics (Table 5).
Of the five species considered, we assigned the highest degree of specialisation on fungal breeding to D. kuntzei with the remaining species ordered D. phalerata > D. busckii > D.funebris > D. subobscura.

DISCUSSION
Fungal breeding Drosophila species, or at least those studied here, distribute their eggs in similar ways over A. bisporus sporophores.All three species had an overall preference for the cap despite D. phalerata and, particu larly, D. busckii, also favouring the upper stipe.In similar experiments Shorrocks & Wood (1973) also detected no difference between the egg distributions of D. phalerata and D. busckii but their preferences cannot be determined because, although both laid more eggs on the stipe, no correction was made for the areas of the two regions.However, on the fungus Phallus impudicus, D. subob scura eggs were laid all over the sporophore whereas Table 5. Rankings of the specific characteristics of five spe cies of Drosophila exploiting fungi in Europe showing the extent of agreement between the rankings of different charac teristics.Ovipositor ranking is the average rank obtained for ovipositor shape and hardness (Table 4), degree of specialism is imputed on the basis of the literature and our field collections, percentage eggs covered (Table 3) and the last two columns are rankings derived from the proportion of eggs laid during single species experiments in, respectively, the stipe, the toughest part of the mushroom, and in the toughest part of the cap, the top (Table 1 those of D. phalerata were confined to the peridium and the lower stipe enclosed by it (Shorrocks & Charlesworth, 1982).Phallus impudicus is atypical, however, since the peridium contains a gelatinous substance quite unlike anything found in other species of fungi.Thus there is little evidence for evolved consistent specific differences in egg distribution that would lower interspecific larval competition.This is unlike the situation for insects feeding in plants (Jermy, 1984) or for the flower breeding drosophilid genus Colocasiomyia (Yafuso, 1994) where each species of plant is shared by two spe cies of fly each of which consistently oviposit in different parts of the flower.Neither did females reduce interspe cific larval competition by changing oviposition patterns in the presence of allospecifics although our study did not address the possibility that females may change oviposi tion sites in response to pre-existing allospecific eggs.
There was also no tendency in our two-species experi ments for females to oviposit together although it might be adaptive at the low densities we used (Wertheim et al., 2002).Such facilitative effects are disputed, however (Hoffmeister & Rohlfs, 2001), but similar experiments at higher female densities would be of interest.
Our evidence suggests that within-sporophore resource partitioning through species-specific patterns of egg dis tribution is unlikely.Differential egg distributions might not evolve in Drosophila because, since, unlike insects developing in plants, the larvae often exhaust all the food in a single sporophore (Grimaldi & Jaenike, 1984), they would eventually compete wherever the eggs had been laid originally.
Within-sporophore patterns of oviposition are therefore more likely to be innate responses to sporophore structure rather than labile behavioural responses.However, except in D. busckii, there is no response to substrate toughness even though the ovipositors of each species show concor dant suites of characters.One possible reason for this is that egg insertion is not common as only D. funebris and D. phalerata often fully insert their eggs.If eggs are not inserted substrate toughness can have no effect on ovipo sition.A further possibility is that the difficulty of laying an egg into fungal tissue may be inadequately represented by the penetrometer measurements we made.Drosophila females do not insert eggs by simply pushing in the ovipositor but actively excavate a hole in the sub strate.The concordant ovipositor characteristics might therefore be adaptations to the resistances of different parts of the sporophore to excavation and we measured resistance to penetration not to excavation.The disjunc tion between substrate toughness and ovipositor charac teristics might also arise because ovipositor characteristics are determined by phylogeny rather than by adaptation.Drosophila kuntzei and D. phalerata belong to the same quinaria-species group and the close similarity of their ovipositors may be a simple conse quence of this relatedness.Each of the other species tested comes from a different species group within the Drosophila genus and therefore their individual charac teristics might merely represent the characteristics of the species group to which they belong.Our experimental design does not allow the separation of adaptive and phy logenetic characteristics (cf.Harvey & Pagel, 1991).None of these reasons excludes the additional possibility that the lack of a relationship between egg distribution and substrate toughness within sporophores and the lack of links between egg distribution and ovipositor charac teristics may be more fundamentally due to the degrees of specialization on fungi shown by each species.None of the species tested here are confined to fungi in the wild (Offenberger & Klarenberg, 1992;Law et al., 1997) and so the other substrate types exploited may condition ovi positor characteristics.Specialist species can evolve spe cialist ovipositors suited to a particular substrate but generalists, ovipositing in substrates with very varied characteristics, cannot.This relationship appears to be present among the species we examined but many more, phylogenetically independent, species would have to be tested to confirm it.
The lack of overt micro-resource partitioning in the fungal feeding Drosophila we examined confirms the impression that Drosophila communities are not struc tured by competitive resource partitioning at any level (Shorrocks et al., 1984;Shorrocks & Sevenster, 1995;Wertheim et al., 2000).

Fig. 1 .
Fig. 1.Division of an Agaricus bisporus sporophore into cap and stipe and into the regions: T -cap top; S -cap side; F -cap fringe; U -upper third of stipe; M -middle third of stipe; Llower third of stipe.

Table 2 .
Mean proportions of all eggs laid on different regions of Agaricus bisporus sporophores by three Drosophila species when foraging in the presence of allospecifics.None of the Dmax values for Kolmogorov Smirnov 2-sample tests of the distribution of eggs when foraging alone and when foraging with allospecifics was significant atp< 0.05.

Table 3 .
Proportion of all eggs laid in single species experi ments that were not inserted, partially inserted or fully inserted into the mushroom substrate by four species of Drosophila

Table 4 .
Rankings of four ovipositor characteristics, and the overall ranking, for five species of Drosophila exploiting fungi in Europe.All four measures of ovipositor structure for the five species examined were highly concordant (Kendal W45= 4.175, p <<< 0.01).Overall ranking is the rank of the mean ranking for all four ovipositor characteristics.