The pollens consumed by common green lacewings Chrysoperla spp . ( Neuroptera : Chrysopidae ) in cabbage crop environment in western France

The pollens consumed by common green lacewings Chrysoperla spp. in crop environment in western France, were analysed. The diverticulum contents including quantity of pollen of Chrysoperla adults were analysed to compare a feeding behaviour between the sexes. Females consume more pollen than males. The feeding behaviour of the two main species present was compared, yielding new insights into differences between species within the complex Chrysoperla carnea (Stephens). 547 * = Chrysoperla affinis (Stephens, 1836) sensu Thierry MATERIAL AND METHODS Investigations were conducted in the Loire Valley in western France, in agricultural zones of vegetable crops (market gardening). The different sampling sites are managed with different cultural techniques: conventional, IPM, and strictly organic farming. All sites include cabbage crops, three sites had seed production and one site had vegetable production with organic farming. Lacewings were sampled regularly during the growing season, from March to August 2004. In spring, Chrysoperla spp. live and feed in the arborescent stratum. In summer, they stay in their resting places in the trees and fly to the low vegetation during the twilight for feeding, copulation and oviposition (Duelli, 1986). Therefore, every 15 days, from March to April, we collected lacewings during the day, and from June to August we collected during the day in the arborescent stratum, and during the twilight in the herbaceous stratum. Survey of Chrysopidae For the herbaceous stratum, we used a mobile vacuum device for 10 min at each sampling occasion. In the canopy, we made two hundred hand net sweeps.


INTRODUCTION
The need for food of high quality for the West European market has been significantly increasing so that Integrated Pest Management (IPM) and organic farming are now becoming a concern for agriculture.Crops must be harvested free of chemicals and their protection against phytophagous pests must be increasingly managed by alternative methods.In this way, naturally occurring generalist predators may play a key role in crop protection.Barbosa (1998) emphasised the need to attract natural enemies to crop environments (Conservation Biological Control).Field crops are temporary habitats.Part of the year, they are inhospitable.The strategy of field-crop specialist insects can be migration within a heterogeneous habitat (Ronce & Kirkpatrick, 2001).A more heterogeneous vegetation bordering or surrounding cultivated areas supplies sites for oviposition and overwintering and a wider food choice for beneficial organisms, serving them as shelters (Thies & Tscharntke, 1999;Nicholls & Altieri, 2002) in part through diversification of plant species (Theunissen & Den Ouden, 1980;Baliddawa, 1985).Thus, Van Emden (1965) noted that Brassicaceae bordering uncultivated fields were less damaged by Brevicoryne brassicae because syrphid predators were more abundant.The presence of weeds and wild plants in agroecosystems provides significantly more pollen and nectar as food for natural enemies in and around fields (Cowgill et al., 1993;Colley & Luna, 2000).
The present study contributes to the clarification of chrysopid flight behaviour in the crop environment.Consideration of sources of pollen found in the lacewing diverticulum reveals amplitude and frequency of Chrysoperla migration in the crop environment in western France.The feeding behaviour is analysed, yielding new insights concerning the status and behaviour of species within the complex Chrysoperla carnea.

MATERIAL AND METHODS
Investigations were conducted in the Loire Valley in western France, in agricultural zones of vegetable crops (market gardening).The different sampling sites are managed with different cultural techniques: conventional, IPM, and strictly organic farming.All sites include cabbage crops, three sites had seed production and one site had vegetable production with organic farming.Lacewings were sampled regularly during the growing season, from March to August 2004.In spring, Chrysoperla spp.live and feed in the arborescent stratum.In summer, they stay in their resting places in the trees and fly to the low vegetation during the twilight for feeding, copulation and oviposition (Duelli, 1986).Therefore, every 15 days, from March to April, we collected lacewings during the day, and from June to August we collected during the day in the arborescent stratum, and during the twilight in the herbaceous stratum.

Survey of Chrysopidae
For the herbaceous stratum, we used a mobile vacuum device for 10 min at each sampling occasion.In the canopy, we made two hundred hand net sweeps.

Pollen analysis
Specimens collected were killed at -80°C and kept in a freezer.Thus, the pollen was not damaged.We extracted the oesophagal diverticulum of adults, which is the part of the gut where the pollen is stored and not yet digested (Fig. 1).The diverticulum was opened on a microscope slide.After lipid extraction with diethyl ether, the pollen grains from the diverticulum were placed on a slide with glycerin jelly (Kaiser fourn, Merck Eurolab) containing basic fuchsin as a stain.
It is necessary to use only as much jelly as occupies exactly the space beneath the coverslip without undue pressure being applied (Hyde & Adams, 1958).Pollen grains were observed with an optic microscope ×400 (Olympus BH-2 with objective SPlan40, ×40).
The pollen grains were determined to family or species level at ×400.To quantify the amount of pollen, the slides were observed at ×20 and the total number of pollen grains in the diverticulum was estimated, taking in account its size, which ranges from 2 mm (empty crop) to 3 mm (full crop).

Statistical analysis
An analysis of variance was used to compare the means of two samples (Scherrer, 1984).The size of samples can be large or small.The logarithm of the quantity of pollen grains was analysed (to meet assumptions of normality in the analyses).

Pollens in the food ingested
From March to August 2004, 157 Chrysoperla adults were collected, and therefore 157 slides were observed.The number of pollen grains varied from 0 to 12,000 grains with a mean number of 432 (SD = 1647) per adult diverticulum.Pollen of about 40 plant species were observed (Table 2), but the mean number of plant species observed was only 1.99 per adult (maximum six).The diverticula never contained a large variety of pollen.

Feeding during the flight periods and during the day From March to May
From March to May, Chrysoperla spp.adults feed in the arborescent and herbaceous strata.About 35% of the specimens collected during this period had an empty diverticulum.The quantity of pollen grains found in diverticula adults collected from the two strata was not significantly different: 1715 (mean number) in the arborescent stratum and 575 in the herbaceous stratum (Fig. 2) (after log transformation Fc = 0.95 << F = 4.54, = 0.05, df = 16).Among the lacewings with diverticula that contained pollen grains, we observed pollen of arbored plants in 89% of the specimens collected in the trees, versus in only 37% of those specimens collected in the herbaceous stratum.It seems difficult to conclude that the adults feed in the stratum they live in.
In the arborescent stratum, the main plant family represented by pollen grains in Chrysoperla diverticula was Rosaceae, and included such plants as Prunus sp., Spiraea sp., Pyrus pyraster and Amelanchier sp.(42% of specimens with Rosaceous pollen contents).The other families represented were Caprifoliaceae (Sambucus sp.), Betulaceae (Corylus avellana), Taxaceae and Pinaceae (Pinus sp.).In March-April, the flowering trees in the crop environment are mainly Rosaceae.

From June to August
From June to August, the pollen analysis seems to confirm the observations of Principi & Canard (1984) and Duelli (1986): Chrysopidae are considered mainly nocturnal fliers in the herbaceous stratum whereas during the day the arborescent stratum is their resting place.The mean pollen quantity in specimens collected during the day is very low because the pollen have apparently been digested (Fc = 5.26 >> F = 3.95, = 0.05, df = 85) (Fig. 3).

Food according to sex status
Diverticula of Chrysoperla females contained significantly more pollen grains (mean number = 1137) than did those of males (mean number = 44) (after log transformation Fc = 8.87 >> F = 3.96, = 0.05, df = 84, significant difference) (Fig. 4).Furthermore, 59% of empty diverticula belong to males.The results of pollen analyses sug-gest that male energetic needs are limited to precopulatory searching.

Comparison between the sibling species
The ecological demands and biological traits are similar for the two species of lacewings present in crops of western France.Chrysoperla lucasina is a holomediterranean species (Henry et al., 1996;Thierry et al., 1996;Aspöck et al., 2001) and is more adapted to high temperatures, while the other species is a North European species    (Thierry et al., 1996).Both species feed in low vegetation (field and herbaceous stratum) beginning in later spring.Our results show a difference in food consumed between these green lacewings (Table 2).The number of different plant species consumed is very low compared to the total possible range.According to Monserrat & Marín (1994), Chrysoperla carnea s. lat.has no preference for plant family or species because it has been found in all types of vegetation.However, separating sibling species of the carnea-complex reveals differences in feeding habits and behaviour.The high ecological versatility of these lacewings facilitates colonisation of various habitats (arborescent and/or herbaceous strata).However, we can see a difference in the range of plant species chosen by the two species, with the common green lacewing feeding on more species of plants than does Ch.lucasina.We found a total of 22 plant families and about 40 different species for the comon green lacewing, versus 11 plant families and about ten species for Ch.lucasina (Table 2).So Ch. carnea sensu Henry is a more polyphagous species than is Ch.lucasina.
Both species consume the pollen of flowering plants such as leek (A.porrum), carrot (D. carota), cabbage or rape (Brassica sp.) (Table 2).However, Ch. lucasina seems to prefer field crops and other herbaceous plants while Ch.affinis seems to consume the pollens of trees and herbaceous plants.Furthermore, 88% of Ch. lucasina were collected in June when the low vegetation is flowering.Ch. lucasina is then specialised in field crops.In previous studies, e.g.Monserrat & Marín (1994), it was not possible to separate the species among carnea group, and that may explain the differences in our results versus those of these previous studies.
According to Paulian (2001), Ch. lucasina is more frequent and abundant in field crops and on herbaceous plants, where it is the first species to appear in spring, while the common green lacewing is rather a more frequent tree canopy inhabitant.
Because much variability occurs among individuals in all analyses, our conclusions must confirmed by further studies.In addition, the presence of pollen in diverticula can also be misleading because lacewings adults feed frequently on honeydew in which pollen grains have become stuck and are therefore also consumed (Duelli, 1999).
Alternative methods to attract beneficial organisms for conservation biological control may consist in growing plants that are preferred as pollen sources for green lacewings in the field neighbourhood.These might include Rosaceae, Caprifoliaceae, Betulaceae and Pinaceae as trees and Caryophyllaceae, Asteraceae, Gramineae and Chenopodiaceae as herbaceous plants for Ch.carnea sensu Henry; and mainly herbaceous plants as Brassi-caceae, Gramineae, Apiaceae and Asteraceae for Ch.lucasina.
We assume that the adult chrysopids feed on a narrow range of pollen species, that occur mainly in the habitats in which they spend most time.Chrysopidae seem to feed near their resting-place with the bordering cropvegetation being reservoir habitats for Chrysoperla spp.To test this hypothesis further, relations between plants occurring in the agroecosystems and the pollens consumed by Chrysoperla, and the possible choice or selection of food by these lacewing species will be investigated.In addition, pollen analysis can be used to study the Chrysoperla migration flight from March to October.

Fig. 2 .
Fig. 2. Comparison of the mean of the log number of pollen grains observed in 32 Chrysoperla spp.adults diverticulum contents (50% empty).Adults collected in the arborescent (mean of log number = 2.70, SD = 1.82) and herbaceous (mean of log number = 1.60,SD = 1.19) strata, from March to May 2004 in field environments in western France.

Fig. 3 .
Fig. 3. Comparison between the diverticulum contents of 125 Chrysoperla adults (32% empty) collected in the arborescent stratum (mean of log number = 1.20,SD =1.13) during the day and in the herbaceous stratum (mean log number = 0.59, SD = 0.60) during the twilight, from June to August 2004 in field environments in western France.

Fig. 4 .
Fig. 4. Comparison between the mean number of pollen grains present in the diverticula of male (SD = 0.72) and female (SD = 1.23) of 157 Chrysoperla (77 males, 80 females) collected from March to August 2004 in field environments in western France.

TABLE 2 .
Comparison of diverticulum contents (pollen) of 114 adults of Chrysoperla carnea sensu Henry and 43 adults of Ch. lucasina: plant families and species observed; proportion (%) of adults consuming the family plant.Plant families ranked by alphabetic order.