Interaction between Apis mellifera L . and Baccharis dracunculifolia DC , that favours green propolis production in Minas Gerais

In Minas Gerais, green propolis is produced from the collection of resinous substance found in shoot apices of Baccharis dracunculifolia. The aim of this study was to evaluate the biological parameters associated with the interaction Apis mellifera x Baccharis dracunculifolia, to elucidate the supply of resin for green propolis production in Minas Gerais. We selected male and female individuals of two populations of Baccharis dracunculifolia located on São Judas Tadeu Farm – FSJT, in the municipality of Betim, MG and the Experimental Garden of the Ezequiel Dias Foundation – HORTO, located in an urban area in Belo Horizonte, MG. We made weekly observations, from June 2007 to June 2008, and evaluated in both populations: richness and abundance of insect visitors; resin collecting visits of Apis mellifera; presence of Baccharopelma dracunculifoliae galls; growth of individuals and phenological phases. Statistical analyses were made using R software. The rainy season showed the highest number of visitors. A. mellifera collected resin in shoot apices of Baccharis dracunculifolia from August to April, only in the FSJT population, where galls of B. dracunculifoliae were also present. Ovoposition of gall inductor on host plants occurs during the rainy season, when there is a peak of visitants and resin collecting visits of honeybees. This fact stimulates plant defense strategies against parasitoids and predators, which includes the production of several secondary metabolites, and ultimately reduces competition for food by inhibiting the attack of other phytophagous insects, not adapted to the chemical environment of plant tissues. Green propolis production in Minas Gerais is related to the abundant supply of resin by Baccharis dracunculifolia, when they are parasitised by B. dracunculifoliae galls. They induce plant production of defense exudates, which attract Apis mellifera bees to collect resin and consequently favour the production of green propolis.


Introduction
Baccharis is the largest genus of the subtribe Baccharidinae (Asteraceae) and is widely represented in South America with more than 500 native species (Barroso, 1976).Some species are typically invasors and occur mainly in denuded and empty areas, like margin of roads and railroads, plowed and not planted fields, pastureland, margins of ditches, fences and gaps (Palmer, 1987;Palmer and Bennett, 1988;Gagné and Boldt, 1989;Boldt and Robbins, 1987;Palmer, 1987).Species are dioecious, 0.5 to 4 m high, mostly evergreen shrubs and native from New World (Barroso, 1976).Baccharis dracunculifolia, popularly known as vassourinha, alecrim do campo or alecrim de vassoura, occurs from the southeast to the southern region of Brazil and extends into Argentina, Uruguay, Paraguay and Bolivia (Barroso, 1976).Plants occur in fields or open areas and constitute dense and dominant formations on uncultivated or degraded land (Araújo et al., 1995).In this species and others of the genus, blooming occurs after the rainy season (Boldt et al., 1988).
They have the highest diversity of galling insects in the tropics (Fernandes et al., 1994).Studies on insect-plant interactions consider that richness in herbivorous insect communities is influenced by the host plant characteristics, such as range of occurrence and structural complexity (Strong et al., 1984).
Quality and quantity of resources offered by the host plant also play an important role in herbivorous insect diversity (Price, 1992;Araújo et al., 1995;Marques et al., 2002;Espírito-Santo et al., 2004).According to Herms and Mattson (1992), plants adapted to stressful habitats (nutrient-poor) have better defenses against herbivore attack and when there are limited resources, allocation of these is selective on whether used for growing or for chemical defenses production.As nutrient deficiency limits growth, plants with small growth rates favour secondary compounds production (Herms and Mattson, 1992).
These compounds have many roles, and their presence is considered part of the plant defense strategy against herbivorous insects attack (Mani, 1964;Price et al., 1987Price et al., , 1990)).However, it is known that some herbivores use these secondary metabolites of plants to stop their own predators as proposed by Herms and Mattson (1992).Defense is not the only role of these substances, that can be also associated with pollinator attraction, ultraviolet light protection, structural support, temporary storage of nutrients and location of host plants (Price et al., 1987).
Several studies on insect fauna associated with Baccharis dracunculifolia sought knowledge of population control using potential species herbivores (Tilden, 1951;Kraft and Denno, 1982;Boldt and Robbins, 1987;Palmer, 1987).These herbivores frequently associated with Baccharis dracunculifolia in Brazil are galling insects of nine different and still unidentified species Cecidomyiidae (Diptera), and a species of Psyllidae, Baccharopelma dracunculifoliae (Hemiptera,Psyllidae) (Lara and Fernandes, 1994;Fagundes et al., 2001;Arduin et al., 2005;Carneiro et al., 2009).Baccharopelma dracunculifoliae is restricted to the Neotropical Region (Burckhardt, 1987) and its biology and natural history are largely unknown.All hosts of B. dracunculifoliae belong to Baccharis genus (Burckhardt, 1987) and, in general, the psyllids are host specific of family, genus and species (Burckhardt et al., 2004).
Biological association of galling insects and plants do not imply plant benefits.On the other hand, galling insects have many advantages like shelter, nutrition, and spread.The adaptive value of this complex relationship -gall versus plant -has been widely discussed (Mani, 1964;Price et al., 1987;Isaias, 1998;Espírito Santo et al., 2004;Carneiro et al., 2009).
Galls are like drains, nutritionally supplied by direct vascular tissue connections of the host plant, where it assimilates direct flow, without any plant control (Hartley, 1998).Besides primary metabolites clearly related to inductor nutrition, galls are related to compounds originated from secondary metabolism, among which the most studied are phenolic derivatives.These substances are considered part of plant defense strategy to herbivorous insects attack (Mani, 1964;Price et al., 1987Price et al., , 1990)), and it seems that damaged tissues promote changes in metabolism phenolic derivatives, and also benefit insect gall.Increased production of these substances by the host plant could directly or indirectly protect galling insects from its parasitoids and predators, and moreover, inhibiting attacks of other phytophagous insects not adapted to this chemical environment of plant tissues, could reduce competition for food (Janzen, 1977;Cornell, 1983;Soares et al., 2000;Espírito-Santo et al., 2004).
Apis mellifera propolis has several functions in hives such as gap filling, reduction of entrance and exit spaces, mummification of dead insects for prevention of decay and putrefaction, covering hive inner walls, and inside broodcomb cells to prevent infestation from microorganisms (Ghisalberti, 1979).
The main goal of this study was to evaluate biological parameters associated to Apis mellifera x Baccharis dracunculifolia interaction, in order to discover how resin production by this plant species contributes to green propolis production in the state of Minas Gerais.

Areas of study
Two Baccharis dracunculifolia populations were used in this study: Fazenda São Judas Tadeu-FSJT (UTM 584413 and 7802335), municipality of Betim, MG, presenting heterogeneous vegetation composed of native species, invasive exotic, ornamental and fruit trees.We studied 18 plants, 10 male and 8 female located in a successional area.
Experimental Garden of Ezequiel Dias Foundation-HORTO (UTM 606202 e 7796212), in an urban area in Belo Horizonte, MG, where we studied 11 plants, 4 male and 7 female.

Field observations
Individuals of Baccharis dracunculifolia studied were identified by comparison with material from the herbarium of the Botany Department UFMG (BHCB 41 396).Individuals were randomly selected, numbered, sexed and marked with signs.Field observations comprised the period from June 2007 to May 2008 in FSJT population and from July 2007 to June 2008 in the HORTO population.
We evaluated the following parameters in both Baccharis dracunculifolia populations studied: 1) richness and abundance of visitor insects; 2) number of resin collecting visits of Apis mellifera; 3) presence of galls of Baccharopelma dracunculifoliae; 4) growth of Baccharis dracunculifolia individuals throughout the study period by measuring their stem diameter; and 5) Phenophases (flowering and vegetative state) of individuals.
Stem diameter, phenophases, and number of galls Baccharopelma dracunculifoliae were collected, recorded monthly.Galls estimation was performed according to the phases of development proposed by Arduim et al. (2005), and only the second phase was used in data analysis, because it was easier to recognise in the field (Figures 1 and 2).
Weekly observations were made between 10:30 AM and 3:00 PM, lasting 20 minutes on each individual (Ferracini, 1995), this, divided into two rounds of 10 minutes.

Statistical analysis
Field data, like number and types of visitors, especially Apis mellifera and the presence of galls Baccharopelma dracunculifoliae, may present a difference due to season (rainy or dry), environment (both studied populations) and gender of individuals (male and female).
To assess differences in the medians of the variables: number of visitors, number of B. dracunculifoliae galls phase two, stem diameter and resin collecting visits of A. mellifera, among dry and rainy seasons, and gender of individuals, we used the Mann-Whitney test.Correlation between variables was addressed by the Spearman coefficient.
Possible associations with A. mellifera visits (response variable) and explanatory variables: season, visitors, galls, gender and size of individuals, were verified using Chi-square test.Significance level was 5%.Analyses were performed with the software R Development Core Team (2009).

Baccharis dracunculifolia visitors
A total of 1645 visitors were observed in HORTO and 3875 visitors in FSJT.Of these, 1123 belong to the following orders: Coleoptera, Hemiptera, Homoptera, Orthoptera, Thysanoptera and Lepidoptera.In both populations, the order Hymenoptera was represented by 10-45% among the total number of visitors, the order Coleoptera 5-10% and orders, Hemiptera, Homoptera, Lepidoptera, Mantodea, Odonata, Orthoptera, Neuroptera had a frequency <1% (Table 1).In both populations, the number of visitors observed was significantly different between dry and rainy seasons (Mann-Whitney test, p < 0.05).

Resin collecting visits of Apis mellifera
For the FSJT population, resin collecting visits of A. mellifera (Figure 3) began in August and continued until April (Figure 4).Individuals from the HORTO population were not visited by Apis mellifera in order to collect resin.
In the FSJT population, there was a significant positive correlation of resin collecting visits of Apis mellifera with other visitors (r = 0.177, n = 201, p < 0.05) and ants (r = 0.141, n = 201, p < 0.05).

Presence of Baccharopelma dracunculifoliae galls
Baccharopelma dracunculifoliae galls were present only in the FSJT population and during the dry season (from April to July), had greater abundance of phase 2 galls (Figure 5).

Growth of Baccharis dracunculifolia individuals
Individual size showed significant positive correlation with other visitors in the HORTO population.In the FSJT population, individual size was positively correlated with Baccharopelma dracunculifoliae galls and with resin collecting visits of Apis mellifera (Table 2).

Phenological phases (flowering and vegetative state)
The flowering period was from December to May and vegetative period from June to November, in both populations studied.During the flowering period, resin    collecting visits of Apis mellifera to vegetative apices occurred in the FSJT population.
There was no significant difference of biological parameters of (Baccharopelma dracunculifoliae galls, resin collecting visits of Apis mellifera, and number of others visitors) among male and female individuals (Table 3).
Only size showed difference between gender in FSJT population and number of ants in the HORTO population (Table 3).

Discussion
The highest number of visitors was recorded in the rainy season in both populations, in agreement with Araújo et al. (1995), who reported that the greater abundance of herbivores on B. dracunculifolia during the rainy season and reproductive state is related with a major resource offered.
After this period, most of the leaves fall, giving the shrub a dry aspect.The sprouting of new leaves comes with the rain, and consequently, the abundance of herbivorous insects.
Resin collecting visits honeybees (Figure 4) occurred from August to April, and in December, January, February and April there was the highest frequency.The rainy season also showed higher visitation of ants and other insects.
The genus Baccharis produces exudates that attract a variety of flying insects (Boldt et al., 1988;Boldt and Robbins, 1990) and these substances occur naturally in the plant or its production is stimulated by insect injured leaves, stems or other plant parts.
Green propolis harvest in the state of Minas Gerais, starts in August and lasts until April (Lima, 2006), and the results show that during this same period, honeybees collect resin on B. dracunculifolia individuals.Previous studies on seasonal production of essential oils in populations of B. dracunculifolia showed higher rates of production from February to April and the lowest during May, June and July (Souza, 2007), in agreement with results reported here.
Galls caused by the psyllid Baccharopelma dracunculifoliae were present only in the population of FSJT, where resin collecting visits of A. mellifera occurred.The second phase of the gall, which was used in this work to count for them, was higher from April to July, during the dry season (Figure 5).
Therefore, it is during the rainy season that the oviposition of gall inductor on host plants occurs, with the peak of visitants and resin collecting visits of honeybees.This fact stimulates plant defense strategies against parasitoids and predators, which includes the production of several secondary metabolites, and ultimately reduces competition for food by inhibiting the attack of other phytophagous insects, not adapted to the chemical environment of plant tissues (Janzen, 1977;Cornell, 1983;Isaias, 1998;Soares et al., 2000).
The most commonly studied compound of plant secondary metabolism are phenolic derivates.The presence of these substances is considered part of the plant defense strategy against attack by herbivorous insects (Mani, 1964;Price et al., 1987;Price et al., 1990).
Contrarily the, HORTO population did not develop these galls.Probably this is due to its isolation within an urban area, and according to previous studies (Araújo et al.,  1995; Fernandes et al., 1994;Burckhardt et al., 2004;Julião et al., 2005), the psyllid has little ability to fly and a small time interval is required between the colonisation of a particular site by the host plant and the arrival of this gall inductor.Additionally, the degree of the host plant population isolation also represents another important factor for the arrival of this insect.In this population, the absence of resin collecting visits of A. mellifera as well as B. dracunculifoliae galls, confirms the relationship between presence of the psyllid and the production of exudates by host plants as a defense strategy.
Individual size showed positive significant correlation with other visitors, with the presence of Baccharopelma dracunculifoliae galls and with resin collecting visits of A. mellifera in FSJT, which are all in agreement with the literature (Espírito-Santo and Fernandes, 1998;Collevatti and Sperber, 1997).
During blooming period, from December to May, there was no resin collecting visits of A. mellifera, and these are not significantly different among male and female plants.This shows that resin production by Baccharis dracunculifolia plants is not related to its phenological aspects, but with the presence of metabolic inductors like galling insects.

Conclusion
During the rainy season, Baccharis dracunculifolia populations secretes a number of exsudates that attract ants, others insects and Apis mellifera bees for resin collecting visits.
These resin collecting visits on vegetative apices occur on Baccharis dracunculifolia individuals that are oviposited by Baccharopelma dracunculifoliae galls, during the rainy season.Bees collect independently of individuals gender (male or female) and its phenological state (flowering or vegetative), and the period of high resin collecting visits coincides with the harvest of green propolis in Minas Gerais.

Figures 1- 2 .
Figures 1-2.Baccharopelma dracunculifoliae gall.a) Gall development phases: Phase 1 -Swelling phase, leaf swelling in response to gall-inducing insect oviposition.Phase 2 -Folding phase, the leaf lamina folded upward alongside the midrib and the edges of the upper portion of the leaf approached each other, forming a longitudinal slit.Phase 3 -Dehiscence phase, Phase 4 -Senescence phase.b) Gall phase 2 among leaves (arrow) used in statistical analysis because of its easy recognition in the field.

Figure 4 .
Figure 4. Number of resin collecting visits of Apis mellifera monthly recorded on B. dracunculifolia individuals, for the Fazenda São Judas Tadeu (FSJT) population.Horizontal line represents the mean.

Table 1 .
Total number and percentage of visitors recorded in Baccharis dracunculifolia Individuals in Horto (n = 11) and in FSJT population (n = 18) during the study (from 2007 to 2008).
*This group includes only other bees (except for A. mellifera) and wasps.**Visitors whose identification was not possible.

Table 3 .
Results of Mann-Whitney test of biological parameters: herbivores, ants, size of individuals, galls of B. dracunculifoliae and resin collecting visits of A. mellifera, among gender (male and female).All null hypothesis were that male and female medians were equal and alternative hypothesis were that male and female medians were different for each parameter.
*Resin collecting visits of A. mellifera.