Pollen analysis in honey samples from the two main producing regions in the Brazilian northeast

GENI DA S. SODRÉ1, LUÍS C. MARCHINI2, CARLOS A.L. DE CARVALHO1 and AUGUSTA C. DE C.C. MORETI3 1Centro de Ciências Agrárias, Ambientais e Biológicas, Laboratório de Entomologia, Universidade Federal do Recôncavo Baiano, 44380-000 Cruz das Almas, BA, Brasil 2Departamento de Entomologia, Fitopatologia e Zoologia Agrícola, ESALQ/USP, Av. Pádua Dias, 11, 13418-900 Piracicaba, SP, Brasil 3Instituto de Zootecnia/ APTA, SAA, Rua Heitor Penteado, 56, 13460-000 Nova Odessa, SP, Brasil


INTRODUCTION
The identification of plants used by honeybees assumes great importance, since it provides beekeepers with indications about adequate and abundant nectar and pollen supply sources (Howes 1953), especially considering that beekeeping in Brazil is targeted at an improved use of the richness offered spontaneously by the natural vegetation.
Knowledge about the pollen morphology of honey plants is important in the identification of plant species which contribute toward composition of honey. Preserving and multiplying those plants is also necessary, aiding in the establishment of a sustainable apiculture (M.C. Santos Júnior and F.A.R. Santos, personal communication). The pollen involuntarily collected by bees at nectar collection time is an important indicator of its botanical and especially geographical origin (Barth 1989, E.M.A.F. Bastos, personal communication).
The qualitative and quantitative palynological survey of a honey sample constitutes its pollen spectrum. This spectrum relates to nectar-producing plants, nonnectar-producing plants, contaminations, tainted honey, and mixtures (Barth 1989). Pollen qualitative analysis can provide important data, especially for honey characterization with reference to its geographic origin, occurrence of nectariferous plants, collection season, and at a later time with regard to honey determination from an unknown or questionable origin (Barth 1989). By means of quantitative pollen grain analysis, the honey makeup can be established, and the nectariferous botanical species which gave origin to it can be determined (Iwama and Melhem 1979). Studies developed by Santos (1963Santos ( , 1978, Barth (1969Barth ( , 1970Barth ( a,b,c, 1971aBarth ( , 1989Barth ( , 1990Barth ( , 2004, Cortopassi-Laurino and Gelli (1991), B.M. Freitas (personal communication), Aires and Freitas (2001), Sodré et al. (2001), Arruda et al. (2005, and Persano-Oddo et al. (2004) demonstrated the importance of pollen found in honeys in furnishing knowledge about the botanical and geographic origin of the resource-providing plants, thus increasing the use of these trophic resources.
The present study means to identify the pollen types occurring in 58 samples of honey produced in two states of the northeastern region of Brazil, Piauí (38 samples) and Ceará (20 samples), and to verify the potential of honey plants during certain seasons of the year.
All honey samples were prepared using the acetolysis method (Erdtman 1952) and then submitted to both a qualitative and a quantitative analysis.

QUALITATIVE ANALYSIS
The qualitative analysis, referring to the pollen types present in the samples, was determined by comparison with a reference slide collection (600 species) and the descriptions of pollen grains obtained from the specialized literature (Barth 1970(Barth a,b,c, 1971a(Barth ,b, 1989(Barth , 1990.

QUANTITATIVE ANALYSIS
The quantitative analysis was performed by consecutively counting 300 pollen grains/replicate/sample. Percentages and classes of occurrence were determined. According to Louveaux et al. (1978) these are the dominant pollen (> 45% of total grains) (Dp), accessory pollen (16 to 45%) (Ap), important isolated pollen (3 to 15%) (Iip), and occasional isolated pollen (< 3%) (Oip). The results presented here correspond to the average of two counts.
For some plant species, the term "pollen type" was assigned, due to the little knowledge available about the honey species in the study region. We resorted to pollen type, which comprises all species that have equal or similar pollen grains, either belonging or not to a species in the same genus (Barth 1970b).

RESULTS AND DISCUSSION
The pollen analysis results are shown in Tables I and II  The qualitative pollen analysis of the 58 honey samples showed a great variety of pollen grain types. Fortyone pollen types were found, distributed through 17 botanical families in the State of Ceará, while 39 pollen types, distributed through 19 botanical families, were found in the State of Piauí (Tables I and II). This information may contribute toward the characterization of honeys from these states with regard to their botanical origin.
The quantitative pollen analysis detected five different pollen types occurring as dominant pollen in the 20 samples of the State of Ceará that were analyzed: Mimosa caesalpiniaefolia (Mimosaceae) (50.0%), M. verrucosa (Mimosaceae) (5.0%), Borreria verticillata (Rubiaceae) (10.0%), Serjania sp. (Sapindaceae) (5.0%), and Fabaceae type (Fabaceae) (5.0%) ( Table I) Dp = dominant pollen greater than 45% of total pollen grains, Ap = accessory pollen from 16 to 45%, Iip = important isolated pollen from 3 to 15%, and Oip = occasional isolated pollen smaller than 3%.   (Table II). It can be seen that the samples in both states are quite similar with regard to their floral origins, as they were collected in the same season and at nearby geographic regions. The quantitative pollen analysis of the honeys demonstrated the important participation of M. caesalpiniaefolia in the makeup of honeys from the State of Ceará. In the State of Piauí, in addition, the strong participation of the genus Piptadenia sp. was detected. These two pollen types reiterate the beekeeping potential of these plants in the states under study. Barth (1989)  Based upon the quantitative analysis of pollen grains, we were able to observe a great participation of accessory pollen in the samples, corresponding to 55.0% of the samples for the State of Ceará and 58% of the samples for Piauí. Barth (1970b) mentioned that honeys containing accessory pollen occur frequently, due to the fact that Brazilian beekeeping is carried out in a primitive manner with regard to bee pasture. The author highlighted the importance of accessory and dominant pollen with reference to the amount of nectar supplied. In the present work, we observed samples containing accessory pollen without the presence of dominant pollen for the State of Ceará in samples 8, 10, 13, 15, and 19 (25.0% of samples) (Table I), and for the State of Piauí in samples 11, 12, 19, 22, 28, and 29 (15.8%) (Table II).
With regard to important or occasional isolated pollen, their presence was observed in almost all samples studied (Tables I and II). This pollen type has little importance with regard to the amount of nectar supplied; however, it provides information with regard to the source and geographic origin of the sample (Barth 1989).

CONCLUSIONS
Typical honeys from the State of Ceará contain mostly M. caesalpiniaefolia as dominant pollen; in turn, honeys from the State of Piauí contain Piptadenia sp., demonstrating the importance of these plants in those states during the months of February through August.
Species like M. verrucosa, M. scabrella, Serjania sp., C. urucurana, and two species in the genus Borreria were observed in both states studied and have potential as honey plants during certain seasons of the year.