With the objective of knowing the kind of honey produced in Uruguay, periodical meetings with beekeeping groups from different areas of the country were performed. Sixty samples of honey were obtained from apiaries located in states of higher production: Artigas, Rivera, Salto, Paysandú, Río Negro, Soriano, Colonia, San José, Canelones and Florida (Figure 1).

Figure 1.
Study area. Sampling sites are represented by dots.

The vegetation of interest for beekeeping in these zones are: Eucalyptus forestation, prairies cultivated for cattle industry (Lotus spp., Medicago sativa, Trifolium pratense y Trifolium repens), natural forest, fruit trees (Citrus spp., Malus domestica) and natural prairies (Baccharis spp, Senecio spp., Eryngium spp.).

For obtaining mono-floral honey the honeycombs were collected directly from apiary and were extracted at the laboratory with plastic press. The honeycombs arrived at the laboratory in a box specially designed for this purpose, with the aim of avoiding contamination with metals in the transport or extraction. Once the honey samples arrived at the laboratory, the extraction of the honey was done through pressing from small pieces of honeycomb discarding zones with pollen reserves. After this, the honey was filtered to separate the wax pieces (mesh of 1 mm size) and was stored in glass jars until analysis. Honey samples from private honey extraction factories, authorized by the MGAP (Livestock, Agriculture and Fisheries Ministry from Uruguay) were also received. The samples were identified by the name of the beekeeper, locality, date of the harvest, partner vegetation in the apiary and botanical origin of honey proposed by beekeepers.

A pollen collection was produced for reference in the botanical characterization of honeys, with botanical material collected in several sampling fields around apiaries that provided honey for this study. For pollen processing, the technique of pollen mounting from the microscopical analysis of sediment, from the Spanish Food Legislation, was used with some modifications. The flower antenna were extracted and washed with ethylic ether for degreasing the pollen removed. The ether was evaporated in the gas extraction chimney and once dried 10 mL of distillated water was added, removing the surplus parts of the flowers. It was centrifuged for 10 m at 2500 rpm and the surplus was moved away for making the mounting of the pollen residue in the hot glycerin over a microscope slide. It was covered with a slide cover and sealed with paraffin over a hot iron for conservation.

For the botanical determination, 20 g of honey were processed with distilled water according to Louveaux et al. (1978). After the final centrifugation, the residual was mounted over a microscope slide to be analyzed under a microscope (400X augment) and the counting of 600 pollen grains was done according to Von der Ohe et al. (2004) for determining the relative abundances of each pollen type.

Honeys that presented dominance of a particular kind of pollen with a percentage of pollen higher or equal than 45% were considered mono-floral. When there was no dominance of a floral type, the honey was considered to be poly-floral (Louveaux, et al., 1978; Bogdanov, et al., 1997). Exceptions were considered for the mono-floral honey of citrus and Eucalyptus, being of 18% and 75% respectively.

The electric conductivity (μsm/cm) was determined in a solution of honey, the equivalent of 20 g of dry matte, in 100 mL of deionized water and using a conductimeter. (Orion Model 105) (Bogdanov, 1999). For determining the honey color, the sample was placed in a manual colorimeter cell, in its liquid state at environmental temperature. The sample color was visually compared with the color scale installed at the colorimeter (Pfund scale, Fell, 1978).

It was possible to collect 198 botanical samples that were used for the pollen extraction. Samples of repeated species, feminine specimens, flower buttons immature or too mature (without pollen) were discarded. A collection of pollen of 99 botanical species was conformed. This collection may be used in future studies and is stored at the Facultad de Ciencias, Universidad de la República, Montevideo, Uruguay.

35% of the species present in the collection corresponded to the category natural prairies, 31% to natural forest, 19% to species of agriculture interest (harvested), 12% corresponded to the ornamental category and 2% corresponded to forestry species.

In the palynological analysis, 26 botanical species were registered in the total of the samples, and 6 pollen types were not identified. In 41,7% of the cases, the pollen result determined at the laboratory did not coincide with the botanical origin proposed by the beekeeper that sent the samples.

The majority of the honeys were mono-floral (78,3%). The kind of mono-floral honey most common was from Lotus spp. (Lotus), from the 47 mono-floral samples 34 corresponded to this kind (72%) and from all the honey samples, Lotus occurred in 49 (81,7%). The second most represented mono-floral honey (13,3%) corresponded to Eucalyptus spp. The rest of the mono-floral honeys corresponded to Brassica napus (Colza), Baccharis spp. (“Chirca”) and Citrus sp. (Citrus). These results highlight the importance of the floral fodder species to produce honey in Uruguay. Results coincide with Díaz and Raudovinche (2010), who signal that the extensive cattering progress that is occurring at present (plantations of Glycine max, soybean) has decreased the area destined to fodder grass affecting the apiary performance (Table 1).

Table 1.
Samples analyzed in this study.

From the natural forest honeys (multi-floral) the botanic species more represented were Blepharocalyx salicifolius (Arrayán) and Myrcianthes cisplatensis (red Guava tree), grouped in the table as native Mirtaceas.

It was also verified that the botanical origin of one of the honeys was not floral. In this sample that was sent by a beekeeper there were no pollen grains that allowed to identify it, but there were abundant elements of honeydew that would indicate an extrafloral origin.

Conductivity values ranged between 184 and 1306 µsm/cm. Darker honeys came from Eucaliptus forests and natural forest, and these also present high values of electric conductivity, higher than 800 µS/cm (Table 2).

The majority of the honeys that were analyzed came from the west of the country (Figure 1), area characterized by prairies, hence the honeys were not dark in this study.

Taking into consideration only honeys extracted directly from honeycombs at laboratory, under proper conditions that did not affect the composition of the honey (n=28), the conductivity values ranged between 184 and 889 µS/cm. Honeys with values over 800µ S/cm came from Eucalyptus spp., honeydew, multifloral (with high content of Eucalyptus spp.) and Lotus spp. (with high content of Eryngium spp.) (Table 2).

Considering only honeys extracted at the laboratory, it was observed that the color values oscillated between 12,8 and 88,9 mm Pfund (Extra White and Ambar respectively), with a mean value of 54 mm Pfund, that corresponds to Light Amber. This color category was the most represented between the honeys (Figure 2)

Table 2.
Pollen content from honeys extracted at the laboratory, with electric conductivity values higher than 800 µS/cm.

1-Lotus spp., 2-Medicago sativa, 3-Trifolium pratense, 4-Trifolium repens, 5-Eucaliptus spp., 6-Senecio spp., 7-Echium plantagineum,

8-Eryngium spp., 9-Blepharocalyx salicifolius, 10-Schinus longifolius, 11-Bacharis sp., 12-Ammi biznaga, 13-Acacia sp., 14-Salix sp., 15-Prosopis alba, 16-Brassicaceae, 17-Malus domestica.

Figure 2.
Quantity of honeys for color category.

Table 3.
Brassica napus honey.

Regarding the botanical origin, the results showed that honey of Lotus sp. present some color variation, but always in the lower colours in the Pfund scale. Only one sample was from the Amber category (89 mm Pfund) and its darker color (more than the other Lotus sp.) was explained because of the partner flora registered in the pollinic analysis: Eryngium spand Blepharocalyx salicifolius) with important proportions (Table 2). It was verified that at higher purity in the percentage of pollen, the lighter the honey classified in the categories Extra White and White.

The same occurred with two samples obtained of colza (Brassica napus). Among them, there was high variability of the color, instead of coming from the same region and dates, approximately. However, it was observed that the companion flora in the pollinic analysis presented a high proportion of Eucalyptus spp. that makes natural color and the conductivity of the Colza honey variated (Table 3).

From the ordination analysis performed with samples of Lotus sp. it was possible to observe that the samples from Soriano presented higher dispersion than samples from Río Negro. The variables that contributed to the Principal Component Analysis (PCA) were color and conductivity. The cumulated variance in the first axis was 47,1% and in the second axis 67,4%. Samples with higher content of Echium plantagineum, Medicago sativa and Trifolium pratense, as companion flora, presented minor color and conductivity. Samples with higher values of color presented higher content of Eucalyptus sp., sunflower and soybean (Figure 3).

Figure 3.
Principal component analysis of the honey samples from Uruguay (k: conductivity, Euca: Eucalyptus spp., borr: Echium plantagineum, gir: sunflower, alf: Medicago sativa, rojo: Trifolium pratense).

Figure 4.
Linear regression and confidence interval between electric conductivity (K) and color value of all the samples (A) and from samples from honeycombs (B).

On the other hand, Eucalyptus sp. honeys presented light amber color, similar to a sample of natural forest and Baccharis trimera sample. Despite the low number of these floral type samples, we might propose that the color do not vary from what we obtained, since samples were monofloral.

Honey color and electric conductivity presented a positive relationship (Figure 4).

The darker the color of the honey, the higher the electric conductivity it presented, independently of the locality or botanical origin.