J. Agric. Food Chem.

2002, 50, 2633−2637 2633

Analysis of Volatiles from Spanish Honeys by Solid-Phase

Microextraction and Gas Chromatography−Mass Spectrometry
ROSA A. PEÄ REZ,† CONSUELO SAÄ NCHEZ-BRUNETE,‡ ROSA M. CALVO,§ AND
JOSEÄ L. TADEO*,‡
Departamento de Protección Vegetal, Departamento de Medio Ambiente, and Servicio de Biometrı́a,
INIA, Apdo 8111, 28080 Madrid, Spain

Headspace solid-phase microextraction (SPME), followed by gas chromatography (GC)-mass

spectrometry (MS) determination, has been used for the analysis of honey volatiles. Two SPME
fibers were employed to study the composition of volatiles from various types of Spanish honeys.
The best results were obtained with the Carboxen/PDMS fiber, using a homogenization time of 1 h
at 70 °C and a sampling period of 30 min. A total of 35 compounds were detected, most of them
identified by GC-MS and quantified using external standards. Differences in the composition of honey
volatiles were obtained, and these results allowed the differentiation of honeys. However, further
studies are necessary to confirm the utility of this technique as an alternative tool for the
characterization of the floral origin of honeys.

KEYWORDS: Honey; volatiles; solid-phase microextraction; gas chromatography; mass spectrometry

INTRODUCTION (6), although other techniques such as extraction with organic

solvents (9, 12), column extraction (8), and dynamic headspace
The floral origin of honeys is a very important characteristic
analysis (5, 7, 11) have also been used.
of the quality of these food products. Unifloral honeys possess
Solid-phase microextraction (SPME) is a solventless extrac-
distinctive flavors, mainly derived from their nectar sources,
tion technique based on the exposure of an immobilized
indicating the presence of volatile components. Nevertheless,
stationary phase into the matrix containing the analytes, which
the floral origin of honeys is usually determined by pollen
could be liquid, solid, or gaseous, followed by thermal desorp-
analysis, physicochemical parameters, and organoleptic proper-
tion of the analytes in the injector of a gas chromatograph (13).
ties. In fact, the characterization of various Spanish honeys has
This technique has been successfully used for the analysis of
been reported based on palynological properties, physicochem-
volatile flavor compounds (14-17). However, data on the
ical parameters, and composition of sugars (1, 2). However,
application of SPME to the analysis of volatile compounds
the presence of the specific pollen is low in some types of honey
present in honey are scarce in the available literature (18).
(citrus, lavender, and rosemary) making a wider range of
markers necessary to ascertain the origin of honeys. The aim of the present work was to study the determination
of volatiles from honey using headspace SPME and gas
A large number of organic compounds have been described
chromatography with mass spectrometric detection (GC-MS).
as components of different types of honeys (3-10). The main
This method was applied to determine the main volatile
components or source specific honey volatiles belong, in general,
components of various Spanish unifloral honeys (orange,
to three principal categories such as terpenes, norisoprenoids,
eucalyptus, rosemary, lavender, and thyme) and to evaluate the
and benzene derivatives (9). Some of these substances have been
possibility of using this technique for the characterization of
described as characteristics of the floral source, and other
honeys.
compounds, like some alcohols, branched aldehydes, and fural
derivatives, may be related to the microbial purity of processing
MATERIALS AND METHODS
and storage conditions of honey (11).
The identification and quantification of volatile compounds Materials. Dimethyl sulfide, 1-hydroxy-2-propanone, 3-methyl-3-
from a complex mixture such as honey are difficult. Analysis buten-1-ol, 3-hydroxy-2-butanone, 3-methyl-1-butanol, 2-methyl-1-
of volatile and semivolatile components of honey has been butanol, 2,3-butanediol, 2-furaldehyde (furfural), 2-furanmethanol,
carried out by simultaneous steam distillation-solvent extraction 1-hexanol, benzaldehyde, benzyl alcohol, benzene acetaldehyde, 2-phe-
nyl ethanol, and 3,5,5-trimethyl-2-cyclohexen-1-one were purchased
from Sigma-Aldrich Chimica. Ethanol, acetone, and acetic acid were
* To whom correspondence should be addressed. Tel.: 34-913476821. obtained from Panreac (Spain).
Fax: 34-913572293. E-mail: [email protected]. Four replicate samples of honeys from various producers and regions
† Departamento de Protección Vegetal.
‡ Departamento de Medio Ambiente. were collected for this study from commercial sources. The honeys
§ Servicio de Biometrı́a. studied were orange (Citrus spp.), eucalyptus (Eucalyptus spp.),

10.1021/jf011551r CCC: $22.00 © 2002 American Chemical Society

Published on Web 03/23/2002
2634 J. Agric. Food Chem., Vol. 50, No. 9, 2002 Pérez et al.

rosemary (Rosmarinus officinalis L.), lavender (LaVandula latifolia

Med.), and thyme (Thymus spp.). Another eight replicates of orange
honeys were obtained directly from beekeepers of Valencia (Spain);
these honeys were collected in a zone where the main varieties were
Navel and Naveline oranges (Citrus sinensis, Osbeck). All honey
samples originated from Spain.
Carboxen/PDMS and PDMS/DVB fibers (Supelco, Spain) were used
to extract headspace volatiles of honey. Before analysis, the fibers were
preconditioned in the injection port of a gas chromatograph according
to the instructions provided by the supplier.
GC-MS. Analyses were performed using a Hewlett-Packard 6890
(Waldbronn, Germany) gas chromatograph equipped with a mass
spectrometric detector (MSD) model HP 5973. Samples were injected
splitless, and volatiles were separated using a fused silica capillary
column (HP-5MS), diphenyl dimethylpolysiloxane as nonpolar station-
ary phase (30 m × 0.25 mm i.d.), and 0.25 µm film thickness, supplied
by Agilent (Madrid, Spain) with helium as carrier gas at a flow rate of
1 mL/min. The injector port temperatures were 270 °C for Carboxen/
PDMS and 250 °C for PDMS/DVB. The interface temperature was
250 °C. The oven temperature was maintained at 50 °C for 4 min,
programmed at 10 °C/min to 200 °C, held for 0.5 min, programmed at
20 °C/min to 230 °C, and held 1 min. Figure 1. Chromatograms obtained by headspace SPME analysis of
The mass spectrometer was operated in the electron impact mode
eucalyptus honey with Carboxen/PDMS and PDMS/DVB fibers.
with an electron energy of 70 eV; source temperature, 230 °C;
quadrupole temperature, 150 °C; mass range m/z 25-500; scan rate, high signal of 5-hydroxymethyl-2-furancarboxaldehyde was
3.62 s per scan; and EM voltage, 1150. detected, as a result of the fructose decomposition. This signal
Compounds were identified based on NIST mass spectra library has been described as an indicator of loss of honey quality by
search. Most of these compounds were further confirmed by comparing overheating during processing or storage or by honey adultera-
their mass spectra and retention times with those obtained for standards. tion with acid-hydrolyzed inverted sugar (12).
Headspace SPME Analysis. Analyses of volatile compounds of The effect of the other experimental conditions, the equilibra-
honeys were carried out by weighing 1 g of honey in 4 mL vials (4.4
tion and sampling times, was studied at 70 °C. To evaluate the
cm height × 1.1 cm i.d.) with PTFE/silicone septa and a stirring bar.
The vials and septa were previously heated, at 150 °C during 24 h, to
equilibration time, SPME analyses were carried out after 30 or
remove undesirable chromatographic signals able to interfere with honey 60 min of equilibration. The obtained results showed a higher
volatiles. Samples were maintained and magnetically stirred for 1 h at number of chromatographic signals after 60 min of equilibration
70 °C to allow equilibration. Sampling of the volatile honey compounds (data not shown). This equilibration time is in agreement with
was done by inserting the sheathed fiber through the septum and the results previously reported by Bartelt for headspace SPME
exposing it to the headspace for 30 min. The fiber was then retracted analysis of various organic compounds (21). The effect of the
and transferred to the injector port of the gas chromatograph where sampling time on the extraction efficiency was evaluated by
the compounds were desorbed for 5 min. sampling the headspace during 30 or 60 min, after 1 h of
Quantification. The concentrations of honey volatiles were deter- equilibration, and similar peak areas were obtained in both
mined by comparing the ratios of the peak areas in the sample with conditions.
those found for a known concentration of these compounds in the
These results indicated that a stirring time of 1 h at 70 °C,
external standard mixture, prepared in ethanol with the compounds
indicated above. The quantification of compounds not present in the followed by a sampling period of 30 min, were adequate
standard mixture was performed with respect to benzyl alcohol, conditions for SPME analysis of honey volatiles.
assuming a response factor equal to 1. The detector response was linear Type of Fiber. Two different SPME fibers, Carboxen/PDMS
in the range of concentrations found. The limit of detection was 0.1 ng and PDMS/DVB, were evaluated in the conditions described
of benzyl alcohol, considering a signal higher than three times the above. These fibers were selected based on the results previously
background noise. obtained in our laboratory in the SPME analysis of organic
Statistical Analysis. The concentrations of the components detected volatiles (22). Both fibers were used in the headspace SPME
in honeys were statistically tested. Univariate analysis (ANOVA, analysis of honey samples with different floral origins. The
Kolmogorov-Smirnov test of good fitness) and multivariate analysis results obtained with the Carboxen/PDMS fiber were better in
(stepwise discriminant analysis, canonical discriminant analysis, M. De
all cases, due to the higher number and concentration of honey
Box test) were used in the analyses. These were conducted using the
BMDP7M and CANDIS programs from BMDP Statistical Software
volatiles, mainly those with retention times between 0 and 4
release 7 (19) and SAS version 8 (20), respectively. min. Figure 1 shows representative chromatograms obtained
with these fibers for eucalyptus honey. As a result of these
assays, the Carboxen/PDMS fiber was selected for the analysis
RESULTS AND DISCUSSION
of volatile compounds from honey.
SPME Analysis of Honey Volatiles. Equilibration and Analysis of Honey Volatiles. Five commercially available
Sampling. Preliminary assays with Carboxen/PDMS were car- Spanish honeys of different floral origin (orange, eucalyptus,
ried out in order to establish the experimental conditions for rosemary, lavender, and thyme) were analyzed by headspace
headspace SPME analysis of honey volatiles, particularly the SPME and GC-MS. Table 1 shows the volatile compounds
temperature and the equilibration and sampling times. detected in the studied honeys. These compounds were quanti-
Three temperatures were evaluated for headspace analysis: fied, and the results obtained are summarized in that table. The
50, 70, and 85 °C. The best results were obtained at 70 °C (data quantification of honey volatiles has been carried out by
not shown). At this temperature, the number of chromatographic comparing their peak areas with those of an external standard
signals was higher than those obtained at 50 °C. When the mixture. Figure 2 shows representative total ion chromatograms
analysis was carried out at 85 °C, or at a higher temperature, a of the analyzed samples.
SPME Analysis of Honey Volatiles J. Agric. Food Chem., Vol. 50, No. 9, 2002 2635

Table 1. SPME Analysis of Volatiles from Spanish Honeys Using Carboxen/PDMS Fiber

mean amount extracted by SPME, ng (SD)

retention honey samples
peak no. Ida compd time (min) orange eucalyptus rosemary lavender thyme
1 A ethanol 1.24 14.4 (1.0) 30.0 (7.1) 105.8 (10.1) 21.3 (6.0) 98.3 (7.7)
2 A acetone 1.27 32.4 (2.3) 6.0 (2.2) 10.1 (2.5) 11.5 (1.2) 11.4 (5.0)
3 A dimethyl sulfide 1.30 4.7 (0.5) 5.3 (1.9) 8.0 (2.2) 81.6 (5.6) 42.0 (5.0)
4 A acetic acid 1.40 36.0 (7.3) 61.2 (18.0) 44.0 (3.4) 57.5 (4.0) 77.7 (5.9)
5 B 2-butanol 1.46 11.9 (1.4)
6 B 2-methyl-3- 1.49 20.1 (2.3) 3.0 (0.8)
buten-2-ol
7 B 2-methyl-1- 1.54 2.4 (0.8) 2.8 (0.2) 3.1 (0.4)
propanol
8 B 3-methyl butanal 1.65 2.8 (1.1)
9 A 1-hydroxy-2- 1.69 28.4 (5.0) 17.7 (3.4) 27.6 (4.7) 41.1 (5.7) 43.8 (3.5)
propanone
10 A 3-hydroxy-2- 1.95 2.2 (0.9) 117.6 (11.1) 2.0 (0.2) 3.4 (0.1) 14.5 (3.9)
butanone
11 A 3-methyl-3- 2.13 1.8 (0.9) 2.7 (0.9) 33.2 (2.5) 2.8 (0.8)
buten-1-ol
12 A 3-methyl- 2.15 4.6 (1.8) 2.1 (0.6) 5.5 (0.6)
1-butanol
13 A 2-methyl- 2.18 1.7 (0.4) 2.7 (1.2) 1.6 (0.5) 2.6 (0.6)
1-butanol
14 B dimethyl disulfide 2.26 1.6 (0.7) 1.8 (0.3)
15 B 2-methyl-2- 2.53 4.2 (0.6)
buten-1-ol
16 B 3-methyl-2- 2.55 5.2 (0.1)
buten-1-ol
17 A 2,3-butanediol 2.64; 2.78 10.2 (2.1) 132.7 (37.4) 29.0 (3.4) 853.0 (57.3) 115.9 (14.7)
18 A furfural 3.51 55.0 (5.7) 15.6 (1.9) 39.4 (4.4) 24.6 (5.9) 35.6 (2.6)
19 B 3-methyl butanoic 3.74 12.4 (2.6)
acid
20 A 2-furanmethanol 4.03 3.0 (0.9) 23.4 (0.7) 6.0 (0.2)
21 A 1-hexanol 4.34 20.4 (2.4)
22 unknown 1 6.06 8.8 (1.7)
23 unknown 2 6.17 9.6 (1.3)
24 A benzaldehyde 6.46 4.2 (0.8) 6.4 (1.7) 4.2 (1.0) 16.8 (4.4) 3.8 (0.8)
25 A benzene acetaldehyde 8.17 2.2 (0.7) 4.3 (1.2) 10.3 (4.2) 8.2 (1.1) 13.7 (5.8)
26 unknown 3 8.82 25.3 (2.2) 2.3 (0.6) 5.2 (1.0) 2.2 (0.7)
27 B 3,7-dimethyl-1,5,7- 9.27 12.7 (1.1)
octatrien-3-ol
28 A 2-phenyl ethanol 9.45 3.7 (1.6) 1.8 (0.4) 4.5 (1.0) 9.6 (2.5) 1.6 (0.1)
29 A 3,5,5-trimethyl-2- 9.53 11.5 (2.0)
cyclohexen-1-one
30 B lilac aldehyde 1 9.85 5.2 (1.5)
31 B 2,6,6-trimethyl-2- 9.93 0.9 (0.2) 1.1 (0.4) 1.1 (0.2)
cyclohexene-1,4-dione
32 B lilac aldehyde 2 10.07 14.8 (5.0)
33 B lilac aldehyde 3 10.31 7.8 (2.4)
34 B 2-amino methyl 12.95 4.3 (1.2)
benzoate
35 B 4,5,6,7-tetrahydro- 16.76 1.5 (0.2)
3,6-dimethyl
benzofuran

aIdentification: (A) comparison of retention time and mass spectrum with that of an authentic sample recorded under the same conditions. (B) Tentative identification
by comparison of mass spectrum with NIST library (computer) spectrum.
A total of 35 signals were identified in the studied honeys three signals of lilac aldehydes (R,5-dimethyl-5-ethylenyl-2-
and 11 of these compounds were detected in all of the honeys: tetrahydrofuran acetaldehydes), floral odor, seems to be a
ethanol, acetone, dimethyl sulfide, acetic acid, 1-hydroxy-2- particular characteristic of the orange honey. Lavender honey
propanone, 3-hydroxy-2-butanone, 2,3-butanediol, furfural, ben- could be distinguished by the detection of 3,7-dimethyl-1,5,7-
zaldehyde, benzene acetaldehyde, and 2-phenyl ethanol. Some octatrien-3-ol and 3,5,5-trimethyl-2-cyclohexen-1-one, together
of these 11 compounds, ethanol, furfural, benzene acetaldehyde, with a high amount of 2,3-butanediol and a certain presence of
acetone, and dimethyl sulfide, have been reported as common 2-methyl-3-buten-2-ol, 3-methyl-3-buten-1-ol, 1-hexanol, 2-meth-
components of various unifloral honeys at variable concentra- yl-2-buten-1-ol, and 3-methyl-2-buten-1-ol. A high content of
tions (5, 11). The amount of these compounds found in the 3-hydroxy-2-butanone and the presence of dimethyl disulfide,
studied Spanish honeys was quite different depending on the 3-methylbutanoic acid, and 4,5,6,7-tetrahydro-3,6-dimethyl ben-
floral origin. zofuran seem to be characteristic of the eucalyptus honey. The
Some of the compounds shown in Table 1 can be used to chromatograms obtained for rosemary honey, Figure 2, showed
distinguish the different types of honey. Thus, the presence of few representative signals; the high content of ethanol and the
2636 J. Agric. Food Chem., Vol. 50, No. 9, 2002 Pérez et al.

Figure 3. Plots of discriminant canonical functions. Honey codes: (1)

orange, (2) eucalyptus, (3) rosemary, (4) lavender, and (5) thyme.

Table 2. Canonical Structure Values

variable Can 1 Can 2 Can 3 Can 4

ethanol −0.3935 −0.3552 0.7577 −0.3759
2-butanol −0.1767 −0.3655 0.7223 0.5579
3-hydroxy-2-butanone −0.3094 0.9316 0.0259 0.1875
Figure 2. Representative chromatograms obtained by headspace SPME 3-methyl-2-buten-1-ol 0.9988 0.0437 0.0076 −0.0164
analysis with Carboxen/PDMS for orange, eucalyptus, rosemary, lavender, 2,3-butanediol 0.9676 0.1452 0.1064 0.0583
and thyme honeys. lilac aldehyde −0.2140 −0.3983 −0.8201 0.2670

presence of dimethyl disulfide were the most representative.

Thyme honey also showed a high content of ethanol and few the original variables, called canonical structures (Table 2), were
chromatographic signals at retention times higher than 3 min, used in conjunction with plots of discriminant canonical
although the presence of 2-butanol and 3-methyl butanal has functions to aid interpretation of group differences. The plots
been detected only in this honey type. obtained are shown in Figure 3.
Various factors, such as plant source, seasonal and climatic Lavender (number 4) and eucalyptus (number 2) honeys were
conditions, and processing or storage circumstances, may affect clearly separated from the other honeys in plot A. Lavender
the composition of honey volatiles. Therefore, additional SPME honey was positioned at the extreme positive side of the first
analyses were carried out with samples of orange honeys canonical function (Can 1). The high positive canonical structure
obtained directly from beekeepers, to compare the volatile values (Table 2) were observed for 3-methyl-2-buten-1-ol and
compounds detected for these honeys with those of the 2,3-butanediol, and according to this, both compounds are
previously analyzed commercial orange honeys. In these important to distinguish lavender honey from the others. Other
analyses, three signals identified as lilac aldehydes were compounds, like 3,7-dimethyl-1,5,7-octatrien-3-ol and 3,5,5-
observed in the samples, as found previously in the commercial trimethyl-2-cyclohexen-1-one, may also be used in the charac-
orange honey. Therefore, these components seem to be a terization of lavender honey. Eucalyptus honey was positioned
characteristic of Spanish orange honeys. at the extreme positive side of the second canonical function
Differentiation of Spanish Honey Types. The results of (Can 2); 3-hydroxy-2-butanone was the characteristic variable
SPME-GC analysis, summarized in Table 1, were used as of this honey. The high content of 3-hydroxy-2-butanone is in
variables in the statistical analysis. Previously to the discriminant agreement with the results obtained by D’Arcy (9) in unmethy-
analysis, the Kolmogorov-Smirnov test was done to evaluate lated eucalyptus Australian honeys.
the normality. As a result of this previous test and the M. De Plots B and D allow the differentiation of thyme (number 5)
Box test, no transformation was done. Canonical discriminant and orange (number 1) honeys from the others, considering the
and stepwise discriminant analyses were made in order to third discriminant canonical function (Can 3). As a result from
evaluate the studied honeys and to establish the characteristic the canonical structure values of function 3, orange honey was
variables of each honey type (19, 20, 23). defined by the variable corresponding to lilac aldehyde and
The univariate and multivariate analyses showed very sig- thyme honey by the ethanol and 2-butanol variables.
nificant differences (Pr < 0.0001) among the average values Finally, rosemary honey (number 3) could be distinguished
for all of the variables in relation to honey type. The canonical from the other honeys, in plots C and D, by the discriminant
discriminant analysis showed that the first four discriminant canonical function 4 (Can 4), which shows high negative values
functions provided a good summary of the original data of the for ethanol in this honey type. Although, thyme also has a high
considered variables. Thus, the proportion of the total ac- ethanol content, these honey types can be differentiated by the
cumulated dispersion with the two and three first functions were 2-butanol content. Nevertheless, other markers are necessary
0.90 and 0.96, respectively, and the squared canonical correla- to better characterize these honeys.
tions were 0.998, 0.990, 0.980, and 0.968 for these four The five honeys studied were correctly classified by using
functions. The correlations between the canonical variables and these canonical functions. In addition, a Jack-knifed classifica-
SPME Analysis of Honey Volatiles J. Agric. Food Chem., Vol. 50, No. 9, 2002 2637

tion matrix (19) was used as a special case of the general cross- (8) Shimoda, M.; Wu, Y.; Osajima, Y. Aroma compounds from
validation, and again, all of the honeys were correctly classified. aqueous solution of haze (Rhus succedanea) honey determined
Finally, eight samples of orange honey from apiaries placed in by adsorptive column chromatography. J. Agric. Food Chem.
a citrus zone were considered. These honeys were classified 1996, 44, 3913-3918.
using the canonical functions obtained above, and all of the (9) D’Arcy, B. R.; Rintoul, G. B.; Rowland, C. Y.; Blackman, A. J.
Composition of Australian honey extractives. 1. Norisoprenoids,
samples were correctly classified as orange honeys.
monoterpenes, and other natural volatiles from Blue Gum
The headspace SPME-GC analysis of different types of (Eucalyptus leucoxylon) and Yellow Box (Eucalyptus mel-
Spanish honeys described above allows the identification and liodora). J. Agric. Food Chem. 1997, 45, 1834-1843.
quantification of a wide range of volatile components of honey. (10) Guyot, C.; Bouseta, A.; Scheirman, V.; Collin, S. Floral origin
The best results were obtained with the Carboxen/PDMS fiber. markers of chestnut and lime tree honeys. J. Agric. Food Chem.
A temperature of 70 °C, 1 h of sample equilibration, and 30 1998, 46, 625-633.
min of SPME sampling were established as adequate conditions (11) Bouseta, A.; Collin, S.; Dufour, J.-P. Characteristic aroma profiles
for SPME analysis of honey volatiles. A total of 35 compounds of unifloral honeys obtained with a dynamic headspace GC-MS
were detected, most of them identified by GC-MS and system. J. Apic. Res. 1992, 31, 96-109.
quantified using external standards. (12) Graddon, A. D.; Morrison, J. D.; Smith, J. F. Volatile constituents
The results obtained show that SPME followed by GC-MS of some unifloral Australian honeys. J. Agric. Food Chem. 1979,
27, 832-837.
can be successfully used in the analysis of volatile components
(13) Pawliszyn, J. Solid-Phase Microextraction. Theory and Practice;
of honey samples. Differences in the composition of volatiles Wiley-VCH: New York, 1997.
from various unifloral Spanish honeys were observed, and the (14) Sostaric, T.; Boyce, M. C.; Spickett, E. E. Analysis of the volatile
comparative analysis of the volatile composition shows that components in vanilla extracts and flavorings by solid-phase
some compounds can be used for the characterization of the microextraction and gas chromatography. J. Agric. Food Chem.
floral source. The results obtained so far indicate that this 2000, 48, 5802-5807.
technique may be a useful tool for the authentication of the (15) Steffen, A.; Pawliszyn, J. Analysis of flavor volatiles using
floral origin of honeys. Further studies including other honey headspace solid-phase microextraction. J. Agric. Food Chem.
types are however necessary in order to confirm the utility of 1996, 44, 2187-2193.
this technique as an alternative tool for the characterization of (16) Song, J.; Fan, L.; Beaudry, R. M. Application of solid-phase
honeys. microextraction and gas chromatography/time-of-flight mass
spectrometry for rapid analysis of flavor volatiles in tomato and
strawberry fruits. J. Agric. Food Chem. 1998, 46, 3721-3726.
LITERATURE CITED
(17) Ibáñez, E.; López-Sebastián, S.; Ramos, E.; Tabera, J.; Reglero,
(1) Serra Bovehı́, J.; Ventura Coll, F. Characterization of citrus honey G. Analysis of volatile fruit components by headspace solid-
(Citrus spp.) produced in Spain. J. Agric. Food Chem. 1995, phase microextraction. Food Chem. 1998, 63, 281-286.
43, 2053-2057. (18) Guidotti, M.; Vitali, M. Identificazione di composti organici
(2) Mateo, R.; Bosch Reig, F. Sugar profiles of Spanish unifloral volatili in mieli mediante SPME e GC/MS. Ind. Aliment. 1998,
honeys. Food Chem. 1997, 60 (1), 33-41. 37, 351-356.
(3) Tan, S. T.; Wilkins, A. L.; Holland, P. T.; McGhie, T. K. (19) Dixon, W. J., Ed. BMDP Statistical Software Release 7;
Extractives from New Zealand honeys. 3. Unifloral thyme and University of California Press: Los Angeles, 1992; pp 363-
willow honey constituents. J. Agric. Food Chem. 1990, 38, 387.
1833-1838. (20) SAS Institute Inc. SAS/STAT User’s Guide, version 8. 2001.
(4) Wilkins, A. L.; Lu, Y.; Tan, S. T. Extractives from New Zealand (21) Bartelt, R. J. Calibration of a commercial solid-phase microex-
honeys. 4. Linalool derivatives and other components from traction device for measuring headspace concentrations of organic
nodding thistle (Carduus nutans) honey. J. Agric. Food Chem. volatiles. Anal. Chem. 1997, 69, 364-372.
1993, 41, 873-878. (22) Pérez, R. A.; Sánchez-Brunete, C.; Esteban-Durán, J. R.; Tadeo,
(5) Overton, S. V.; Manura, J. J. Flavor and aroma in commercial J. L. SPME analysis of potential attractants for palm weevils.
bee honey. A purge-and-trap thermal desorption technique for Int. J. EnViron. Anal. Chem. 2001, 79, 229-240.
the identification and quantification of volatiles and semivolatiles (23) Hair, J. F., Jr.; Anderson, R. E.; Tatham, R. L.; Black, W. C.
in honey. Am. Lab. 1994, 26, 45, 47-53. Análisis MultiVariante, 5th ed.; Prentice Hall Iberia: Madrid,
(6) Bouseta, A.; Scheirman, V.; Collin, S. Flavor and free amino Spain, 1999.
acid composition of lavender and eucalyptus honeys. J. Food
Sci. 1996, 61, 683-687. Received for review November 26, 2001. Revised manuscript received
(7) Bouseta, A.; Collin, S. Optimized Likens-Nickerson methodology February 1, 2002. Accepted February 1, 2002.
for quantifying honey flavors. J. Agric. Food Chem. 1995, 43,
1890-1897. JF011551R