European Journal of Soil Science, June 2008, 59, 551–558 doi: 10.1111/j.1365-2389.2008.01022.

Decomposition in soil and chemical characteristics

of pollen

E. A. W EBSTER a , E. L. T ILSTON b , J. A. C HUDEK c & D. W. H OPKINS a , d

a
School of Biological and Environmental Sciences, University of Stirling, Stirling FK9 4LA, bDepartment of Soil Science, University of
Reading, Whiteknights, Reading RG6 6DW, cCollege of Life Sciences, University of Dundee, Dundee DD1 4HN, and dScottish Crop
Research Institute, Invergowrie, Dundee DD2 5DA, UK

Summary
The input to soils made by pollen and its subsequent mineralization has rarely been investigated from a soil
microbiological point of view even though the small but significant quantities of C and N in pollen may
make an important contribution to nutrient cycling. The relative resistance to decomposition of pollen
exines (outer layers) has led to much of the focus of pollen in soil being on its preservation for archaeo-
logical and palaeo-ecological purposes. We have examined aspects of the chemical composition and
decomposition of pollen from birch (Betula alba) and maize (Zea mays) in soil. The relatively large N
contents, small C-to-N ratios and large water-soluble contents of pollen from both species indicated
that they would be readily mineralized in soil. When added to soil and incubated at 16°C an amount of
C equivalent to 22–26% of the added pollen C was lost as CO2 within 22 days, with the Z. mays pollen
decomposing faster. For B. alba pollen, the water-soluble fraction decomposed faster than the whole
pollen and the insoluble fraction decomposed more slowly over 22 days. By contrast, there were no sig-
nificant differences in the decomposition rates of the different fractions from Z. mays pollen. Solid-state
13
C nuclear magnetic resonance (NMR) revealed no gross chemical differences between the pollen of
these two species, with strong resonances in the alkyl- and methyl-C region (0–45 p.p.m.) indicative of
aliphatic compounds, the O-alkyl-C (60–90 p.p.m.) and the acetal- and ketal-C region (90–110 p.p.m.)
indicative of polysaccharides, and the carbonyl-C region indicative of peptides and carboxylic acids. In
addition, both pollens gave a small but distinct resonance at 55 p.p.m. attributed to N-alkyl-C. The reso-
nances attributed to polysaccharides were lost completely or substantially reduced after decomposition.

Introduction is readily recycled by soil micro-organisms (Perez-Moreno &

Read, 2001). Even though this amount of N could exceed the
A large quantity of pollen from wind-pollinated plants fails to
annual N recycled through litter fall in some forests (Greenfield,
reach stigmata and is therefore unsuccessful in terms of plant
1999), information on the mineralization of pollen is rather lim-
reproduction (Greenfield, 1999; Perez-Moreno & Read, 2001).
ited (Greenfield, 1999; Hopkins, 2000). In addition to its role in
Much of this ‘unsuccessful’ pollen falls on the soil surface and
nutrient cycling, the fate of pollen in soil merits investigation
may be incorporated through the activities of soil invertebrates
because of: its possible allelopathic role (Mallik & Williams,
(Davidson et al., 1999; Van Mourik, 2003), where it contrib-
2005); its potential importance as a reservoir of DNA from
utes to nutrient cycling as a source of readily available
transgenic plants (Lynch et al., 2004); and the apparent con-
nutrients (Greenfield, 1999; Perez-Moreno & Read, 2001). For
tradiction between, on the one hand, the rapid biological
example, Greenfield (1999) estimated that Pinus radiata plan-
exploitation of resources in pollen by micro-organisms and
tations may produce 1–3 t pollen ha
1. Pollen has an N con-
invertebrates (Goldstein, 1960; Scott & Stojanovich, 1963;
tent of 1–8% and is enriched in both N and P relative to other
Hutchison & Barron, 1997; Greenfield, 1999; Perez-Moreno &
plant materials (Greenfield, 1999; Perez-Moreno & Read,
Read, 2001) and, on the other hand, the relative recalcitrance of
2001). Thus the annual deposit of pollen into soils may repre-
components of pollen (Hedges & Oades, 1997). This apparent
sent recycling of approximately 20 kg N ha
1, much of which
contradiction may be explained in part at least by the emphasis
Correspondence: D. W. Hopkins. E-mail: [email protected] in palynological investigations on the integrity of pollen as seen
Received 15 February 2007; revised version accepted 2 January 2008 by microscopy (Sangster & Dale, 1961, 1964), compared with

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science 551
552 E. A. Webster et al.

release of nutrient resources from the entire pollen grain, Decomposition of pollen and pollen fractions in soil
which is more relevant to functional ecology. Although the
Aliquots (125 mg dry weight equivalent) of pollen and of the
sporopollenin-containing exine (outer layer) of pollen is usually
insoluble and the water-soluble fractions from 125 mg of pollen
sufficiently stable in anoxic soils and sediments to provide a
(dry weight) were added separately to triplicate portions of 25 g
record of previous plant communities (Hedges & Oades, 1997),
(dry weight equivalent) soil and mixed in with a spatula. There
the stability differs with environmental conditions and between
were also unamended (control) soil samples. Water was added to
species. The consequent problem of under-representation of
the soil samples to bring them to 50% water holding capacity.
some species of pollen in profiles is well recognized (Sangster &
CO2 released from both the soil and the soil-pollen mixtures
Dale, 1961, 1964; Havinga, 1971; Wilmshurst & McGlone,
during incubation at 16°C in the dark was measured using
2005), but the underlying biochemical reasons for the differ-
a Respicond IV respirometer (Nordgren Innovations, Umeå,
ences in persistence of pollen from some species are poorly
Sweden; Nordgren, 1988). This instrument measures CO2 by
understood.
trapping it in KOH and measuring the change in conductance
There have been few quantitative investigations of the decom-
of the KOH traps at regular intervals. Daily measurements of
position of pollen in soil. The objectives of this work were to
CO2 production were collected for 22 days, by which time the
explore the hypothesis that pollen inputs to soil provide a rapid
rates of CO2 production from the pollen-amended and the
‘fix’ of energetically rich compounds for soil micro-organisms,
unamended control soils were not significantly different from
which may in turn promote further nutrient recycling (Greenfield,
each other. In a parallel experiment, triplicate samples of pol-
1999). We have done this by determining the rate of mineraliza-
len and the insoluble fraction of pollen were enclosed in
tion of C from pollen, by determining the effects of supplementing
1  2 cm mesh bags with 25 mm apertures before incubation
the soil with different C and N sources and inoculating the soil
in soil (as above). At the end of the incubation period, the
with different decomposer fungi, and by chemically characteriz-
remaining pollen was recovered from the mesh bags, dried (as
ing the decomposing fractions of pollen.
above), sub-samples analysed for total C and N (as above)
and, along with fresh pollen, analysed by solid-state 13C
nuclear magnetic resonance (NMR) spectroscopy (see below).
Materials and methods
Soil and pollen samples
Effect of soil amendment and inoculation on pollen
Soil from the 0–10 cm depth beneath arable fields near Seaton decomposition
(2°32.2¢W, 56°34.0¢N) in the east of Scotland was used. The soil
is a free-draining iron podzol of the Vinny series (Laing, 1976; Before addition of either Z. mays or B. alba pollen, the soil was
USDA Haplorthod). The soil pH was 5.9, it contained 57 mg amended with either 2 mg NH4NO3 g
1 soil, or 2 mg glucose
total C g
1 (standard deviation ¼ 2.2), 3.1 mg total N g
1 g
1 soil or 2 mg glycine g
1 soil, or inoculated with a culture
(standard deviation ¼ 0.20) and had sandy-loam texture of the lignolytic fungus, Phanerochaete chrysosporium (IMI
(Hopkins et al., 2001). The soil contained 0.78 mg biomass 232175, ATCC 34540), or the cellulolytic fungus, Chaetomium
C g
1 soil and the basal respiration rate was 0.032 mmol CO2 globosum (IMI 185462). The fungi were obtained from CABI
g
1 soil hour
1 (Hopkins et al., 2001). The soil was sieved to Bioscience (Egham, UK) as freeze-dried spores. Inocula were
pass a 2-mm aperture sieve, and stones and large fragments of prepared by the addition of a few drops of sterile water and
plant material were removed by hand. The soil was kept in cultures grown for 14 days at 20°C on potato dextrose agar,
the field-moist state at 20°C for 20 days prior to the start of and then 1-cm diameter discs were cut from the leading edges
experimental work. of the actively growing mycelia. Four discs of each fungus
Pollen from Betula alba and Zea mays was obtained from were added to 100 ml conical flasks filled with sand-cornmeal
Sigma (Poole, UK). The pollen was fractionated into the medium (SCM; 50 g washed sand, 1 g ground maize and
water-soluble and the water-insoluble fractions by shaking 17 ml water; Butler, 1953), and incubated for 3 weeks at 20°C
pollen samples in deionized water (125 mg pollen to 5 ml with periodic shaking to distribute the mycelium within the
water) for 30 minutes at 20°C and then centrifuging for medium. Another set of soil samples were inoculated with
15 minutes at 1200 g and decanting the water-soluble fraction. either 2% (w/w) SCM or autoclaved damp sand as a control.
Samples of the whole pollen and the insoluble fraction were The soils were then incubated in the respirometer and CO2
dried (60°C for 20 hours) in an oven and re-weighed to deter- production determined as described above.
mine the water-soluble mass and then ground in a ball mill.
The C and N contents were determined using a CHN analyser
Nuclear magnetic resonance spectroscopy
(Carlo-Erba Instruments, Milan, Italy). The C and N contents
of the water-soluble fraction were calculated by difference. Solid-state cross-polarization (CP) magic angle spinning
The different pollen fractions were used for subsequent experi- (MAS) 13C NMR spectra for pollen samples were recorded
mental work without further treatment or storage. using a Varian Chemagnetics CMX LITE 300 MHz NMR

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558
 Decomposition of pollen in soil 553

spectrometer (Varian Ltd, Oxford, UK) (1H, 300.06 MHz; 13C Table 1 Properties of pollen. All values are the mean of three measure-
75.46 MHz) and 7.5 mm diameter Pencil# rotors made from ments and the standard deviations are shown in brackets. The amounts
zirconium with Kel-F caps. The contact time was 1 ms, the of C and N in the water-soluble fraction were determined by difference
relaxation delay was 2 s, the 13C pulse was 4 ms (90°), having between the whole pollen and the insoluble fraction
set the Hartmann-Hahn condition with adamantine, and the Water
samples were spun at 4 kHz MAS for 10 000 scans. Dipolar Total C Total N soluble
dephased (DDP) spectra were collected on the samples by /mg C g
1 /mg N g
1 fraction
using the same parameters as the conventional CP MAS spec- pollen pollen C-to-N /%
tra with a dephasing delay (T1) of 40–50 mS and acquisition
Zea mays
time of 4 ms. Liquid-state 13C CP NMR spectra of the water- Whole pollen 483 (36.10) 37.3 (2.80) 12.9 22.5 (4.50)
soluble pollen fractions were obtained on a Bruker DPX300 Insoluble fraction 374 (25.9) 29.7 (2.23) 12.6
300 MHz spectrometer (Bruker, Coventry, UK) (1H, 299.90 MHz; Soluble fraction 109 7.6 14.3
13
C 75.42 MHz) fitted with a 5 mm probe. D2O was added Betula alba
to the solutions as a field lock and the number of scans was Whole pollen 483 (6.95) 39.0 (0.19) 12.4 27.7 (0.19)
16 384. The 13C NMR spectra were referenced to an external Insoluble fraction 409 (2.10) 35.4 (0.20) 11.6
tetramethylsilane standard and interpreted by reference to Soluble fraction 74 3.6 20.6
the following chemical shift limits (Wilson, 1987): chemical
shifts between 0 and 45 p.p.m. were assigned to methyl and
alkyl-C indicative of aliphatic compounds including amino the soluble fractions were much greater (C-to-N ¼ 14.3 and
acids, lipids and waxes; chemical shifts between 45 and 20.6 for the soluble fraction from Z. mays and B. alba pollen,
60 p.p.m. were assigned to N-alkyl-C and methoxyl-C indic- respectively; Table 1).
ative of lignin substituents, amino acids and amino sugars; Significantly (P < 0.01) more CO2 was released from the
chemical shifts between 60 and 90 p.p.m. were assigned to pollen-amended soil compared with unamended soil and sig-
O-alkyl-C indicative of sugars; chemical shifts between 90 nificantly (P < 0.05) more CO2 was released from the soil
and 110 p.p.m. were assigned to acetal- and ketal-C indica- amended with Z. mays pollen than from the soil amended with
tive of sugars; chemical shifts between 110 and 160 p.p.m. B. alba pollen (Figure 1). Following pollen addition, no delay
were assigned to aromatic-C indicative of phenyl compounds in the onset of increased CO2 production was detected, indi-
including lignin and tannins; and chemical shifts between 160 cating that the soil micro-organisms responded to the pollen
and 200 p.p.m. were assigned to carbonyl-C indicative of addition within 1 day. Initially there was no significant differ-
organic acids and peptides. ence between the CO2 production from either of the pollen-
amended soils, but by the fourth day of incubation, the rate of
Statistical analyses CO2 release was significantly (P < 0.05) greater for the Z. mays
pollen than for the B. alba pollen (Figure 1). The resulting
Data from incubation experiments were analysed by analysis of difference in C mineralization persisted until the end of the
variance. The significant differences were assessed using Tukey’s experiment, when the rates of CO2 production from the two
honestly significant difference (HSD). pollen-amended treatments were not significantly different

Results
1.4
Pollen from both species contained 483 mg C g
1 and similar
1.2
amounts of N (37–39 mg N g
1), leading to C-to-N ratios of
CO2 propduction rate

12.9 and 12.4 for the Z. mays and the B. alba pollen, respec- 1.0
/ mg C g-1 soil

tively (Table 1). A significantly (P < 0.05) greater proportion 0.8

of the B .alba pollen was water-soluble compared with Z. mays
0.6
pollen (27.7% compared with 22.5%). Although the water- Soil
soluble mass and the water-soluble C contents of the Z. mays Zea mays
0.4
Betula alba
pollen were both 22.5%, only about 15% of the C in B. alba
0.2
pollen was water-soluble compared with 27.7% of the overall
mass (Table 1). By contrast, smaller fractions of the N were 0.0
0 5 10 15 20 25
water-soluble (20.4 and 9.2% for the Z. mays and the B. alba,
Time / days
respectively) compared with both the C and the total mass.
Because extraction with water preferentially removed C, the Figure 1 CO2 production from soil amended with intact pollen
C-to-N ratio of the insoluble fraction was slightly smaller than grains from Zea mays and Betula alba. Each point is the mean three
that of the whole pollen, and the calculated C-to-N ratios of replicates and the vertical bars are  standard errors.

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558
554 E. A. Webster et al.

from the control (Figure 1), but the net CO2-C released from 0.3
the pollen treatments (i.e. after subtraction of the control) was
Betula alba

Net CO2 propduction rate

equivalent to 26.3% of the C in the Z. mays pollen, but only
Insoluble Betula alba

/µg C g-1 soil day-1

22.1% of the C in the B. alba pollen (Table 2). Soluble Betula alba
0.2
The respiratory response of the soil micro-organisms was fast- Insoluble+soluble
est for the addition of whole pollen and the water-soluble frac-
tions, with the maximum rates being observed after only 1 day
for both pollens. Removal of the soluble fraction delayed the 0.1
maximum rate of CO2 production for both pollens, but only by
approximately 1 day (Figure 2). Thereafter, the rate of CO2
production for the whole pollen and both fractions declined to
the basal rate over the next 20 days (Figure 2). As proportions 0.0
of the C added, the net CO2-C released did not differ signifi-
0.4
cantly between the whole pollen and either of the two fractions
for the Z. mays pollen (Table 2). However, for the B. alba pol- Zea mays

Net CO2 propduction rate

len, a significantly (P < 0.05) smaller proportion of the C in 0.3 Insoluble Zea mays

/µg C g-1 soil day-1

the insoluble fraction and a significantly greater (P < 0.05) Soluble Zea mays
fraction of the C in the soluble fraction was apparently miner- Insoluble+soluble

alized (Table 2). Much of the difference in C mineralization 0.2

from the soluble fractions of the two pollens occurred during
the first 3 days, when the net rates of CO2 production were
significantly (P < 0.05) greater for the B. alba compared with 0.1
the Z. mays soluble pollen fractions (Figure 2). The sum of the
CO2 released from the two fractions exceeded the CO2 release
0.0
from the whole (non-fractionated) pollen from both species 0 5 10 15 20 25
(Figure 2). Time / days
Addition of both glucose and glycine increased the CO2 pro-
duction from pollen-amended soils, but not by significantly Figure 2 Rate of CO2 production from intact pollen grains, the
more than was produced by their addition to soil without pol- insoluble fraction from pollen grains and the water-soluble fraction
len (Figure 3). The addition of NH4NO3 had no significant from pollen from Zea mays and Betula alba added to soil. Each point
effect on CO2 production in the presence or absence of pollen. is the mean of three replicates.
Inoculation with C. globosum had no significant effect on the
mineralization of C in soil amended with either pollen, and regions indicative of carbohydrates, and the carbonyl-C region
inoculation with P. chrysosporium had no significant effect indicative of peptides and carboxylic acids. In addition, both
on C mineralization in soil amended with B. alba pollen pollens gave small but distinct resonance between 45 and
(Figure 3). However, in the soil amended with Z. mays pollen, 55 p.p.m., attributed to either methoxyl- or N-alkyl-C. The
inoculation with P. chrysosporium led to significantly (P < 0.05) presence of strong carbonyl-C resonances and the relatively
greater C mineralization (Figure 3). large N contents of the pollens suggested that the 55 p.p.m.
Pollen from both species had similar solid-state 13C NMR signal was due to N-alkyl-C rather than methoxyl-C. Pollen
spectra (Figure 4). The spectra for the whole pollen samples from both species had broad and very weak signals in the aro-
had strong resonances in the alkyl- and methyl-C region matic region (110–160 p.p.m.); however, spinning sidebands
(0–45 p.p.m.) indicative of aliphatic compounds, the O-alkyl-C from the carbonyl-C (marked with asterisks in Figures 4 and
(60–90 p.p.m.) and the acetal- and ketal-C (90–110 p.p.m.) 5) will have contributed substantially to these signals.
DDP MAS NMR revealed those functional groups with high
Table 2 Net cumulative CO2-C released after 22 days’ incubation in molecular mobility or those that were un-protonated (Figure 5).
soil expressed as a percentage of the pollen C added Pollen from both species had distinct, but small, resonances in
the alkyl- and methyl-C region and especially strong resonances
Pollen treatment Percentage of added C lost
in the carbonyl region, which contributed spinning sidebands in
Z. mays whole pollen 26.3 (2.96) the aromatic region, as mentioned above. The main difference
Z. mays insoluble fraction 26.6 (1.13) between the DDP MAS spectra for the two species was the
Z. mays soluble fraction 25.2 (1.54) strong resonance in the N-alkyl- and methoxyl-C region for
B. alba whole pollen 22.1 (0.32)
the Z. mays compared with the B. alba pollen. DDP MAS spec-
B. alba insoluble fraction 18.5 (0.94)
tra allow N-alkyl-C and methoxyl-C to be distinguished
B. alba soluble fraction 28.1 (2.45)
because the N-alkyl-C decays rapidly whilst leaving a sharp

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558
 Decomposition of pollen in soil 555

12

CO2 produced / mg C g-1 soil

10

m
m

m
m

n
il

um
So

lle

lle
iu

iu
su

iu
su

os
or

po

or

po

or
bo

bo
sp

sp

sp

ob
s
lo

ba
lo
so

ay

so

so

l
.g

.g

.g
al
ry

ry

ry

C
C

B.
C
ch

ch

ch
Z.

+
+

P.

n
P.

il

P.

n
+

il

lle
So

lle

So
il

+
+

+
So

po
po

n
il

lle
So

lle

ba
s

po
ay
po

al
m

B.
ba
s
ay

Z.

+
al
m

B.

il
+

So
il
Z.

+
So

il
+

So
il
So

12
CO2 produced / mg C g-1 soil

10

Figure 3 Effect of addition of glucose, gly- 0

e

n
e

n
n
n
3
il

in

lle

lle

cine or NH4NO3 on CO2 production, and

lle

lle
os

lle

lle
lle
lle
O
So

yc
4N

po

po

po

po

po

po
po
uc

po
gl
gl

the effect of inoculation with either Phanaer-

H

ba

ba

ba
s
s
+

ay

ay

ay
N

b
ay
+

al

al

al

al
il

m
+

m
m
il

So

ochaete chrysosporium or Chaetomium globo-

So

B.

B.

B.

B.
il

Z.

Z.

Z.
Z.
So

l+

+
+

+
+

sum on C mineralization in soil amended

3

se

e
i
il

So
e
se

in
So

lin

o
4N

yc
co

uc
4N

yc

with either Zea mays and Betula alba pollen.

gl
H
u

gl
gl
H

gl

+
+
N

Each bar is the mean three replicates and the

l+

il
+

il

So
+

il

So
il
i

So
So
il

So
So

vertical bars are  standard errors.

signal from methoxyl-C. The loss of most of the signal inten- (Figure 4). Both the liquid-state spectra also showed reso-
sity around 50 p.p.m. is attributed to the loss of the N-alkyl-C nances in the methyl- and alkyl-C region at 24 and 29 p.p.m.,
signal, which is consistent with the large N contents of the pol- and a single distinct resonance at 47 p.p.m. in the methoxyl-
len and indicative of amino acids. The sharp residual signal in and N-alkyl-C region, which cannot be attributable to amino
this region was likely to be due to methoxyl-C. acids because of the absence of a corresponding carbonyl sig-
Extraction of the water-soluble fraction led to large reduc- nal in the pollen spectra (Figure 6). The insoluble fractions
tions in the intensities of the O-alkyl-C solid-state resonances had increased alkyl-C and methyl-C and carbonyl-C contents
for both pollens, indicating selective removal of polysaccharide- by comparison with the whole pollen (Figure 6). The N-alkyl-C
rich material. This is consistent with the liquid-state 13C NMR signal was also more intense for the insoluble fractions
spectra, which were very similar for the two species and compared with the whole pollen (Figure 4). Following decom-
showed a series of strong resonance in the O-alkyl-C and the position, the intensities of all the resonances were substantially
acetal- and ketal-C regions characteristic of carbohydrates less except for the alkyl-C and methyl-C signal. In the case of

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558
556 E. A. Webster et al.

Figure 4 CP MAS 13C NMR spectra of

whole and insoluble fractions of Zea mays
and Betula alba pollen before and after incu-
bation in soil. Spinning sidebands are indi-
cated by asterisks (*).

the whole pollen, weak O-alkyl-C and carbonyl-C resonances tion, as reflected by the different decomposition rates and the
were detectable after decomposition, but for the insoluble frac- differential effects of the ligninolytic fungus, P. chrysosporium,
tion only the alkyl-C and methyl-C resonance remained. on pollen breakdown.
The small C-to-N ratio and the relative large N contents of the
pollen are consistent with the results of Greenfield (1999), as are
Discussion the relatively rapid decomposition and the lack of an effect of N
addition on decomposition of pollen that we saw. Greenfield
The main findings of this work are that pollen from Z. mays
(1999) assessed decomposition from mass loss (and net N immo-
and B. alba represents a high-quality resource for soil micro-
bilization) during incubation in soil for 30 days at between 8 and
organisms by supplying sugars and nitrogenous compounds
20°C to be in the range 21–72% and we report C mineralization
with a small overall C-to-N ratio in water-soluble forms.
equivalent to 22–27% of the pollen C over 22 days at 16°C. The
Despite similarities between the different species, as seen by
water-soluble fraction of the pollen (22–28%) is in the same
NMR, there are functional differences in pollen decomposi-
range (24–60%) as reported by Greenfield (1999) and Nielson
et al. (1955) for pollen from a larger number of plant species.
Furthermore, the presence of N-alkyl-C and carbonyl-C reso-
nances indicative of amino acids and polypeptides supports
Greenfield’s (1999) analysis in which he showed that a-amino-
N was the largest fraction of organic N in pollen identifiable
by acid hydrolysis and steam distillation.
Our estimates of the amount of pollen N that was water-sol-
uble (9–20%) are, however, substantially less than those of
Greenfield (1999), who reported data in the range 40–80%.
However, Greenfield’s (1999) method involved extraction over
a much longer period with a possibly more stringent extractant
(water containing a trace of CHCl3 to arrest microbial growth)
and was likely, therefore, to have been a much more thorough
extraction. The fact that we observed pronounced N-alkyl-C
and carbonyl-C resonances in the insoluble pollen fraction
indicates the presence of a substantial amount of N that was
not extracted by our more rapid procedure. The possibility
Figure 5 DDP MAS 13C NMR spectra of whole Zea mays and Betula that a significant fraction of the potentially soluble material
alba pollen. Spinning sidebands are indicated by asterisks (*). was not extracted in our procedure probably explains the fact

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558
 Decomposition of pollen in soil 557

alkyl-, methyl- and aromatic-C resonances, indicating that

these components had been selectively preserved or synthe-
sized de novo during decomposition. Our spectra for degraded
pollen show a similar shift in signal from O-alkyl-C to alkyl-,
methyl- and aromatic-C. The rapid loss of polysaccharide
material during the decomposition of plant material is com-
mon (e.g. Hopkins et al., 1993; Webster et al., 2001). This indi-
cates that a fraction of the plant polysaccharide is protected
from mineralization and that some of the plant alkyl-C is pre-
served during decomposition or that microbial synthesis dur-
ing colonization contributes to aliphatic-C in the alkyl- and
methyl-C resonance detected by NMR (Hopkins et al., 1997).
In the case of pollen, most of the polysaccharide was lost rap-
idly, indicating that it was not protected from microbial
attack, as occurs for the cellulose in the component of ligno-
cellulose (e.g. Webster et al., 2005).
Whereas Greenfield (1999) focused on the contribution to
rapid N turnover in soil that pollen may make, we offer direct
evidence for a complimentary contribution to microbial respi-
ration that pollen can make. This observation extends observa-
tions of the subtle role that pollen plays in linking above- and
below-ground processes, potentially by driving the energetic
metabolism of the soil micro-organisms and by contributing to
both C and N for biosynthesis.

Acknowledgements
We are grateful to L. C. English and R. L. MacKay for technical
assistance, to both L. G. Greenfield (University of Canterbury,
Christchurch, New Zealand) and P. E. J. Wiltshire for useful
discussion on palynology, and to the UK Natural Environment
Research Council for financial support.

References
Butler, F.C. 1953. Saprophytic behaviour of some cereal root-rot
13 fungi. I. Saprophytic colonization of wheat straw. Annals of Applied
Figure 6 Liquid-state C NMR for glycine, alanine and glucose, and
Biology, 40, 284–297.
the water-soluble fractions from Zea mays and Betula alba pollen.
Davidson, D.A., Carter, S., Boag, B., Long, D., Tipping, R. & Tyler,
A. 1999. Analysis of pollen in soils: processes of incorporation and
the insoluble fraction decomposed relatively rapidly. Never- redistribution of pollen in five soil profile types. Soil Biology and
theless, the fact that we were able to extract up to 20% of the Biochemistry, 31, 643–653.
pollen N in less than 1 hour indicates the great solubility of Goldstein, S. 1960. Degradation of pollen by phycomycetes. Ecology,
pollen N, consistent with Greenfield’s (1999) general conclu- 41, 543–545.
Greenfield, L.G. 1999. Weight loss and release of mineral nitrogen
sion that leachates from pollen can provide a readily available
from decomposing pollen. Soil Biology and Biochemistry, 31, 353–
substrate for soil micro-organisms.
361.
The differences in NMR spectra between species were small, Guilford, W.J., Schneider, D.M., Labovitz, J. & Opella, S.J. 1988.
which is consistent with other observations (Guilford et al., High resolution solid state 13C NMR of sporopollenins from dif-
1988; Hemsley et al., 1992, 1993, 1996). However, unlike ferent plant taxa. Plant Physiology, 86, 134–136.
Hemsley et al. (1993), the spectra we recorded did not have Havinga, A.J. 1971. An experimental investigation into the decay of
especially large signals in the alkyl- and methyl-C region. The pollen and spores in various soil types. In: Sporopollenins (ed.
NMR spectra for degraded or fossilized pollen reported by J. Brooks), pp. 446–473. Academic Press, New York.
Hemsley et al. (1993) showed no O-alkyl-C resonances (indica- Hedges, J.I. & Oades, J.M. 1997. Comparative organic geochemistries
tive of polysaccharides and sugars), but relatively enriched of soils and marine sediments. Organic Geochemistry, 27, 319–361.

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558
558 E. A. Webster et al.

Hemsley, A.R., Chaloner, W.G., Scott, A.C. & Groombridge, C.J. 1992. Nielson, N., Grommer, J. & Lunden, R. 1955. Investigations on the
Carbon-13 solid-state nuclear magnetic resonance of sporopollenins chemical composition of pollen from some plants. Acta Chimica
from modern and fossil plants. Annals of Botany, 69, 545–549. Scandinavica, 9, 1100–1106.
Hemsley, A.R., Barrie, P.J., Chaloner, W.G. & Scott, A.C. 1993. The Nordgren, A. 1988. Apparatus for the continuous, long-term moni-
composition of sporopollenin and its use in living and fossil plant toring of soil respiration rate in large numbers of samples. Soil
systematics. Grana Supplement, 1, 2–11. Biology & Biochemistry, 20, 955–957.
Hemsley, A.R., Scott, A.C., Barrie, P.J. & Chaloner, W.G. 1996. Perez-Moreno, J. & Read, D.J. 2001. Exploitation of pollen by
Studies of fossil and modern spore wall biomacromolecules using mycorrhizal mycelial systems with special reference to nutrient
C-13 solid-state NMR. Annals of Botany, 78, 83–94. recycling in boreal forests. Proceedings of the Royal Society of Lon-
Hopkins, D.W. 2000. Interfaces of soil biology with archaeological don Series, B-Biological Sciences, 268, 1329–1335.
investigations. In: Soil Biochemistry, Vol 10 (eds J.-M. Bollag & Scott, H.G. & Stojanovich, C.J. 1963. Digestion of juniper pollen by
G. Stotzky), pp. 483–512. Marcel Dekker, New York. collembola. Florida Entomologist, 46, 189–191.
Hopkins, D.W., Chudek, J.A. & Shiel, R.S. 1993. Chemical charac- Sangster, A.G. & Dale, H.M. 1961. A preliminary study of differen-
terisation and decomposition of organic matter from two contrast- tial pollen grain preservation. Canadian Journal of Botany, 39, 35–
ing grassland soil profiles. Journal of Soil Science, 44, 147–157. 43.
Hopkins, D.W., Chudek, J.A., Webster, E.A. & Barraclough, D. Sangster, A.G. & Dale, H.M. 1964. Pollen grain preservation of
1997. Following the decomposition of ryegrass labelled with 13C underrepresented species in fossil spectra. Canadian Journal of
and 15N in soil by solid-state nuclear magnetic resonance spectros- Botany, 42, 437–449.
copy. European Journal of Soil Science, 48, 623–631. Van Mourik, J.M. 2003. Life cycle of pollen grains in mormoder
Hopkins, D.W., Webster, E.A., Chudek, J.A. & Halpin, C. 2001. humus forms of young acid forest soils: a micromorphological
Decomposition in soil of tobacco plants with genetic modifications approach. Catena, 54, 651–663.
to lignin biosynthesis. Soil Biology and Biochemistry, 33, 1455–1462. Webster, E.A., Hopkins, D.W., Chudek, J.A., Haslam, S.F.I., Šimek,
Hutchison, L.J. & Barron, G.L. 1997. Parasitism of pollen as a nutri- M. & Picek, T. 2001. The relationship between the size of the soil
tional source for lignicolous Basidiomycota and other fungi. Myco- microbial community and the resource quality of soil carbon. Jour-
logical Research, 101, 191–194. nal of Environmental Quality, 30, 147–150.
Laing, D. 1976. Soils from the Country Round Perth, Arbroath and Dun- Webster, E.A., Halpin, C., Tilston, E.L. & Hopkins, D.W. 2005.
dee. Soil Survey of Scotland, Her Majesty’s Stationery Office, Decomposition in soil of soluble, insoluble and lignin-rich frac-
Edinburgh, UK. tions of plant material from tobacco with genetic modifications
Lynch, J.M., Benedetti, A., Insam, H., Nuti, M.P., Smalla, K., to lignin biosynthesis. Soil Biology and Biochemistry, 37, 751–
Torsvik, V. et al. 2004. Microbial diversity in soil: ecological 760.
theories, the contribution of molecular techniques and the impact Wilmshurst, J.M. & McGlone, M.S. 2005. Corroded pollen and
of transgenic plants and transgenic microorganisms. Biology & spores as indicators of changing lake sediment sources and catch-
Fertility of Soils, 40, 363–385. ment disturbance. Journal of Paleolimnology, 34, 503–517.
Mallik, M.A.B. & Williams, R.D. 2005. Allelopathic growth stimula- Wilson, M.A. 1987. NMR Techniques and Applications in Geochemistry
tion of plants and microorganisms. Allelopathy Journal, 16, 175–198. and Soil Chemistry. Pergamon Press, Oxford.

# 2008 The Authors

Journal compilation # 2008 British Society of Soil Science, European Journal of Soil Science, 59, 551–558