Directory UMM :Data Elmu:jurnal:T:Tree Physiology:Vol15.1995:
Tree Physiology 15, 519--526
© 1995 Heron Publishing----Victoria, Canada
Effects of supplemental mass pollination (SMP) in a young and a
mature seed orchard of Pinus sylvestris
U. ERIKSSON,1 G. JANSSON,1 R. YAZDANI2 and L. WILHELMSSON1
1
Forestry Research Institute, Glunten, S-751 83 Uppsala, Sweden
2
Department of Forest Genetics, The Swedish University of Agricultural Sciences, P.O. Box 7027, S-750 07 Uppsala, Sweden
Received May 31, 1994
Summary The effects of supplemental mass pollination
(SMP) were studied in two Pinus sylvestris seed orchards
differing in pollen production. Pollen was dusted over the
whole tree during the period of peak female receptivity using
a pressurized backpack sprayer. The success of SMP was
assessed by means of allozyme markers. In the orchard with
high pollen production, detectable differences in SMP success
rate were found between clones, but the success rate was not
influenced by the number of pollinations per day. The average
estimated success rate of SMP was 19%. In the orchard with
low pollen production, no significant differences in SMP success rate were found between years (22 versus 34%) or between
clones. The SMP success rate in the low pollen production
orchard varied between 14 and 69%.
Keywords: isozymes, pollen production, Scots pine, seed orchard management.
Introduction
Based on height growth, the average genetic superiority of
first-generation plus-tree seed orchards of Pinus sylvestris L.
in Sweden is estimated to be between 6 and 8% compared with
unselected stand material (Danell 1991, 1993). However, the
genetic potential of seed orchards has not been entirely exploited mainly because it has not yet been possible to eliminate
pollen contamination from unselected sources outside the orchards (e.g., Savolainen 1991, Di-Giovanni and Kevan 1991).
Estimates of pollen contamination in P. sylvestris seed orchards range between 17 and 74% (El-Kassaby et al. 1989,
Harju and Muona 1989, Yazdani and Lindgren 1991, Wang et
al. 1991, Paule 1991). Pollen contamination affects the growth,
quality and hardiness of the seedling output. Serious pollen
contamination in southern orchards producing seed intended
for northern regions with a harsh climate makes the seed crops
unsuitable for the intended forest regeneration area (Lestander
and Lindgren 1985, Andersson and Westin 1990, Savolainen
1991). Beside pollen contamination, differences in random
mating and gamete production capacity within and among
clones may affect the genetic quality of the output from seed
orchards (e.g., Eriksson et al. 1973, Jonsson et al. 1976, Bhumibhamon 1978, Chung 1981). Thus, there is scope to develop
methods that facilitate a higher recovery of the potential genetic gain of seed orchards.
In seed orchards of Pinus taeda L. and Pseudotsuga menziesii (Mirb.) Franco, having limited pollen supply, supplementary mass pollination (SMP) (Bridgwater and Trew 1981),
defined as the broadcast application of pollen to unisolated
female strobili, has been used to increase the yield of sound
seeds (see review by Bridgwater et al. 1993). The success rate
of SMP has been estimated by using pollen with unique biochemical marker alleles (Wheeler and Jech 1985, Yazdani et al.
1986). Based on this technique, high success rates of SMP
have been reported for seed orchards of P. menziesii (Wheeler
and Jech 1985), P. taeda (Blush 1987) and P. sylvestris
(Eriksson et al. 1994). The latter investigation showed average
success rates between 66 and 84% when individual strobili
were pollinated; however, in an operational study, in which
whole grafts were pollinated, the success rate declined to
7--26%.
Besides competition by pollen from other sources, the success rate of SMP may be influenced by other factors including
the prevailing environmental conditions at flowering and the
number of occasions a graft must be pollinated in order to
achieve a satisfactorily high success rate. Eriksson et al. (1994)
concluded that, for a single strobilus, timing rather than the
number of pollinations is critical but, because of the gradual
maturation of female strobili, whole grafts need to be pollinated more than once a day during the flowering period.
Bridgwater et al. (1993) concluded that a single pollination at
peak female receptivity is enough to achieve fairly high success rates in P. taeda. However, differences in success rates
among clones have been reported for P. taeda (Bridgwater and
Williams 1983, Blush 1987) and P. menziesii (Wheeler and
Jech 1985) that can be explained mainly by differences in
flower phenology in relation to main pollen shedding. Bridgwater et al. (1993) concluded that pollination of clones that
flower before maximum pollen flight provides the best opportunity for successful SMP, but did not exclude the possibility
of successful use of SMP for clones that flower during maxi-
520
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
mum pollen flight. Differences in SMP success rate between
clones could also be expected for P. sylvestris because there are
large variations in flower phenology among and within clones
of this species. Furthermore, because the course of flowering
varies between years as a result of differing weather conditions
during the flowering period (Sarvas 1962, Jonsson et al. 1976),
annual variations in flowering phenology could also influence
SMP success rate. Therefore, we have tested the hypothesis
that the success rate of SMP in P. sylvestris seed orchards is
influenced by clone, weather and frequency of application.
to June 12, 1988). Pollen from one clone, not represented in
the orchard, was used in both years. The pollen clone (F) is
heterozygous at the GOT-B locus with the genotype B1/B2,
where B1 is the rare allele. Nine randomly selected clones
(G--O), not represented in Experiment 1, were used as mother
clones. Two individual grafts were sampled per clone, one for
each treatment. The same grafts and clones were pollinated
both years. The two pollination treatments in Experiment 2
were: (i) wind-pollinated control (Control), and (ii) SMP of
whole trees twice a day during the period of peak receptivity.
Materials and methods
Application of pollen
Experiment 1
Experiment 1 was carried out in the spring of 1987 in a mature
P. sylvestris seed orchard with high pollen production (Table 1). Trees were pollinated for 8 days between May 28 and
June 4, 1987. A mix with equal portions of pollen from two
clones not cultivated in the orchard, designated A and B, was
used. The clones were chosen because they had rare isozyme
markers. Clone A is heterozygous at the LAP-B locus with the
genotype B1/B2, and clone B is heterozygous for the GOT-B
locus with the genotype B1/B3. B1 is the rare allele at both
loci.
Four grafts from each of three mother clones (C--E) were
each subjected to four pollination treatments as follows: (i)
wind-pollinated control (Control), (ii) SMP of whole trees
once per day during the period of peak receptivity, (iii) SMP
of whole trees up to three times per day during the period of
peak receptivity, and (iv) SMP of whole trees up to six times
per day during the period of peak receptivity.
Weather and amount of pollen
Experiment 2
Experiment 2 was carried out during both 1987 and 1988 in a
P. sylvestris seed orchard with low pollen production (Table 1).
Supplemental mass pollination was carried out during 11 days
in both 1987 and 1988 (June 15 to June 26, 1987, and June 1
Table 1. Description of the seed orchards.
Characteristic
Experiment 1
Experiment 2
Name
Latitude
Longitude
Altitude
Area
Year of establishment
Spacing
Mean height (approx.)
No. of clones
Soil texture
Seed yield, 1979--19881
Pollen production, 19872
Pollen production, 19882
493 Askerud
59°53′ N
13°10′ E
80 m
14 ha
1966--1969
5×5m
5m
43
Clay
52 kg ha − 1
40 kg ha − 1
39 kg ha − 1
123 Klocke
62°54′ N
18°16′ E
75 m
16 ha
1968--1972
5.6 × 5.6 m
3.5 m
60
Silt
3.6 kg ha − 1
-0.5 kg ha − 1
1
2
In both experiments, the pollen mix was dusted over clusters
of receptive, unisolated female strobili throughout the whole
graft in an attempt to pollinate as many receptive strobili as
possible. Pollen was delivered to the strobili from a pollination
wand activated by compressed air from a tube mounted on a
modified backpack sprayer (Eriksson et al. 1994). About 3 ml
of pollen was dusted over each graft in each pollination. The
SMP was started when approximately 20% of the female
strobili on a graft were judged to be receptive, according to the
classification system presented by Jonsson et al. (1976). No
pollinations were done during precipitation. The intention was
to pollinate each graft for 3--5 days during the period of peak
female strobilus receptivity. As a consequence of frequent
precipitation, which slowed maturation of female strobili, the
pollination period was extended to 4 to 8 days per graft in
Experiment 1. For the same reason, the pollination period in
Experiment 2 was extended to about 11 days in 1987 and to
between 7 and 11 days in 1988.
Eriksson and Palmér (1991).
Eriksson and Wilhelmsson (1991).
For both experiments, temperatures were estimated as the daily
mean temperatures recorded by the Swedish Meteorological
and Hydrological Institute. For Experiment 1, we used the data
collected at Station 9240 Arvika, situated 40 km southwest of
the orchard. The station closest to the orchard in Experiment 2
was Station 6311 Skagsudde, situated 40 km north of the
orchard. The occurrence of precipitation was observed daily
during the experimental periods in both experiments. The
amount of pollen in the air during the experimental periods was
assessed with a pollen-catching device (Sarvas 1962).
Pollen handling, cone collection and seed extraction
In 1986, pollen was extracted under controlled temperature
and humidity conditions as described by Eriksson (1993). The
pollen lots were stored in sealed glass jars at −20 °C. All cones
on the treated grafts were collected in the autumn of 1988
(Experiments 1 and 2) and 1989 (Experiment 2). Seeds were
extracted and stored at −4 °C until isozyme analyses were
done. Numbers of empty seed and filled seeds per cone were
determined for all treatments. Two clones in Experiment 2
produced no seed after the pollinations in 1987, and one clone
produced no seed after the pollinations in 1988. These clones
were excluded from subsequent analyses.
SUPPLEMENTAL MASS POLLINATION IN PINE SEED ORCHARDS
Marker detection
The analyses of SMP success rate were carried out by isozyme
separation and starch gel electrophoresis. In Experiment 1, two
enzymes, leucine aminopeptidase (LAP) (Rudin 1977) and
glutamate oxalate transaminase (GOT) (Rudin 1975), were
analyzed in the diploid embryo and in the haploid megagametophyte. One enzyme (GOT) was analyzed in Experiment 2.
With one exception, 100 seeds were analyzed per clone and
treatment in Experiment 1 (only 40 seeds were available for
clone D in treatment 3P). In Experiment 2, with a few exceptions, about 100 seeds per clone and treatment were used in the
subsequent studies.
The genotypes of the mother clones in the orchard with high
pollen production were unknown at the time of the experiment.
In a multilocus analysis of all clones in the orchard, mother
clone E appeared to have the same rare allele (B1) in the GOT
system as father clone B.
Statistical analysis
Statistical analyses of frequencies were performed on the logit
transformations of y (Ashton 1972), as described by Eriksson
et al. (1994). Briefly, the observations y were the observed
proportions of SMP success rates and empty seed frequencies
on each graft. In both experiments, the empirical logits lij of the
responses in the treatment × clone subclasses were obtained as:
yij
lij = ln
.
1 − yij
(1)
aij = (oij − cj)/(1 − cj),
In Experiment 2, the total contribution rates from SMP were
computed as:
B1
yij = aGOT
,
ij
lij = µ + ti + sj + eij ,
(5)
where lij is the logit value of success rate in subclass ij, µ is the
overall mean, ti is the fixed effect of the ith treatment where i
= 1, 2 or 3, sj is the fixed effect of the jth clone where j = 1, 2
or 3, and eij is the random residual effect of ijth observed logit
value assumed to be individually and independently distributed (IID) (0, π2/3). The weight used for the ijth observation
was:
nij × yij (1 − yij ),
where nij is the number of observations in the ijth class. The
following model was used for Experiment 2:
(6)
where lij is the logit values of success rate in subclass ij, µ is
the overall mean, si is the fixed effect of the ith clone where i
= 1, ..., 8, uj is the fixed effect of the jth year where j = 1 or 2,
and eij is the random residual effect of the ijth observed logit
value assumed IID (0, π2/3). The weight, used for each observation in the least squares equation, was the same as in Equation 5.
The estimated responses in both experiments were backtransformed from the underlying scale to the visible frequency
scale by means of Equation 7:
(2)
p^ =
where aij is the adjusted observed frequency of LAP B1 or
GOT B1 in the seed from treatment i and clone j, oij is the
corresponding observed LAP B1 or GOT B1 frequency, and cj
is the observed allele frequency for the pooled control data.
Values of a less than 0 are unrealistic and have therefore been
given the value 0.
In Experiment 1, the total contribution rates from SMP were
computed as:
B1
B1
yij = (aLAP
+ aGOT
),
ij
ij
(4)
where yij is the success rate of SMP pollen fertilization for a
B1
is the observed
particular clone × year subclass, and aGOT
ij
rate adjusted according to Equation 2.
The statistical analyses of the logit values were performed
with the weighted least squares procedure available in the
GLM procedure of the SAS program (SAS Institute Inc., Cary,
NC). The following model was used for Experiment 1:
lij = µ + si + uj + eij ,
Because the logit transformations are undefined when the
observed and adjusted yij values are 0 or 1, values of yij equal
to 0 or 1 were excluded from the subsequent analysis (Harville
and Mee 1984).
The observed values of LAP B1 and GOT B1 frequencies
were adjusted for the contribution from the background pollen
to the marker allele frequency by means of Equation 2. This
equation assumes that the background contributions of LAP
B1 and GOT B1 are proportional to the part not fertilized by
the SMP pollen:
521
(3)
where yij is the success rate of SMP pollen fertilization for a
B1
B1
particular treatment × clone subclass, and aLAP
and aGOT
ij
ij
are the observed rates adjusted according to Equation 2.
1
_^ ,
1
+
e r)
(
(7)
where p^ is the estimated frequency corresponding to a leastsquares mean r^ measured in the logit scale.
Because all father clones were heterozygous for the locus
with the rare marker allele, the estimated frequencies p^ were
multiplied by 2. In both experiments, the open-pollinated controls were included as one of the treatments in the analyses of
empty seed frequency.
The ordinary least squares procedure of the SAS software
package was used to analyze numbers of filled seeds per cone.
The following model was used for Experiment 1:
z ij = µ + ti + sj + eij,
(8)
522
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
where zij is the number of filled seeds per cone in subclasses ij,
µ is the overall mean, ti is the fixed effect of the ith treatment
where i = 1, ..., 4, sj is the fixed effect of the jth clone where j
= 1, 2 or 3, and eij is the random residual effect of ijth observed
value assumed IID (0, σ2e ). The following model was used for
Experiment 2:
zijk = µ + ti + sj + uk + eijk ,
(9)
where zijk is the number of filled seeds per cone in the subclasses ijk, µ is the overall mean, ti is the fixed effect of the ith
treatment where i = 1 or 2, sj is the fixed effect of the jth clone
where j = 1, ..., 8, uk is the fixed effect of the kth year where k
= 1 or 2, and eijk is the random residual effect of ijkth observed
value assumed IID (0, σ2e ). The wind-pollinated controls were
included as a treatment in the analyses of the number of filled
seeds per cone.
Results
Experiment 1
From May 29 to May 31, the weather was dry and sunny, but
turned rainy from June 1 to the end of the experimental period
(Figure 1). The average daily mean temperature was about
10 °C, and only small amounts of pollen were trapped during
the experimental period. The main pollen shedding in the
orchard started on June 4 (Figure 1).
The results from the ANOVAs of SMP success rate, empty
seed frequency and number of filled seeds per cone are presented in Table 2. The number of pollinations per day (pollination treatments) did not significantly affect the success rate.
However, significant differences in success rate were found
between clones. The pollination treatments had a significant
effect on empty seed frequency, but there was no significant
difference in empty seed frequency among clones. There were
no significant effects of either pollen treatment or clone on
number of filled seeds per cone. Estimated least squares means
of SMP success rate, empty seed frequency and number of
filled seeds per cone for treatments and clones are presented in
Table 3.
Experiment 2
During the 1987 experimental period, the weather was mostly
rainy and cold with an average daily mean temperature of
about 10 °C (Figure 2). Two small peaks of pollen in the air
were detected (Figure 2). Unfortunately, the device for pollen
catching was defective from June 24, 1987, to the end of the
experimental period, but another device for pollen catching in
another part of the orchard indicated that there were only small
amounts of pollen in the air during this time.
Figure 1. Daily mean temperatures
(dashed line) recorded by the Swedish
Meterological and Hydrological Institute
at their station 9240 Arvika, situated
40 km southwest of the orchard, and the
amounts of pollen in the air (solid line), assessed with a pollen catching device, in
Experiment 1, 1987. The bar shows the period of female strobili receptivity on the
treated grafts, i.e., the experimental period.
Table 2. Analyses of variance for success rate, empty seed frequency and number of filled seeds per cone in Experiments 1 and 2.
Source of variation
Success rate
Empty seed frequency
Number of filled seeds per cone
df
MS
F-value
P
df
MS
F-value
P
df
MS
Experiment 1
Treatment
Clone
Error
2
2
3
1.48
3.15
0.16
9.32
19.81
0.052
0.019
3
2
6
229.1
9.5
8.2
27.93
1.16
< 0.001
0.37
3
2
6
11.2
71.4
16.7
Experiment 2
Year
Treatment
Clone
Error
1
-7
4
8.10
-3.15
5.05
1.60
-1.64
0.27
-0.33
1
1
2
6
37.3
21.0
253.2
45.0
0.83
0.47
5.62
0.37
0.50
0.001
1
1
2
6
204.1
5.5
140.1
12.4
F-value
0.67
4.28
16.4
0.4
11.3
P
0.6
0.07
< 0.001
0.52
< 0.001
SUPPLEMENTAL MASS POLLINATION IN PINE SEED ORCHARDS
523
Table 3. Success rate of SMP, empty seed frequency and number of filled seeds per cone for treatments in Experiment 1. The frequency and standard
error range for success rate and empty seed frequency are expressed as back-transformed (probability scale) values and estimated least squares
means. The number of filled seeds per cone are estimated least squares means.
Success rate
Empty seed frequency
Number of filled seeds per cone
Frequency
Error range
Frequency
Error range
Treatment-wise
Max 1 SMP per day
Max 3 SMP per day
Max 6 SMP per day
Control
Average
0.16
0.24
0.18
-0.19
0.15--0.17
0.22--0.26
0.17--0.19
--
0.22
0.36
0.21
0.13
0.22
0.20--0.24
0.33--0.40
0.19--0.24
0.11--0.15
16.74
12.88
13.81
16.50
14.98
Clone-wise
Clone C
Clone D
Clone E
Average
0.26
0.11
0.23
0.19
0.24--0.27
0.10--0.13
0.22--0.25
0.27
0.18
0.22
0.22
0.23--0.31
0.14--0.22
0.21--0.23
11.16
14.26
19.52
15.34
Figure 2. Daily mean temperatures
(dashed line) recorded by the Swedish
Meterological and Hydrological Institute
at their station 9311 Skagsudde, situated
40 km north of the orchard, and the
amounts of pollen in the air (solid line), assessed with a pollen catching device, in
Experiment 2, 1987. The bar shows the period of female strobili receptivity on the
treated grafts, i.e., the experimental period.
Figure 3. Daily mean temperatures
(dashed line) recorded by the Swedish
Meterological and Hydrological Institute
at their station 9311 Skagsudde, situated
40 km north of the orchard, and the
amounts of pollen in the air (solid line), assessed with a pollen catching device, in
Experiment 2, 1988. The bar shows the period of female strobili receptivity on the
treated grafts, i.e., the experimental period.
During the first half of the 1988 experimental period, the
weather was rainy and cold with an average daily mean temperature of about 8 °C (Figure 3). The weather during the latter
part of the period was mostly sunny with only a few showers
and an average daily mean temperature of about 12 °C. Small
amounts of pollen in the air were found before June 8 when the
main pollen shedding started (Figure 3).
The ANOVAs for SMP success rate, empty seed frequency
and number of filled seeds per cone are shown in Table 2. No
significant effects, for either years or clones, were found in the
analysis of success rate. Significant differences in empty seed
frequency were found between clones, but not between years
524
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
or pollination treatments. The difference in number of filled
seeds per cone was significant between years but not between
pollination treatments. Furthermore, clones had significantly
different numbers of filled seeds per cone. Estimated least
squares means of SMP success rate, empty seed frequency and
number of filled seeds per cone for years, treatments and
clones are presented in Table 4.
Discussion
Success rate and number of pollinations per day
Eriksson et al. (1994) concluded that a single SMP of a
P. sylvestris strobilus is enough to achieve a high success rate,
provided that the strobilus is at peak receptivity when pollinated, whereas whole trees should be pollinated more than
once a day, because not all female strobili on a tree reach peak
receptivity at the same time (Sarvas 1962). However, we found
that increasing the number of pollinations per day caused only
a small, nonsignificant increase in SMP success rate in Experiment 1. The absence of pollination treatment effects in our
study may be associated with the slow maturation of female
strobili as a result of the cold and rainy weather. Thus, only a
limited number of new unpollinated strobili may have become
receptive between pollinations. If so, this suggests that SMP of
a graft should be carried out relatively more often during a hot
day than during a cold day.
Success rate and weather conditions
The average SMP success rate in Experiment 1 was higher than
that reported for a comparable study in 1986 (Eriksson et al.
1994). One explanation could be that the weather conditions
differed between 1986 and 1987. In contrast to 1987, the
weather in 1986 was mostly sunny with little or no precipitation. The finding that there was no significant difference in
SMP success rate between years in Experiment 2 does not
exclude the possibility that there could be differences between
years with more variable weather conditions. Even though the
weather in 1988 improved during the second part of the experimental period, both flowering periods studied in Experiment 2
were rainy and cold. Frequent precipitation and high humidity
could have contributed to SMP success rate in two ways. First,
precipitation reduces the amount of competing pollen in the air
by washing it to the ground. Sarvas (1962) showed that abundant precipitation during the flowering period destroys as
much as 50% of the total annual pollen catch. Andersson
(1955) and Eklund-Ehrenberg and Simak (1957) also observed
that humid weather conditions prevent pollen from shedding
in P. sylvestris. Second, rain drops could serve as agents in the
transport of pollen from the bract scales to the micropyle. In a
study of the pollination mechanism in P. taeda, it was concluded that the force of falling raindrops can cause the movement of pollen from the micropylar arms to the micropyle
(Brown and Bridgwater 1987). In P. taeda, Greenwood (1986)
found that the transfer of pollen through the micropyle to the
nucellus is carried out by either precipitation or the pollen
drop, whichever appears first. Sweet et al. (1992) studied
artificial pollination of Pinus radiata D. Don and found significantly more pollen grains in the micropyles after pollinations
with pollen suspended in water compared with dry pollinations. Sarvas (1962) claims that rain water has little chance of
penetrating between the bud scales in P. sylvestris; however,
Eklund-Ehrenberg and Simak (1957) reported that already
captured pollen grains were washed from the strobili when
pollinated P. sylvestris strobili were rinsed with water. They
concluded that rain not only prevents pollen shedding, but it
also reduces the possibility of already captured pollen reaching
the pollen chamber. Furthermore, both Sarvas (1962) and Eklund-Ehrenberg and Simak (1957) reported that pollen vigor is
rapidly lost when pollen is exposed to humid conditions.
Table 4. Success rate of SMP, empty seed frequency and number of filled seeds per cone for years and clones in Experiment 2. The frequency and
standard error range for success rate and empty seed frequency are expressed as back-transformed (probability scale) values and estimated least
squares means. The number of filled seeds per cone are estimated least squares means.
Success rate
Empty seed frequency
Number of filled seeds per cone
Frequency
Error range
Frequency
Error range
Year-wise
1987
1988
Average
0.34
0.22
0.27
0.27--0.44
0.16--0.29
0.40
0.38
0.39
0.38--0.44
0.36--0.39
9.68
15.08
12.38
Clone-wise
Clone G
Clone H
Clone I
Clone J
Clone K
Clone L
Clone M
Clone N
Average
0.34
0.22
0.42
0.69
0.18
0.24
0.14
0.20
0.26
0.22--0.50
0.14--0.34
0.31--0.57
0.55--0.85
0.08--0.39
0.13--0.42
0.05--0.37
0.10--0.37
0.36
0.31
0.34
0.44
0.59
0.43
0.34
0.31
0.39
0.32--0.40
0.29--0.33
0.30--0.37
0.39--0.50
0.54--0.64
0.39--0.46
0.30--0.38
0.29--0.33
5.45
20.83
10.05
11.23
12.30
11.85
6.05
21.25
12.38
SUPPLEMENTAL MASS POLLINATION IN PINE SEED ORCHARDS
Clonal differences in SMP success rate
In Experiment 1, differences in SMP success rate among
clones cannot be explained by differences in the timing of peak
receptivity because we observed only marginal differences
between the grafts in time of peak receptivity. Furthermore, all
pollinations were carried out before the main pollen release
occurred. Another explanation could be that preferential fertilization occurred and the degree to which it did so varied among
clones.
525
pollen was not a limiting factor in determining seed yields in
this study.
Supplemental mass pollination is a practical means for improving the genetic gain both in young and in mature
P. sylvestris seed orchards. However, to improve the gain further as well as to reduce the cost per harvested seed, it will be
necessary to combine the use of SMP on selected clones with
flower stimulation techniques. New propagation methods will
also have to be developed if complete control of the parentage
is to be achieved in mass propagation of improved P. sylvestris.
Empty seed frequency and number of filled seeds per cone
Compared with the control trees, all SMP-treated trees in
Experiment 1 had a higher empty seed frequency. Because
formation of empty seed in the Pinus genera always requires
the presence of a germinating pollen grain in the ovule
(McWilliam 1959, Sarvas 1962, Plym-Forshell 1974, Owens
et al. 1981), these findings suggest that the applied pollen
germinated but was not able to complete the fertilization process as successfully as pollen from competing sources. In Experiment 2, the empty seed frequency in the control and
SMP-treated trees was similar, indicating that the pollen supplied in Experiment 2 was of similar vigor to that of the
competing wind-borne pollen. We conclude that pollen used in
SMP must be highly viable to compete successfully with
wind-borne pollen (Bridgwater et al. 1993). Other factors that
could cause formation of empty seeds in P. sylvestris are
homozygosity of sublethal genes causing degeneration of the
embryo or disturbances during embryo development induced
by climate or poor nutrition (e.g., Sarvas 1962, Plym-Forshell
1974). The latter is the most probable explanation for the
difference in number of filled seeds observed between years in
Experiment 2.
Success rate and pollen production in the seed orchard
Although SMP success rate was higher in the orchard with low
pollen production than in the orchard with high pollen production (cf. Tables 2 and 3), the effects of orchard pollen production may explain only some of this difference. First, almost all
of the pollinations in the study were carried out before the time
for main pollen shedding. Moreover, the small amount of
pollen shed during the experimental period was probably
washed away by the frequent rains. Second, the grafts were
taller in the orchard with high pollen production than in the
orchard with low pollen production, and therefore, female
strobili at the top of the taller grafts were more difficult to
detect than those at the top of the smaller grafts. This interpretation supports the idea that relatively small, densely spaced
grafts offer the best prospects of successful SMP of
P. sylvestris. Finally, if the amount of seed orchard and background pollen was insufficient for an adequate pollination in
the orchard with low pollen production, the number of filled
seeds per cone would have been higher in SMP-treated trees
than in control trees. It is unlikely that the supply of pollen
from within and outside the orchard was limited during the
flowering period. Hadders (1973) showed that seed production
in a young seed orchard was normal despite the emasculation
of all male strobili in the orchard. Thus, we conclude that
Acknowledgments
The authors thank Anders Ramström, Mats Eriksson, Roger Granbom
and Inga-Britt Carlsson for skillful assistance. We are also grateful to
Prof. Öje Danell for fruitful discussions and good advice, as well as to
Prof. Gösta Eriksson for valuable comments on the manuscript. Financial support was provided by the Swedish Forestry Research Foundation (SSFf) and by the research Committee of the Swedish National
Board of Forest Research.
References
Andersson, E. 1955. Pollenspridning och avståndsisolering av
fröplantager. [Pollen dispersal and isolation of seed orchards by
distance.] Norrlands Skogsvårdsförbunds Tidsskr. 1955:35--100.
(Summary in English).
Andersson, B. and J. Westin. 1990. Results of freezing tests on Scots
pine (Pinus sylvestris L.) from Swedish and Finnish seed orchards
and natural stands. Report No. 17, The Institute for Forest Tree
Improvement, Uppsala, Sweden, 31 p.
Ashton, W.D. 1972. The logit transformation. Charles Griffin and Co.
Ltd., London, U.K., 88 p.
Bhumibhamon, S. 1978. Studies on Scots pine seed orchards in Finland with special emphasis on the genetic composition of the seed.
Commun. Inst. For. Fenn. 94(4):1--118.
Blush, T.D. 1987. An operational trial of supplemental mass pollination in a loblolly pine seed orchard. In Proc. 19th South. For. Tree
Improve. Conf. College Station, Texas, USA, pp 240--246.
Bridgwater, F.E. and I.F. Trew. 1981. Supplemental mass pollination.
In Pollen Management Handbook. Ed. E.C. Franklin. USDA For.
Serv. Agric. Handbook No. 587, Washington DC, USA, pp 52--56.
Bridgwater, F.E. and C.G. Williams. 1983. Feasibility of supplemental
mass pollination to increase genetic gains from seed orchard. In
Proc. 17th South. For. Tree Improve. Conf. Athens, GA, USA, pp
78--83.
Bridgwater, F.E., T.D. Blush and N.C. Wheeler. 1993. Supplemental
mass pollination. In Advances in Pollen Management. Eds. D.
Bramlett, G.R. Askew, T.D. Blush, F.E. Bridgwater and J.B. Jett.
USDA For. Serv. Agric. Handbook No. 587, Washington DC, USA,
pp 69--79.
Brown, D.B. and F.E. Bridgwater. 1987. Observations on pollination
in loblolly pine. Can. J. For. Res. 17:229--303.
Chung, M.S. 1981. Flowering characteristics of Pinus sylvestris L.
with special emphasis on the reproductive adaptation to local temperature factor. Acta For. Fenn. 169:1--69.
Danell, Ö. 1991. Possible gains in initial stages of a national tree
breeding programme using different techniques. For. Tree Improve.
23:11--30.
Danell, Ö. 1993. Breeding gain and genetic diversity. In Production of
Genetically Improved Seed. Eds. L. Wilhelmsson, U. Eriksson and
Ö. Danell. Redogörelse No. 3, The Forestry Research Institute,
Uppsala, Sweden, 52 p (In Swedish, summary in English).
526
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
Di-Giovanni, F. and P.G. Kevan. 1991. Factors affecting pollen dynamics and its importance to pollen contamination. Can. J. For. Res.
21:1155--1170.
Eklund-Ehrenberg, C. and S. Simak. 1957. Flowering and pollination
in Scots pine (Pinus sylvestris L.). Medd. Statens Skogsforskningsinst. 46:12--27.
El-Kassaby, Y.A., D. Rudin and R. Yazdani. 1989. Levels of outcrossing and contamination in two Pinus sylvestris L. seed orchards in
Northern Sweden. Scand. J. For. Res. 4:41--49.
Eriksson, G., A. Jonsson and D. Lindgren. 1973. Flowering in a clone
trial of Picea abies. Karst. Stud. For. Suec. 110:1--45.
Eriksson, U. 1993. System for large scale extraction of Scots pine
pollen. In Advances in Pollen Management. Eds. D. Bramlett, G.R.
Askew, T.D. Blush, F.E. Bridgwater and J.B. Jett. USDA For. Serv.
Agric. Handbook No. 587, Washington DC, USA, pp 95--97.
Eriksson, U. and C.-H. Palmér. 1991. Forest tree seed orchards in
Sweden 1990. In Yearbook 1990. Föreningen Skogsträdsförädling,
Institutet för skogsförbättring, Uppsala, Sweden, pp 46--60. (Summary in English).
Eriksson, U. and L. Wilhelmsson. 1991. Slutrapport till skogsbrukets
forskningsfond för anslag nr 128 ‘‘Styrd pollinering och kontrollerad fröframställning.’’ Work Report No. 243, Institutet för
Skogsförbättring, Uppsala, Sweden, 27 p.
Eriksson, U., R. Yazdani, L. Wilhelmsson and Ö. Danell. 1994. Success rate of supplemental mass pollination in a seed orchard of
Pinus sylvestris L. Scand. J. For. Res. 9:60--67.
Greenwood, M.S. 1986. Gene exchange in loblolly pine: the relation
between pollination mechanism, female receptivity and pollen
availability. Am. J. Bot. 73:1443--1451.
Hadders, G. 1973. Kontroll av inkorsningen i en tallplantage In Yearbook 1972. Föreningen Skogsträdsförädling, Institutet för skogsförbättring, Uppsala, Sweden, pp 47--64.
Harju, A. and O. Muona. 1989. Background pollination in Pinus
sylvestris seed orchards. Scand J. For. Res. 4:513--520.
Harville, D.A. and R.W. Mee. 1984. A mixed model procedure for
analysing ordered categorical data. Biometrics 40:393--408.
Jonsson, A., I. Ekberg and G. Eriksson. 1976. Flowering in a seed
orchard of Pinus sylvestris L. Stud. For. Suec. 135:1--38.
Lestander, T. and D. Lindgren. 1985. The genetic efficiency of seed
orchards. In Reforestation Material for Harsh Northern Sites. Ed.
J.-E. Nilsson. Report No. 4, Department of Forest Genetics and
Plant Physiology, Swedish University of Agricultural Sciences,
Umeå, Sweden, pp 111--127. (Summary in English).
McWilliam, J.R. 1959. Interspecific incompatibility in Pinus. Am. J.
Bot. 46:425--433.
Owens, J.N., S.J. Simpson and M. Molder. 1981. Sexual reproduction
of Pinus contorta. 1. Pollen development, the pollination mechanism, and early development. Can. J. Bot. 39:1828--1843.
Paule, L. 1991. Clone identity and contamination in a Scots pine seed
orchard. In Proc. Meet. of the Nordic Group for Tree Breeding in
Sweden. Ed. D. Lindgren. Report No. 10, Department of Forest
Genetics and Plant Physiology, Swedish University of Agricultural
Sciences, Umeå, Sweden, pp 22--32.
Plym-Forshell, C. 1974. Seed development after self pollination and
cross-pollination of Scots-pine, Pinus sylvestris L. Stud. For. Suec.
118:1--37.
Rudin, D. 1975. Inheritance of glutamate oxalate transaminase (GOT)
from needles and endosperms of Pinus sylvestris. Hereditas
80:296--300.
Rudin, D. 1977. Leucine-amino-peptidase (LAP) from needles and
endosperms of Pinus sylvestris L. A study of inheritance of allozymes. Hereditas 85:219--226.
Sarvas, R. 1962. Investigations on the flowering and seed crop of
Pinus sylvestris. Commun. Inst. For. Fenn. 53: 4--198.
Savolainen, O. 1991. Background pollination in seed orchards. In
Proc. Meet. of the Nordic Group for Tree Breeding in Sweden. Ed.
D. Lindgren. Report No. 10, Department of Forest Genetics and
Plant Physiology, Swedish University of Agricultural Sciences,
Umeå, Sweden, pp 6--13.
Sweet, G.B., B.D. Dickson, B.D. Donaldson and H. Litchwark. 1992.
Controlled pollination without isolation----a new approach to the
management of radiata pine seed orchards. Silvae Genet. 41:95--99.
Wang, X., D. Lindgren, A.E. Szmidt and R. Yazdani. 1991. Pollen
migration into a seed orchard of Pinus sylvestris and the methods of
its estimation using allozyme markers. Scand. J. For. Res. 6:379-386.
Wheeler, N.C. and S.J. Jech. 1985. Estimating supplemental mass
pollination (SMP) success electorphoretically. In Proc. 20th Can.
Tree Improve. Assoc., Part 2. Eds. F. Caron, A.G. Corriveau and
T.J.B. Boyle. Quebec City, Quebec, Canada, pp 111--120.
Yazdani, R., G. Hadders and A.E. Szmidt. 1986. Supplemental mass
pollination in a seed orchard of Pinus sylvestris L. investigated by
isozyme analyses. Scand. J. For. Res. 1:309--315.
Yazdani, R. and D. Lindgren. 1991. Variation of pollen contamination
in a Scots pine seed orchard. Silvae Genet. 40:243--246.
© 1995 Heron Publishing----Victoria, Canada
Effects of supplemental mass pollination (SMP) in a young and a
mature seed orchard of Pinus sylvestris
U. ERIKSSON,1 G. JANSSON,1 R. YAZDANI2 and L. WILHELMSSON1
1
Forestry Research Institute, Glunten, S-751 83 Uppsala, Sweden
2
Department of Forest Genetics, The Swedish University of Agricultural Sciences, P.O. Box 7027, S-750 07 Uppsala, Sweden
Received May 31, 1994
Summary The effects of supplemental mass pollination
(SMP) were studied in two Pinus sylvestris seed orchards
differing in pollen production. Pollen was dusted over the
whole tree during the period of peak female receptivity using
a pressurized backpack sprayer. The success of SMP was
assessed by means of allozyme markers. In the orchard with
high pollen production, detectable differences in SMP success
rate were found between clones, but the success rate was not
influenced by the number of pollinations per day. The average
estimated success rate of SMP was 19%. In the orchard with
low pollen production, no significant differences in SMP success rate were found between years (22 versus 34%) or between
clones. The SMP success rate in the low pollen production
orchard varied between 14 and 69%.
Keywords: isozymes, pollen production, Scots pine, seed orchard management.
Introduction
Based on height growth, the average genetic superiority of
first-generation plus-tree seed orchards of Pinus sylvestris L.
in Sweden is estimated to be between 6 and 8% compared with
unselected stand material (Danell 1991, 1993). However, the
genetic potential of seed orchards has not been entirely exploited mainly because it has not yet been possible to eliminate
pollen contamination from unselected sources outside the orchards (e.g., Savolainen 1991, Di-Giovanni and Kevan 1991).
Estimates of pollen contamination in P. sylvestris seed orchards range between 17 and 74% (El-Kassaby et al. 1989,
Harju and Muona 1989, Yazdani and Lindgren 1991, Wang et
al. 1991, Paule 1991). Pollen contamination affects the growth,
quality and hardiness of the seedling output. Serious pollen
contamination in southern orchards producing seed intended
for northern regions with a harsh climate makes the seed crops
unsuitable for the intended forest regeneration area (Lestander
and Lindgren 1985, Andersson and Westin 1990, Savolainen
1991). Beside pollen contamination, differences in random
mating and gamete production capacity within and among
clones may affect the genetic quality of the output from seed
orchards (e.g., Eriksson et al. 1973, Jonsson et al. 1976, Bhumibhamon 1978, Chung 1981). Thus, there is scope to develop
methods that facilitate a higher recovery of the potential genetic gain of seed orchards.
In seed orchards of Pinus taeda L. and Pseudotsuga menziesii (Mirb.) Franco, having limited pollen supply, supplementary mass pollination (SMP) (Bridgwater and Trew 1981),
defined as the broadcast application of pollen to unisolated
female strobili, has been used to increase the yield of sound
seeds (see review by Bridgwater et al. 1993). The success rate
of SMP has been estimated by using pollen with unique biochemical marker alleles (Wheeler and Jech 1985, Yazdani et al.
1986). Based on this technique, high success rates of SMP
have been reported for seed orchards of P. menziesii (Wheeler
and Jech 1985), P. taeda (Blush 1987) and P. sylvestris
(Eriksson et al. 1994). The latter investigation showed average
success rates between 66 and 84% when individual strobili
were pollinated; however, in an operational study, in which
whole grafts were pollinated, the success rate declined to
7--26%.
Besides competition by pollen from other sources, the success rate of SMP may be influenced by other factors including
the prevailing environmental conditions at flowering and the
number of occasions a graft must be pollinated in order to
achieve a satisfactorily high success rate. Eriksson et al. (1994)
concluded that, for a single strobilus, timing rather than the
number of pollinations is critical but, because of the gradual
maturation of female strobili, whole grafts need to be pollinated more than once a day during the flowering period.
Bridgwater et al. (1993) concluded that a single pollination at
peak female receptivity is enough to achieve fairly high success rates in P. taeda. However, differences in success rates
among clones have been reported for P. taeda (Bridgwater and
Williams 1983, Blush 1987) and P. menziesii (Wheeler and
Jech 1985) that can be explained mainly by differences in
flower phenology in relation to main pollen shedding. Bridgwater et al. (1993) concluded that pollination of clones that
flower before maximum pollen flight provides the best opportunity for successful SMP, but did not exclude the possibility
of successful use of SMP for clones that flower during maxi-
520
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
mum pollen flight. Differences in SMP success rate between
clones could also be expected for P. sylvestris because there are
large variations in flower phenology among and within clones
of this species. Furthermore, because the course of flowering
varies between years as a result of differing weather conditions
during the flowering period (Sarvas 1962, Jonsson et al. 1976),
annual variations in flowering phenology could also influence
SMP success rate. Therefore, we have tested the hypothesis
that the success rate of SMP in P. sylvestris seed orchards is
influenced by clone, weather and frequency of application.
to June 12, 1988). Pollen from one clone, not represented in
the orchard, was used in both years. The pollen clone (F) is
heterozygous at the GOT-B locus with the genotype B1/B2,
where B1 is the rare allele. Nine randomly selected clones
(G--O), not represented in Experiment 1, were used as mother
clones. Two individual grafts were sampled per clone, one for
each treatment. The same grafts and clones were pollinated
both years. The two pollination treatments in Experiment 2
were: (i) wind-pollinated control (Control), and (ii) SMP of
whole trees twice a day during the period of peak receptivity.
Materials and methods
Application of pollen
Experiment 1
Experiment 1 was carried out in the spring of 1987 in a mature
P. sylvestris seed orchard with high pollen production (Table 1). Trees were pollinated for 8 days between May 28 and
June 4, 1987. A mix with equal portions of pollen from two
clones not cultivated in the orchard, designated A and B, was
used. The clones were chosen because they had rare isozyme
markers. Clone A is heterozygous at the LAP-B locus with the
genotype B1/B2, and clone B is heterozygous for the GOT-B
locus with the genotype B1/B3. B1 is the rare allele at both
loci.
Four grafts from each of three mother clones (C--E) were
each subjected to four pollination treatments as follows: (i)
wind-pollinated control (Control), (ii) SMP of whole trees
once per day during the period of peak receptivity, (iii) SMP
of whole trees up to three times per day during the period of
peak receptivity, and (iv) SMP of whole trees up to six times
per day during the period of peak receptivity.
Weather and amount of pollen
Experiment 2
Experiment 2 was carried out during both 1987 and 1988 in a
P. sylvestris seed orchard with low pollen production (Table 1).
Supplemental mass pollination was carried out during 11 days
in both 1987 and 1988 (June 15 to June 26, 1987, and June 1
Table 1. Description of the seed orchards.
Characteristic
Experiment 1
Experiment 2
Name
Latitude
Longitude
Altitude
Area
Year of establishment
Spacing
Mean height (approx.)
No. of clones
Soil texture
Seed yield, 1979--19881
Pollen production, 19872
Pollen production, 19882
493 Askerud
59°53′ N
13°10′ E
80 m
14 ha
1966--1969
5×5m
5m
43
Clay
52 kg ha − 1
40 kg ha − 1
39 kg ha − 1
123 Klocke
62°54′ N
18°16′ E
75 m
16 ha
1968--1972
5.6 × 5.6 m
3.5 m
60
Silt
3.6 kg ha − 1
-0.5 kg ha − 1
1
2
In both experiments, the pollen mix was dusted over clusters
of receptive, unisolated female strobili throughout the whole
graft in an attempt to pollinate as many receptive strobili as
possible. Pollen was delivered to the strobili from a pollination
wand activated by compressed air from a tube mounted on a
modified backpack sprayer (Eriksson et al. 1994). About 3 ml
of pollen was dusted over each graft in each pollination. The
SMP was started when approximately 20% of the female
strobili on a graft were judged to be receptive, according to the
classification system presented by Jonsson et al. (1976). No
pollinations were done during precipitation. The intention was
to pollinate each graft for 3--5 days during the period of peak
female strobilus receptivity. As a consequence of frequent
precipitation, which slowed maturation of female strobili, the
pollination period was extended to 4 to 8 days per graft in
Experiment 1. For the same reason, the pollination period in
Experiment 2 was extended to about 11 days in 1987 and to
between 7 and 11 days in 1988.
Eriksson and Palmér (1991).
Eriksson and Wilhelmsson (1991).
For both experiments, temperatures were estimated as the daily
mean temperatures recorded by the Swedish Meteorological
and Hydrological Institute. For Experiment 1, we used the data
collected at Station 9240 Arvika, situated 40 km southwest of
the orchard. The station closest to the orchard in Experiment 2
was Station 6311 Skagsudde, situated 40 km north of the
orchard. The occurrence of precipitation was observed daily
during the experimental periods in both experiments. The
amount of pollen in the air during the experimental periods was
assessed with a pollen-catching device (Sarvas 1962).
Pollen handling, cone collection and seed extraction
In 1986, pollen was extracted under controlled temperature
and humidity conditions as described by Eriksson (1993). The
pollen lots were stored in sealed glass jars at −20 °C. All cones
on the treated grafts were collected in the autumn of 1988
(Experiments 1 and 2) and 1989 (Experiment 2). Seeds were
extracted and stored at −4 °C until isozyme analyses were
done. Numbers of empty seed and filled seeds per cone were
determined for all treatments. Two clones in Experiment 2
produced no seed after the pollinations in 1987, and one clone
produced no seed after the pollinations in 1988. These clones
were excluded from subsequent analyses.
SUPPLEMENTAL MASS POLLINATION IN PINE SEED ORCHARDS
Marker detection
The analyses of SMP success rate were carried out by isozyme
separation and starch gel electrophoresis. In Experiment 1, two
enzymes, leucine aminopeptidase (LAP) (Rudin 1977) and
glutamate oxalate transaminase (GOT) (Rudin 1975), were
analyzed in the diploid embryo and in the haploid megagametophyte. One enzyme (GOT) was analyzed in Experiment 2.
With one exception, 100 seeds were analyzed per clone and
treatment in Experiment 1 (only 40 seeds were available for
clone D in treatment 3P). In Experiment 2, with a few exceptions, about 100 seeds per clone and treatment were used in the
subsequent studies.
The genotypes of the mother clones in the orchard with high
pollen production were unknown at the time of the experiment.
In a multilocus analysis of all clones in the orchard, mother
clone E appeared to have the same rare allele (B1) in the GOT
system as father clone B.
Statistical analysis
Statistical analyses of frequencies were performed on the logit
transformations of y (Ashton 1972), as described by Eriksson
et al. (1994). Briefly, the observations y were the observed
proportions of SMP success rates and empty seed frequencies
on each graft. In both experiments, the empirical logits lij of the
responses in the treatment × clone subclasses were obtained as:
yij
lij = ln
.
1 − yij
(1)
aij = (oij − cj)/(1 − cj),
In Experiment 2, the total contribution rates from SMP were
computed as:
B1
yij = aGOT
,
ij
lij = µ + ti + sj + eij ,
(5)
where lij is the logit value of success rate in subclass ij, µ is the
overall mean, ti is the fixed effect of the ith treatment where i
= 1, 2 or 3, sj is the fixed effect of the jth clone where j = 1, 2
or 3, and eij is the random residual effect of ijth observed logit
value assumed to be individually and independently distributed (IID) (0, π2/3). The weight used for the ijth observation
was:
nij × yij (1 − yij ),
where nij is the number of observations in the ijth class. The
following model was used for Experiment 2:
(6)
where lij is the logit values of success rate in subclass ij, µ is
the overall mean, si is the fixed effect of the ith clone where i
= 1, ..., 8, uj is the fixed effect of the jth year where j = 1 or 2,
and eij is the random residual effect of the ijth observed logit
value assumed IID (0, π2/3). The weight, used for each observation in the least squares equation, was the same as in Equation 5.
The estimated responses in both experiments were backtransformed from the underlying scale to the visible frequency
scale by means of Equation 7:
(2)
p^ =
where aij is the adjusted observed frequency of LAP B1 or
GOT B1 in the seed from treatment i and clone j, oij is the
corresponding observed LAP B1 or GOT B1 frequency, and cj
is the observed allele frequency for the pooled control data.
Values of a less than 0 are unrealistic and have therefore been
given the value 0.
In Experiment 1, the total contribution rates from SMP were
computed as:
B1
B1
yij = (aLAP
+ aGOT
),
ij
ij
(4)
where yij is the success rate of SMP pollen fertilization for a
B1
is the observed
particular clone × year subclass, and aGOT
ij
rate adjusted according to Equation 2.
The statistical analyses of the logit values were performed
with the weighted least squares procedure available in the
GLM procedure of the SAS program (SAS Institute Inc., Cary,
NC). The following model was used for Experiment 1:
lij = µ + si + uj + eij ,
Because the logit transformations are undefined when the
observed and adjusted yij values are 0 or 1, values of yij equal
to 0 or 1 were excluded from the subsequent analysis (Harville
and Mee 1984).
The observed values of LAP B1 and GOT B1 frequencies
were adjusted for the contribution from the background pollen
to the marker allele frequency by means of Equation 2. This
equation assumes that the background contributions of LAP
B1 and GOT B1 are proportional to the part not fertilized by
the SMP pollen:
521
(3)
where yij is the success rate of SMP pollen fertilization for a
B1
B1
particular treatment × clone subclass, and aLAP
and aGOT
ij
ij
are the observed rates adjusted according to Equation 2.
1
_^ ,
1
+
e r)
(
(7)
where p^ is the estimated frequency corresponding to a leastsquares mean r^ measured in the logit scale.
Because all father clones were heterozygous for the locus
with the rare marker allele, the estimated frequencies p^ were
multiplied by 2. In both experiments, the open-pollinated controls were included as one of the treatments in the analyses of
empty seed frequency.
The ordinary least squares procedure of the SAS software
package was used to analyze numbers of filled seeds per cone.
The following model was used for Experiment 1:
z ij = µ + ti + sj + eij,
(8)
522
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
where zij is the number of filled seeds per cone in subclasses ij,
µ is the overall mean, ti is the fixed effect of the ith treatment
where i = 1, ..., 4, sj is the fixed effect of the jth clone where j
= 1, 2 or 3, and eij is the random residual effect of ijth observed
value assumed IID (0, σ2e ). The following model was used for
Experiment 2:
zijk = µ + ti + sj + uk + eijk ,
(9)
where zijk is the number of filled seeds per cone in the subclasses ijk, µ is the overall mean, ti is the fixed effect of the ith
treatment where i = 1 or 2, sj is the fixed effect of the jth clone
where j = 1, ..., 8, uk is the fixed effect of the kth year where k
= 1 or 2, and eijk is the random residual effect of ijkth observed
value assumed IID (0, σ2e ). The wind-pollinated controls were
included as a treatment in the analyses of the number of filled
seeds per cone.
Results
Experiment 1
From May 29 to May 31, the weather was dry and sunny, but
turned rainy from June 1 to the end of the experimental period
(Figure 1). The average daily mean temperature was about
10 °C, and only small amounts of pollen were trapped during
the experimental period. The main pollen shedding in the
orchard started on June 4 (Figure 1).
The results from the ANOVAs of SMP success rate, empty
seed frequency and number of filled seeds per cone are presented in Table 2. The number of pollinations per day (pollination treatments) did not significantly affect the success rate.
However, significant differences in success rate were found
between clones. The pollination treatments had a significant
effect on empty seed frequency, but there was no significant
difference in empty seed frequency among clones. There were
no significant effects of either pollen treatment or clone on
number of filled seeds per cone. Estimated least squares means
of SMP success rate, empty seed frequency and number of
filled seeds per cone for treatments and clones are presented in
Table 3.
Experiment 2
During the 1987 experimental period, the weather was mostly
rainy and cold with an average daily mean temperature of
about 10 °C (Figure 2). Two small peaks of pollen in the air
were detected (Figure 2). Unfortunately, the device for pollen
catching was defective from June 24, 1987, to the end of the
experimental period, but another device for pollen catching in
another part of the orchard indicated that there were only small
amounts of pollen in the air during this time.
Figure 1. Daily mean temperatures
(dashed line) recorded by the Swedish
Meterological and Hydrological Institute
at their station 9240 Arvika, situated
40 km southwest of the orchard, and the
amounts of pollen in the air (solid line), assessed with a pollen catching device, in
Experiment 1, 1987. The bar shows the period of female strobili receptivity on the
treated grafts, i.e., the experimental period.
Table 2. Analyses of variance for success rate, empty seed frequency and number of filled seeds per cone in Experiments 1 and 2.
Source of variation
Success rate
Empty seed frequency
Number of filled seeds per cone
df
MS
F-value
P
df
MS
F-value
P
df
MS
Experiment 1
Treatment
Clone
Error
2
2
3
1.48
3.15
0.16
9.32
19.81
0.052
0.019
3
2
6
229.1
9.5
8.2
27.93
1.16
< 0.001
0.37
3
2
6
11.2
71.4
16.7
Experiment 2
Year
Treatment
Clone
Error
1
-7
4
8.10
-3.15
5.05
1.60
-1.64
0.27
-0.33
1
1
2
6
37.3
21.0
253.2
45.0
0.83
0.47
5.62
0.37
0.50
0.001
1
1
2
6
204.1
5.5
140.1
12.4
F-value
0.67
4.28
16.4
0.4
11.3
P
0.6
0.07
< 0.001
0.52
< 0.001
SUPPLEMENTAL MASS POLLINATION IN PINE SEED ORCHARDS
523
Table 3. Success rate of SMP, empty seed frequency and number of filled seeds per cone for treatments in Experiment 1. The frequency and standard
error range for success rate and empty seed frequency are expressed as back-transformed (probability scale) values and estimated least squares
means. The number of filled seeds per cone are estimated least squares means.
Success rate
Empty seed frequency
Number of filled seeds per cone
Frequency
Error range
Frequency
Error range
Treatment-wise
Max 1 SMP per day
Max 3 SMP per day
Max 6 SMP per day
Control
Average
0.16
0.24
0.18
-0.19
0.15--0.17
0.22--0.26
0.17--0.19
--
0.22
0.36
0.21
0.13
0.22
0.20--0.24
0.33--0.40
0.19--0.24
0.11--0.15
16.74
12.88
13.81
16.50
14.98
Clone-wise
Clone C
Clone D
Clone E
Average
0.26
0.11
0.23
0.19
0.24--0.27
0.10--0.13
0.22--0.25
0.27
0.18
0.22
0.22
0.23--0.31
0.14--0.22
0.21--0.23
11.16
14.26
19.52
15.34
Figure 2. Daily mean temperatures
(dashed line) recorded by the Swedish
Meterological and Hydrological Institute
at their station 9311 Skagsudde, situated
40 km north of the orchard, and the
amounts of pollen in the air (solid line), assessed with a pollen catching device, in
Experiment 2, 1987. The bar shows the period of female strobili receptivity on the
treated grafts, i.e., the experimental period.
Figure 3. Daily mean temperatures
(dashed line) recorded by the Swedish
Meterological and Hydrological Institute
at their station 9311 Skagsudde, situated
40 km north of the orchard, and the
amounts of pollen in the air (solid line), assessed with a pollen catching device, in
Experiment 2, 1988. The bar shows the period of female strobili receptivity on the
treated grafts, i.e., the experimental period.
During the first half of the 1988 experimental period, the
weather was rainy and cold with an average daily mean temperature of about 8 °C (Figure 3). The weather during the latter
part of the period was mostly sunny with only a few showers
and an average daily mean temperature of about 12 °C. Small
amounts of pollen in the air were found before June 8 when the
main pollen shedding started (Figure 3).
The ANOVAs for SMP success rate, empty seed frequency
and number of filled seeds per cone are shown in Table 2. No
significant effects, for either years or clones, were found in the
analysis of success rate. Significant differences in empty seed
frequency were found between clones, but not between years
524
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
or pollination treatments. The difference in number of filled
seeds per cone was significant between years but not between
pollination treatments. Furthermore, clones had significantly
different numbers of filled seeds per cone. Estimated least
squares means of SMP success rate, empty seed frequency and
number of filled seeds per cone for years, treatments and
clones are presented in Table 4.
Discussion
Success rate and number of pollinations per day
Eriksson et al. (1994) concluded that a single SMP of a
P. sylvestris strobilus is enough to achieve a high success rate,
provided that the strobilus is at peak receptivity when pollinated, whereas whole trees should be pollinated more than
once a day, because not all female strobili on a tree reach peak
receptivity at the same time (Sarvas 1962). However, we found
that increasing the number of pollinations per day caused only
a small, nonsignificant increase in SMP success rate in Experiment 1. The absence of pollination treatment effects in our
study may be associated with the slow maturation of female
strobili as a result of the cold and rainy weather. Thus, only a
limited number of new unpollinated strobili may have become
receptive between pollinations. If so, this suggests that SMP of
a graft should be carried out relatively more often during a hot
day than during a cold day.
Success rate and weather conditions
The average SMP success rate in Experiment 1 was higher than
that reported for a comparable study in 1986 (Eriksson et al.
1994). One explanation could be that the weather conditions
differed between 1986 and 1987. In contrast to 1987, the
weather in 1986 was mostly sunny with little or no precipitation. The finding that there was no significant difference in
SMP success rate between years in Experiment 2 does not
exclude the possibility that there could be differences between
years with more variable weather conditions. Even though the
weather in 1988 improved during the second part of the experimental period, both flowering periods studied in Experiment 2
were rainy and cold. Frequent precipitation and high humidity
could have contributed to SMP success rate in two ways. First,
precipitation reduces the amount of competing pollen in the air
by washing it to the ground. Sarvas (1962) showed that abundant precipitation during the flowering period destroys as
much as 50% of the total annual pollen catch. Andersson
(1955) and Eklund-Ehrenberg and Simak (1957) also observed
that humid weather conditions prevent pollen from shedding
in P. sylvestris. Second, rain drops could serve as agents in the
transport of pollen from the bract scales to the micropyle. In a
study of the pollination mechanism in P. taeda, it was concluded that the force of falling raindrops can cause the movement of pollen from the micropylar arms to the micropyle
(Brown and Bridgwater 1987). In P. taeda, Greenwood (1986)
found that the transfer of pollen through the micropyle to the
nucellus is carried out by either precipitation or the pollen
drop, whichever appears first. Sweet et al. (1992) studied
artificial pollination of Pinus radiata D. Don and found significantly more pollen grains in the micropyles after pollinations
with pollen suspended in water compared with dry pollinations. Sarvas (1962) claims that rain water has little chance of
penetrating between the bud scales in P. sylvestris; however,
Eklund-Ehrenberg and Simak (1957) reported that already
captured pollen grains were washed from the strobili when
pollinated P. sylvestris strobili were rinsed with water. They
concluded that rain not only prevents pollen shedding, but it
also reduces the possibility of already captured pollen reaching
the pollen chamber. Furthermore, both Sarvas (1962) and Eklund-Ehrenberg and Simak (1957) reported that pollen vigor is
rapidly lost when pollen is exposed to humid conditions.
Table 4. Success rate of SMP, empty seed frequency and number of filled seeds per cone for years and clones in Experiment 2. The frequency and
standard error range for success rate and empty seed frequency are expressed as back-transformed (probability scale) values and estimated least
squares means. The number of filled seeds per cone are estimated least squares means.
Success rate
Empty seed frequency
Number of filled seeds per cone
Frequency
Error range
Frequency
Error range
Year-wise
1987
1988
Average
0.34
0.22
0.27
0.27--0.44
0.16--0.29
0.40
0.38
0.39
0.38--0.44
0.36--0.39
9.68
15.08
12.38
Clone-wise
Clone G
Clone H
Clone I
Clone J
Clone K
Clone L
Clone M
Clone N
Average
0.34
0.22
0.42
0.69
0.18
0.24
0.14
0.20
0.26
0.22--0.50
0.14--0.34
0.31--0.57
0.55--0.85
0.08--0.39
0.13--0.42
0.05--0.37
0.10--0.37
0.36
0.31
0.34
0.44
0.59
0.43
0.34
0.31
0.39
0.32--0.40
0.29--0.33
0.30--0.37
0.39--0.50
0.54--0.64
0.39--0.46
0.30--0.38
0.29--0.33
5.45
20.83
10.05
11.23
12.30
11.85
6.05
21.25
12.38
SUPPLEMENTAL MASS POLLINATION IN PINE SEED ORCHARDS
Clonal differences in SMP success rate
In Experiment 1, differences in SMP success rate among
clones cannot be explained by differences in the timing of peak
receptivity because we observed only marginal differences
between the grafts in time of peak receptivity. Furthermore, all
pollinations were carried out before the main pollen release
occurred. Another explanation could be that preferential fertilization occurred and the degree to which it did so varied among
clones.
525
pollen was not a limiting factor in determining seed yields in
this study.
Supplemental mass pollination is a practical means for improving the genetic gain both in young and in mature
P. sylvestris seed orchards. However, to improve the gain further as well as to reduce the cost per harvested seed, it will be
necessary to combine the use of SMP on selected clones with
flower stimulation techniques. New propagation methods will
also have to be developed if complete control of the parentage
is to be achieved in mass propagation of improved P. sylvestris.
Empty seed frequency and number of filled seeds per cone
Compared with the control trees, all SMP-treated trees in
Experiment 1 had a higher empty seed frequency. Because
formation of empty seed in the Pinus genera always requires
the presence of a germinating pollen grain in the ovule
(McWilliam 1959, Sarvas 1962, Plym-Forshell 1974, Owens
et al. 1981), these findings suggest that the applied pollen
germinated but was not able to complete the fertilization process as successfully as pollen from competing sources. In Experiment 2, the empty seed frequency in the control and
SMP-treated trees was similar, indicating that the pollen supplied in Experiment 2 was of similar vigor to that of the
competing wind-borne pollen. We conclude that pollen used in
SMP must be highly viable to compete successfully with
wind-borne pollen (Bridgwater et al. 1993). Other factors that
could cause formation of empty seeds in P. sylvestris are
homozygosity of sublethal genes causing degeneration of the
embryo or disturbances during embryo development induced
by climate or poor nutrition (e.g., Sarvas 1962, Plym-Forshell
1974). The latter is the most probable explanation for the
difference in number of filled seeds observed between years in
Experiment 2.
Success rate and pollen production in the seed orchard
Although SMP success rate was higher in the orchard with low
pollen production than in the orchard with high pollen production (cf. Tables 2 and 3), the effects of orchard pollen production may explain only some of this difference. First, almost all
of the pollinations in the study were carried out before the time
for main pollen shedding. Moreover, the small amount of
pollen shed during the experimental period was probably
washed away by the frequent rains. Second, the grafts were
taller in the orchard with high pollen production than in the
orchard with low pollen production, and therefore, female
strobili at the top of the taller grafts were more difficult to
detect than those at the top of the smaller grafts. This interpretation supports the idea that relatively small, densely spaced
grafts offer the best prospects of successful SMP of
P. sylvestris. Finally, if the amount of seed orchard and background pollen was insufficient for an adequate pollination in
the orchard with low pollen production, the number of filled
seeds per cone would have been higher in SMP-treated trees
than in control trees. It is unlikely that the supply of pollen
from within and outside the orchard was limited during the
flowering period. Hadders (1973) showed that seed production
in a young seed orchard was normal despite the emasculation
of all male strobili in the orchard. Thus, we conclude that
Acknowledgments
The authors thank Anders Ramström, Mats Eriksson, Roger Granbom
and Inga-Britt Carlsson for skillful assistance. We are also grateful to
Prof. Öje Danell for fruitful discussions and good advice, as well as to
Prof. Gösta Eriksson for valuable comments on the manuscript. Financial support was provided by the Swedish Forestry Research Foundation (SSFf) and by the research Committee of the Swedish National
Board of Forest Research.
References
Andersson, E. 1955. Pollenspridning och avståndsisolering av
fröplantager. [Pollen dispersal and isolation of seed orchards by
distance.] Norrlands Skogsvårdsförbunds Tidsskr. 1955:35--100.
(Summary in English).
Andersson, B. and J. Westin. 1990. Results of freezing tests on Scots
pine (Pinus sylvestris L.) from Swedish and Finnish seed orchards
and natural stands. Report No. 17, The Institute for Forest Tree
Improvement, Uppsala, Sweden, 31 p.
Ashton, W.D. 1972. The logit transformation. Charles Griffin and Co.
Ltd., London, U.K., 88 p.
Bhumibhamon, S. 1978. Studies on Scots pine seed orchards in Finland with special emphasis on the genetic composition of the seed.
Commun. Inst. For. Fenn. 94(4):1--118.
Blush, T.D. 1987. An operational trial of supplemental mass pollination in a loblolly pine seed orchard. In Proc. 19th South. For. Tree
Improve. Conf. College Station, Texas, USA, pp 240--246.
Bridgwater, F.E. and I.F. Trew. 1981. Supplemental mass pollination.
In Pollen Management Handbook. Ed. E.C. Franklin. USDA For.
Serv. Agric. Handbook No. 587, Washington DC, USA, pp 52--56.
Bridgwater, F.E. and C.G. Williams. 1983. Feasibility of supplemental
mass pollination to increase genetic gains from seed orchard. In
Proc. 17th South. For. Tree Improve. Conf. Athens, GA, USA, pp
78--83.
Bridgwater, F.E., T.D. Blush and N.C. Wheeler. 1993. Supplemental
mass pollination. In Advances in Pollen Management. Eds. D.
Bramlett, G.R. Askew, T.D. Blush, F.E. Bridgwater and J.B. Jett.
USDA For. Serv. Agric. Handbook No. 587, Washington DC, USA,
pp 69--79.
Brown, D.B. and F.E. Bridgwater. 1987. Observations on pollination
in loblolly pine. Can. J. For. Res. 17:229--303.
Chung, M.S. 1981. Flowering characteristics of Pinus sylvestris L.
with special emphasis on the reproductive adaptation to local temperature factor. Acta For. Fenn. 169:1--69.
Danell, Ö. 1991. Possible gains in initial stages of a national tree
breeding programme using different techniques. For. Tree Improve.
23:11--30.
Danell, Ö. 1993. Breeding gain and genetic diversity. In Production of
Genetically Improved Seed. Eds. L. Wilhelmsson, U. Eriksson and
Ö. Danell. Redogörelse No. 3, The Forestry Research Institute,
Uppsala, Sweden, 52 p (In Swedish, summary in English).
526
ERIKSSON, JANSSON, YAZDANI AND WILHELMSSON
Di-Giovanni, F. and P.G. Kevan. 1991. Factors affecting pollen dynamics and its importance to pollen contamination. Can. J. For. Res.
21:1155--1170.
Eklund-Ehrenberg, C. and S. Simak. 1957. Flowering and pollination
in Scots pine (Pinus sylvestris L.). Medd. Statens Skogsforskningsinst. 46:12--27.
El-Kassaby, Y.A., D. Rudin and R. Yazdani. 1989. Levels of outcrossing and contamination in two Pinus sylvestris L. seed orchards in
Northern Sweden. Scand. J. For. Res. 4:41--49.
Eriksson, G., A. Jonsson and D. Lindgren. 1973. Flowering in a clone
trial of Picea abies. Karst. Stud. For. Suec. 110:1--45.
Eriksson, U. 1993. System for large scale extraction of Scots pine
pollen. In Advances in Pollen Management. Eds. D. Bramlett, G.R.
Askew, T.D. Blush, F.E. Bridgwater and J.B. Jett. USDA For. Serv.
Agric. Handbook No. 587, Washington DC, USA, pp 95--97.
Eriksson, U. and C.-H. Palmér. 1991. Forest tree seed orchards in
Sweden 1990. In Yearbook 1990. Föreningen Skogsträdsförädling,
Institutet för skogsförbättring, Uppsala, Sweden, pp 46--60. (Summary in English).
Eriksson, U. and L. Wilhelmsson. 1991. Slutrapport till skogsbrukets
forskningsfond för anslag nr 128 ‘‘Styrd pollinering och kontrollerad fröframställning.’’ Work Report No. 243, Institutet för
Skogsförbättring, Uppsala, Sweden, 27 p.
Eriksson, U., R. Yazdani, L. Wilhelmsson and Ö. Danell. 1994. Success rate of supplemental mass pollination in a seed orchard of
Pinus sylvestris L. Scand. J. For. Res. 9:60--67.
Greenwood, M.S. 1986. Gene exchange in loblolly pine: the relation
between pollination mechanism, female receptivity and pollen
availability. Am. J. Bot. 73:1443--1451.
Hadders, G. 1973. Kontroll av inkorsningen i en tallplantage In Yearbook 1972. Föreningen Skogsträdsförädling, Institutet för skogsförbättring, Uppsala, Sweden, pp 47--64.
Harju, A. and O. Muona. 1989. Background pollination in Pinus
sylvestris seed orchards. Scand J. For. Res. 4:513--520.
Harville, D.A. and R.W. Mee. 1984. A mixed model procedure for
analysing ordered categorical data. Biometrics 40:393--408.
Jonsson, A., I. Ekberg and G. Eriksson. 1976. Flowering in a seed
orchard of Pinus sylvestris L. Stud. For. Suec. 135:1--38.
Lestander, T. and D. Lindgren. 1985. The genetic efficiency of seed
orchards. In Reforestation Material for Harsh Northern Sites. Ed.
J.-E. Nilsson. Report No. 4, Department of Forest Genetics and
Plant Physiology, Swedish University of Agricultural Sciences,
Umeå, Sweden, pp 111--127. (Summary in English).
McWilliam, J.R. 1959. Interspecific incompatibility in Pinus. Am. J.
Bot. 46:425--433.
Owens, J.N., S.J. Simpson and M. Molder. 1981. Sexual reproduction
of Pinus contorta. 1. Pollen development, the pollination mechanism, and early development. Can. J. Bot. 39:1828--1843.
Paule, L. 1991. Clone identity and contamination in a Scots pine seed
orchard. In Proc. Meet. of the Nordic Group for Tree Breeding in
Sweden. Ed. D. Lindgren. Report No. 10, Department of Forest
Genetics and Plant Physiology, Swedish University of Agricultural
Sciences, Umeå, Sweden, pp 22--32.
Plym-Forshell, C. 1974. Seed development after self pollination and
cross-pollination of Scots-pine, Pinus sylvestris L. Stud. For. Suec.
118:1--37.
Rudin, D. 1975. Inheritance of glutamate oxalate transaminase (GOT)
from needles and endosperms of Pinus sylvestris. Hereditas
80:296--300.
Rudin, D. 1977. Leucine-amino-peptidase (LAP) from needles and
endosperms of Pinus sylvestris L. A study of inheritance of allozymes. Hereditas 85:219--226.
Sarvas, R. 1962. Investigations on the flowering and seed crop of
Pinus sylvestris. Commun. Inst. For. Fenn. 53: 4--198.
Savolainen, O. 1991. Background pollination in seed orchards. In
Proc. Meet. of the Nordic Group for Tree Breeding in Sweden. Ed.
D. Lindgren. Report No. 10, Department of Forest Genetics and
Plant Physiology, Swedish University of Agricultural Sciences,
Umeå, Sweden, pp 6--13.
Sweet, G.B., B.D. Dickson, B.D. Donaldson and H. Litchwark. 1992.
Controlled pollination without isolation----a new approach to the
management of radiata pine seed orchards. Silvae Genet. 41:95--99.
Wang, X., D. Lindgren, A.E. Szmidt and R. Yazdani. 1991. Pollen
migration into a seed orchard of Pinus sylvestris and the methods of
its estimation using allozyme markers. Scand. J. For. Res. 6:379-386.
Wheeler, N.C. and S.J. Jech. 1985. Estimating supplemental mass
pollination (SMP) success electorphoretically. In Proc. 20th Can.
Tree Improve. Assoc., Part 2. Eds. F. Caron, A.G. Corriveau and
T.J.B. Boyle. Quebec City, Quebec, Canada, pp 111--120.
Yazdani, R., G. Hadders and A.E. Szmidt. 1986. Supplemental mass
pollination in a seed orchard of Pinus sylvestris L. investigated by
isozyme analyses. Scand. J. For. Res. 1:309--315.
Yazdani, R. and D. Lindgren. 1991. Variation of pollen contamination
in a Scots pine seed orchard. Silvae Genet. 40:243--246.