INTERSPECIFIC AND INTERGENERIC FERTILIZATION BARRIERS IN
BROMELIACEAE
E. Parton, I. Vervaeke, R. Deroose and M.P. De Proft
Laboratory of Plant Culture
Catholic University of Leuven
W. De Croylaan 42
B-3001 Heverlee
Belgium
[email protected]
Keywords: Bromeliaceae, pollen germination, pollentube growth, prefertilization barrier,
post-fertilization barrier
Abstract
Bromeliaceae are tropical ornamentals with a variety of colours and inflorescence shapes.
To remain the interest of the public, breeders create new cultivars using interspecific and
intergeneric hand-pollinations. In our breeding studies, 4 genera were used: Aechmea (2
species), Guzmania (3 species), Tillandsia (1 species) and Vriesea (3 species). In
preparation of crossing experiments, two important flower characteristics were studied:
pollen quality and stigma receptivity. Pollen germination of different flower stages was
determined by an in vitro germination test. All species, except two of the Guzmania
species, possessed viable pollen and germination was always highest at anthesis. Stigma
receptivity was studied by pollinating flowers before, at and after anthesis. Pollen
germination and pollentube growth on stigma were visualised with fluorescence
microscopy (aniline blue staining). All species showed optimal stigma receptivity and
ovule penetration at anthesis. Interspecific and intergeneric hand-pollinations were
performed with pollen and stigma at anthesis stage. Pollen germination on stigma,
pollentube growth in style and ovary and fertilization of ovules was evaluated using
fluorescence microscopy. Five species were self-compatible, making emasculation
necessary. Half of the interspecific and all of the intergeneric crosses were classified as
incongruent because no ovules were fertilised (prefertilization barriers) or no seedlings
developed (postfertilization barriers). Prefertilization barriers occurred in the style, mostly
the lower part of the style. Only in crosses with the self-incompatible Aechmea species as
mother or the extremely short-styled Tillandsia species as father, pollentube growth
stopped in the upper part of the style. To understand more about these style barriers, the
morphology of the pistil was studied. Styles of all species consisted of a hollow channel,
covered with protein-rich cells. These ‘feeder’-cells also appeared on the placenta and the
ovule funiculus. In this way, pollen tubes could be guided towards the micropyles.
Further, Ca-oxalate and starch grains were observed in the style.
1. Introduction
The monocot family of the Bromeliaceae comprises more or less 50 genera and
3000 species, native to Central and South America. Based on evolutionary relationships,
deduced from floral characteristics, the family is divided in three subfamilies:
Pitcairnioideae, Tillandsioideae and Bromelioideae (Benzing, 1980). Several bromeliad
genera are actually used as ornamentals. Their success can be explained by three
characteristics: a great colour and form variety, longevity of the inflorescence (one month
and more) and an easy adaptation to growing conditions indoor.
To maintain the public’s interest, it is important for the breeders to renew the
bromeliad assortment by making hybrids. However, not all desired combinations are
Proc. XX EUCARPIA Symp. on New Ornamentals
Eds. J. Van Huylenbroeck et al.
Acta Hort. 552, ISHS 2001
43
possible due to fertilization barriers. Until now, the existence and nature of interspecific
and intergeneric crossing barriers in Bromeliaceae has not been well understood.
Crossing barriers are the result of incompatibility and incongruity. Incompatibility
is the phenomena where the activity of S-alleles results in an intraspecific barrier on
stigma (sporophytic SI) or style (gametophytic SI) (Knox et al., 1986). Incongruity occurs
in interspecific crosses as a result of lack of genetic information in one partner necessary
to complete pre- and postfertilization processes in the other (Hogenboom, 1973).
Prefertilization barriers prevent sperm cells from fertilizing egg cells. This can be caused
by lack of pollen germination on stigma or by arrest or disorientation of pollentube
growth in style or ovary. Postfertilization barriers can cause premature degeneration of
the embryo and / or endosperm (Shivanna and Sawhney, 1997).
The aim of this study was to identify pre- and postfertilization barriers in
interspecific and intergeneric crosses between different bromeliad species by studying
pollen germination on the stigma, pollentube growth in the pistil and pollentube
penetration of the ovules.
2. Materials and methods
2.1. Plant material
In our breeding studies, 4 genera were used: Aechmea (2 species A1 and A2),
Guzmania (3 species G1, G2 and G3), Tillandsia (1 species T1) and Vriesea (3 species
V1, V2 and V3). The first genus belongs to the Bromelioideae subfamily. The
Tillandsioideae subfamily is represented by the other three genera. All plants were grown
under normal greenhouse conditions suitable for bromeliad production in Belgium:
average day/night temperature of 21/19.5°C, 65% relative humidity and photoperiod 16
hours with light intensity of minimum 200 W/m2 . Flowering was induced with ethephon
(Belchim) or ethylene gas.
2.2. In vitro pollen germination test
Before and after anthesis, flowers were collected in the greenhouse between 8.00
and 12.00 pm. They were brought in Eppendorf tubes filled with 1 ml of liquid pollen
germination medium (20% sucrose and 0,001% H3 BO3 in distilled water). Pollen were
released from the anthers by gently shaking of these tubes. The mixture of anthers and
liquid medium was poured onto solid germination medium (containing 0,5% agar), and
incubated for 24 hours at 21°C and 40 µE/m2 s. Pollen germination rates were determined
using a light microscope (OLYMPUS BX 40). Pollen were considered as viable when
pollen tubes were longer than the pollen diameter after 24 hours of incubation on the
germination medium.
2.3. Stigma receptivity
Flowers were pollinated before, at and after anthesis with compatible or congruent
pollen. Two to four days after pollination, depending on style length, flowers were fixed
in 70% ethanol. A minimum fixation period of 24 hours was needed before pistils were
washed for 15 minutes with tap water, softened with 4N NaOH for one hour and stained
overnight in decolorized aniline blue as described by Shivanna and Rangaswamy (1992).
Ovules were isolated from the ovary and put on a microscope slide together with style and
stigma. In a drop of water they were squashed gently under a coverslip and examined
with an Olympus fluorescence microscope (OLYMPUS BX 40) using blue light
excitation (455 nm). Pollen and pollentubes were located by the specific fluorescence of
callose on cell walls and plugs (Stone et al., 1984). The percentage of pollen germination
on stigma and pollen tube penetration in ovules was determined.
44
2.4. Hand-pollinations
Each of the nine species used, required a specific emasculation and pollination
technique. In case of non-germinative pollen (Guzmania 1 and 2), self-incompatibility
(Aechmea 1) or short anthers not reaching the stigma (Vriesea 1 and 2), anthers were not
removed before anthesis. Flowerbuds of Guzmania 3 could not be emasculated without
affecting stigma receptivity. For this reason, this species was not used as mother in our
breeding experiments. Anthers of Vriesea 3 were removed one day before anthesis. To
prevent the stigma from drying, the flowerbud was covered after emasculation. This extra
manipulation was not necessary for Aechmea 2 flowers emasculated one day before
anthesis. Anthers of Tillandsia 1 were removed at anthesis and the pollinated pistil was
covered to ensure a high humidity necessary for pollen germination on the stigma. Intra-,
interspecific and intergeneric hand-pollinations were performed in different seasons and
years by rubbing fresh or stored anthers (Parton et al., 1998) on the stigma of freshly
opened flowers (anthesis). A total of 60 to 300 flowers were pollinated in this way. One
sample of flowers was fixed after 2 to 5 days to examine pollen germination and pollen
tube growth with fluorescence microscopy (as described in 2.3). Pollen germination on
stigmas was determined by counting minimum 200 pollen. Pollentube growth in style
tissue was expressed as the ratio between pollentube length and style length. This gives a
value of 0,0 for a style without pollentubes and 1,0 for pollentubes traversing the entire
length of the style (Liedl et al., 1996). Finally, the fertilization% was determined by
counting ovules with a pollentube growing into the micropyle and evoking fertilization.
These three parameters (germination%, pollentube-style ratio and fertilization%) were
measured on 20 to 40 flowers. The percentage of the flowers that showed fertilized ovules
was called fertilization index (FI). Crossing combinations with an index of minimum 5%
were called compatible (self-crossings) or congruent (interspecific and intergeneric
crossings). Another sample of pollinated flowers was harvested after 4 months to
determine seed formation and germination. In this way, both pre- and postfertilization
barriers could be studied.
2.5. Style morphology
Flower buds of different species and flower stages were preserved in ethanol 70%.
Firstly, plant material was gradually dehydrated using a series of different ethanol,
butanol and xylol solutions. Paraffin slides were stained with fast green (cytoplasm),
safranine (nuclei, lignin, cutin and suberin) and haematoxyline (cell-wall and nuclei).
Evaluation was performed with a light microscope (OLYMPUS BX 40).
3. Results
3.1. Pollen germination in relation to flower development
G1 and G2 pollen did not germinate at or after anthesis (Table 1). Samples before
anthesis were also negative. Results are not shown because it was not possible to
determine the exact number of days before anthesis of the flowerbuds. The other six
species possess viable pollen with an in vitro germination ranging from 37% to 83%.
Pollen germination of A1 is the highest at anthesis. Anthers of A2, V3 and G3 can also be
used in hand-pollinations one day before anthesis. This range of good pollen quality is
even broader for V2 and T1, respectively 5 days before and 4 days after anthesis.
3.2. Stigma receptivity in relation to flower stage
Stigma receptivity at different flower stages was studied using viable, compatible/
congruent pollen. Stigmas of A1, A2, G2 and G3 showed the highest pollen germination
at anthesis (Table 2). V2 stigmas could maintain high pollen germination even 2 days
45
after anthesis. However, the number of pollinated flowers with fertilized ovules (FI)
dropped from 90 to 10% if pollination was done one day after anthesis. Pollen
germination percentage and fertilization index of V3 were sufficiently high from one day
before until one day after anthesis. The percentage of fertilized ovules was highest one
day before and at anthesis.
3.3. Interspecific and intergeneric crosses
Intraspecific pollinations were all compatible, except one. A1 was classified as
incompatible because only short pollentubes were formed on its stigma (Table 3). Pollen
tube abnormalities, such as high callose deposits, occurred on the stigma of V3, a self
compatible species (Fig. 2).
Half of the interspecific combinations was named congruent because pollen
germinated on the stigma, pollentubes grew the entire length of the style, ovules were
penetrated by the pollentubes (Fig. 1) and normal growing seedlings were obtained.
Swollen pollen tube tips were observed in the style of these congruent combinations. In
the case of V1xV2 and V2xV1, seed germination was at least 50%. The 4 incongruent
crosses showed prefertilization barriers at stigma (G2xG3) and lower style (V1xV3) level.
The low number of seeds formed in the A2xA1 cross, failed to germinate. V3xV2
fertilized ovules did not develop into seeds.
All intergeneric crosses within the Tillansioideae subfamily were incongruent.
21% showed inhibited pollentube growth on the stigma (G1xV1, G2xV3 and T1xG3).
Another prefertilization barrier was present in the style of 50% of the crosses. Pollentubes
stopped growing in the upper half of the style in V1xT1, G2xT1, T1xV2 and T1xV3.
Pollentubes reached the lower part of the style in V1xG3, V2xT1 and V3xT1. Another 4
combinations (29%) showed postfertilization barriers between Vriesea and Guzmania
species. In all cases, ovules were fertilized but no seeds were formed. Four different
pollen tube abnormalities, such as curling tubes, high callose deposits, tubes with swollen
tip or with branched character, were observed.
Intergeneric pollinations were also performed between species of different
subfamilies (Tillandsioideae and Bromelioideae). Out of 22 combinations only one
seemed congruent (A2xG3). However, seedling and inflorescence appearance did not
differ from the mother species. Moreover, RAPD analysis showed no presence of father
DNA. The other 21 cross combinations were classified as incongruent. In 27% of the
incongruent combinations, pollentube growth stopped on the stigma of G1, G2 and T1
when pollen of A1 and A2 were applied. A style barrier was present in 55% of the
crosses. Except in the A1 style, pollentube growth stopped in the lower part of the style.
Postfertilization barriers were present in 14% of the crosses. In case of A2xV1 and
V2xA2, fertilized ovules failed to develop into seeds. Seeds of A2xV2 did not germinate.
In this group of cross combinations many pollen tube abnormalities occurred in the style.
The most common were callose deposits and tubes with swollen tips.
3.4. Style morphology
All 9 bromeliad species possessed a hollow style, divided in three compartments
that were mostly fused into a single canal (Fig. 3). The inner wall of this style canal is
covered with papil-like cells that are colored dark blue after a haematoxyline staining.
Those cells were also present in the ovary where they covered the placenta and its
outgrowths on which the ovules were nested. Ca-oxalate crystals were present in large
cells along the entire length of the styles and the ovary wall. Finally, starch grains could
be recognized in the style tissue of V2 and T1.
4. Discussion
Flowers of seven bromeliad species used in our breeding experiments possesed the
46
highest stigma receptivity and pollen germination at anthesis, evaluating a period of 5
days before up to 5 days after anthesis. V3 can also be pollinated 1 day before anthesis
because pollen germination on stigma and fertilization of ovules was comparable to that
at anthesis. Pollen of T1 can be used up to 4 days after anthesis. In this situation, the
arrangement of the corolla as a tube protecting the anthers from outside conditions, can
explain this behavior. Furthermore, unlike for the other species, the corolla of T1 showed
no sign of aging until 4 to 5 days after anthesis. For our 9 bromeliad species, aging of the
petals was an indication for loosing pollen viability. Anthers of young flower buds only
possessed viable pollen in the case of V2 (V1 was not tested).
Fertilization barriers were present in half of the interspecific combinations and all
intergeneric combinations (except A2xG3, but seedlings lacked hybrid character). Crosses
between species of different subfamilies showed more pollentube abnormalities than
crosses within one subfamily. One species, A1, was self-incompatible.
The first level where a prefertilization barrier can occur is the stigma. Due to
insufficient or uncontrolled hydration, lack of specific pollen germination factors
(calcium or boron) (Bednarska, 1991; Gartel, 1974) or different pH (Ganeshaian and
Shaanker, 1988) pollen germination can be inhibited. In our breeding experiment, pollen
germination was never absent. However, on stigmas of G1, G2 and T1 only short
pollentubes were formed and pollen germination was lower than on other stigmas. For
Guzmania this can be explained by an overall bad flower quality (also non viable pollen).
To apply pollen on a T1 stigma, bracts, sepals and petals had to be removed. The naked
pistil was covered to create a humid atmosphere around the pistil, but to avoid fungal
growth this was removed after one day. This may have caused an early dehydration of
stigma and style and explain our results. Pollentube inhibition on the stigma of these three
species was therefore not the result of a barrier between two parents. In general we can
state that the stigmas of the nine bromeliad species are not very selective. They are
probably wet stigmas (both types occur in Bromeliaceae, Heslop-Harrison and Shivanna,
1977) that secrete a nutrient-rich exudate that promotes pollen hydration in an
indiscriminate manner (Wilhelmi and Preuss, 1997), even for pollen belonging to a
different subfamily (Tillandsioideae and Bromelioideae).
Stylar prefertilization barriers were most common in our breeding experiment.
Growing pollentubes utilize stylar nutrients (Labarka and Loewus, 1973). Pollen tube
growth can be arrested in the style due to lack of suitable nutrients in the style or lack of
suitable enzymes in the pollen tubes (Shivanna and Sawhney, 1997).
A difference in the style length of the two parents can also result in a stylar barrier
(Gopinathan et al., 1986; Potts et al., 1987). The inability of pollentubes of the shortstyled parent to reach the ovary of the long-styled parent appears to be due to the intrinsic
inability of pollen tubes to grow beyond the length of their own pistils. In this study, this
phenomenon was illustrated by T1 with a significantly shorter style than the other species
(0.6 cm for T1 in comparison to 4.3 cm for V1, 4.7 cm for V2, 5.7 cm for V3, 4.5 cm for
G1, 4.7 cm for G2 and 4.0 cm for G3). T1 pollentubes stopped growing in the upper half
of most styles, while the barrier in the other crosses was mostly present in the lower half.
This behavior was not present in Aechmea styles, although they also have a short style
(1.4 cm for A1 and 3.0 cm for A2). It is possible that only large differences between style
lengths resulted in early pollentube arrest in the style.
Cross combinations in which A1 is the mother species, showed an arrested
pollentube growth in the upper half of the style. This can be due to its self-incompatibility
system. In interspecific crosses involving one or both self-incompatible parents, pollen
tube inhibition seems to be the result of active recognition of the pollen mediated through
S-gene action (Shivanna and Johri, 1985). The incompatibility system of A1 is probably
of the gametophytic type because short pollentubes were formed. As this is only an
indication, genetic studies would be necessary to proof the gametophytic incompatibility
of A1.
Finally, prefertilization barriers can also occur in the ovary, due to absence of
chemotropic factors in the micropyle (Shivanna and Sawhney, 1997). In the crosses with
47
problems at the seed level, many pollentubes that entered the ovary, failed to grow
towards an ovule. Only a few succeeded. In this way, postfertilization barriers (ovule
fertilization but no seed development, no seed germination) seem to go hand in hand with
a barrier at the ovary level (pollen tubes in the ovary, but not penetrating the micropyles)
in our breeding experiment. Postfertilization barriers could be overcome by embryo
rescue techniques, but this was not yet performed in this study.
The style of the nine bromeliad species in this study was hollow. The stylar canal
through which pollentubes grow, was covered with papil-like cells, densely colored with
haematoxyline. This may be due to its protein-rich contents. The same type of cells were
present in the ovary. They guided the pollentubes towards the micropyle of the ovules.
Acknowledgements
The first author received a PhD scholarship from the Flemish Institute to promote
Scientific and Technological Research in the Industry, Brussels. Parts of the study were
supported by the Ministry of Middle Class and Agriculture of Belgium, Committee of
Research and Development. The authors thank Reginald Deroose bvba for supplying
plant material and N. Plovie and C. De Brouwer for cooperation in this research.
References
Bednarska E., 1991, Calcium uptake from the stigma by the germinating pollen in
Primula officinalis L. and Ruscus aculeatus L., Sex plant reprod, 4 : 36-38
Benzing D.H., 1980, The biology of the bromeliads, Mad River Press Eureka, California,
13-19
Ganeshaian K.N., and Shaanker R.U., 1988, Regulation of seed number and female
incitation of male competition by a pH dependent proteinacious inhibitor of pollen
grain germination in Leucaena leucocephala, Oecologia, 75 : 110-113
Gartel W., 1974, Micronutrients: their significance in vein nutrition with special regard to
boron deficiency and toxicity, Weinberg and Keller, 21 : 435-508
Gopinathan M.C., Babu C.R., and Shivanna K.R., 1986, Interspecific hybridization
between rice bean (Vigna umbelatta) and its wild relative (V. minima): fertility-sterility
relationships, Euphytica, 35: 1017-1022
Heslop-Harrison Y., and Shivanna K.R., 1977, The receptive surface of the Angiosperm
stigma, Ann Bot, 41: 1233-1258
Hogenboom N.G., 1973, A model for incongruity in intimate partner relationships,
Euphytica, 22: 219-233
Knox R.R., Williams E.G., and Dumas C., 1986, Pollen, pistil and reproductive function
in crop plants, Plant breeding reviews, 4: 9-79
Labarka C., and Loewus F., 1973, The nutritional role of pistil exudate in pollentube wall
formation in Lilium longiflorum. Production and utilization of exudate from stigma
and stylar canal, Plant physiology, 52: 87-92
Liedl B.E., McCormick S., and Mutschler M.A., 1996, Unilateral incongruity in crosses
involving Lycopersicon pennellii and L. esculentum is distinct from selfincompatibility in expression, timing and location, Sex Plant Reprod, 9 : 299-308
Parton E., Deroose R., and De Proft M.P., 1998, Cryostorage of Aechmea fasciata pollen,
Cryo-letters, 19: 355-360
Potts B.M., Potts W.C., and Cauvin B., 1987, Inbreeding and interspecific hybridization
in Eucalyptus gunnii, Silvae Genetica, 30 : 194-199
Shivanna K.R., and Johri B.M., 1985, The Angiosperm pollen: structure and function,
New Delhi, Wiley Eastern
Shivanna K.R., and Rangaswamy N.S., 1992, Pollen biology: a laboratory manual, Berlin,
Heidelberg, New York, Springer-Verlag
Shivanna K.R., and Sawhney V.K., 1997, Barriers to hybridization in Pollen
biotechnology for crop production and improvement, Cambridge University Press,
48
London, 261-272
Stone B.A., Evans N.A., Bonig I., and Clarke A.E., 1984, The application of Sirofluor, a
chemically defined fluorochrome from aniline blue, for histochemical detection of
callose, Protoplasma, 122: 191-195
Wilhelmi L.K., and Preuss D., 1997, Pollentube guidance in flowering plants, Plant
Physiology 113, 307-312
49
Table 1: In vitro pollen germination (%) of 7 bromeliad species, before and after anthesis
(A = day of anthesis, e.g. A-1=one day before anthesis). Average and standard error of 10
repetitions are shown, . =not done. For each species / cultivar, means were compared with
a Duncan ‘s Multiple Range Test (Pr ≤ 0.05) within a row. Significantly best values are
underlined.
A-5
A-4
A-3
A-2
A-1
A
*A1 0
0
0
0
5±11
37±20
A2
0
6±11
10±18 48±22
80±14
83±10
V2
56±10 57±10 65±15 64±10
41±23
62±12
V3
0
0
2±6
0
41±11
50±13
G1
.
.
.
.
.
0
G2
.
.
.
.
.
0
G3
0
0
12±19 23±20
63±23
70±24
T1
0
0
0
2±3
53±30
62±11
*: A=Aechmea, G=Guzmania, V=Vriesea, T=Tillandsia
A+1
A+2
A+3
A+4
A+5
8±15
29±23
2±3
32±21
0
0
49±13
43±14
4±5
0
1±1
18±16
0
0
22±27
54±17
3±5
0
0
0
0
0
13±14
46±13
0±1
0
0
1±1
0
0
7±4
46±20
0
0
0
0
0
0
2±2
29±21
Table 2: Stigma receptivity for 6 different bromeliad species, before and after anthesis
(for example, A-1 = one day before anthesis). Pollen germination on stigma (1),
fertilization index (2) and % of fertilized ovules (3) is registered for a compatible /
congruent cross. Average and standard error of 20 repetitions are shown, . =not done. For
each species / cultivar, means were compared with a Duncan ‘s Multiple Range Test (Pr ≤
0.05) within a row. Significantly best values are underlined.
A-5 A-4 A-3
A-2
A-1
A
A+1
A+2
A+3
A+4
A+5
*A1A2
(1)
(2)
(3)
0
0
0
0
0
0
3±8
0
0
4±7
0
0
9±20
5
4±0
72±29
5
1
0±1
0
0
1±2
0
0
0
0
0
0
0
0
0
0
0
A2A2
(1)
(2)
(3)
0
0
0
0
0
0
2±5
0
0
24±33
30
7±4
45±34
60
25±14
90±9
95
36±20
10±17
5
13±0
0±2
0
0
0
0
0
0
0
0
0
0
0
V2V2
(1)
(2)
(3)
0
0
0
0
0
0
7±8
0
0
0
0
0
12±11
0
0
66±18
90
45±24
75±12
10
68
60±23
0
0
.
.
.
.
.
.
.
.
.
V3V3
(1)
(2)
(3)
0
0
0
0
0
0
0
0
0
2±6
0
0
52±35
53
61±29
55±21
60
51±29
59±22
75
26±27
0
0
0
0
0
0
0
0
0
0
0
0
G2V2
(1)
(2)
(3)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
38±32
16
1±1
40±29
7
1
8±14
0
0
2±5
0
0
0
0
0
0
0
0
G3G3
(1)
(2)
(3)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
43±25
94
37±20
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
*: A=Aechmea, G=Guzmania, V=Vriesea, T=Tillandsia
50
Table 3: Pre- and post-fertilization results for intra-, interspecific and intergeneric crosses
between 9 different bromeliad species. ‘Barrier’ refers to the level at which a fertilization
barrier occurred, such as stigma=short pollen tubes on stigma, style=pollen tube
inhibition in style, seed=ovule fertilization but no development of germinative seed. Prefertilization results: PG%=pollen germination% on stigma, PTL/SL=pollentube
length/stylar length, FI=fertilization index (% of pollinated flowers with at least one
fertilized ovule), F%=ovule fertilization % per flower, A=Pollentube abnormalities:
numbers in table refer to photographs in figure 1. Post-fertilization results: SI=seed index
(% of pollinated flowers with at least one seed), NS=number of seeds per flower,
SG%=seed germination %. .=not measured.
Intergeneric, same subfamily
Stigma
Stigma
Style
Seed
Seed
Stigma
Stigma
Stigma
Stigma
Stigma
Style
Style
Style
Style
Style
Seed
Seed
Seed
Seed
Intergeneric, different subfamily
Interspecific
Intraspecific
barrier
Stigma
Stigma
Stigma
Stigma
Stigma
Stigma
Style
Style
Style
Style
Style
Style
mother
*A2
V1
V2
V3
T1
A1
A1
V1
V2
V2
G2
V1
A2
V3
G1
G2
T1
T1
T1
V1
V1
V2
V3
G2
V2
V3
G2
G2
A2
G1
G1
G2
G2
T1
T1
A1
A1
A1
A1
A1
A2
father
A2
V1
V2
V3
T1
A1
A2
V2
V1
V3
G3
V3
A1
V2
V1
V3
V2
V3
G3
G3
T1
T1
T1
T1
G3
G3
V1
V2
G3
A1
A2
A1
A2
A1
A2
V1
V2
V3
G3
T1
V3
PG%
88±8
62±18
55±24
43±18
45±20
56±23
74±22
68±24
59±25
60±21
9±11
49±32
63±23
64±28
26±36
10±13
18±22
3±5
12±17
60±27
62±31
69±14
67±26
37±41
62±22
33±25
30±25
35±32
70±14
20±24
25±22
45±24
39±28
49±20
19±19
50±23
47±23
42±22
65±19
72±16
57±19
PTL/SL
1.0±0.0
0.6±0.2
1.0±0.0
0.9±0.1
0.6±0.4
0.0±0.0
0.8±0.3
0.6±0.2
0.8±0.3
0.9±0.1
0
0.7±0.3
1.0±0.1
1.0±0.0
0
0
0.0±0.1
0.0±0.1
0.0
0.5±0.3
0.3±0.2
0.5±0.2
0.6±0.3
0.1±0.2
0.7±0.3
0.5±0.5
0.6±0.5
0.3±0.4
0.9±0.2
0
0
0.0
0.0
0.0
0.0
0.1±0.1
0.2±0.2
0.2±0.2
0.1±0.0
0.1±0.1
0.6±0.3
FI
91±5
17±23
67±18
75±21
29±13
0
30±20
.
59±30
40±57
0
0
5±5
24
0
0
0
0
0
0
0
0
0
0
10
20
20
15±7
15±10
0
0
0
0
0
0
0
0
0
0
0
0
F%
42±17
18±0
16±15
55±28
10±5
0
3±1
.
13±10
20±15
0
0
1±0
1±1
0
0
0
0
0
0
0
0
0
0
5±5
7±7
2±0
3±1
3±3
0
0
0
0
0
0
0
0
0
0
0
0
A
2,3
3
3
2,3
3
2
1,3
1
1
3,4
2,3
2
3
2
3,5
3
2
SI
77±13
1
63±46
65
.
0
53±7
13
5
19
0
0
8±6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0
NS
93±60
35±52
168±95
216±77
.
0
4±3
102±105
86±44
126±60
0
0
3±1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
51
SG%
79±29
.
15±16
94±3
.
0
61±38
.
.
25±15
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
100
0
0
0
0
0
0
0
0
0
0
0
0
Style
Style
Style
Style
Style
Style
Seed
Seed
Seed
A2
V1
V1
V2
V3
V3
A2
A2
V2
T1
A1
A2
A1
A1
A2
V1
V2
A2
81±12
56±22
70±18
69±16
60±9
55±15
65±19
76±13
72±18
0.5±0.1
0.8±0.1
0.7±0.2
0.6±0.2
0.8±0.3
1.0±0.0
0.7±0.3
0.8±0.3
0.8±0.2
0
0
0
0
6
0
13±15
2±4
2±3
*: A=Aechmea, G=Guzmania, V=Vriesea, T=Tillandsia
52
0
0
0
0
1±0
0
1±0
1±0
36±0
2
1,5
1,5
1,3
2,5
1
3
2,3
1,2,4
0
0
0
0
0
0
0
1±1
0
0
0
0
0
0
0
0
1±0
0
0
0
0
0
0
0
0
0
0
1
2
3
Figure 1: Pollen germination on stigma (20x10) (1), pollentube growth in style (4x10) (2)
and pollentube growth in ovary (4x10) (3) in a congruent cross combination.
1
4
2
5
3
300 µm
6
Figure 2: Pollentube abnormalities in bromeliad crossings: (1) winding pollentube,
V2xA1; (20x10) (2) callose deposits in pollentubes on stigma, V3xV3(20x10); (3)
swollen pollen tube tip, V3xV2 (20x10); (4) branched pollen tube, V2xA2 (20x10); (5)
pollen tube turning back, A1xG3 (20x10); (6) disorientated pollen tube in ovary, A2xA1
(10x10).
1
2
3
Figure 3: Stylar morphology of Aechmea 2. (1) Ca-oxalate crystals in stylar tissue, 40x10;
feeder cells in ovary (2) and style (3), 20x10.
53