HORTSCIENCE 29(12):14871492. 1994.

Flowering Behavior and Pollination

Requirements in Climbing Cacti with
Fruit Crop Potential
Julia Weiss1, Avinoam Nerd2, and Yosef Mizrahi1, 2
Ben-Gurion University of the Negev, P.O.Box 1025, Beer Sheva 84110,
Israel
Additional index words. Hylocereus, Selenicereus megalanthus, pitaya, compatibility
system, fruit set, seeds, pollen viability
Abstract. The reproductive biology of the climbing cacti Hylocereus polyrhizus (Weber)
Britt. & Rose, H. undatus (Haworth) Britt. & Rose, H. costaricensis (Weber) Britt. &
Rose, and Selenicereus megalanthus (Schum. ex Vaupel) Moran (syn. Mediocactus
megalanthus) was studied with the aim of cultivating the cacti in Israel as fruit crops.
Flowering in Hylocereus spp. occurred in two to three waves during the summer,
whereas in S. megalanthus, flowering was concentrated at the end of autumn. Flowers
of all species opened 1 to 1.5 hours before sunset and closed 6 hours after sunrise. In
the Hylocereus spp., H. polyrhizus and H. costaricensis were self-unfruitful, and crosspollination with other species led to high fruit set (100%). Hylocereus undatus was selffruitful, setting fruit with self-pollen. Cross-pollination between the clones of S.
megalanthus led to a high fruit set and each clone was self-fruitful. In contrast to H.
undatus, S. megalanthus clones could set fruit without pollen vector involvement,
although the set was slightly lower than with hand pollination. Pollen source influenced
fruit weight. In the self-fruitful species of Hylocereus, fruit obtained by hand crosspollination with other Hylocereus spp. were significantly heavier than fruit obtained by
hand self-pollination. The largest fruit in each of the Hylocereus spp. were obtained by
specific cross-combinations within the group. Fruit of S. megalanthus had a lower
weight than fruit of the Hylocereus spp. Flowers of all species were visited by day-active
honeybees only. Fruit set and fruit weight with open pollination was lower than with
hand pollination in Hylocereus spp. Since stigma receptiv- ity and pollen germinability
stayed high during anthesis, the low pollination effectivity has to be related to other
factors, such as the short bee visits and the absence of specific adaptation by the bees to
the flower. In S. megalanthus, fruit set and fruit weight with open pollination were
similar to values obtained with hand pollination. This similarity is probably related to
the fact that pollen transfer in open pollination is achieved by bee visits and direct
transfer of pollen to the stigma, which occurs via physical contact between anthers and
stigma during flower closing.

Climbing epiphytic cacti, members of the

genera Hylocereus and Selenicereus, are
found throughout Central America, the West
Indies, and western Mexico (Britton and
Rose, 1963). These cacti are branched
climbers that attach their slender stems to
trees or rocks via adven- titious roots, which
grow along the stems. These roots also serve
to absorb water (Gibson and Nobel, 1986).
Several species of Hylocereus and
Selenicereus, known as strawberry pear
(Morton, 1987), produce medium to large
fruit, traditionally consumed by the local
popula- tion (Cacioppo, 1990). In Latin
America, the plants and their fruit are known
as pitaya or pitahaya (Donaj Ortiz
Hernandez, 1993). The Hylocereus fruit are
characterized by a red peel with large scales
and a white or red pulp, while those of
Selenicereus have a knobby yellow peel with

1 HORTS CIENCE, VOL. 29(12), DECEMBER 1994

spines and a white flesh (Britton

Rose, 1963; Kimnach, 1984;

and

Received for publication 9 Mar. 1994. Accepted

for publication 18 Aug. 1994. The cost of
publishing this paper was defrayed in part by the
payment of page charges. Under postal
regulations, this paper therefore must be
hereby marked advertisement solely to indicate
this fact.
1
Department
of
Life
Sciences.
2
The Institutes for Applied
Research.

cies has begun recently, H. undatus in

Nicara- gua and Vietnam and S.
megalanthus in Co- lombia (Barbeau, 1990;
Cacioppo, 1990).
Studies are rare on the reproductive biology of the climbing cacti. The flowers are
hermaphroditic, nocturnal, and strongly
scented. Each flower is large, up to 30 cm
long, with a long flower tube that fans out
into a funnel-shaped corolla with thin, white
petals (Fig. 1). A nectary chamber is located
at the bottom of the flower tube (Britton and
Rose,
1963). Flowers exhibit characteristics of
those pollinated by night-active bats and
large hawk- moths (Grant and Grant, 1979;
Haber, 1983). Three species, H. ocamponis,
H. undatus, and S. megalanthus, are
probably autogamous and the visits of
pollinators increase fruit set (Cacioppo,
1990; Cullmann et al., 1987; Kimnach,
1967).
Preliminary studies in Israel show that
climbing cacti are susceptible to high solar
irradiation and become bleached when
grown in the open; however, plants develop
well and flower when they are shaded
(Raveh et al.,
1993). In their studies, flowers did not set
fruit, suggesting fertilization failure, possibly
due to a lack of pollen vector activity, or to
com- patibility problems. The objective of the
present study was to examine flowering
behavior and pollination requirements in
selected species of Hylocereus and in S.
megalanthus, with the aim of proposing
agromanagements that will enable them to
be grown commercially.
Materials and Methods
Plant material and study site. Clones of
four species were examined (Table1). For
two of them, H. undatus and S. megalanthus,
two genotypes in each one, differing in
origin and morphological traits, produced
abundant flow- ers during the study period.
The investigation
Seaton, 1991). The juicy, sweet fruit contain
numerous edible, tiny black seeds, and the
fruit are much appreciated when served chilled
(Morton, 1987). Cultivation of selected spe-

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994

1

Fig. 1. Flower of Hylocereus polyrhizus.

2 HORTS CIENCE, VOL. 29(12), DECEMBER 1994

was therefore extended to two clones within

these species. Cuttings for propagation were
obtained from the Huntington Botanical Garden in California, from a private garden in

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994

2

CROP PRODUCTION
Table 1. Plant material, fruit characteristics, and source for climbing cacti.
Species
Hylocereus polyrhizus (Weber)
Britton & Rose

Local clone
designation

H. undatus (Haworth)
Britton & Rose

A
B

H. costaricensis (Weber)
Britton & Rose

Fruit characteristics

Source

Fruit oblong; peel dark-red with

large scales; pulp violet-red

Huntington Botanical Garden, Calif.,

collection no. 15885 HBG

Fruit oblong; peel light-red with

large scales; pulp white
Fruit oblong; peel light-red with
large scales; pulp white

Private garden, Israel

Fruit round; peel dark-red with

large scales; pulp violet-red

Huntington Botanical Garden, Calif.,

collection no. 49639 HBG

Fruit oblong; peel yellow with

tubercles and spines; pulp white
Fruit oblong; peel yellow with
tubercles and spines; pulp white

Plantation, Colombia

Selenicereus megalanthus
(Schum. ex Vaupel) Moran

Syn. Mediocactus megalanthus

Israel, from commercial plantations in Colombia, and from the wild in Ecuador. The
study was carried out in 1992 and 1993 in
Beer Sheva, which is situated in the northern
Negev Desert of Israel. The experiments
were per- formed on 3- to 4-year-old plants
in green- houses under 50% shade, which
were covered with polyethylene during the
coldest months (November to March). Six
plants of each clone were grown in two
greenhouses. The cuttings were planted in
sandy soil at a spacing of 1.5
1.5 m, and a trellis system 1.5 m in height
was installed for supporting the plants. Water
containing 20N20P20K fertilizer was applied by dripper at a concentration of each
70
ppm N (NH3 NO3 + KNO3), P2 O5 (H3PO4 ),
and K2O (KNO 3) every 2 days at a weekly
amount of 5 liters/plant during the hot season
(May to October) and 3 liters/plant per week
during the cold, wet season. At the beginning
of the study, a honeybee hive was placed
inside one of the greenhouses. The
greenhouse was open on two sides to allow
other insects to enter. The average
minimum/maximum temperatures in the
coldest month (January) were 18/22C and in
the hottest month (August) 16/35C in both
greenhouses.
Phenology of flowering, flower behavior,
and insect visits. During the flowering
season, all plants were monitored daily for
new flow- ers. The flowers of three plants
for each clone were observed for 5 days to
determine the phases of anthesis. The
presence of nectar was determined by
probing the flowers with pasteur pipettes at
half-hour intervals.
Breeding system and pollination type.
Three treatments were applied: 1) Hand selfpollina- tionpollen from the same flower
was ap- plied to the stigma after flowers
opened in the evening. To prevent open
pollination, flowers were kept bagged except
during hand pollina- tion. 2) Hand crosspollinationthe stigma was enclosed in a
small cloth bag in the after- noon before
pollen was shed to prevent self- pollination;
pollen gathered from a different concurrently
flowering clone was applied the same
evening. 3) Covered, undisturbed flow- ers
flowers were covered with bags throughout anthesis. Fruit set was recorded in all
treatments, and fruit weight and seed count
were determined. Full skin color change was

Pollen germinability and viability. Pollen

germinability at anthesis and during the postanthesis stages was studied by sampling
flow- ers at 12-h intervals, beginning at
flower open- ing and ending 72 h later. Six
flowers were sampled for each clone. A
sample of 5 mg of pollen was spread on a
glass slide in a thin film of 1% agar
containing 30% sucrose, 100 ppm H3BO4 ,
100 ppm Ca(NO3)2 4H2O, 100 ppm
MgSO4, and 100 ppm KNO3. The slide was
closed in a petri dish at 25C for 24 h, after
which the pollen sample was diluted with 1
ml of 30% sucrose solution; then,
germination was examined under a
microscope. Grains were scored as
germinated when tube length exceeded the
diameter of the grain itself. Pol- len viability
was assessed by a fluorochromatic assay
(Heslop-Harrison et al., 1984).
Effective pollination period. The time interval following anthesis during which pollination will result in a commercial crop is

an indicator for fruit harvest.

Private garden, Israel

Wild, Ecuador

defined as the effective pollination period,

and it is a function of ovule longevity and
the rate of pollen tube growth (Dennis,
1986). The effective pollination period was
studied in two species, H. costaricensis and
H. polyrhizus. Stigmas of flowers were
bagged before anthe- sis and then crosspollinated by hand with pollen of H. undatus
(clone A) at flower open- ing (evening) and
again in the morning and evening of the
following days. [Previous un- published
studies have shown that H. undatus (clone
A) was highly compatible with the two other
species]. Fresh pollen gathered during the
morning or evening, at which times in vitro
pollen germinability was high, was used for
pollination. Seven to 11 flowers were pollinated in each species at a time and fruit set
and fruit weight were determined.
Effect of open pollination on fruit set and
fruit weight. Two treatments were applied in
1993 to flowers on plants growing in the

CROP PRODUCTION
Fig. 2. Flowering periods for various species of
Hylocereus and Selenicereus during the 1992
93 flowering seasons (upper line1992, lower

line1993). The thickened sections of the lines indicate periods of concentrated flowering; at least
four flowers per species were monitored each day.

greenhouse in which the bee hive had been

placed: 1) hand cross-pollination and 2) open
pollination, in which flowers were exposed
to insects. In all treatments, fruit set was
evalu- ated and fruit were picked at full color
change for fruit weight determination.
Hylocereus spp. were studied for two
consecutive years, whereas S. megalanthus
clones were studied only in the second year
due to an inadequate flower set in the first
year.
Results
Flowering phenology, flower behavior,
and insect visits. The number of flowers per
plant averaged 9 7 in Hylocereus spp. and 4
1 in S. megalanthus and ranged between 5
and 25 flowers in the Hylocereus group and
between
3 and 8 in S. megalanthus in the second year
of the study. Flowering in the Hylocereus
spp. occurred in two to three waves, mainly
during
whereas flowering in S. megalanthus was
con- centrated in autumn (October to
November) (Fig. 2).
In all the species, the phases of anthesis
followed a very similar pattern so that a
single representative scheme is presented
(Fig. 3). Flowers stayed open for one night.
They began to open 1 to 1.5 h before sunset,
were fully open by sunset, began to close
1.5 h after sunrise, and were fully closed
by midday.
Anthers dehisced 0.5 to 1 h before flower
opening. The upper part of the anthers in the
Hylocereus spp. were at least 2 cm below the
stigma, whereas in S. megalanthus, they were
at the same height as the stigma, touching it
when the flowers closed. The first nectar secretion was evident when the flowers were
fully open. The flowers were visited by the
honeybee, Apis mellifera L. The bees
foraged for pollen and contacted the stigma
frequently

Fig 3. Anthesis phases and timing of insect visits to flowers in various species of Hylocereus and
Selenicereus.
Table 2. Percentage of fruit set for climbing cacti in relation to the breeding system used.
Fruit set (%)
Species
Hylocereus polyrhizus
H. costaricensis
H. undatus clone A
H. undatus clone B
Selenicereus megalanthus clone A
S. megalanthus clone B
z

pollination
100z
100z
100z
100z
100y
100y

pollination
0
0
50.0
79.6
100
100

flowers
----0
0
73.0
60.0

Crossings with pollen from a different, concurrently flowering Hylocereus clone.

Crossing with pollen from the other clone of S. megalanthus.

Table 3. Effect of pollen source on fruit weight (grams) in Hylocereus spp.

Female
Male
H. polyrhizus
H. undatus clone A
H. undatus clone B
H. costaricensis
Tukeys LSD 0.05, n = 8

H. polyrhizus
--476
410
539
155

H. undatus
clone A
580
182
287
633
90

H. undatus
clone B
567
337
301
495
102

H. costaricensis
384
298
329
--46

Table 4. Effect of pollination method on fruit weight (grams) in Selenicereus megalanthus.

during their visits, which were confined to the

morning, and they did not appear to reach the
nectar at the base of the flower tube.
Breeding system and pollination type. Synchronization of flowering determined the type
of crossing combinations (Table 2). The
Hylocereus clones were crossed with each
other, as were the S. megalanthus clones.
Among the Hylocereus spp., H. polyrhizus
and H. costaricensis were self-unfruitful, and
cross- pollination with other species led to
high fruit set (100%). Hylocereus undatus
was self-fruit- ful, setting fruit with selfpollen, but self- fruitfulness was partial since
a higher fruit set was obtained by crosspollination with the other species (50% to
80% vs. 100%).
Cross-pollination between the S.
megalanthus clones led to a high fruit set and
each clone was self-fruitful. In contrast to H.
undatus, they also set fruit without the involvement of a pollen vector (automatic selfpollination).
Pollen source influenced fruit weight
(Table
3). In the self-fruitful species of Hylocereus,
fruit obtained by hand cross-pollination with
other Hylocereus spp. were significantly
heavier than fruit obtained by hand selfpolli- nation. The largest fruit in each of
the Hylocereus spp. were obtained by
specific

Pollination method
Self
Cross
Covered, undisturbed flowers
Tukeys LSD 0.05, n = 8

S. megalanthus
clone A
77
98
38
44

cross-combinations within the group.

Fruit of S. megalanthus weighed less than
fruit of the Hylocereus spp. (Table 4). Hand
cross-pollination and hand self-pollination in
S. megalanthus produced fruit with similar
fruit weights, which were about twice that
with automatic self-pollination.
Regression analysis showed that the seed
count per fruit explained 75% to 91% of the
variability in fruit weight (Fig. 4). Seed
count was much higher in fruit of the
Hylocereus group than in those of S.
megalanthus: maxi- mum values ranged
between 3918 and 6505 in the Hylocereus
vs. 26 and 446 in S. megalanthus. Ovules
were counted in two species: H. undatus
(clone A) with many seeds and S.
megalanthus (clone A) with relatively few.
The average number of ovules was 7184
613 and 1969 176, respectively.
Pollen germinability and viability. Pollen
germinability in all species was highest
when flowers started to open (Fig. 5).
In H.

S. megalanthus
clone B
86
108
51
50

polyrhizus, a sharp decline was recorded the

next morning when flowers began to close,
while in H. undatus (clone A) and H.
costaricensis, pollen germinability declined
only the following evening, by which time
the flowers had already closed. Pollen of S.
megalanthus sampled when flowers started
to open or 12 h later had very low
germinability
(<4%).
Changing
the
concentration of the medium components or
using other germina- tion methods, such as
hanging-drop, standing- drop, and mass
culture techniques (Calzoni et al., 1979), did
not affect the germination of S. megalanthus
pollen, which remained low (data not
shown). Viability of fresh pollen was high,
90% to 94%, in the Hylocereus spp. and
about one-third of it in S. megalanthus (Fig.
6).
Effective pollination period. Pollination
of H. polyrhizus and H. costaricensis with
fresh pollen of H. undatus (clone A) led to
high fruit set and heavy fruit (Table 5).
Pollination was conducted during the period
between the time

CROP PRODUCTION
Fig. 5. In vitro germination of pollen of
Hylocereus undatus clone A, H. polyrhizus,
and H. costaricensis sampled from flower
opening over 60 h. Values are the average of
counts of 100 pollen grains from six randomly
selected flowers.

Fig. 4. Relationship between fruit weight and seed count in Hylocereus spp. and in Selenicereus
megalanthus.
Fruit included in the analysis were from open- and hand-pollination treatments. Calculations are
based on measurements of 17 to 25 fruit for the various species (P = 0.01).

7 HORTS CIENCE, VOL. 29(12), DECEMBER 1994

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994

7

CROP PRODUCTION
of flower opening in the evening and the evening of the following day. Later pollina- tion
produced significantly lower fruit weights, and no fruit set when pollination was con- ducted
2 days after flower opening.
Effect of open pollination on fruit set and fruit weight. Fruit set with open pollination was
higher in S. megalanthus (80%) than in Hylocereus spp. (12.5% to 50%) (Table 6). Hand
cross-pollination resulted in higher fruit set and heavier fruit than were achieved by open
pollination in the Hylocereus spp. In S. megalanthus, fruit weight in open pollination was
similar to fruit weight after hand pollina- tion. Open-pollinated flowers were visited only by
honeybees, indicating that pollen transfer was related to honeybee activity.
Discussion
Studies on the pollination requirements of orchard crops usually examine the compatibility between cultivars within species. In the case of the climbing cacti under discussion, some
are not cultivated at all, and for the commercially cultivated species, no defined cultivars are
available (Barbeau, 1990). Since preliminary studies had shown us that cross- pollination
between the investigated species was successful, the study set out to examine the possibility
of using particular species as pollinators for other species. For the commer- cial species, H.
undatus and S. megalanthus, the availability of clones differing in origin and in some
morphological traits allowed us to extend the study within these species.
Two species, H. undatus and S. megalanthus, were self-fruitful. For both spe- cies, the
involvement of pollen vectors was important for fruit production: H. undatus set fruit only
when hand self-pollinated, and S. megalanthus, which set fruit by self-pollina- tion without a
pollen vector, produced heavier fruit when hand self-pollinated. The ability of flowers to set
fruit without the involvement of a pollen vector was associated with flower morphology:
upper anthers touched the stigma lobes in S. megalanthus during flower closing, but anthers

8 HORTS CIENCE, VOL. 29(12), DECEMBER 1994

were separated spatially from the stigma in

the Hylocereus spp.
Studies of the breeding system lead to
some practical conclusions with regard to the
desired orchard design: a) S. megalanthus
can be planted alone (one or two clones)
because of the high fruit set obtained by selfand cross- pollination between the clones. b)
Fruit pro- duction in H. undatus would
benefit from planting this species with
pollinizers of two other Hylocereus spp., H.
polyrhizus, and H. costaricensis, which
induced both high fruit set and heavy fruit in
H. undatus. The flower- ing period of the
two proposed pollinizers was parallel to that
of H. undatus. c) In the self- unfruitful
species, H. polyrhizus and H. costaricensis,
any of the Hylocereus spp. may be planted as
pollinizers in relation to their influence on
fruit set and fruit weight and synchronous
flowering.
As is the case for other fruit crops
produc- ing abundant seeds, such as kiwifruit
[Actinidia deliciosa (A. Chev.) C.F. Liang et
R. Ferguson

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994

8

Fig 6. Pollen viability of Hylocereus spp. and Selenicereus megalanthus clones. Pollen viability was
determined by a fluorochromatic assay. Pollen samples were taken shortly after flower opening.
Values are the average of counts of 100 pollen grains from six randomly selected flowers.

var. deliciosa] (Hopping and Jerram, 1979)

or feijoa (Feijoa sellowiana Berg.)
(Patterson,
1990), fruit weight in the cacti was highly
correlated with seed count. Thus, attention
must be paid to obtain efficient pollination in
the plantation. The maximum percentage of
ovules that developed into seeds was 23% in
S. megalanthus (clone A), whereas a
maximum
of 90% of the ovules set seed in H. undatus
(clone A). The low seed set in S.
megalanthus
(clone
A) can be related to the low viability
and germinability of its pollen. Higher seed
set and, therefore, higher fruit weight could
prob-be obtained by pollinating with highly
ably
viable pollen of other clones of S. megalanthus
or other Selenicereus spp.
Although the flowers of both genera displayed the characteristics of those pollinated
by night-active pollinators (Grant and Grant,
1979; Haber, 1983), the only pollen visitor to
Hylocereus and Selenicereus was a day-active
honeybee. The honeybee is a polytropic pollinator and forages on many nocturnal cactus
flowers, including the saguaro [Carnegiea
gigantea (Engelm.) Britt. & Rose], the organ
pipe cactus [Stenocereus thurberi (Engelm.)
Britt. & Rose] (Alcorn et al., 1962;
McGregor et al., 1962; Schmidt and
Buchmann, 1986), and the apple cactus
[Cereus peruvianus (L.) Miller] (Weiss et al.,
1993). The low rate of fruit set and the low
fruit weight recorded for open pollination in
the Hylocereus spp. indi- cates that the bees

9 HORTS CIENCE, VOL. 29(12), DECEMBER 1994

the brevity of the visits and the lack of

specific adaptation to the flower; for
example, bees did not reach the nectar
accumulated at the base of the flower tube.
In S. megalanthus, fruit set and fruit weight
with open pollination were similar to those
obtained in hand pollination. This feature
may be related to the fact that

pollen transfer in open pollination is

achieved by bee visits and by direct transfer
of pollen to the stigma, which occurs via
physical contact between anthers and stigma
during flower closing.
Further studies will have to be made on
improving the efficiency of the bees as
pollinizers for the Hylocereus spp. with
regard to the interaction between the flowers
of the climbing cacti and bee behavior in the
planta- tion. The problems associated with
pollination by bees do not occur when
pollination is car- ried out artificially. Hand
pollination ensured high fruit set and heavy
fruit when conducted during the entire
period of anthesis. Hand pollination can thus
be carried out conve- niently during the
morning hours, if economi- cally feasible.
Other potential pollinators of the climbing
cacti exist in Israel; for example, the large
carpenter bee (Xylocopa pubescens L.)
potentially could be introduced successfully inside the shaded greenhouse.
The results of this study clarify various
aspects of the reproductive biology of
promis- ing climbing cacti, including the
flower phe- nology, the compatibility
system, and the pol- lination type, which will
enable the grower to define the
agromanagement required for pro- ductive
plantations. More studies concerned with
clones are needed to determine if solid
species plantations, consisting only of different clones of one species, may be
established.
Literature Cited
Alcorn, S.M., S.E. MacGregor, and G. Olin. 1962.
Pollination requirements of the organpipe cactus. Cactus Succulent J. (Amer.) 34:134139.
Barbeau, G. 1990. La pitahaya rouge, un nouveau
fruit exotique. Fruit 45:141147.

Table 5. Effect of timing of pollination on percentage of fruit set and fruit weight in two Hylocereus spp.
Hylocereus undatus clone A was the pollen source.
Time after
flower opening
Fruit set
Fruit wt
Speciesz
(h)
(%)
(g)
H. polyrhizus
0
100
392 aby
12
100
396 a
24
100
304 bc
36
38
228 c
48
0
--H. costaricensis
0
100
316 a
12 species under the conditions
100
were not effective pollina- tors for these
of our study. 369
Theabees
24 before flower closing.
100Since pol- len germinability
241 b
visited the flowers only in the morning
36
20
154 bc
and female receptivity were
48
0
---

ity has to be related to other factors, possibly

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994

9

Species
H. polyrhizus
In each species 7 to 11 flowers were pollinated at a
time.
y
Mean separation by Fishers PLSD test, P 0.05.
z

Table 6. Effect of pollination type on fruit set and

fruit weight in climbing cacti. For the
Hylocereus spp., fruit set and fruit weight are
the mean of 2 years of observation.

10 HORTS CIENCE, VOL. 29(12), DECEMBER 1994

H. undatus
clone A
H. costaricensis

Pollination
type
Hand cross
Open

No. flowers
37
24

Fruit set
(% of total)
100
20.8

Hand cross
Open
Hand cross
Open

30
44
30
16

100
43.2
100
12.5

588
120
291
113

15
13
19
12

Open

10

80

121

Fruit wt (g)
599 25
108 11

Selenicereus megalanthus

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994

10

CROP PRODUCTION
Britton, N.L. and J.N. Rose. 1963. The Cactaceae.
vol. I and II. Dover Publications, New York.
Cacioppo, O. 1990. Pitaya: Una de las mejores
frutas producida por Colombia. Informativo
Agroeconomico Febrero:1519.
Calzoni, G.L., A. Speranza, and N. Bagni. 1979. In
vitro germination of apple pollen. Scientia Hort.
10:4955.
Cullmann, W., E. Goetz, and G. Groener. 1987. The
encyclopedia of cacti. Timber Press, Portland,
Ore. p. 340.
Dennis, F.G., Jr. 1986. Apple, p. 144. In: S.P.
Monselise (ed.). Handbook of fruit set and development. CRS Press, Boca Raton, Fla.
Donaj Ortiz Hernandez, Y. 1993. Hylocereus spp:
Plantas natives de potential ornamental y fruticola
en Oaxaca. Investigacin Hoy, Febrero:1314.
Gibson, A.C. and P.S. Nobel. 1986. The cactus
primer. Harvard Univ. Press, Cambridge, Mass.
Grant, V. and K.A. Grant. 1979. The pollination
spectrum in the southwestern American cactus
flora. Plant Syst. Evol. 133:2937.

1492

Haber, W.A. 1983. Hylocereus costaricensis

(pitahaya silvestre, wild pitahaya), p. 252253.
In: D.H. Janzen (ed.). Costa Rican natural history. The Univ. of Chicago Press, Chicago.
Heslop-Harrison, J., Y. Heslop-Harrison, and K.R.
Shivanna. 1984. The evaluation of pollen quality,
and a further appraisal of the fluorochromatic
(FCR)
test
procedure. Theor. Appl. Genet. 67:367375.
Hopping, M.E. and E.M. Jerram. 1979. Pollination
of kiwifruit (Actinidia chinensis Planch.):
Stigmastyle structure and pollen tube growth.
N.Z. J. Bot. 17:233240.
Kimnach, M. 1967. Hylocereus calcaratus. Cactus
Succulent J. (Amer.) 39:102105.
Kimnach, M. 1984. Hylocereus escuintlensis, a new
species from Guatemala. Cactus Succulent J.
(Amer.) 56:177180.
McGregor, S.E., S.M. Alcorn, and G. Olin. 1962.
Pollination and pollinating agents of the saguaro.
Ecology 43:259267.
Morton, J.F. 1987. Cactaceae, strawberry pear, p.
347348. In: C.F. Dowling, Jr. (ed.). Fruit of

warm climates. Morton J.F., Miami.

Patterson, K.J. 1990. Effects of pollination on
fruitset, size, and quality of the feijoja (Acca
sellowiana (Berg.) Burret). N.Z. J. Crop
Hort.
Sci. 18:127131.
Raveh, E., J. Weiss, A. Nerd, and Y. Mizrahi.
1993. Pitayas (genus Hylocereus): A new
fruit for the Negev Desert of Israel, p. 491
crop
495. In: J. Janick and J.E. Simon (eds.). New
crops. Wiley, New York.
Schmidt, J.O. and S.L. Buchmann. 1986. Floral
biology of the saguaro (Cereus giganteus) I.
Pollen harvest by Apis mellifera. Oecologia
(Berlin) 69:491498.
Seaton, L. 1991. Pitaya: A classic case of
novelty
and lack of consumer awareness. Fresh Produce J. May:10.
Weiss, J., A. Nerd, and Y. Mizrahi. 1993. Development of Cereus peruvianus (apple cactus)
aasnew crop for the Negev Desert of Israel, p.
486491. In: J. Janick and J.E. Simon (eds.).
New crops. Wiley, New York.

HORTSCIENCE, VOL. 29(12), D ECEMBER 1994