|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
USE AND EFFECTS OF DIPLOID POLLENIZERS FOR TRIPLOID WATERMELON
[Citrullus lan2atus (Thunberg) Matsumura and Nakai] PRODUCTION
JOSHUA HERBERT FREEMAN
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
O 2007 Joshua Herbert Freeman
To my wife Lindsey for the sacrifices she has made for me to be here, and to my parents for their
Foremost, I thank my advisor, Dr. Stephen Olson for giving me the opportunity to pursue
this degree and for his personal and professional guidance. I also thank the members of my
committee for the time they have dedicated and the knowledge and guidance they have provided.
Special thanks go to Dr. Eric Simonne and Dr. Bill Stall for their professional guidance and
willingness to help me with any aspect of my career. I greatly appreciate the field and office
staff at NFREC who helped throughout my dissertation. Special thanks go to Dr. Powell Smith
for planting a seed in my mind that has led me to where I am today.
None of this would have been possible if not for the sacrifices that my wife Lindsey has
made, putting her career on hold in order to support our family during this time. She has been
the strong foundation that has kept me sane and kept me on point. I will be forever grateful to
my parents, Linda and Herb Freeman for their relentless support and constant prayer. I would
also like to thank my grandparents, Lurie Goff and Blake Freeman for their support of me in
anything I do. I thank God every day for blessing me with the mind and body that have allowed
me to do this and for the family that has supported me. I also thank God for putting me where I
need to be when I need to be there. I hope I have made my wife, my family and my God proud.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .............. ...............4.....
LIST OF TABLES ........._..... ...............7....__........
LIST OF FIGURES .............. ...............8.....
AB S TRAC T ......_ ................. ............_........9
1 INTRODUCTION ................. ...............11.......... ......
2 REVEW OF THE LITERATURE. ............. ...... ...............13..
Review of Diploid Watermelon ............_...... ...............13...
Taxonomy and Bot any .............. ............... 13....
Watermelon Production Practices .............. ...............15....
Watermelon Production Statistics .............. ...............18....
Review of Triploid Watermelon ................. ...............19................
Background Information .............. ...............19....
Seedless Watermelon Market Share ................. ...............20................
Triploid Seedless Watermelon Production ................. ...............21.......__.....
Pollinator Activity and Preference .............. ...............26....
Pollen Effects............... ...............27
Com petition ................. ...............30........ ......
3 CHARACTERISTICS OF DIPLOID POLLENIZERS FOR USE IN TRIPLOID
WATERMELON PRODUCTION ............_...... ...............32...
Introducti on ............. ...... ._ ...............32...
Materials and Methods .............. ...............32....
Re sults............. ...... ._ ...............34...
Discussion ............. ...... ._ ...............36...
4 POLLEN VIABILITY OF DIPLOID WATERMELON POLLENIZER CULTIVARS.......40
Introducti on ..........._.. ....._. ._ ...............40.....
Materials and Methods .............. ...............40....
Re sults.............._ ...... ...............42.....
Discussion .............._ ....... ...............42.....
5 DIPLOID WATERMELON POLLENIZER CULTIVARS EXHIBIT VARYING
DEGREES OF PERFORMANCE WITH RESPECT TO TRIPLOID WATERMELON
Y IELD .............. ...............46....
Introducti on ............. ..... .. ...............46...
M materials and M ethods .............. ...............46....
Re sults............. ..... ...............49...
Discussion ............. ..... ...............50...
6 COMPETITIVE AFFECT OF IN-ROW DIPLOID WATERMELON POLLENIZERS
ON TRIPLOID WATERMELON YIELD ....._ .....___ .........__ ............5
Introducti on ............. ..... .._ ...............57....
M materials and M ethods .............. ...............57....
Re sults............. ...... ...............58....
Discussion ............. ...... ._ ...............59....
7 VARIABILITY IN WATERMELON FLOWER ATTRACTIVENESS TO INSECT
POLLINATORS .............. ...............63....
Introducti on ................. ...............63.................
M materials and M ethods .............. ...............63....
Re sults ................ ...............64.................
Discussion ................. ...............66.................
8 VARIABILITY IN POLLEN PRODUCTION BY DIPLOID WATERMELON
POLLENIZ ERS ............. ...... ._ ...............72...
Introducti on ........._.__....... ._ __ ...............72....
M materials and M ethods .............. ...............72....
Re sults........._.__....... .__ ...............73....
Discussion ........._.__....... .__ ...............74....
LIST OF REFERENCE S ........._.__....... .__. ...............77...
BIOGRAPHICAL SKETCH .............. ...............84....
LIST OF TABLES
3-1 Analysis of variance for pollenizer fruit weight and fruit per plant at Quincy and
Live Oak FL during 2005. ............. ...............37.....
3-2 Interaction effect of location and watermelon pollenizer cultivars on fruit per plant at
Quincy and Live Oak FL during 2005. ............. ...............37.....
3-3 Main effects for pollenizer fruit weights combined over experiments conducted in
Quincy and Live Oak, FL, during 2005. ............. ...............38.....
4-1 Analysis of variance for pollen viability of watermelon pollenizer cultivars tested
during the Spring and Fall of 2006 at Quincy, FL. .............. ...............44....
4-2 Influence of diploid watermelon pollenizer cultivar on pollen viability at Quincy, FL
during the Spring and Fall of 2006. ........._ ....__ ......_ .....__ ............45
5-1 Pollenizer cultivar effect on 'Tri-X 313' yield at Blackville, SC during 2005..................54
5-2 Pollenizer cultivar effect on Supercrisp' watermelon yield and average fruit weight
at Blackville, SC, Citra, FL, and Quincy, FL during 2006. ............... ...................5
5-3 Pollenizer cultivar effect on soluble solids concentration of seedless watermelons at
Blackville, SC during 2005 and Citra, FL, Quincy, FL, and Blackville, SC during
5-4 Pollenizer cultivar effect on hollowheart disorder in Supercrisp' watermelon at
Quincy, FL, and Blackville, SC combined with Citra, FL during 2006. Means are to
be compared within the same column ................. ...............56...............
5-5 Seed sources for various pollenizer cultivars used during 2005 and 2006. .......................56
6-1 Influence of pollenizer cultivar and spacing on triploid watermelon yield during
8-1 Pollen production by four diploid watermelon pollenizer cultivars at Quincy, FL
during the Fall of 2006. ........... ..... .._ ...............76..
LIST OF FIGURES
3-1 Pollenizer staminate flower counts combined over Quincy and Live Oak, FL, during
2005. Means not followed by the same letter are significantly different at (P < 0.05)
by Duncan's Multiple Range Test. ............. ...............39.....
5-1 Field diagram for pollenizer experiments at Blackville, S.C., Citra, FL., and Quincy,
FL. in 2005 and 2006. Columns represent individual rows. The same design was
used for replications three and four. ............. ...............52.....
5-2 Individual three-row plot design for pollenizer experiments at Blackville, S.C., Citra,
FL, and Quincy, FL in 2005 and 2006. Plot shown is using a pollenizer
recommended to be planted at a 1:3 pollenizer to seedless ratio ................... ...............53
7-1 Influence of cultivar on pollinator visitation to staminate watermelon flowers at
Quincy, FL, Spring 2006. Visitation means are to be compared within sample date.
Means not followed by the same letter are significantly different at P I 0.05. .................68
7-2 Interaction effect of cultivar and time on pollinator visitation to staminate
watermelon flowers at Quincy, FL, on 23 May, 2006. Visitation means are to be
compared within sample time. Means not followed by the same letter are
signify cantly different at P I 0.05. ............. ...............69.....
7-3 Influence of cultivar on pollinator visitation to staminate watermelon flowers at
Quincy, FL, Fall 2006. Visitation means are to be compared within sample date.
Means not followed by the same letter are significantly different at P I 0.05. .................70
7-4 Interaction effect of cultivar and time on pollinator visitation to staminate
watermelon flowers at Quincy, FL, on 23 Sept., 2006. Visitation means are to be
compared within sample time. Means not followed by the same letter are
significantly different at P I 0.05. ............. ...............71.....
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
USE AND EFFECTS OF DIPLOID POLLENIZERS FOR TRIPLOID WATERMELON
[Citrullus lanatus (Thunberg) Matsumura and Nakai] PRODUCTION
Joshua Herbert Freeman
Chair: Stephen M. Olson
Major: Horticultural Science
The use of in-row pollenizers has become popular because it allows a greater number of
triploid plants to be planted per hectare. Multiple in-row pollenizer cultivars are available but it
is unclear if any exhibit superior performance with respect to triploid watermelon yield and if so,
what varietal characteristics contribute to a pollenizer' s performance.
Field trials were conducted during the Spring and Fall of 2005 and 2006 at various
locations in FL. and S.C. to determine the performance of various pollenizers and also what
contributed to their success. Of the in-row pollenizers that were tested, Sidekick' resulted in the
greatest triploid watermelon yields. Yields from plots pollenized by 'Patron', 'SP-1', 'Jenny',
and 'Mickylee' were not significantly lower. The use of 'Companion' resulted in significantly
lower yields than plots pollenized 'Sidekick', 'Patron', 'SP-1', and 'Jenny'.
Pollen viability can vary between cultivars of a plant species and the pollen viability of
four pollenizer cultivars was evaluated. Poor pollen viability from a pollenizer could result in
increased fruit abortions and lower yield. No significant differences in pollen viability were
detected between pollenizer cultivars tested. The production of staminate flowers is a crucial
factor for pollenizers as there must be adequate pollen flow during triploid watermelon fruit set.
Flower production for 'Companion', 'Jenny', 'Mickylee', and 'SP-1' was recorded and
'Companion' produced as many flowers as 'Jenny' or 'Mickylee' throughout most of the season.
The flowering period of 'Companion' does appear to be shorter than other cultivars. Pollinator
preference was examined between 'Companion', 'Mickylee', and 'SP-1' and 'Companion' was
found to be the least attractive of the three. The lack of pollinator visitation to 'Companion'
appears to be the greatest contributor to its poor performance.
'Mickylee' is an attractive option to use as a pollenizer because of its lower seed costs.
However, 'Mickylee' has a growth habit that is more vigorous than most other pollenizers.
Studies comparing 'SP-1' and 'Mickylee' showed that the 'Mickylee' competed more with
associated triploid plants and reduced yield. Factors affecting pollenizer performance the most
appear to be pollinator preference, staminate flower production, and competitive effect on
Seedless watermelons now account for 78% of the watermelons sold in the United States.
Triploid watermelon plants produce seedless fruits however; they do not produce sufficient
viable pollen which is necessary for fruit set. To achieve optimal seedless watermelon yields,
rows of diploid watermelon cultivars are planted in the same field as the triploids. These diploid
cultivars account for 20 to 33% of the watermelon plants per hectare. With low demand for
seeded watermelons, it has become less desirable to maintain this much hectarage in diploid
plants. Diploid cultivars (pollenizers) designed to be planted within the row of triploid
watermelons are now available. The use of these in-row pollenizers requires no dedicated space
for the pollenizer plants, thus allowing for an increase in triploid plants per hectare. When in-
row pollenizers first became available there were few options, now there are 11 commercially
available cultivars. Using in-row pollenizers is a new concept and most of the in-row pollenizer
cultivars have been on the market less than two years. The performance of in-row pollenizer
cultivars has not been demonstrated, however earlier studies have reported differences in triploid
watermelon yields due to pollenizer cultivar. It is suspected that the similar results will be
observed with in-row pollenizers.
Available in-row pollenizers have diverse phenotypes and variable plant characteristics
and it is unclear how these may effect the cultivar' s performance. Characteristics such as flower
production, pollen production, pollen viability, attractiveness to pollinators, and plant vigor
could have a marked effect on how triploid plants in association with the pollenizers will yield.
Pollenizer seed/plant costs are highly variable and it is unclear whether the more expensive
pollenizers will provide for greater seedless watermelon yields.
The aim of this body of work is to determine how well in-row pollenizers perform and how
their varietal characteristics influence their performance.
REVIEW OF THE LITERATURE
Review of Diploid Watermelon
Taxonomy and Botany
The watermelon [Citrullus lan2atus (Thunberg) Matsumura & Nakai] was originally
described in Africa as Mamordica lanata by Carl Thunberg in 1794 (Hara, 1969). Schrader
proposed Citrullus vulgaris in 1836 which was commonly used until lan2atus was accepted as the
oldest specific epithet. Hara revealed the original description in 1969 that was published by
Matsumura and Nakai in 1920 and watermelon has since been designated as Citrullus lan2atus
(Thunberg) Matsumura & Nakai (Hara, 1969).
Watermelon is a dicotyledonous angiosperm in the order Violales and family
Cucurbitaceae (Kartesz, 2006). The center of origin for the species is believed to be in southern
Africa, where it was first described by Thunberg. David Livingstone reported seeing
watermelons growing wild in the Kalahari Desert in 1857 (Wehner, 2006). Watermelon is a
warm-season, herbaceous vining annual with angular stems that may reach 9 m long (Wehner,
2006). There are also genotypes that exhibit a bush or dwarf growth habit (due to shorter inter-
nodes) that have shorter stems and are less branched. The main stem or crown of the plant
develops from the seedling stem and may have few to many lateral branches develop depending
on genotype. Most watermelon genotypes have pinnately lobed leaves with three to five pairs of
lobes and are born singly on the stem (McCreight, 1996). Cultivars have been developed that
have a nearly entire leaf which is due to a gene mutation (Wehner, 2006). Branched tendrils are
produced at each node. Watermelon is a monoecious plant with imperfect flowers that are born
singly at the leaf axil (Wehner, 2006). Watermelon flowers are small in comparison with other
cucurbits such as those in the genera Cucurbita and are born on short peduncles. When plants
reach sexual maturity, staminate (male) flowers are produced first and for a period of time before
pistillate (female) flowers are produced. Generally, staminate flowers have three anthers and
pistillate flowers have three stigmatal lobes. Pistillate flowers have inferior ovaries which vary
in size and shape depending on genotype. Staminate flowers are only open for one day and
pistillate flowers are only receptive for one day. The corollas, stamens and pistils of watermelon
flowers are all yellow.
Watermelons are entomophilous plants that are primarily pollinated in cultivation by
honeybee (Apis mellifera L.) (Maynard, 2001). Watermelon entomophily is a symbiotic
relationship between the insect and plant as plants receive pollination and bees obtain pollen
(which is the only protein source for honeybees) and nectar. Pollination is facilitated by the
production of a sticky substance on the surface of the pollen grains called pollenkitt. Pollenkitt
enables adherence to floral visitors in order to disseminate pollen and is primarily seen in
entomophilous angiosperms (Dobson et al., 1996). Pollenkitt is comprised mostly of lipids
which also play a role in attraction of pollinators (Dobson, 1988; Pernal and Currie, 2002). Poor
pollination may result in fruit abortion or misshapen fruit which is unmarketable (Rubatzky and
Yamaguchi, 1997; Stanghellini et al., 1997; Walters, 2005).
Watermelon fruit is a specialized berry with thick skin (rind) known as a pepo (Rubatzky
and Yamaguchi, 1997). Watermelon rind thickness is dependent on genotype and can be very
thin to thick depending on the intended use of the watermelon. The edible portion of watermelon
is the endocarp, although rind and seeds can and are consumed in certain geographic areas.
Watermelon fruit are highly variable in size, shape, and color, and although all three are
genetically determined, size can be altered through production practices. Watermelons may
range from cylindrical to round and may be from two to over 90 kg at maturity (Wehner, 2006).
The largest watermelon recorded was 119 kg (Wehner, 2006). Most cultivars that are currently
grown commercially in the U.S. range from 5.5 to 9.0 kg. Consumer preference has changed
gradually over time and large fruit (> 11.3 kg) that were once preferred are now virtually
unacceptable in most of the U.S. market. Average maturity for watermelon fruit is 80 days from
transplant and 110 days from seed. Watermelon seeds are variable in size, shape, and color but
are usually flattened, teardrop shaped and brown to black in color (Robinson and Decker-
Walters, 1997). Optimum temperatures for seed germination range form 29.4-32.30C (Olson et
al., 2006). Watermelon plants are intolerant to frost and grow best when the average temperature
is above 21.10C (Rubatzky andYamaguchi, 1997).
Watermelon Production Practices
Growing watermelons and producing optimal yields is dependent on many factors
including fertilization, irrigation, soil management, and other cultural practices (Hochmuth et al.,
2001a). The pH of the soil can greatly influence nutrient availability and crop growth.
Watermelon plants can tolerate a range of pH values from 5.5 to 7.5 but 6.0 to 6.5 is optimum
(Hochmuth et al., 2001a). Various forms of lime can be used to correct acidic soil pH, which is
common in the highly weathered soils of the southeast. Calcium is important in the growth and
development of watermelon because it is necessary to maintain the structural integrity of cell
walls. Inadequate calcium in developing watermelon fruit can lead to a disorder known as
blossom-end rot (BER) (Taylor and Locascio, 2004). Blossom-end rot occurs when tissue at the
blossom-end of the fruit begins to collapse. This collapse is visualized as necrotic areas at the
blossom-end which allow pathogens to enter the fruit and eventually leads to decay. Although
BER is a calcium deficiency, factors such as inadequate or inconsistent irrigation, elemental
competition, high salinity, high temperature and high or low transpiration may induce the
problem (Taylor and Locacsio, 2004).
Other elements such as nitrogen, phosphorus, potassium, magnesium, and micro-nutrients
are also necessary for adequate fruit growth and plant development (Hochmuth et al., 2001a).
The amount of these elements needed may vary by soil type and location. Soil testing should be
performed in order to understand what the crop requirements are at a given location. Nitrogen,
phosphorus, and potassium are generally the three most limiting nutrients and recommendations
for the maximum amounts of nutrient addition for watermelon production in Florida are 168N-
72.2P-140K kg/ha (Olson et al., 2006).
Irrigation amounts and frequencies required by the crop can vary depending on soil type,
its water holding capacity, and environmental conditions (Rubatzky and Yamaguchi, 1997). A
minimum of 93 cm of water per hectare is required to successfully grow watermelons (Rubatzky
and Yamaguchi, 1997). Irrigation frequencies are important because some portions of growth
and development require more water and are more sensitive to moisture stress than others.
Water deficit during flowering and early fruit growth can have a greater negative impact on
watermelon yield than water stress during other growth periods (Erdem and Yuksel, 2003).
Watermelon fruit growth and development is triggered by hormones produced by
developing seeds, which are dependent on pollen transfer from staminate to pistillate flowers
(Rubatzky and Yamaguchi, 1997). Pollination can be a yield limiting factor in watermelon
production and in many areas, pollination by feral honeybees is no longer reliable due to the
recent introduction of two parasitic mite species [Acarapis woodi (Rennie) and Varroa jacobsoni
(Oudemans)] (Delaplane and Mayer, 2000). These mites have reduced feral and domesticated
honeybee colonies by 90% and 30% respectively (Stanghellini et al., 1998; Ambrose, 1997;
Harbo and Hoopinger, 1997). The reduction of feral honeybee populations necessitates the
introduction of domesticated honeybees in large hectarage, high density plantings. Commercial
pollinators are available for rent or purchase and honeybees are the most readily available.
Recommendations vary from one to five hives per hectare, which is dependent on hive strength
(Olson et al., 2006). Adlerz (1966) reported that in diploid watermelon, a minimum of eight
honeybee visits and 1000 pollen grains per pistillate flower were necessary for fruit set.
Stanghellini (1997) reported similar findings of a minimum six visits per flower in diploid
watermelon. Visitation rates lower than this resulted in significantly greater fruit abortion rates
Cultural practices used in watermelon production have changed over the years but the ones
having the most effect are the use of polyethylene mulch and drip irrigation. Traditionally
watermelons were grown on bare ground with either overhead or seep irrigation. The use of drip
irrigation decreases the amount of water used and increases the use efficiency (Smaj strla et al.,
2002). It also allows growers to target fertilizer applications directly to the plants through the
use of soluble fertilizer that is pumped through the drip tape (Clark et al., 2005). The use of
polyethylene mulch has many advantages including earliness, increased yields, increased
profitability, irrigation and fertilizer retention, weed suppression, and fruit protection (Lu et al.,
2003; Hochmuth et al., 2001b; Sanders et al., 1999; Lamont, 1993; Bryan, 1966). By covering
the ground with black polyethylene mulch, soil temperatures are raised which increases early
plant growth rate and thus early yield. Mulch has been shown to decrease harvest time in the
spring as much as a month (Bryan, 1966). The earliness of the crop may greatly influence the
profitability by allowing producers to enter the market before other producers or producing
regions. The use of polyethylene mulch has also allowed for the adoption of soil fumigants
which are essential for profitable commercial production in certain areas. The use of
greenhouse-grown containerized watermelon seedlings has also increased which is due mostly to
increased seed prices and the benefit of significantly greater early yields (Nettles, 1963).
Implementing polyethylene mulch and the use of transplants has increased the success of
producers in Florida, Georgia, and South Carolina as it allows them to enter the market before
the July 4th holiday. Generally, watermelon movement in the U.S. is greater and the price is
higher before this holiday (USDA, 2006). These three states accounted for 42% of the
watermelon hectarage in the U.S. in 2005 (USDA-NASS, 2006).
Watermelon plant spacing has changed over the years; most likely due to adoption of other
cultural practices that make production more efficient and a decrease in size preference by
consumers (Hochmuth et al., 2001a). Current recommendations for watermelon spacing range
from 1.1 to 2.7 m2 per plant (Daniello, 2003, Olson et. al, 2006). Recent research has shown that
greater yields per hectare may be achieved on polyethylene mulch with spacings as small as 1.0
m2 per plant (Goreta et al., 2005; Sanders et al., 1999). Plant spacing is also dependent on plant
architecture, as some genotypes have a more vigorous growth habit than others (Edelstein and
Nerson, 2002; Kultur et al., 2001; Reiners and Riggs, 1997).
Watermelon Production Statistics
Watermelons are grown throughout the world in tropical and subtropical climates. World
watermelon production has ranged from 81 billion kg in 2001 to 93.2 billion kg in 2004 (Arney
et al., 2006). The world's largest producer of watermelons is China which usually accounts for
over half of the world production (Arney et al., 2006). The U. S. generally ranks fourth in world
production and produces on average 1.8 billion kg (Arney et al., 2006).
Florida consistently ranks among the top three watermelon-producing states in the nation
in both hectares harvested and crop value and produces about 20% of the watermelons grown in
the United States (USDA, 2005). Over the last five years, Florida harvested on average ten
thousand hectares of watermelons annually worth an average of 72. 1 million dollars. In Florida,
watermelons account for 3.9 % of the cash receipts for vegetables and 0.91% of the cash receipts
for agriculture (USDA, 2002). Although watermelon hectarage in the U.S. has declined over the
last five years, total production has remained stable as a result of increased production per
hectare. One factor that has changed is the types of watermelons that are being produced.
Review of Triploid Watermelon
Triploid seedless watermelon was first described in the United States in 1951 based on
work that had been conducted in Japan since 1939 (Maynard and Elmstrom, 1992; Kihara,
1951). Seedless watermelons are produced by watermelon plants that are genetically triploid
(3n). Watermelon plants are naturally diploid (2n). Triploid plants are grown from triploid seed
which is produced by crossing a tetraploid (4n) female parent with a diploid male parent. The
female parent is produced by treating a diploid seedling with colchicine, a chemical that binds to
tubulin and inhibits the formation and activity of microtubules in plant cells. This inhibits the
separation of chromosomes during mitosis which doubles the number of chromosomes in the
plant and creates a tetraploid (Wehner, 2006). The triploid offspring are sterile which is why
fruits produced by these plants have no seeds. Fruit growth and enlargement in watermelons is
signaled by hormones produced by developing seeds. Since there are no developing seeds in
triploid fruit, these developmental signals are obtained from pollen tube growth and ovule
fertilization (Rhodes et al., 1997; Robinson and Decker-Walters, 1997; Maynard and Elmstrom,
1992). Ovules abort shortly after fertilization but may remain in the flesh as small rudimentary
white seeds (Maynard and Elmstrom, 1992; Kihara 1951). Though fruit growth and
development in triploids is signaled by pollen tube growth, triploid plants produce little if any
viable pollen (Rhodes et al., 1997; Robinson and Decker-Walters, 1997; Maynard and Elmstrom,
1992). A diploid watermelon cultivar must be planted in close proximity to the triploid to
provide sufficient viable pollen. This diploid plant is referred to as a pollenizer.
The first hybrid triploid watermelon cultivars produced by Kihara were finished in 1951
(Wehner, 2006). Triploid watermelon cultivars have been commercially available for nearly 35
years but interest from consumers and growers remained low until the late 20th century. There
was little interest in early triploid cultivars due to erratic and poor performance in the field and
high seed costs as compared to diploid cultivars (Maynard and Elmstrom, 1992). Seed for early
triploid cultivars were quoted at $135 per 1000 seed which was 900 times higher than hybrid
diploid cultivars at the time (Maynard and Elmstrom, 1992). Reluctance in adoption of triploids
may have also been due to the necessity of using transplants and the increased input costs
associated with them. A survey of over 1300 people conducted in 1992 indicated that while 74%
knew of seedless watermelons only 31% had ever purchased one (Wiemann, 1992). Marr and
Gast (1991) surveyed consumers and indicated that they were willing to pay 50% more for
seedless watermelons and that there was no differential preference between taste of seeded and
seedless. The authors suggested that the response seen was on appearance alone as participants
were shown cut fruit of both types. In 1990, Karst (1990) estimated that 5% of the U.S.
watermelon market was seedless but they had the potential to gain up to 50% market share.
Since this time, seedless watermelons have gained popularity in the marketplace and also with
watermelon growers. Cultivars that would be considered modern cultivars became commercially
available in the early 1990s and several are still considered industry standards today (Maynard
and Elmstrom, 1992).
Seedless Watermelon Market Share
Before 2002, watermelon market data was not separated into seedless and seeded
categories so it is difficult to locate accurate statistics on the production of either. To date, there
is still no separation of the production area grown in the U. S. but there are now reliable data on
seedless and seeded watermelons sold in the U.S. Seedless watermelons accounted for 78% of
the watermelons sold in the U. S. in 2006 which is up from 50% of the U. S. market in 2002
(USDA, 2006). The portion of Florida' s watermelon production that is triploid has increased
from 42% in 2002 to 79% in 2006 (USDA, 2006). There is some incentive for growers to
produce seedless watermelons because they typically receive 4.5 to 11.0 cents per kg premium
and it has become increasingly harder to market seeded watermelons since their market share has
decreased (USDA, 2006).
Triploid Seedless Watermelon Production
The cultural practices used in triploid seedless watermelon production are similar to
diploid production with respect to plant spacing, fertilization, and irrigation. Average maturity is
also similar between the two. Growth habit of triploids is similar to diploids but there are also
genotypes that exhibit compact growth and plant spacing may need to be altered to produce fruits
of desirable size. Triploid seed are more difficult to germinate than diploids and require precise
environmental conditions. While diploid cultivars may germinate in as low as 12.7 OC, triploid
seed will not germinate below 26.6 OC and optimum temperature is between 29.4 OC and 32.2 OF
(Hochmuth et al., 2001a). Seed coat adherence is also a problem with triploids and may
negatively affect seedling growth. To avoid this, seeds must be planted with the radicle end up
at 450 to 900 (Maynard and Elmstrom, 1992; Maynard, 1989). With these requirements and the
high cost of triploid seed, the use of transplants in spring triploid production is necessary as soil
temperatures are too cold for direct seeding.
Pollination is also a necessity in triploid plants and the introduction of domesticated
pollinators, such as honeybees or bumblebees, may be more important in triploid production
(Walters, 2005). Stanghellini (1997) and Alderz (1966) both reported that pistillate diploid
flowers required a minimum of six to eight honeybee visits for optimal fruit set and visitation
rates lower than this significantly increased fruit abortion. Walters (2005) conducted
experiments on triploid watermelons in which honeybee visitations were controlled in order to
determine the number of visits necessary for optimal fruit set. Research plots contained a 33%
pollenizer ratio in order to mimic commercial production. Findings suggested that between 16
and 24 honeybee visits were required to achieve maximum seedless watermelon fruit set. These
visitation numbers are two to four times higher than what is needed in diploid plants and it was
suggested that this is due to a dilution of viable diploid pollen with non-viable triploid pollen.
The most crucial difference between diploid and triploid production systems is the addition
of the diploid pollenizer in the triploid field. Kihara (195 1) suggested that one diploid plant
should provide enough pollen to achieve adequate fruit set in four to five triploid plants. By
these recommendations, 16 to 20% of the plants per hectare should be diploid. Until 2001, there
were no scientific data on which pollenizer ratio would maximize seedless watermelon yield.
Maynard and Elmstrom (1992) indicated that a pollenizer ratio of 33% had produced acceptable
seedless yields and other sources recommended ratios of 20 to 33% (Robinson and Decker-
Walters, 1997; Rubatzky and Yamaguchi, 1997). The method for introducing diploid
watermelons into the field at this time was to plant solid rows of diploid plants between rows of
triploid plants. The diploid cultivar must be different than the triploid in size, shape, or rind
pattern in order to facilitate efficient harvest. NeSmith and Duval (2001) used distance of a
triploid row from a pollenizer row to make inferences on pollenizer frequencies. 'Genesis'
triploid was used and 'Ferrari' was used as the pollenizer. Their results showed the greatest
seedless yields in rows 3.0 m from the pollenizer row with yields in rows farther away declining.
Yield estimates produced by NeSmith and Duval (2001) suggested that the greatest seedless
watermelon yield per hectare would be achieved with a 1:4 pollenizer to seedless ratio when 1.5
m between-row spacing is used. In this scenario, pistillate triploid flowers would never be more
than 3.8 m away from a staminate diploid flower.
Fiacchino and Walters (2003) conducted the same type of experiment but used isolated
fields with different pollenizer ratios. The plot design used by Fiacchino and Walters more
closely resembled a commercial watermelon field. Plots consisted of raised, plastic mulched
beds with 1.5 m between-row spacing. This experiment also used dedicated rows of pollenizers
at ratios of 11, 20, and 33%. In this study, multiple pollenizer cultivars were used to determine if
cultivar, as well as frequency, had an effect on seedless watermelon yield. 'Millionaire' triploid
was used and 'Crimson Sweet' and 'Fiesta' were used as pollenizers. Fiacchino and Walters
(2003) found plots containing a 33% pollenizer ratio did not have greater yield than those with
20%, but both 20 and 33% plots had greater seedless yields than the plots with an 11% pollenizer
ratio. Though there was no difference in yield between the 20 and 33% plots, a field with 20%
pollenizer ratio would have greater seedless watermelon yield on a per hectare basis due to a
higher number of triploid plants. These researchers also reported a significant difference in
seedless yield between the two pollenizer cultivars used, with plots pollenized by 'Crimson
Sweet' having greater yield. When 'Fiesta' was used as a pollenizer cultivar there was
significantly greater hollow heart disorder present in the seedless watermelons.
Previous to Fiacchino and Walters (2003), pollenizer choice was based on marketing
concerns and not how it affected the triploid crop. In this study, watermelon yields per hectare
were greatest in plots where pistillate triploid flowers never exceeded 3.8 m away from a
staminate diploid flower, which is in agreement with Nesmith and Duval (2001).
The consensus between Nesmith and Duval (2001) and Fiacchino and Walters (2003)
reinforced the non-scientific recommendations of Maynard and Elmstrom (1992) who suggested
that a 1:2 pollenizer to triploid ratio. The maximum distance between pistillate triploid and
staminate diploid flowers was never greater than 3.9 or 4.5 m because 2.75 and 3.0 m between-
row spacing's were used (Maynard and Elmstrom, 1989, 1992). This research was conducted
when seedless watermelons held less than 50% of the U. S. market so the production scheme of
using dedicated pollenizer rows allowed producers to be diversified in the marketplace. During
the early years of commercial seedless production, it may have been more economically
beneficial for growers to grow at a 1:1 pollenizer to triploid ratio as seedless melons held such a
small market share. With the growth in popularity of seedless watermelons and their increased
market share, it has become less desirable to grow seeded watermelons. Under previous triploid
production schemes, as much as a third of a grower' s hectarage needed to be in diploid
A new cultural management system has recently been developed that allows for an
increase in the number of triploid plants per hectare. New diploid cultivars have been developed
specifically for the role of pollenizer and these cultivars, commonly called special pollenizers,
are designed to be planted within the row of triploid plants without changing in-row spacing.
Special pollenizer cultivars became available in the early 2000s and were used on large hectarage
beginning in 2004. As there is no dedicated space for the pollenizer, triploid plants can be
planted at 100% stand. Common practices are to punch plant holes and transplant the field solid
with triploid seedlings then go back through the field and transplant pollenizers between triploid
seedlings at the appropriate density. Diploid cultivars produce flowers sooner than triploid
cultivars so transplanting the pollenizer several days later may more closely synchronize
blooming in the two types of plants (Freeman and Olson, 2007). This system of pollenizer
arrangement increases triploid plant numbers by 20 to 33% per hectare, thus increasing the
number of seedless fruits harvested per hectare. Most of these pollenizer cultivars are not
intended to be harvested which, allows for harvest of only seedless fruit. This can avoid the
confusion of having multiple types of harvestable watermelons in the field and the added labor
costs of multiple harvests. Though most special pollenizers were not intended to be harvested,
the following cultivars produce marketable fruit: 'Jenny', 'Minipol', 'Pinnacle', 'Polimore'.
Any small fruited diploid cultivar could be used as a pollenizer if it produces adequate staminate
flowers and pollen. Some producers may have a market for the seeded pollenizer fruits and may
be able to benefit economically by using one of these cultivars.
There are two different types of special pollenizer cultivars available; highly branched
plants with reduced foliage and thin vines or short inter-node bush-type plants. The thin-vine
types have foliage and vines that are smaller than standard watermelon plants by varying
degrees. These types also exhibit some degree of increased branching which increases the
number of terminals and therefore the number of male flowers produced. The reduction in vine
and foliage size is intended to reduce the negative effects that may occur when decreasing the
area per plant by introducing the pollenizer in-row. Thin-vine special pollenizers currently
available are: 'Increase' (Southwestern Seeds), 'Jenny' (Nunhems USA, Inc., Acampo, CA),
'Minipol' (Hazera Seeds, Inc., Coconut Creek, FL), 'Patron' (Zeraim Gedera Seed Co, Ltd.,
Palm Desert, CA), 'Pinnacle' (Southwestern Vegetable Seed, LLC., Casa Grande, AZ),
'Polimore' (Hazera Seeds, Inc., Coconut Creek, FL), 'Sidekick' (Harris Moran Seed Co.,
Modesto, CA), 'SP-1' (Syngenta Seeds, Inc., Boise, ID), 'SP-4' (Syngenta Seeds, Inc., Boise,
ID). The bush-type pollenizers have a compact growth habit with short internodes and a
branching pattern more similar to standard watermelons. These cultivars also have a nearly
entire leaf with highly reduced lobes. Bush-type pollenizers currently available are 'Companion'
(Seminis, Inc., Oxnard, CA) and 'Stud' (Abbott and Cobb, Inc., Feasterville, PA).
Most fruits produced by special pollenizers are small and usually weigh less than 2.5 kg
(Freeman and Olson, 2007). Special pollenizer cultivars produce fruits with one of two types of
rind patterns, solid grey to light green or light green with a dark green stripe, and vary in shape
from round to oblong and blocky. As with the dedicated-row pollenizer arrangement, a
pollenizer cultivar that has fruits easily distinguishable from the seedless fruits should be chosen.
Most special pollenizer fruits are substantially smaller than medium and large seedless fruits
which aides in their distinction. However, when personal size seedless watermelons (< 3.2 kg)
are produced, a cultivar with a distinct rind should be chosen as separation based on size may not
be possible. The thin-vine pollenizers are recommended to be planted at a 1:3 pollenizer to
triploid ratio while the bush-type cultivars are recommended at a 1:2 pollenizer to triploid ratio.
Pollinator Activity and Preference
Cultivated watermelon crops require pollination and domesticated honeybees (Apis
mellifera L.) are the most important pollinator (Free, 1993). Walters (2005) illustrated that
increased honeybee visitation to pistillate triploid flowers is required for fruit set due to the
dilution of viable diploid pollen with non-viable triploid pollen. Both triploid and diploid
watermelon plants produce visually similar staminate flowers and triploid flowers produce pollen
although it is not viable.
Honeybee foraging habits are controlled by both visual and olfactory cues but it is unlikely
that they can visually distinguish between triploid and diploid flowers (Butler, 1951; von Frishch
1967). These cues are processed during pre-alighting inspection and determine whether the
flower will be foraged. It has been shown that floral structures such as petals, sepals,
gynoecium, and pollen have distinct volatile emissions that are species and genotype specific
(Dobson et al., 1996; Dobson, 1991; Dobson et al., 1987). The volatiles from pollen (which are
derived from pollenkitt) are the most important factors when honeybees decide to forage a flower
or not (Pernal and Currie, 2002). Although a hierarchy of pollen preference was shown by Olsen
et al. (1979), no differences were observed between the species used by Pernal and Currie
(2002). The olfactory cues from pollen also appear to be quantitative and decreasing emissions
throughout the day indicate less reward to foragers (Dobson et al., 1996).
It has been suggested that pollen odor may be distinct between male-fertile and male-
sterile flowers of the same species (Dobson et al., 1996). Preference for male-fertile over male-
sterile potato flowers has been shown in bumblebee which may have been due to pollen odor
(Arndt et al., 1990; Batra, 1993). Wolf et al. (1999) conducted pollinator preference experiments
in which honeybees were placed in a Hield with two watermelon cultivars, one Citrulhts
colocynthis accession and one C. colocynthis x C. lan2atus hybrid 'BAG', with the number of
visitations being recorded. Significantly greater bee visitation was seen in the watermelon
cultivar BAG and the C. colocynthis accession, neither of which had nectar volume, pollen
quantity, or flower size that was different from the other genotypes tested. Wolf et al. (1999)
found a positive correlation between honeybee visitation and sugar concentration of nectar which
is what greater visitation was attributed to. This conclusion is in contrast to Pernal and Currie
(2002) who illustrated that pollen odor was more important than forage quality for honeybees.
Triploid seedless watermelon fruit growth and development is dependent on pollination of
the pistillate triploid flowers with viable diploid pollen (Kihara, 1951; Maynard and Elmstrom,
1992). It has been shown in other genera that genotype can have a significant effect on pollen
viability and that variations in viability can affect the reproductive success of the individual
receiving the pollen. Parties et al. (2005) reported significant differences in pollen viability
within and between species of barley however, these differences were only evident after the
pollen was subj ected to incubation treatments. These results indicate more of a difference in
pollen longevity as opposed to viability.
Fortescue and Turner (2004) investigated the pollen viability within and among multiple
banana species and among ploidy levels within a single species. This study reported significant
differences in pollen viability between species, within species, between ploidy levels within a
single species, and within ploidy levels of a single species. The differences in pollen viability
were as great as 100% between cultivars of the same ploidy level and same species (Fortescue
and Turner, 2004). Pollen source has been investigated in mandarin orange and significant
effects were reported on fruit quality parameters. Vithanage (1991) investigated the pollen donor
effects on 'Ellendale' mandarin using six different pollenizers. Vithanage reported that fruit
weights of 'Ellendale' were significantly greater when 'Murcott' and 'Emperor' were used as
pollenizers. Wallace and Lee (1999) conducted experiments in which 'Ellenor' mandarin was
pollinated by 'Murcott', 'Imperial', and 'Ellenor'. This study found that fruits from 'Ellenor'
had significantly greater size and sugar content when 'Murcott' was used as a pollenizer. Lavi et
al. (1996) reported significant differences in pollen viability among cultivars of macadamia but
found no correlation between pollen viability and fruitlet retention.
Nikkanen et al. (2000) illustrated that pollen viability within Picea abies was significantly
effected by individual pollen donor and germination conditions, and that there was an interaction
between these two factors. These results show that individuals of the same species, within a
geographic area, may require specific environmental conditions for reproductive success. Brevis
et al. (2006) reported significant differences in pollen viability among rabbiteve blueberry
cultivars, although all cultivars had a high average viability. This study suggested that while
pollen viability was statistically significant, it may not be biologically significant and is not
thought to contribute to reproductive failure in blueberry. Pollen viability data presented by
Brevis et al. (2006) was similar to previous findings in rabbiteve and southern highbush
blueberry (Cockerham and Galleta, 1976; Lang and Parrie, 1992).
There is only one published study on pollen viability in the family Cucurbitaceae. Nepi
and Pacini (1993) investigated various aspects of pollination in a single cultivar of Cucurbita
pepo. They reported that average pollen viability at anthesis was 92% which decreased to 75%
within 6 h and further decreased to 20% at 11 h after anthesis. This decrease in viability was
attributed to dehydration of the pollen grain. There is no published data on pollen viability of
Variation in pollen production has been reported between genera and among species of the
same genera in the Poaceae, and among species of the same genera in the Cupressaceae (Hidalgo
et al., 1999; Prieto-Baena et al., 2003). In diploid watermelon, Stanghellini and Schultheis
(2005) investigated 27 cultivars and found significant differences in production of pollen grains
per flower and pollen grains per plant. Pollen production ranged from 134,206 grains per plant
per day for 'Jamboree' to 264,589 grains per plant per day for 'Summer Flavor 800'.
The time period over which diploid cultivars produce pollen may be as important as the
amount of pollen produced. Diploid watermelon cultivars begin to produce staminate flowers
about seven days before triploid plants begin to flower and it is essential that the diploids
continue producing staminate flowers throughout triploid fruit set (Freeman and Olson, 2007).
Significant differences in total staminate flower production, as well as flowering longevity, have
been reported in diploid watermelon cultivars (Freeman and Olson, 2007; Stanghellini and
Schultheis, 2005). Greater staminate flower and pollen production by a diploid cultivar may
improve its performance as a pollenizer by reducing the dilution effect of viable pollen that is
created by triploid plants.
Competition has been defined as the negative interaction between two organisms (Connell,
1990). In plants, this competition is for light, water, nutrients, and space, and can be inter-
specifie (between two species) or intra-specific (between individuals of the same species). Inter-
specific competition from weed species as well as intra-specific competition from neighboring
crop plants can reduce the survivability of plants and the yield and quality of plant products
(Firbank and Watkinson, 1990). Intra-specific competition in cropping systems is regulated by
planting density which is intended to maximize production per unit area. Maximum production
per unit area occurs when plant population and yield per plant are in correct proportions.
Experimental models have shown that plant yields increase with plant density to a maximum
point and then plateau or decline as density continues to increase (Holliday, 1960). Intra-specific
competition and plant density can also be used as tools to manipulate yield parameters such as
size distribution of fruit (Motsenbocker and Arancibia, 2002; Reiners and Riggs, 1999; Sanders
et al., 1999).
Intra-specific competition of crop plants is investigated through studies that examine the
effect of planting density or spatial arrangement on crop yield. In vining cucurbits such as
muskmelon (Cucumis melo L.), pumpkin (Cucurbitapepo L.), and watermelon, it has been
shown that increasing plant density increases total yield but decreases yield per plant (Ban et al.,
2006; Duthie et al., 1999a; Duthie et al., 199b; Goreta et al., 2005; Kultur et al., 2001; Maynard
and Scott, 1998; Reiners and Riggs, 1999, 1997; Sanders et al., 1999). The increases in yields
per unit area reported in these studies were due to increased fruit numbers per unit area.
Although the use of in-row diploid pollenizers has increased, it has not been determined if the
increased competition on neighboring triploid plants will be deleterious. When planted at a 1:3
pollenizer to triploid ratio, the pollenizer will directly impact 2/3 of the plants per hectare by
decreasing in-row spacing by 1/4. This would reduce area per plant from 2.2 m2 to 1.6 m2 foT
plants grown on 2.4 m between-row spacing and 0.9 m in-row spacing. Though reduced, this
area is still greater than the 1.0 m2 per plant which has been shown to produce greatest
watermelon yields per hectare (Goreta et al., 2005; Sanders et al., 1999). Results from other
studies in watermelon do not provide insight as the phenotype of the pollenizer and triploid are
different. As pollenizer growth will not impact all triploid plants per hectare, the appropriate
study to investigate pollenizer competition effect is the neighborhood (area of influence) study in
which the performance of a single individual is measured as a function of distance from the
competitor (Radosevich and Roush, 1990).
CHARACTERISTICS OF DIPLOID POLLENIZERS FOR USE IN TRIPLOID
With triploid seedless watermelons now occupying 78% of the United States market, it is
suspected that the use of in-row pollenizer cultivars will increase (USDA, 2006). There are no
published studies that compare important characteristics of diploid watermelon pollenizers such
as staminate flower production, flowering period, and fruit production. Pollenizer flower
production may be a strong indicator of how a cultivar will perform and flowering period is
critical to the type of production system the pollenizer is used in. Using a pollenizer cultivar
with low fruit production and easily distinguishable fruit could also increase efficiency in
harvesting operations. The objectives of this study were to determine staminate flower
production, flowering period, and quantity and size of fruit production of several commercially
available diploid pollenizers.
Materials and Methods
Experiments were conducted at the North Florida Research and Education Center, Quincy,
FL, and the North Florida Research and Education Center-Suwannee Valley, Live Oak, FL. In
Quincy, the soil type was a Norfolk Loamy Fine Sand (fine-loamy, kaolinitic, thermic, Typic
Kandiudults) and in Live Oak the soil was a Lakeland fine sand (thermic, coated Typic
Quartzipsamments). At both locations, the experiment was arranged as a randomized complete
block design with four replications. Transplants were produced in a greenhouse at Quincy in
expanded polystyrene flats of the inverted pyramid design which were 3.75 x 3.75 x 6.25 cm
using soil-less media. Prior to the laying of the mulch, pre-plant fertilizer was applied at
recommended rates and incorporated into the soil (Olson et al., 2004). All fertilizer was applied
pre-plant in Quincy and one-fourth was applied pre-plant in Live Oak. Weekly fertigation was
used to apply the remainder of the fertilizer in Live Oak. Watermelon plants were irrigated as
On 1 Apr. 2005, 5-week-old seedlings of 'SP-1', 'Companion', 'Jenny', and 'Mickylee'
were transplanted into raised beds fumigated with methyl bromide and chloropicrin (67/33) and
covered with black polyethylene mulch. In Live Oak, plots consisted of two rows; beds were 0.6
m wide by 10 m long on 2. 1 m centers, with in-row spacing of one meter. Pollenizer cultivars
were planted with the triploid cultivar 'Tri-X 313'. A pollenizer plant was planted at the
beginning and end of each plot, and between every third and fourth 'Tri-X 313' plant in the plot.
Plots at Live Oak consisted of 18 triploid plants and eight pollenizers. In Quincy, plots consisted
of two rows; beds were 0.9 m wide by 13.2 m long on 2.4 m centers, with an in-row spacing of
one meter. 'Tri-X 3 13' was also used in Quincy and placement and spacing of pollenizer
cultivars were the same as for Live Oak. Plots at Quincy consisted of 24 triploid plants and 10
Plants were sampled after the onset of male flowering by pollenizer cultivars, and data
were collected twice a week at both locations. Once 70% of pollenizer plants in each plot had at
least one open male flower (flowering threshold), the plot was considered to have begun
flowering. Numbers of male flowers per plant were recorded from the beginning of flowering
(29 Apr.) until the end of the fruit set period (3 June). Early season flower counts were obtained
by counting open male flowers on all pollenizer plants in each plot. However, after plants began
to vine heavily (making flower counting difficult), flower counts were obtained from a single
row in each plot. Fruit from pollenizers were harvested at or near maturity and weighed on two
dates per location. 'Tri-X 313' was not harvested for yield and only served to provide intra-
specific competition for the pollenizers. Without this competition, growth habit of the
pollenizers would be different and not reflect flower counts that would be seen under commercial
Statistical analysis was performed using the GLM procedures of SAS (SAS Institute, Inc.,
Cary, NC). For statistical analysis, location was added and the experiment was analyzed as a
factorial experiment with two factors, location and cultivar. If an interaction was present,
LSMEANS analysis and LSD were was used to explain results, otherwise means separation was
performed using Duncan's multiple range test.
The first pollenizer cultivar to reach the threshold for flowering in all plots was SP-1', on
2 May, 2005 in Quincy. All cultivars had reached the flowering threshold by 10 May and 12
May for Quincy and Live Oak, respectively. Peak fruit set by 'Tri-X 313' began 14 May and
ended 3 June at both locations. By 1 June, 'Tri-X 313' had set all commercially harvestable
fruit, and by 10 June, mature 'Tri-X 313' melons were present at both locations. Location did
not significantly affect male flower counts, and there was no significant interaction between
location and cultivar (P > 0.05). Data were combined over locations and analyzed.
On 29 Apr., SP-1' produced 0.65 flowers per plant (fpp), which was not significantly
different from 'Companion' or 'Mickylee' which had 0.65 and 0.21 fpp, respectively (Fig. 3-1).
'Jenny' had 0. 12 fpp which was significantly lower than that of 'Companion' or 'SP-1', but not
'Mickylee'. On 3 May, 'SP-1' had 0.73 fpp which was greater than 'Companion' at 0.52 fpp or
'Jenny' at 0.38 fpp. 'Mickylee' had 0.25 fpp which was significantly lower than that of
'Companion' or 'SP-1', but not 'Jenny'. On 9 May, 'Jenny' had 1.95 fpp which was greater than
'Companion' at 1.00 fpp. However, fpp for 'Jenny' was not different than 'SP-1' or 'Mickylee',
which were 1.65 and 1.50 fpp, respectively. On 13 May, 'SP-1' had 6.36 fpp which was not
different than 'Jenny' at 5.76 fpp or 'Mickylee' at5.52 fpp, however, 'SP-1' was significantly
higher than 'Companion', 4.75 fpp. On 16 May ('Tri-X 313' had begun producing female
flowers), 'SP-1' had greater flower counts compared to the other pollenizer cultivars; 'SP-1' had
9.72 fpp compared to 7.88, 7.56, and 7.5 fpp for 'Mickylee', 'Companion' or 'Jenny',
respectively. This trend continued throughout the remainder of the season as most mid (16 May)
to late (1 June) season male flower counts showed that 'Companion', 'Jenny', and 'Mickylee'
had similar numbers of flowers but were less than 'SP-1'. Flower numbers increased and peaked
on 26 May when 'SP-1' had 35.5 fpp which was greater than 'Companion', 'Mickylee', and
'Jenny' which had 14.50, 13.80 and 12.80 fpp, respectively.
Analysis of pollenizer fruit counts per plot indicated a significant location effect and an
interaction between location and cultivar (Table 3-1). Distribution of fruit per plant by cultivars
was slightly different at the two locations (Table 3-2). At Quincy, fruit set was greater for
'SP-1', with 5.8 melons per plant (mpp) compared to all other cultivars. 'Companion' set the
least number of fruit, with 1.8 mpp, which was less than 'Jenny' at 4.3 mpp or 'Mickylee' at3.7
mpp. At Live Oak, 'Jenny' produced 3.4 mpp and 'SP-1' produced 3.3 mpp, which were greater
than 'Mickylee' at 1.7 mpp or 'Companion' at 1.2 mpp. The interaction between cultivar and
location is due to the different rankings of fruit production by cultivar at each location. In
Quincy, fruit production by 'Jenny' and 'Mickylee' were not significantly different, however
both produced less fruit than 'SP-1'. At Live Oak, fruit production by 'SP-1' and 'Jenny' were
not significantly different but both produced more fruit than 'Mickylee'. At Quincy, 'Mickylee'
produced more fruit than 'Companion' but at Live Oak fruit production was similar between the
Data analysis indicated that location had a significant effect on pollenizer fruit weight, and
that fruit weights in Live Oak (avg. = 2.80 kg) were lower than those at Quincy (avg. = 3.06 kg)
(P < 0.05). There was no interaction between location and cultivar. 'Mickylee' had the highest
fruit weight, 4.34 kg/fruit, which was higher than for 'Companion' or 'Jenny' which were 2.61
and 2.56 kg/fruit, respectively (Table 3-3). 'SP-1' had the lowest fruit weight at 2.21 kg/fruit,
which was less (P < 0.05) than of all other pollenizer cultivars.
'SP-1' and 'Companion' produces non-edible fruits while 'Jenny' and 'Mickylee' produce
edible melons. When used as pollenizers, fruit production is not desirable because melons from
pollenizers can confuse the harvesting process (with mixing of seeded and seedless fruit) as well
as hinder harvesters from moving efficiently through the field. 'Companion' and 'Mickylee'
produce easily distinguishable melons based on the grey to pale green color and no rind pattern.
'SP-1' produces a light green melon with very thin light green broken stripes while 'Jenny'
produces melons that have a medium green background with dark green stripes. All four
pollenizer cultivars would be easily distinguishable by size from most commercial melons, other
than personal-size seedless watermelons which generally weigh between 1.8 and 2.2 kg.
All four pollenizers had male flowers present during peak seedless watermelon fruit
setting. Results were similar to Dittmar et al. (2005) with 'SP-1' producing significantly more
male flowers than 'Companion', 'Jenny', or 'Mickylee'. Although flower production by 'SP-1'
was highest other cultivars may also provide more than enough pollen to accomplish optimal
seedless watermelon fruit set. Stanghellini and Schultheis (2005) reported that pollen production
is also variable between diploid watermelon cultivars so flower production may not be
completely indicative of a cultivars male reproductive output. All cultivars appear to be viable
options for use as pollenizers in triploid watermelon production as all were producing male
flowers during fruit set of seedless watermelons.
Table 3-1. Analysis of variance for pollenizer fruit weight and fruit per plant at Quincy and Live
Oak FL during 2005.
Source df MS P-value
Replication 3 0.10217 0.848
Location 1 2.67961 0.015
Cultivar 3 35.3234 <0.0001
Location*Cultivar 3 0.17825 0.708
Error 21 0.38212
Fruit per plant
Replication 3 0.33208 0.207
Location 1 17.7012 <0.0001
Cultivar 3 14.2712 <0.0001
Location*Cultivar 3 1.75791 0.0006
Error 21 0.20113
Table 3-2. Interaction effect of location and watermelon pollenizer cultivars on fruit per plant at
Quincy and Live Oak FL during 2005.
Location x cultivar Fruit (no./plant)
Quincy SP-1 5.8
Jenny 4.3 **
Mickylee 3.7 Ns
Companion 1.8 **
Live Oak Jenny 3.4
SP-1 3.3 Ns
Mickylee 1.7 **
Companion 1.2 Ns
Ns, *,** non-si nificant, or si nificant at P<0.05 or P<0.01, Least S uares Means anal sis.
Table 3-3. Main effects for pollenizer fruit weights combined over experiments conducted in
Quincy and Live Oak, FL, during 2005.
Cultivar Avg fruit wt (kg)
Mickylee 4.34 az
Companion 2.61 b
Jenny 2.56 b
SP-1 2.21 c
z Means followed by the same letter are not significantly different at (P < 0.05) by
Duncan's multiple range test.
-M-Co mpa union
b b b
29-Apr 3-May 9-May 13-May 16-May 19-May 23-May 26-May 30-May 3-Jun
Figure 3-1. Pollenizer staminate flower counts combined over Quincy and Live Oak<, FL, during 2005. Means followed by the same
letter are not significantly different at (P <0.05) by Duncan' s multiple range test.
POLLEN VIABILITY OF DIPLOID WATERMELON POLLENIZER CULTIVARS
Pollen produced by diploid watermelon pollenizers is important because it is necessary for
fruit set and flesh fill in associated triploid watermelon crops (Kihara, 1951; Maynard and
Elmstrom, 1992). Viable pollen produced by pollenizers is diluted with non-viable triploid
pollen which increases pollinator visitation rates required by pistillate triploid flowers.
Pollenizers with poor pollen viability could further increase required visitation from pollinators
and reduce stigmatal area available to viable pollen; both of which could have negative effects
on seedless watermelon yield and quality. Significant differences in pollen viability have been
reported between cultivars of the same species in several genera. The objective of this project
was to determine pollen viability of four pollenizer cultivars to investigate possible effects on
Materials and Methods
On 3 Apr. and 1 Aug. 2006, 4-week old watermelon seedlings were transplanted into
raised beds. The beds were covered with black polyethylene mulch in the spring and white
polyethylene mulch in the fall. Experiments were performed at the North Florida Research and
Education Center (NFREC) in Quincy, FL. Soil type present at NFREC is Norfolk loamy sand
(fine-loamy, kaolinitic, thermic Typic Kandiudults). Experimental design both seasons was a
randomized complete block with four replications. Four diploid pollenizer cultivars,
'Companion', 'Jenny', 'Mickylee', and 'SP-1', were used to determine pollen viability.
Experimental plots were 4.57 m long with an in-row spacing of 0.91 m and between-row spacing
of 2.43 m. Three seedlings were planted in each plot. Fertilization, irrigation, and pesticide
application practices recommended by the University of Florida Institute of Food and
Agricultural Sciences were followed (Olson et al., 2006).
Sampling was initiated on 17 May and samples were taken on 24 May and 31 May in the
spring. Sampling was initiated on 31 Aug. and samples were taken on 7 Sept. and 14 Sept. in the
fall. Sampling was initiated when other triploid watermelons at NFREC that were transplanted
on the same dates began to set fruit. The sampling period was scheduled to mimic peak fruit set
in triploid watermelons. The fruit set in this time frame would be the maj ority of the fruit that
would be available for commercial harvesting schemes that are typical for FL.
On sampling dates, watermelon flowers were removed from the plant before anthesis. This
was to insure that pollinators would not remove pollen and an adequate supply would be
available for analysis. Three flowers were removed from each plot and placed into plastic cups
and covered to exclude pollinators. Flowers were taken to the lab and allowed to open. After
anther dehiscence had occurred (verified with a hand lens) sample analysis was initiated. Pollen
was removed from the anthers and placed on a slide. Viability was determined using the
diaminobenzidine (DAB) protocol for peroxidase activity in pollen (Dafni et al., 2005).
Rodriguez-Riano and Dafni (2000) compared the results of four vital dyes versus pollen
germination results and illustrated the superiority of peroxidase tests over other commonly used
vital dyes. This test utilizes a dye that creates a color differential between viable and non-viable
Four 100 pollen grain sub-samples were analyzed from each plot using a compound
microscope. A pollen grain was considered viable if it had turned dark brown or black. All
pollen samples were analyzed on the same day pollen was collected. It is important with
watermelon pollen that the pollen be thoroughly mixed with the dye on the slide in order to have
adequate contact between the pollen grains and dye. If large clumps of pollen are not broken up,
dye may not infiltrate the pollen and false negatives may be observed. On each sample date heat
killed pollen (two h at 800C) was used to check the efficacy of the dye. New dye was prepared
for each sample date. A square-root transformation was performed on the data and analysis of
variance and means separation (Duncan's multiple range test) were performed using the GLM
procedures of SAS (SAS Institute Inc., Cary, NC)
There were no sampling dates in either season where pollenizer cultivar had a significant
(P < 0.05) effect on pollen viability (Table 4-1). There was also no significant interaction
between pollenizer cultivar and sampling date (Table 4-1). In the spring trial, sampling date had
a significant (P < 0.05) effect on pollen viability with 31 May having greater average viability
than 24, or 17 May (Table 4-2). Pollen viability on 31 May was 98.6% which was significantly
greater than 97.4 or 97% for 17 May and 24 May, respectively. The average pollen viability
over all cultivars and all dates for spring and fall were 97.7 and 97.9%, respectively. There was
very little variation within the data and the coefficient of variation was never higher than 0.84%.
The results of this study illustrate that there is no significant variation in pollen viability
within the cultivars tested and that pollen viability is high and changes very little if any
throughout the growing season, at least not within the critical period of triploid fruit set. The
sample date in the spring with higher viability is more likely due to environmental conditions
than cultivar characteristics. Freeman (2007) observed that seedless watermelon yield was
significantly lower when 'Companion' was used as a pollenizer versus 'Jenny' or 'SP-1'. The
results of this study suggest that pollen viability was not a contributing factor in the varying
degrees of performance of these pollenizers. Factors such as floral attractiveness to pollinators,
timing and total production of staminate flowers, and pollen production may be more important
characteristics of pollenizers.
The small amount of variation in pollen viability between the cultivars tested suggests that
there may be little within the species or at least in cultivated varieties. Nepi and Pacini (1993)
reported that the pollen viability of 'Greyzini' (Cucurbita pepo L.) averaged 92% which is
similar to the findings of this study. Pollen viability appears to play no role in the performance
of the pollenizers tested and may not be an important characteristic of pollenizer cultivars.
Table 4-1. Analysis of variance for pollen viability of watermelon pollenizer cultivars tested
during the Spring and Fall of 2006 at Quincy, FL.
Source df MS P-value
Sampling Date 2 0.02827 0.003
Replication 3 0.00410 0.410
Cultivar 3 0.00856 0.124
Date* Cultivar 6 0.00281 0.669
Error 33 0.00415
Sampling Date 2 0.00302 0.367
Replication 3 0.00290 0.408
Cultivar 3 0.00281 0.422
Date* Cultivar 6 0.00257 0.519
Error 33 0.00292
Table 4-2. Influence of diploid watermelon pollenizer cultivar on pollen viability at Quincy, FL
during the Spring and Fall of 2006.
Cultvar 17 May 24 May 31 May 31 Aug. 7 Sept. 14 Sept.
Mikyee 97.8 Nsz 98.0 Nsz 98.7 Nsz 97.4 Nsz 98.2 Nsz 97.8 N
Companion 97.4 97.2 99.2 97.6 98.1 98.8
Jenny 97.3 97.2 98.3 97.8 97.7 97.0
SP-1 97.2 95.5 98.1 97.9 98.7 97.6
Date Means 97.4 b Y 97.0 b 98.6 a 97.7 NS y 98.2 97.8
z P = 0.05 Means are compared within the same column. Y P = 0.05 Means from the same
season are compared within the row.
DIPLOID WATERMELON POLLENIZER CULTIVARS EXHIBIT VARYING DEGREES OF
PERFORMANCE WITH RESPECT TO TRIPLOID WATERMELON YIELD
With seeded watermelons only holding about 20% of the U. S. market, there is interest in
using pollenizers that do not require dedicated field space (USDA, 2006). Traditionally
pollenizers occupied 20-33% of the land area in a field. New pollenizers have been developed to
be planted in-row with triploid plants without altering spacing. There are now multiple in-row
pollenizer cultivars available that exhibit varying growth habits. Previous research has reported
differences in production and timing of staminate flowers by pollenizer cultivars. The seed costs
of in-row pollenizers are greater than open-pollinated and hybrid diploids and most in-row
pollenizers are not intended to be harvested. There is interest in using open-pollinated cultivars
in-row to reduce input costs but it is unclear if acceptable yields will result. These experiments
were conducted to determine if there was a difference in performance between in-row pollenizer
cultivars and if a standard open-pollinated cultivar could be used in-row with similar success.
Materials and Methods
These experiments were performed at one location (Blackville, SC) in 2005 and three
locations (Blackville, SC, Citra, FL, Quincy, FL) in 2006. The experimental design used was a
randomized complete block with four replications. Experimental plots consisted of three raised
bed rows that were spaced 2.43 m center-to-center and covered with black polyethylene mulch.
Watermelon plants were spaced 0.91 m in-row. Replications consisted of three rows 116 m long
with a 7.62 m buffer between replications. A diagram of the experimental layout is shown in
Figures 5-1 and 5-2. The two outside rows were planted with 'Tri-X Palomar' and the interior
row was planted with 'Tri-X 313' in 2005 and 'Supercrisp' in 2006. In 2005, pollenizer cultivars
used were 'Jenny', 'Mickylee', and 'SP-1' with 'Tri-X Palomar' as a control. In 2006, the
pollenizer cultivars used were 'Companion', 'Jenny', 'Mickylee', 'Patron', 'Pinnacle',
'Sidekick', and 'SP-1' with 'Tri-X Palomar' as a control. Pollenizer seed sources are listed in
table 5-5. In order to reduce pollen contamination from neighboring plots, an eight plant buffer
(7.3 m) of 'Tri-X Palomar' was planted in the center row between each plot (Figure 5-2). It has
been demonstrated that distance from a diploid pollenizer of 6.0 m or greater will greatly reduce
the triploid fruit set (NeSmith and Duval, 2001). 'Tri-X Palomar' was chosen as the buffer
cultivar and control plot "pollenizer" because it does not produce viable pollen and its rind
coloration is distinctly different than the harvested cultivars, Tri-X 313 and Supercrisp. Eight
triploid watermelon plants were transplanted into each plot including the control or check plot.
Three plants of a pollenizer cultivar were planted in each plot except the control plot where 'Tri-
X Palomar' was planted in place of a pollenizer. Control plots were in place in order to observe
if pollen was moving from plot to plot. 'Jenny', 'Mickylee', 'Patron', 'Pinnacle', 'Sidekick', and
'SP-1' were planted at a 1:3 pollenizer to triploid ratio while 'Companion' was planted at a 1:2
pollenizer to triploid ratio. These ratios are recommended by producers of the various
pollenizers. Three plants of the 1:3 ratio pollenizers, and four plants of the 1:2 ratio pollenizer
were included in each plot in the same row as the harvested watermelon.
Soil type at the Edisto Research and Education Center (EREC) in Blackville, SC was
Dothan loamy sand (fine-loamy, kaolinitic, thermic Plinthic Kandiudults). Soil type at the North
Florida Research and Education Center (NFREC) in Quincy, FL was Norfolk loamy sand (fine-
loamy, kaolinitic, thermic Typic Kandiudults). Soil type at the Plant Science Research and
Education Unit (PSREU) in Citra, FL was Hague sand (loamy, siliceous, semiactive,
hyperthermic Arenic Hapludalfs). Drip tapes (1.89 1-min. 1/30.48 m at 68 kPa; 30.48 cm emitter
spacing) were laid under and concurrently with the polyethylene mulch. Beds were fumigated
with methyl bromide/chloropicrin 67:33 at a rate of 448 kg-ha-l broadcast at EREC in 2005, and
PSREU and NFREC in 2006. Fertilizer recommendations for EREC were 156N-0P-130.2K
kg-ha-l in 2005 and 2006 (Franklin, 1998). Fertilizer recommendations for PSREU and NFREC
were 168N-48P-140.2K kg-ha-l and 183.6N-24P-152.3K kg-ha- respectively (Olson et al.,
2006). All fertility recommendations were based on soil test results. Four-week-old watermelon
plants were transplanted at EREC on 27 Apr., 2005 and 17 Apr., 2006. Four-week-old seedlings
were transplanted at PSREU and NFREC on 21 Mar. and 3 Apr., 2006, respectively.
Plots were sprayed with fungicides and insecticides as recommended (Olson et al., 2006,
Sanders et al., 2006). Pesticide applications were timed so that there was minimal effect on
pollinators. One honeybee (Apis mellifera L.) hive was located near the center of each
replication at Citra and Quincy, FL in 2006 while at Blackville, SC in both years a grouping of
twenty honeybee hives was maintained 300 ft. north of test plots. At all locations in 2005 and
2006, watermelons were harvested once per week for three weeks. At the last harvest, all
marketable melons were harvested. The center of each fruit was sampled for total soluble solids
using a hand-held refractometer. Hollow heart measurements were taken by measuring the
length and width of hollow cavities in watermelons that had been cut longitudinally from stem
end to blossom end. Soluble solids and hollow heart data were taken from three melons per plot
during the first harvest at all locations in 2006. Soluble solids data were taken from three
watermelons per plot during the first harvest at Blackville, SC in 2005 but no hollow heart data
Yield and soluble solids data from 2005, and hollow-heart data from 2006 were analyzed
using the GLM procedures and means separation was accomplished using Duncan's multiple
range test in the SAS system (SAS Institute, Inc., Cary, NC). In 2006, there were multiple
locations and as location was not of primary interest in this study, location was set as a random
effect. The MIXED procedure was used to analyze cultivar effect on fruit yield (kg-ha- fruit/ha,
and kg/fruit) and soluble solids. This allows for greater inference of the results and how they
may relate to many locations as compared with setting location as a Eixed effect (Cushman et al.,
2003; Schabenberger and Pierce, 2002). Pollenizer cultivar was set as a Eixed effect and
location, replication, and location by cultivar interaction were set as random effects.
Treatments with pollenizer cultivars had significantly greater yield of triploid watermelons
at all locations and in both years compared to the check (Tables 5-1 & 5-2). In addition there
were significant differences among pollenizer cultivars in 2006 (Table 5-2). There were no
significant differences in triploid watermelon yields among pollenizer cultivars in 2005 (Table 5-
1). In 2006, plants pollenized by 'Sidekick' yielded 65,242 kg-ha-l but were not significantly
different than plants pollenized by 'Patron', 'SP-1', 'Jenny', or 'Mickylee' which yielded
63,677, 61,766, 61,751, and 59,599 kg-ha- respectively (Table 5-2). Plants pollenized by
'Companion' had the lowest yields at 49,976 kg-ha- which were significantly lower than those
pollenized by 'Jenny', 'SP-1', 'Patron' or 'Sidekick' but not significantly different than plants
pollenized by 'Pinnacle' or 'Mickylee' which yielded 53,333 and 59,599 kg-ha- respectively.
Plots containing 'Pinnacle' had significantly lower yields than those containing 'Sidekick' but
were not significantly different than plots containing 'Mickylee', 'SP-1', 'Jenny', or 'Patron'
(Table 5-2). Pollenizers had a significant effect on number of triploid watermelons compared to
the check. All plots with pollenizer cultivars had significantly greater numbers of melons per
hectare than the control plots at all locations in both years (Tables 5-1 & 5-2). There were no
significant differences in fruit production between the pollenizer cultivars in 2005 and 2006. In
2006, plants pollenized by 'Patron' produced 9,616 fruit/ha which was not significantly greater
than 'Companion' which produced 7,565 fruit/ha. Pollenizer cultivars had a significant effect on
average triploid watermelon fruit weight in 2006, but not in 2005 (Tables 5-1 & 5-2). Pollenizer
cultivars did not have a significant effect on soluble solids in both years (Table 5-3). In 2006,
pollenizer cultivars did not have a significant effect on hollowheart at the Citra, FL and
Blackville, SC locations (Table 5-4). Pollenizer cultivars did have a significant effect on hollow
heart at Quincy, FL with all plots with pollenizers having significantly less hollow heart in the
triploid watermelons when compared with the control plots (Table 5-4). There were no
significant differences in hollow heart incidence between pollenizer cultivars.
This research shows that some pollenizer cultivars tested can be expected to perform better
than other cultivars, and do so at diverse locations. Similar results were reported by Fiacchino
and Walters (2003) in which triploid watermelon yields were significantly different due to
pollenizer cultivar used.
The only cultivar that showed questionable performance was 'Companion'. Due to its
growth and flowering habit it may not produce enough staminate flowers and pollen at the end of
fruit setting in the triploid crop. 'Companion' is a short internode plant that becomes overgrown
by triploid plants near the end of the season which may lead to staminate flowers that are not
readily detectable by pollinators. Differences in staminate flower production by pollenizer
cultivars have been reported however, it does not appear that flower production is the
determining factor of a pollenizer' s performance (Dittmar et al., 2005; Freeman and Olson,
2007). In both of these studies, SP-1' produced greater numbers of staminate flowers when
compared to 'Jenny' or 'Mickylee', however, data presented here indicates no difference in
triploid watermelon yields between these pollenizer cultivars. Pollenizers must be able to
continue growing and producing flowers throughout the production cycle.
There were significant differences in severity of hollow heart at Quincy between plots
containing pollenizer cultivars and the control but not between the pollenizers. Unfortunately,
this does not help to elucidate the cause of hollow heart as it may have been caused by reduced
pollination in control plots or excessive growth of the few existing watermelons. The incidence
of hollow heart at Blackville, SC and Citra, FL was low overall and this may be why there was
no effect by the pollenizers. The experimental design was successful in reducing pollen flow out
of experimental plots as indicated by minimal fruit set in control plots. This experimental design
spaced the triploid watermelon from a pollenizer cultivar by 7.3 m. NeSmith and Duval (2001)
illustrated that when distance of a triploid from a pollenizer was six meters or greater, triploid
fruit numbers diminished substantially. Triploid pistillate flowers ('Tri-X Palomar') in plot
buffers served to filter viable diploid pollen before pollinators entered another plot.
Of the cultivars tested, it appears that the pollenizers 'Jenny', 'Mickylee', 'Patron',
'Pinnacle', Sidekick', and SP-1' would be good choices. Some of the tested pollenizers
('Mickylee', 'Jenny', 'Pinnacle') can be harvested and sold if the grower has a market for seeded
watermelons. If growers have a strong market for seeded melons then there may be no reason to
plant pollenizers in-row. The pollenizers' costs vary greatly, so this must also be taken into
consideration. Of the pollenizer cultivars that were shown to perform adequately ('Jenny',
'Mickylee', 'Patron', 'Pinnacle', 'Sidekick, 'SP-1'), selection should be based on seed/plant cost
and distinctness between pollenizer and market melon.
Plot 1 Plot 1
B Plot 2 Plot 2
B B B
U Plot 3 Plot 3
U U U
F Plot 4 F 7.6 mn Drive Row F Plot 4 F
EPlot 5 F F Plot 5 F
E E E
R Plot 6 Plot 6
R R R
Plot 7 Plot 7
Plot 8 Plot 8
Figure 5-1. Field diagram for pollenizer experiments at Blackville, SC, Citra, FL, and Quincy,
FL in 2005 and 2006. Columns represent individual rows. The same design was
used for replications three and four.
B Data melon B
R Data melon R
R ~Data melon
Figure 5-2. Individual three-row plot design for pollenizer experiments at Blackville, SC, Citra,
FL, and Quincy, FL in 2005 and 2006. Plot shown is using a pollenizer
recommended to be planted at a 1:3 pollenizer to seedless ratio.
Table 5-1. Pollenizer cultivar effect on 'Tri-X 313' yield at Blackville, SC during 2005
Pollenizer Yield Fruit Avg wt
cultivar (kg-ha-l) (o/a knfit
Tri-X Palomar x
z Yield estimates are based on plant populations of 4483 plants per hectare. Means with the same letter are
not significantly different at (P 5 0.05) by Duncan's multiple range test. x Triploid cultivar serving as check
against pollen contamination from neighboring plots.
Table 5-2. Pollenizer cultivar effect on Supercrisp' watermelon yield and average fruit weight
at Blackville, SC, Citra, FL, and Quincy, FL during 2006.
Pollenizer cultivar Y ie zri zv w
(kg-ha- )" (no./ha)" (kg/fruit)
Sidekick 65,242 a 9,386 a 7.3 a b
Patron 63,677 ab 9,616 a 7.0 b
SP-1 61,766 ab 9, 106 a 7.0 b
Jenny 61,751 ab 9, 195 a 6.9 b
Mickylee 59,599 a bc 9, 106 a 6.7 b
Pinnacle 53,333 b c 7,845 a 7.2 ab
Companion 49,976 c 7,565 a 6.9 b
Tri-X Palomar 8,545 d 1,074 b 7.8 a
LSD x 10,494 2, 116 0.7
z Yield estimates are based on plant populations of 4483 plants per hectare. Triploid cultivar serving as
check against pollen contamination from neighboring plots. x P = 0.05
Table 5-3. Pollenizer cultivar effect on soluble solids concentration of seedless watermelons at
Blackville, SC during 2005 and Citra, FL, Quincy, FL, and Blackville, SC during
Soluble solids concentration (%)
Blackville, SC 2005 Combined locations 2006
Tri-X Palomar Y
z Means with the same letter are not significantly different at (P I 0.05) by Duncan's multiple range test.
Triploid cultivar serving as check against pollen contamination from neighboring plots. x P = 0.05
Table 5-4. Pollenizer cultivar effect on hollowheart disorder in Supercrisp' watermelon at
Quincy, FL, and Blackville, SC combined with Citra, FL during 2006. Means are to
be compared within the same column.
Hollowheart area (cm2)
Blackville, SC &
Pollenizer cultivar Quincy, FL
Tri-X Palomar z187.0 aY 4.7 Ns
Patron 73.3 b 12.2
Jenny 70.2 b 5.8
Sidekick 68.2 b 9.7
Companion 58.1 b 3.3
Mickylee 54.7 b 2.6
SP-1 53.5 b 15.7
Pinnacle 37.5 b 10.7
z Triploid cultivar serving as check against pollen contamination from neighboring plots Means with the
same letter are not significantly different at (PI 0.05) by Duncan's multiple range test.
Table 5-5. Seed sources for various pollenizer cultivars used during 2005 and 2006.
Pollenizer cultivar Company
Patron Zeraim Gedera Seed Co., Ltd. (Palm Desert, CA)
Jenny Nunhems USA, Inc. (Acampo, CA)
Sidekick Harris Moran Seed Co. (Modesto, CA)
Companion Seminis Vegetable Seed, Inc. (Oxnard, CA)
MickyLee Many sources
SP-1 Syngenta Seeds, Inc.(Boise, ID)
Pinnacle Southwestern Vegetable Seed, LLC. (Casa Grande, AZ.)
COMPETITIVE EFFECT OF IN-ROW DIPLOID WATERMELON POLLENIZERS ON
TRIPLOID WATERMELON YIELD
Due to the high cost of in-row diploid pollenizer seed, there is interest in using standard
watermelon cultivars in-row. One cultivar that is being evaluated for in-row use is 'Mickylee'
which produces an easily distinguishable fruit, a crucial characteristic for diploid pollenizers.
Most in-row pollenizer cultivars have reduced foliage in order to compete less with triploid
plants. 'Mickylee' is an attractive option as a pollenizer because of the low seed costs but it is
unclear whether the more vigorous growth habit will negatively impact seedless watermelon
yields. The obj ective of this study was to determine if pollenizer growth habit and pollenizer to
triploid spacing would have an effect on triploid watermelon yield.
Materials and Methods
Experiments were conducted at the North Florida Research and Education Center
(NFREC) in Quincy, FL. and the Plant Science Research and Education Unit (PSREU) in Citra,
FL. in the Spring of 2006 and also at NFREC in the Fall of 2006. Soil type at NFREC was
Norfolk loamy sand (fine-loamy, kaolinitic, thermic Typic Kandiudults) and at PSREU was
Hague sand (loamy, siliceous, semiactive, hyperthermic Arenic Hapludalfs). At NFREC (spring
and fall) all fertilizer was incorporated pre-plant at a rate of 183.6N-24P-152.3K kg-ha- At
PSREU two thirds of the fertilizer was applied pre-plant and the remainder of the
recommendation was fulfilled through weekly fertigation. Total fertilizer applied at PSREU was
168N-48P-140.2K kg-ha- Fertilization was based on soil test results and University of Florida
recommendations (Olson et al., 2006). At both locations, irrigation was provided through drip
tape (1.89 1-min. 1/30.48 m at 68 kPa; 30.48 cm emitter spacing) which was laid concurrently
with black polyethylene mulch in the spring and with white on black polyethylene mulch in the
fall. Soil was fumigated at plastic laying with methyl bromide/chloropicrin 67:33 at a rate of 448
kg-ha l. Beds were spaced 2.43 m center-to-center.
Experimental design was a factorial with four replications and two factors, pollenizer
cultivar and in-row spacing. In-row spacing's between one pollenizer plant and one triploid
plant were 0.2, 0.4, 0.6, 0.8, and 1.0 m. Plots consisted of six pairs (one pollenizer, one triploid)
of plants with equal pollenizer to triploid spacings. Each pair of plants was spaced 1.0 m in-row
from the next pair. On 21 Mar., 3 Apr., and 1 Aug., 4-week old watermelon seedlings were
transplanted. 'Sugarheart' triploid was used along with 'Mickylee' (standard vining habit) and
'SP-1' (reduced foliage, thin-vines) as pollenizers. Harvests of spring trials were made on 8 and
19 June at PSREU and on 21 and 28 June, and 7 July at NFREC. Fall harvests were on 11 and
25 Oct. Insecticides and fungicides were applied as needed to maintain plant health (Olson et al.,
2006). Groupings of honey bee (Apis mellifera L.) hives were placed in close proximity to all
experiments. The GLM procedures of SAS (SAS Institute, Inc., Cary, NC) were used to analyze
the effects of pollenizer cultivar on Sugarheart' watermelon yield and interactions between
pollenizer cultivar and spacing. The GLM procedures were also used to analyze the polynomial
effects of pollenizer to triploid spacing.
Pollenizer cultivar and plant spacing had significant effects on seedless watermelon yield
at Citra and Quincy (P < 0.05) in the spring but not at Quincy in the fall (P > 0.05). Cultivar and
plant spacing had significant effects on fruits per plant (fpp) at both locations in the spring,
however, the significance at Citra for cultivar and plant spacing were P = 0.0983 and P =
0.0633, respectively. Plant spacing had significant linear effects on seedless watermelon yield
and fpp at both locations in the spring (Table 6-1). No significant effects on average watermelon
weight were observed in any of the experiments. Cultivar and plant spacing had no effect on
seedless watermelon yield or yield components at Quincy in the fall. No significant interaction
effects were observed between cultivar and plant spacing. A significant interaction between
location and pollenizer cultivar was detected in the spring therefore locations will be presented
Seedless watermelon yields from plants paired with 'Mickylee' at Citra and Quincy were
24.6 and 33.5 kg/plant, respectively. Yields from plants paired with SP-1' at Citra and Quincy
were 27.7 and 43.3 kg/plant, respectively, which were significantly greater than yields from
plants paired with 'Mickylee'. At Quincy, plants paired with 'Mickylee' produced 3.74 fpp
which was significantly lower than 4.85 fpp produced by plants paired with 'SP-1'. At Citra,
plants paired with 'Mickylee' produced 3.68 fpp which was lower than 4.04 fpp produced by
plants paired with SP-1'; however this was only statistically significant at P = 0.1~0.
Seedless watermelon yields and fpp increased linearly with increased spacing at both
locations in the spring. Yields from plots with 0.2 m in-row spacing's were 22. 1 and 32.0
kg/plant at Citra and Quincy, respectively. Yields from plots with 1.0 m in-row spacing were
30.8 kg/plant at Citra and 47.8 kg/plant at Quincy.
These results appear to indicate that the vigorous growth habit of pollenizer 'Mickylee'
negatively impacted yield of the triploid cultivar Sugarheart compared to the reduced foliage of
pollenizer SP-1'. Seedless watermelon yields (kg/plant) from plants paired with 'Mickylee'
yielded 11.4 and 22.4% less than plants paired with SP-1' at Citra and Quincy, respectively. At
current pollenizer ratios recommended by seed producers, pollenizer competition could affect
two thirds of the triploid plants per hectare. These yield reductions can be expected at all in-row
spacings as there was no interaction between cultivar and spacing. Extrapolations of yield to a
per hectare basis were not performed because it is unclear how many pollenizer plants would be
necessary at these plant populations which ranged from 10,089 plants-hal at the closest
spacing's to 2,017 plants-hal at the widest. There was also no yield data from plants that were
considered to be unaffected by pollenizer competition. During the fall experiment at Quincy,
there was a severe outbreak of gummy stem blight (Didymella bryoniae (Auersw.) Rehm) which
was not controlled by the weekly fungicide applications. This no doubt reduced seedless
watermelon yields which affected the statistical outcome.
The trends in fruit yield and fruit number observed in this study are similar to other reports
in watermelon where increased plant population, and thus increased competition, changes yield
through fruit number and not average fruit mass (Brinen et al., 1979; Duthie et al., 1999a, 1999b;
NeSmith, 1993). Competition studies investigating weed species effect on watermelon yield
have also shown increasing competition lowers watermelon yield as a function of fruit number
and not fruit size (Buker et al., 2003; Monks and Schultheis, 1998).
The results from this proj ect are not in agreement with Freeman (2007) who evaluated the
performance of in-row pollenizers and found no difference in seedless watermelon yield between
plots containing 'Mickylee' or 'SP-1'. The experimental design used was intended to compare
pollenizer cultivars as a function of pollen provided by each cultivar and not its competitive
effect. This study collected yield data by plot which does not provide insight into pollenizer to
triploid competition as some plants were located directly beside pollenizers and some were not.
It is unclear why a significant reduction in yield caused by 'Mickylee' in this study is not in
agreement with the results reported by Freeman (2007).
'Mickylee' is an attractive option as a pollenizer because of the low seed costs. However,
the results of this study indicate that seedless watermelon yields and fruits per plant will
significantly decrease when 'Mickylee' is used as a pollenizer as compared with 'SP-1'
regardless of in-row spacing.
Table 6-1. Influence of pollenizer cultivar and spacing on triploid watermelon yield during 2006.
Citra, FL Spring 2006 Quincy, FL Spring 2006 Quincy, FL Fall 2006
Treatment Fruit Yield Avg. wt. Fruit Yield Avg. wt. Fruit Yield Avg. wt.
(no./plant) (kg/plant) (kg/fruit) (no./plant) (kg/plant) (kg/fruit) (no./plant) (kg/plant) (kg/fruit)
SP-1 4.04 27.7 6.8 4.85 43.4 8.8 3.04 19.0 6.2
Mickylee 3.68 24.5 6.7 3.74 33.5 8.8 2.85 17.4 6.0
Sinfiace*** Ns *** *** Ns Ns Ns Ns
0.2 3.33 22.1 6.5 3.57 32.0 8.9 2.98 18.2 6.0
0.4 3.68 24.9 6.7 3.46 29.7 8.4 2.85 17.5 6.0
0.6 3.92 26.3 6.7 4.52 40.4 8.8 2.86 18.4 6.4
0.8 4.02 26.6 6.7 4.85 43.9 9.0 3.18 19.3 6.0
1.0 4.35 30.8 7.1 5.28 47.8 9.0 2.86 17.6 6.1
Significance Lil L*** Ns L~i L1** Ns Ns Ns Ns
' Non-significant or significant at P I 0.10, 0.05 or 0.01, respectively.
VARIABILITY INT WATERMELON FLOWER ATTRACTIVENESS TO INSECT
In a field that is producing seedless watermelons, there must be both diploid and triploid
plants (Kihara, 1951; Maynard and Elmstrom, 1992). For fruit set to occur in triploid plants,
pollen must be moved from the staminate diploid flower to the pistillate triploid flower. Triploid
and diploid plants produce staminate flowers which bear pollen, however, triploid pollen is non-
viable. It is unlikely that insect pollinators can visually distinguish between the two and foraging
of staminate triploid flowers dilutes the flow of viable pollen within a field. If staminate flowers
produced by the diploid pollenizer are more attractive than triploid staminate flowers, a greater
proportion of viable pollen could be moved by pollinators which may lead to greater
reproductive success in triploid fruits. The floral attractiveness of a pollenizer could also impact
its performance. The objective of this study was to determine the floral attractiveness of
staminate flowers of three pollenizer cultivars and one triploid cultivar.
Materials and Methods
Field experiments were conducted at the North Florida Research and Education Center
(NFREC) in Quincy, FL. during the Spring and Fall of 2006. Soil type at NFREC is Norfolk
loamy sand (fine-loamy, kaolinitic, thermic, Typic Kandiudults). The experimental design was a
randomized complete block with eight replications. Experimental plots were 4.57 m long with
in-row spacing of 0.46 m and between row spacing of 4.9 m. The experiment consisted of two
rows 73.2 m long. On 3 Apr. and 1 Aug. 2006, 4-week old watermelon seedlings were
transplanted into raised beds covered with black polyethylene mulch in the spring and white on
black polyethylene mulch in the fall. Three watermelon plants were transplanted into each plot.
Number of pollinator visitations was recorded for four watermelon cultivars, 'Companion',
'Intruder', 'Mickylee' and 'SP-1'. Three cultivars are diploid pollenizers and one ('Intruder') is
a triploid. Fertilization, irrigation, and pesticide application practices recommended by the
University of Florida Institute of Food and Agricultural Sciences were followed (Olson et al.,
2006). A grouping of two honeybee (Apis mellifera L.) hives was placed near the center of the
Sampling was performed on five occasions in the spring and three in the fall. Sampling
started when plants began to produce staminate and pistillate flowers. On sampling dates,
sampling was initiated at anthesis. Five staminate flowers were chosen in each plot and
visitations from honeybees and bumblebees (Bombus spp. Cresson) were counted for two
minutes. Three to four sampling repetitions were performed on each sampling date and are
referred to as sampling time. Previous research has illustrated an interaction between cultivar
attractiveness and sampling time, for this reason sampling times were kept succinct and sampling
time was considered a main effect. Two individuals recorded visitations in order to keep
repetition time under 45 minutes. Analysis of variance was performed using the GLM
procedures of SAS (SAS Institute, Inc., Cary, NC) to determine significance of main and
interaction effects and Duncan's multiple range test was used for means separation.
Watermelon cultivar and sampling time had a significant effect (P < 0.05) on floral
visitation by insect pollinators on six of eight sampling dates. Cultivar or sampling time did not
influence pollinator visitation on 11 May or 29 Sept. Significant interactions (P < 0.05) between
time and cultivar were detected on 23 May and 22 Sept. Visitation of a diploid cultivar was
significantly greater than the triploid cultivar on six of eight sampling dates.
On 16 May, 'SP-1' had 2.4 visits per plot (vpp) which was significantly greater than
'Mickylee', 'Companion', or 'Intruder' which had 1.1, 1.0, and 0.8 vpp, respectively (Fig. 7-1).
Pollinator visitation of 'Mickylee', 'Companion', or 'Intruder' was not significantly different.
Pollinator visitations to 'Mickylee' and 'SP-1' were 2.9 vpp which was significantly greater than
'Companion' or 'Intruder' which had 1.5 and 1.0 vpp, respectively, on 19 May. An interaction
between sampling time and cultivar occurred on 23 May. During the first sampling time, 'SP-1'
had 4.2 vpp which was not significantly greater than 'Companion' or 'Mickylee' which had 3.0
and 2.75 vpp, respectively (Fig. 7-2). However, all three cultivars had significantly greater
visitation than 'Intruder' at 0.3 vpp. During the second sampling time, 'Mickylee' had 5.5 vpp
which was not significantly different than SP-1' at 2.8 vpp but was significantly greater than
'Intruder' and 'Companion' at 1.5 and 1.3 vpp, respectively. Visitation of 'SP-1', 'Intruder', and
'Companion' were not significantly different. There were no significant differences between
cultivars during the third and fourth sampling times. On 25 May, 'Mickylee' had 2.8 vpp which
was significantly greater than 'SP-1', 'Companion', or 'Intruder' at 1.8, 1.4 and 0.4 vpp,
respectively (Fig. 7-1). SP-1' and Companion' were not significantly different but both had
greater visitation than 'Intruder'.
On 15 Sept, 'SP-1' had 7.7 vpp which was greater than 'Mickylee', 'Intruder' or
'Companion' at 6.1, 5.8, and 3.7 vpp, respectively (Fig. 7-3). 'Mickylee' and 'Intruder' were not
significantly different but both had greater visitation than 'Companion'. An interaction between
sampling time and cultivar occurred on 22 Sept. During the first sampling time, SP-1' had 2.7
vpp which was not greater than 'Mickylee' at 1.8 vpp but was greater than 'Companion' and
'Intruder' at 0.7 and 0.5 vpp, respectively (Fig. 7-4). 'Mickylee', 'Companion', and 'Intruder'
were not significantly different. During the second sampling time, 'SP-1' had 6.8 vpp which was
similar to 'Mickylee' at 5.2 vpp and both had greater visitation than 'Companion' or 'Intruder'
which had 2.3 and 2. 1 vpp, respectively. 'Companion' and 'Intruder' were not significantly
different. There was no significant difference between cultivars during the third sampling time.
On 29 Sept. main effects did not influence pollinator visitation (Fig. 7-3). No data were taken
from 'Companion' on 29 Sept. because it had ceased producing staminate flowers. Complete
floral visitation data are shown in figures 7-1 7-4.
Either 'Mickylee', 'SP-1', or both received greater floral visitation than 'Companion' or
'Intruder'. Previous research as shown that staminate flower production by 'SP-1' is greater than
that of 'Mickylee' or 'Companion' which were similar (Freeman and Olson, 2007). The number
of flowers that were used in sampling was held constant in order to determine the relative
attractiveness of each cultivars staminate flower. This research illustrates differential
attractiveness between cultivars and ploidy levels. Visitation at the whole plant level for SP-1'
could be higher than the other cultivars due to greater numbers of staminate flowers but a
staminate flower produced by SP-1' is not necessarily more attractive than a staminate flower
produced by 'Mickylee'. Other researchers have reported differences in pollinator visitation
between watermelon cultivars and between Citrullus lan2atus and Citrullus colocynthis which
was attributed to nectar sugar concentration (Wolf et al., 1999). It has been shown that floral
volatiles emitted from pollen are the most important close-range cue during foraging by
honeybees, however; visual stimuli are important long-range cues (Pernal and Currie, 2002).
'Companion' has a nearly entire leaf with reduced lobes and produces staminate flowers with
short peduncles. These factors tend to obstruct the view of 'Companion' s staminate flowers
which may be why 'Companion' was generally visited less than 'SP-1' and 'Mickylee'.
A diploid watermelon cultivar was preferred over the triploid on sampling dates where
cultivar affected pollinator visitation. Triploid watermelon plants produce mostly non-viable,
aborted pollen which may be covered with less pollenkitt than viable pollen. Pollenkitt produces
volatiles that are important in foraging decisions of pollinators. These volatiles are an indicator
of pollen reward that is available in a flower and reduced volatile emissions may indicate
reduced reward (Dobson et al., 1996). A reduction in pollenkitt produced in triploid staminate
flowers could represent reduced reward and result in flowers that are less attractive to
pollinators. Previous research has shown that Bombus spp. preferentially foraged potato flowers
that produced viable pollen over ones producing non-viable pollen which may be due to
differences in volatile emissions (Batra, 1993).
When the performance of multiple diploid watermelon pollenizer cultivars were compared,
triploid watermelon yields from plots pollenized by 'Companion' were significantly less than
plots pollenized by 'SP-1' or 'Mickylee'. The lower preference of 'Companion' may lead to less
viable pollen transported by pollinators and thus less fruit on triploid plants. A diploid pollenizer
cultivar with staminate flowers which are more attractive than triploid staminate flowers could
increase the movement of viable pollen within a field and possibly increase seedless watermelon
Figure 7-1. Influence of cultivar on pollinator visitation to staminate watermelon flowers at Quincy, FL., Spring 2006. Visitation
means are to be compared within sampling date. Means followed by the same letter are not significantly different at P I
ina ab O Companion
a M Mickylee
ns I Intruder
Time 1 Time 2 Time 3 Time 4
Figure 7-2. Interaction of cultivar and time on pollinator visitation to staminate watermelon flowers at Quincy, FL., on 23 May, 2006.
Visitation means are to be compared within sampling time. Means followed by the same letter are not significantly
different at P < 0.05.
Figure 7-3. Influence of cultivar on pollinator visitation to staminate watermelon flowers at Quincy, FL., Fall 2006. Visitation means
are to be compared within sampling date. Means followed by the same letter are not significantly different at P I 0.05.
I I IIO Mickylee
LO I SP-1
Time 1 Time 2 Time 3
Figure 7-4. Interaction of cultivar and time on pollinator visitation to staminate watermelon flowers at Quincy, FL., on 23 Sept., 2006.
Visitation means are to be compared within sampling time. Means followed by the same letter are not significantly
different at P < 0.05.
VARIABILITY IN POLLEN PRODUCTION BY DIPLOID WATERMELON POLLENIZERS
Pollen produced by diploid watermelon pollenizers is necessary for fruit set and flesh fill
in associated triploid watermelon crops (Kihara, 1951; Maynard and Elmstrom, 1992). Viable
pollen produced by pollenizers is diluted with non-viable triploid pollen which increases
pollinator visitation rates required by pistillate triploid flowers. It is important that there be
adequate viable diploid pollen moved throughout a field and pollenizers that produce the greatest
amounts of pollen may perform better with respect to triploid watermelon yield. Significant
differences in pollen production have been reported between cultivars of diploid watermelon as
well as in other plant species. The obj ective of this proj ect was to determine the amount of
pollen produced by four pollenizer cultivars to investigate possible effects on pollenizer
Materials and Methods
Experiments were conducted at the North Florida Research and Education Center
(NFREC) in Quincy, FL during the fall of 2006. Soil type at NFREC is Norfolk loamy sand
(fine-loamy, kaolinitic, thermic, Typic Kandiudults). The experimental design was a randomized
complete block with four replications. Experimental plots were 4.57 m long with in-row spacing
of 0.91 m and between row spacing of 2.43 m. On 1 Aug. 2006, 4-week old watermelon
seedlings were transplanted into raised beds covered with white on black polyethylene mulch.
Beds were fumigated with methyl bromide/chloropicrin 67:33 at a rate of 448 kg-hal broadcast.
Three seedlings were planted in each plot. Fertilization, irrigation, and pesticide application
practices recommended by the University of Florida Institute of Food and Agricultural Sciences
were followed (Olson et al., 2006).
Sampling was initiated on 31 Aug. with additional samples taken on 8 Sept. and 14 Sept.
On sampling dates, watermelon flowers were removed before anthesis, placed into plastic
containers and covered to exclude pollinators. Two flowers per cultivar were analyzed from
each replication. Flowers were allowed to open and anther dehiscence was verified before
flowers were processed. Flower petals were excised with a razor blade and peduncles were
removed before flowers were placed into separate vials containing 5 mL of a solution of 70:30
ethanol/ethyl acetate mixture. This solution served to preserve the pollen and also to remove
pollenkitt from the exterior of the pollen grains. The removal of pollenkitt was necessary to
obtain uniform pollen distribution in the liquid media. Sample vials were sonicated in a water-
bath sonicator to remove all pollen from anthers. After the anthers and any other remaining
floral structures were removed, vials were sonicated again to ensure the removal of all pollenkitt.
Three 50 CIL sub-samples were taken from each vial and pollen grains were counted on a grided
microscope slide. Sub-sample values were then extrapolated to obtain total number of pollen
grains per flower.
Pollen sub-sample counts were averaged for each flower. Analysis of variance was
performed using the GLM procedures of SAS (SAS Institute Inc., Cary, N.C.) and means
separation was performed using Duncan's multiple range test.
Pollenizer cultivar and sampling date had significant effects (P < 0.05) on pollen
production per flower. There was no significant (P > 0.05) interaction between cultivar and
sampling date. On 31 Aug., 'Mickylee' produced 49,975 pollen grains per flower (gpf) which
was not significantly greater than 'Companion' at 44,800 gpf but both were greater than 'SP-1'
and 'Jamboree' at 37,813 and 36,700 gpf, respectively (Table 1). Pollen production by
'Companion' was not significantly greater than 'SP-1' or 'Jamboree'. On 8 Sept., 'Companion'
produced 62,275 gpf which was not significantly greater than 'Mickylee' at 60,250 gpf or 'SP-1'
at 53.138 gpf but was greater than 'Jamboree' at 44,975 gpf. Pollen production by 'SP-1' and
'Jamboree' were not significantly different on 8 Sept. On 14 Sept., 'Mickylee' produced 58,900
gpf which was significantly greater than 'Companion', 'Jamboree', and 'SP-1' at 50,825, 48,475,
and 42,250, respectively. Seasonal averages for pollen production are shown inTable 1.
The results of this study illustrate that the amount of pollen produced by diploid
watermelon plants differs by cultivar and also changes throughout the season. Pooled data for
pollen production illustrates that 'Mickylee' and 'Companion' produce significantly greater
amounts of pollen per flower than 'SP-1' or 'Jamboree'. However, the pollen output per flower
may not accurately represent the pollen output per plant. Previous research on flower production
by pollenizers illustrated that SP-1' produced significantly greater numbers of staminate flowers
when compared to 'Mickylee' or 'Companion' (Freeman and Olson, 2007). The increased
production of staminate flowers by SP-1' may result in greater pollen output per plant when
compared with the other cultivars.
The trend observed was that pollen production starts low, increases and then decreases as
the season progresses. These results are in contrast the Stanghellini and Schultheis (2005) who
reported that sampling day did not affect pollen production per flower. However, the duration of
sampling used by Stanghellini and Schultheis (2005) was not stated and flowers sampled over a
shorter time frame may not exhibit the variation observed here. As watermelon plants begin to
branch and produce flowers on secondary terminals, it has been observed that flower size
decreases. The decrease in flower size apparently correlates with the decrease in pollen
production per flower. Although pollen production per flower may decrease as the season
progresses, it is likely that pollen production per plant will increase to a point as the number of
staminate flowers produced increases.
In theory, pollenizers that produce the greatest amount of pollen could outperform other
pollenizers with lower pollen output. Greater pollen output by a pollenizer may also facilitate
the reduction in pollenizer plant numbers used per unit area. However, other experimental
results suggest that pollen production alone is not a reliable indicator of pollenizer performance
as the use of 'SP-1' resulted in significantly greater triploid watermelon yield when compared to
the use of 'Companion'. There are many factors that influence the performance of diploid
watermelon pollenizers and pollen production is likely a strong contributor. However, it is not a
reliable factor on which to judge a pollenizers potential.
Table 8-1. Pollen production by four diploid watermelon pollenizer cultivars at Quincy, FL
during the Fall of 2006.
Pollen grains per flower
Cultivar 31 Aug. 8 Sept. 14 Sept. Combined
Mickylee 49,775 az 60,250 az 58,900 az 56,308 az
Companion 44,800 a b 62,275 a 50,825 b 52,633 a
SP-1 37,813 b 53,138 ab 42,550 c 44,500 b
Jamboree 36,700 b 44,975 b 48,475 bc 43,383 b
z Means followed by the same letter are not significantly different at P I 0.05 by
Duncan's multiple range test.
LIST OF REFERENCES
Adlerz, W.C. 1966. Honey bee visit numbers and watermelon pollination. J. Econ. Entomol.
Ambrose, J.T. 1997. The importance of honey bees in North Carolina. N.C. Coop. Ext. Serv.
Beekeeping Note #1A.
Arndt, G.C., J.L. Rueda, H.M. Kidane-Mariam, and S.J. Peloquin. 1990. Pollen fertility in
relation to open pollinated true seed production in potatoes. Amer. Potato J. 67:499-505.
Arney, M., S.R. Fore, and R. Brancucci. 2006. Watermelon reference book. National
Watermelon Promotion Board. Orlando, Fla.
Ban, D., S. Goreta, and J. Borosic. 2006. Plant spacing and cultivar affect melon growth and
yield components. Scientia Hort. 109:238-243.
Batra, S.W.T. 1993. Male-fertile potato flowers are selectively buzz-pollinated only by Bombus
terricola Kirby in upstate New York. J. Kansas Ent. Soc. 66:252-254.
Brevis, P.A., D.S. NeSmith, H.Y. Wetzstein, D.B. Hausman. 2006. Production and viability of
pollen and pollen-ovule ratios in four rabbiteve blueberry cultivars. J. Amer. Soc. Hort.
Brinen, G.H., S.J. Locascio, and G.W. Elmstrom. 1979. Plant and row spacing, mulch, and
fertilizer rate effects on watermelon production. J. Amer. Soc. Hort. Sci. 104:725-726.
Bryan, H.H. 1966. Effect of plastic mulch on the yield of several vegetable crops in North
Florida. Proc. Fla. State Hort. Soc. 79:139-146.
Buker, R.S., W.M. Stall, S.M. Olson, D.G. Schilling. 2003. Season-long interference of yellow
nutsedge (Cyperus esculentus) with direct-seeded and transplanted watermelon (Citrullus
lan2atus). Weed Tech. 17:751-754.
Butler, C.G. 1951. The importance of perfume in the discovery of food by the worker honeybee
(Apis mellifera L.). Proceedings of the Royal Society of London, Series B, 138:403-413.
Clark, G.A., D.Z. Haman, and F.S. Zazueta. 2005. Injection of chemicals into irrigation systems:
rates, volumes, and injection periods. Univ. Fl. Coop. Ext. Serv. BUL250. 9 Dec. 2006.
(http:.//edi s.ifas.ufl. edu/AE 11 6)
Cockerham, L.E. and G.J. Galleta. 1976. A survey of pollen characteristics in certain Vaccinium
species. J. Amer. Soc. Hort. Sci. 101:671-676.
Connell, J. H. 1990. Apparent versus "real" competition in plants, p. 9-23. In: J.B. Grace and D.
Tilman (eds.) Perspectives on Plant Competition. Academic Press. New York.
Cushman, K.E., D.H. Nagel, T.E. Morgan, P.D. Gerard. 2004. Plant populationaffects pumpkin
yield components. HortTechnology 14:326-331.
Dafni, A., M. Neppi, and E. Pacini. 2005. Pollen and stigma biology, p. 83-147. In A. Dafni.,
P.G. Kevan and B. Husband (eds.) Practical Pollination Biology. Enviroquest, Cambridge,
Daniello, F.J. 2003. Watermelon. Tex. Coop. Ext. Horticulture Crop Guide Series. 10 Dec. 2006.
Delaplane, K.S., and D.F. Mayer. 2000. Crop pollination by bees. CAB Intl., Wallingford, U.K.
Dittmar, P.J., J.R. Schultheis, and D.W. Monks. 2005. Characterization of the growth and
development of commercially available watermelon pollenizers. HortScience 40:872
Dobson, H.E.M. 1988. Survey of pollen and pollenkitt lipids Chemical cues to flower visitors?
American Journal of Botany 75:170-182.
Dobson, H.E.M. 1991. Pollen and flower fragrances in pollination. Acta. Hort. 288:313-320.
Dobson, H.E.M., J. Bergstrom, G. Bergstrom, and I. Groth. 1987. Pollen and flower volatiles in
two Rosa species. Phytochemistry 26:3171-3173
Dobson, H.E.M., I. Groth, and G. Bergstrom. 1996. Pollen advertisement: Chemical contrasts
between whole-flower and pollen odors. American Journal of Botany 83:877-885.
Duthie, J.A., B.W. Roberts, J.V. Edelson, and J.W. Shrefler. 1999a. Plant density-dependent
variation in density, frequency, and size of watermelon fruits. Crop Sci. 39:412-417
Duthie, J.A., J.W. Shrefler, B.W. Roberts, and J.V. Edelson. 1999b. Plant density-dependent
variation in marketable yield, fruit biomass, and marketable fraction in watermelon. Crop
Edelstein, M. and H. Nerson. 2002. Genotype and plant density affect watermelon grown for
seed consumption. HortScience 37:981-983.
Erdem, Y. and A.N. Yuksel. 2003. Yield response of watermelon to irrigation shortage. Sci.
Fiacchino, D.C. and S.A. Walters. 2003. Influence of diploid pollenizer frequencies ontriploid
watermelon quality and yields. HortTechnology 13:58-61.
Firbank, L.G. and A.R. Watkinson. 1990. On the effects of competition: from monocultures to
mixtures, p. 165-192. In: J.B. Grace and D. Tilman (eds.) Perspectives on Plant
Competition. Academic Press. New York.
Fortescue, J.A. and D.W. Turner. 2004. Pollen fertility in M~usa: viability in cultivars grown in
southern Australia. Australian J. of Agr. Res. 55:1085-1091
Franklin, R. 1998. Nutrient management for South Carolina. Clemson Univ. Coop. Ext. Serv.
Extension Circular 476.
Free, J.B. 1993. Insect pollination of crops, 2nd ed. Academic, London.
Freeman, J.H. 2007. Use and effects of diploid pollenizers for triploid watermelon [Citrullus
lanatus (Thunberg) Matsumura & Nakai] production. Univ. FL., Gainesville, PhD Diss.
Freeman, J.H. and S.M. Olson. 2007. Characteristics of watermelon pollenizer cultivars for use
in triploid production. Int. J. Veg. Sci. In Press
von Frisch, K. 1967. The Dance Language and Orientation of Bees. Belknap Press. Cambridge,
Gillaspy, G., H. Bendavid, W. Gruissem. 1993. Fruits a developmental perspective. Plant Cell
Goreta, S., S. Perica, G. Dumicic, L. Bucan, and K. Zanic. 2005. Growth and yield of
watermelon on polyethylene mulch with different spacings and nitrogen rates. HortScience
Hara, H. 1969. The correct author's name for Citrullus lan2atus (Cucurbitaceae). Taxon 18:346-
Harbo, J.R. and R.A. Hoopinger. 1997. Honey bees (Hymenoptera: Apidae) in the United States
that express resistance to Varroa jacobsoni (Mesostigmata: Varroidae). J. Econ. Entomol.
Hidalgo, P.J., C. Galan, and E. Dominguez. 1999. Pollen production in the genus Cupressus.
Hochmuth, G.J, E. Kee, T.K. Hartz, F.J. Dainello, and J.E. Motes. 2001. Cultural management,
p. 78-97. In: D.N. Maynard (ed.) Watermelons characteristics, production and marketing.
ASHS Press. Alexandria, VA.
Hochmuth, G.J., R.C. Hochmuth, and S.M. Olson. 2001. Polyethylene mulching for early
vegetable production in north Florida. Univ. Fla. Coop. Ext. Serv. Circular 805. 9 Dec.
2006. (http://edi s.ifas.ufl. edu/CV213)
Holiday, R. 1960. Plant population and crop yield. Nature 186:22-24
Karst, T. 1990. Seedless watermelon sure to grow. The Grower 23:61.
Kartesz, J.K. 2006. PLANTS profile for Citrullus lan2atus. U.S. Dept. Agric. NRCS. 3 Dec. 2006.
Kihara, H. 1951. Triploid watermelons. Proc. Amer. Soc. Hort. Sci. 58:217-230.
Kultur, F., H.C. Harrison, and J.E. Staub. 2001. Spacing and genotype affect fruit sugar
concentration, yield, and fruit size of muskmelon. HortScience 36:274-278.
Lamont, W.J., Jr. 1993. Plastic mulches for the production of vegetable crops. HortTechnology
Lang, G.A. and E.J. Parrie. 1992. Pollen viability and vigor in hybrid southern highbush
blueberries (Vaccinium corymbosum L. x spp.). HortScience 27:425-427.
Lavi, U., S. Nachman, I. Barucis, D. Gaash, A. Kadman. 1996. The effect of pollen donors and
pollen viability on fruitlet drop in Macadamia integrifolia (Maiden & Betche). Tropic.
Lu, W., J.V. Edelson, J.A. Duthie, and B.W. Roberts. 2003. A comparison ofyield between high
and low intensity management for three watermelon cultivars.HortScience 38:351-356.
Marr, C.W. and K.L.B. Gast. 1991. Reactions by consumers in a farmers' market to prices for
seedless watermelon and ratings of eating quality. HortTechnology 1:105-106.
Maynard, D.N. 2001 An introduction to the watermelon, p. 9-20. In: D.N. Maynard (ed.)
Watermelons characteristics, production and marketing. ASHS Press. Alexandria, VA.
Maynard, D.N. 1989. Triploid watermelon seed orientation affects seedcoat adherence on
emerged cotyledons. HortScience 24:603-604.
Maynard, D.N. 1992. Growing seedless watermelon. Univ. Fla. Coop. Ext. Serv. Fact sheet
HS687. 12 Oct. 2006. (http://edis.ifas.ufl.edu/CV006)
Maynard, D.N. and G.W. Elmstrom. 1989. Evaluation of triploid watermelon cultivars in central
and southwest Florida. Proc. Fla. State Hort. Soc. 102:313-319.
Maynard, D.N. and G.W. Elmstrom. 1992. Triploid watermelon production practices and
varieties. Acta Hort. 318:169-173.
Maynard, E.T. and W.D. Scott. 1998. Plant spacing affects yield of 'Superstar' muskmelon.
McCreight, J.D. 1998. Botany and culture, p. 2-6. In: T.A. Zitter, D.L. Hopkins, and C.E.
Thomas (eds.) Compendium of cucurbit diseases. APS Press. St. Paul, MN.
Monks, D.W. and J.R. Schultheis. 1998. Critical weed free period for large crabgrass (Digitaria
sanguinalis) in transplanted watermelon. Weed Sci. 46:530-532.
Motsenbocker, C.E. and R.A. Arancibia. 2002. In-row spacing influences triploidwatermelon
yield and crop value. HortTechnology 12:437-440.
Nepi, M. and E. Pacini. 1993. Pollination, pollen viability and pistil receptivity in Cucurbita
pepo. Annals of Botany. 72:527-536.
NeSmith, D.S. 1993. Plant spacing influences watermelon yield and yield components.
NeSmith, S. and J. Duval. 2001. Fruit set of triploid watermelon as a function of distance from a
diploid pollenizer. HortScience 36:60-61.
Nettles, V.F. 1963. Planting and mulching studies with cucurbits. Proc. Fla. State Hort. Soc.
Nikkanen, T., T. Aronen, H. Haggman, and M. Venalainen. 2000. Variation in pollen viability
among Picea abies genotypes potential for unequal paternal success. Theor. Appl. Genet.
Olsen, L., R. Hoopinger, and E.C. Martin. 1979. Pollen preferences of honeybees sited on four
cultivated crops. J. Apicultural Res. 18:196-200.
Olson, S.M., E.H. Simonne, W.M. Stall, P.D. Roberts, S.E. Webb, T.G. Taylor, and S.A. Smith.
2006. Cucurbit production in Florida, p. 191-237. In: S. M. Olson and E. H. Simonne
(eds.) Vegetable production handbook for Florida. Univ. Fla. Coop. Ext. Serv. and Vance
Publishing. Lenexa, KS.
Parties, H.K., F. Schnaithmann, and H.H. Geiger. 2005. Pollen viability of Hordeum spp
genotypes with different flowering characteristics. Euphytica 145:229-235.
Pernal, S.F and R.W. Currie. 2002. Discrimination and preferences for pollen-based cues by
foraging honeybees, Apis mellifera L. Animal Behaviour 63:369-390.
Prieto-Baena, J.C., P.J. Hidalgo, E. Dominguez, and C. Galan. 2003. Pollen production in the
Poaceae family. Grana 42: 153-160.
Radosevich, S.R. and M.L. Roush. 1990. The role of competition in agriculture, p. 341-363. In:
J.B. Grace and D. Tilman (eds.) Perspectives on Plant Competition. Academic Press. New
Reiners, S. and D.I.M. Riggs. 1997. Plant spacing and variety affect pumpkin yield and fruit size
but supplemental nitrogen does not. HortScience 32:1037-1039.
Reiners, S. and D.I.M. Riggs. 1999. Plant population affects yield and fruit size of pumpkin.
HortScience 34: 1076-1078.
Rhodes, B., K.B. Gruene, and W.M. Hood. 1997. Honey bees waste time on triploid male
flowers. Cucurbit Genet. Coop. Rpt. 20:45.
Robinson, R.W. and D.S. Decker-Walters. 1997. Cucurbits. CAB Intl., Wallingford, U.K.
Rodriguez-Riano, T. and A. Dafni. 2000. A new procedure to asses pollen viability. Sex. Plant
Rubatzky, V.E. and M. Yamaguchi. 1997. World vegetables, 2nd ed. Chapman & Hall Publ.,
Sanders, D.C. (ed.), J.M. Kemble, E.J. Sikora, R.L. Hassell, G. Miller, T. Keinath, J.K.
Norsworthy, P. Smith, G.E. Boyhan, W.T. Kelley, D.B. Langston, A.S. Culpepper, A.S.
Sparks, J.E. Boudreaux, J.M. Cannon, D.H. Nagel, R.G. Snyder, D. Ingram, M.R.
Williams, B.O. Layton, J.D. Byrd, M.W. Shankle, A. Rankins, R.B. Batts, M.E. Clough,
N.G. Creamer, J.M. Davis, W.R. Jester, D.W. Monks, L.M. Reyes, J.R. Schultheis, A.
Thornton, G.T. Roberson, K.A. Sorensen, J.F. Walgenbach, D.B. Orr, D.R. Tarpy, C.W.
Averre, M.A. Cubeta, K. Ivors, G.J. Holmes, K.M. Jennings, F.J. Louws, D.F. Ritchie,
C.R. Crozier, G.D. Hoyt, D.N. Maynard, R.S. Mylavarape, and H.J. Savoy. 2006.
Vegetable crop guidelines for the southeastern U.S. Vance Publ. Corp., Lincolnshire, Ill.,
in cooperation with the N.C. Veg. Growers Assn., Raleigh, N.C.
Sanders, D.C., J.D. Cure, and J.R. Schultheis. 1999. Yield response of watermelon to planting
density, planting pattern, and polyethylene mulch. HortScience 34: 1221-1223.
Schabenberger, O. and F.J. Pierce. 2002. Contemporary statistical models for the plant and soil
sciences. CRC Press. Boca Raton, Fla.
Smaj strla, A.J., B.J. Bowman, G.A. Clark, D.Z. Haman, D.S. Harrison, F.T. Izuno, D.J. Pitts, and
F.S. Zazueta. 2002. Efficiencies of florida agricultural irrigation systems. Univ. Fl. Coop.
Ext. Serv. Fact Sheet BUL247. 10 Dec. 2006. (http://edis.ifas .ufl.edu/AE 110)
Stanghellini, M. S., J.T. Ambrose, and J.R. Schultheis. 1997. The effects of honey bee and
bumble bee pollination on fruit set and abortion of cucumber and watermelon. Amer. Bee
Stanghellini, M. S., J.T. Ambrose, and J.R. Schultheis. 1998. Seed production in watermelon: a
comparison between two commercially available pollinators. HortScience 33:28-30.
Stanghellini, M.S. and J.R. Schultheis. 2005. Genotypic variability in staminate flower and
pollen grain production of diploid watermelons. HortScience 40:752-755.
Taylor, M.D. and S.J. Locascio. 2004. Blossom-end rot: a calcium deficiency. J. Plant. Nutr.
U. S. Department of Agriculture. 2006. National watermelon report. U. S. Dept. Agr. Agricultural
Marketing Service. Thomasville, Ga. 15 Dec. 2006 (http://www.ams.usda.gov)
U. S. Department of Agriculture. 2005. Quick stats. U. S. Dept. Agr. Natl. Agr. Statistics Serv.
Washington, D.C. 12 Dec. 2006. (http ://www.nass.usda.gov/QuickStats/
CreateFederal All.j sp)
U. S. Department of Agriculture. 2002. Census of agriculture. U. S. Dept. Agr. Fla. Ag. Statistics
Serv. Tallahassee, Fla. 8 Dec. 2006. (http://www. nass.usda.gov/Stati stics by_State/
Vithanage, V. 1991. Effect of different pollen parents on seediness and quality of 'Ellendale'
tangor. Scientia Hort. 48:253-260.
Wallace, H.M. and L.S. Lee. 1999. Pollen source, fruit set and xenia in mandarins. J. Hort. Sci.
Walters, S.A. 2005. Honey bee pollination requirements for triploid watermelon. HortScience
Wehner, T.C. 2006. Watermelon crop information taxonomy, morphology, physiology. Hort.
Sci. Dept. NCSU. 16 Dec. 2006
(http:.//cuke.hort.ncsu. edu/cucurbit/wmelon/wmelonmain. html)
Wiemann, S. (ed.) 1992. The Packer Fresh Trends. Vance Publishing Corp., Lincolnshire, IL
Wolf, S., Y. Lensky, and N. Paldi. 1999. Genetic variability in flower attractiveness to
honeybees (Apis mellifera L.) within the genus Citrullus. HortScience 34:860-86
Joshua Herbert Freeman was born on October 24, 1980 to Linda and Herbert Freeman of
Columbia, S.C. He first became interested in agriculture while working on cattle farm near his
home. After graduating from Ridge View High School he attended Clemson University where
earned a bachelor' s degree in entomology. In the fall of 2002, he enrolled in the University of
Florida and pursued a doctor of plant medicine degree. His career aspirations changed and in the
fall of 2004, he began working on doctorate of philosophy in horticultural science at the
University of Florida.