Cucurbit Genetics Cooperative
Other Crop Genetics Cooperatives
Home About Membership Reports Gene Lists Conferences Links Search NCSU
Cucurbit Genetics Cooperative Report 22:34-37 (article 13) 1999

Triploid and Tetraploid Watermelon (Citrullus lanatus (Thunb.) Matsum. and Nakai) Seed Size and Weight

Marietta Loehrlein

Department of Agriculture, Western Illinois University, Malcomb, IL 61455-1390

Dennis T. Ray

Department of Plant Sciences, the University of Arizona, Tucson, AZ 85721

Introduction. Triploid watermelons are theoretically seedless and are the result of crosses between tetraploid female and a diploid male plants (1). Tetraploid plants are placed in crossing blocks with diploids, and if reproduction is by open-pollination, the resultant seed can be either tetraploid (4n) due to self-pollination, or triploid (3n) from pollination by a diploid. A sorting method was developed by Shimotsuma and Matsumoto (4) to distinguish 3n from 4n watermelon seed based on seed weight and thickness. They found that 3n seed were thinner and lighter than 4n seed, but both were thicker and heavier than diploid seed,

In the U.S. well into the 1990s, 3n watermelon seed was produced predominately by open-pollination, with the resultant mix of 3n and 4n seed separated by size before distribution. In addition, the resulting 3n and 4n plants could be distinguished in the field by the use of a genetic marker for fruit color (3, 5). The diploid parents have dark green (D) fruit, which is dominant to the light green (d) fruit of the tetraploid parents. Triploid plants resulting from this cross will have striped green (ds) fruit .Tetraploid plants resulting from self-pollination will have light green fruit and can be culled from production fields, leaving the triploid plants with striped green fruit.

It was observed by the junior author, by use of the genetic marker system described above, that in some production fields up to 30% of the plants were tetraploid. While Shimotsuma and Matsumoto were able to separate 3n and 4n seed by thickness some 40 years ago, we questioned whether there had been inadvertent selection for thinner 4n seed since crossing blocks were direct-seeded, thus indirectly selecting for earliness in germination, emergence and development in tetraploid plants. In this study we took seed from a 4n x 2n cross to determine if we could separate the seed effectively by size as was done by Shimotsuma and Matsumoto (4).

Methods. Open-pollinated seed from a 4n x 2n cross were obtained from the late Mr. Herb Partridge (Munday Vegetable Growers Co-Op, Munday, TX). Seed were separated by thickness using a hand-held Manostat with accuracy to 0.1 mm. Each millimeter increment between 1.7 and 2.5 mm was considered a group. thirty-three seed from each group were selected, tested for germination, the germinated seed were transplanted to Speedling trays (Speedling, Inc.) in a greenhouse, and then transplanted to a field at The University of Arizona, Campus Agricultural Center (Tucson). Fewer plants were tested in the 1.4 to 1.6 mm and 2.6 to 3.1 mm ranges due to the small numbers of seed in these groups. In the field pollination was allowed to occur naturally, with the plants being visited by a variety of insects, but predominately by honeybees. For the plants that survived to maturity, fruit were scored for ploidy level using the genetic marker system describe above (light green 4n fruit and striped green 3n fruit). A t-test was used to determine whether there were significant (P#0.05) differences in seed thickness between 3n and 4n seed population.

The same seed lot was separated into seed weight groups. Each seed was weighted to the nearest 1.0 mg on a Mettler balance. Seed ranged from 21 to 110 mg, and 18 groups were formed, each consisting of 5 mg increments. Varying numbers of seed were germinated and transplanted as described for seed thickness. For most groups we were able to test 30 seed, but the smaller and larger weight groups had fewer seed available for testing. Ploidy level was determined on surviving plants by the genetic marker system described above, and differences determined utilizing a t-test.

Results. Separation of seed by thickness In the size groups 1.7 to 2.5 mm, germination ranged between 18 and 55%. In the smaller and larger groups germination was 80 to 100%. Both 3n and 4n seed were observed in each size category. Separation of 3n and 4n seed by thickness was not possible by thickness; there were no significant differences between populations.

Separation by seed weight: Germination tended to increase with increased seed size (Table 2). Triploid and 4n seed were found in essentially all weight groups, and there were no significant differences observed between 3n and 4n seed by weight. The mean weights and standard deviations were essentially equal between 3n and 4n seed.

Efforts to separate 3n and 4n seed by size, as was done by Shimotsuma and Matsumoto (4), were unsuccessful in this study. This may be due to the commercial practice of direct-seeding tetraploid parent lines in crossing blocks, resulting in indirect selection for earliness in germination, emergence and fruit production. This inadvertent selection may have resulted in thinner 4n seed coats in these parent lines. In fact, thinner seed coats in 4n seed might be an advantage in that the resulting 3n seed would also have thinner seed coats. This might help with other cultural problems in triploid watermelons production, such as low germination level and slow germination rates (2).

Most seedless watermelon lines today are produced by controlled hand-pollinations, thus separation of 3n and 4n seed is not necessary. If lines are produced by open pollination it may be more beneficial to develop a genetic marker system for some seedling characteristic rather than a fruit characteristic that requires the expenditure of resources toward plants producing non-marketable fruit. If triploid plants are being transplanted into a production field, the development of a seedling selection system would allow for the elimination of unwanted plants before planting.

Table 1. Germination percentage and ploidy level of seed produced from a tetraploid x diploid cross and separated by thickness.

Size group (mm)
No. of seed
Germination (5)
No. of seed scored as
3n
4n
1.4 - 1.6
15
80
4
6
1.7
33
55
4
7
1.8
33
42
8
3
1.9
33
39
4
5
2.0
33
27
4
9
2.1
33
30
4
7
2.2
33
18
1
5
2.3
33
21
4
5
2.4
33
30
3
4
2.5
33
30
5
5
2.6 - 2.8
15
80
5
2
2.9 - 3.0
6
100
1
4
Total no.
47
62
Mean thickness
2.1 + 0.4
2.1 + 0.4z
mm + s.d.
z t-value = 0.00, P0.5 = 1.98, P > 0.90

Table 2. Germination percentage and ploidy level of seed produced from a tetraploid x diploid cross and separated by weight.

No. of seed
Germination (%)
No. of seed scored as
3n
4n
21-25
3
0
0
0
26-30
5
40
1
0
31-35
13
69
3
2
36-40
27
74
7
6
41-45
30
63
6
4
46-50
30
57
6
2
51-55
30
57
5
6
56-60
30
67
4
5
61-65
30
83
6
6
66-70
30
93
4
7
71-75
30
97
9
4
76-80
30
100
9
5
81-85
30
100
4
9
86-90
30
100
3
2
91-95
30
100
7
4
96-100
30
100
5
6
101-105
24
100
4
3
106-110
3
67
1
1
Total no.
84
72
   
Mean weight
67 + 21.4
69.5 + 20.4Z
(mg + s.d.)
z t-value = 0.50, P 0.5 = 1.96, P > 0.90

Literature Cited

  1. Kihara, H. 1951. Triploid watermelons. Proc. Amer. Soc. Hort Sci. 58:217-230.
  2. Partridge, H. 1979. Growing seedless hybrid (triploid) watermelons by direct seeding. Cucurbit Genetic Coop. Rpt. 2:27.
  3. Rhodes, B. and X. Zhang. 1995. Gene list for watermelon. Cucurbit Genetics Coop. Rpt. 18:69-84.
  4. Shimotsuma, M. and K. Matsuma. 1957. Comparative studies on the morphology of polyploid watermelon seeds. Seiken Ziho 8:67-74.
  5. Wall, J.R. 1960. Use of a marker gene in producing triploid watermelons. Proc. Amer. Soc. Hort. Sci. 76:577-581.
Home About Membership Reports Gene Lists Conferences Links Search NCSU
Department of Horticultural Science Box 7609North Carolina State UniversityRaleigh, NC 27695-7609919-515-5363
Page citation: Wehner, T.C., Cucurbit Genetics Cooperative;
Created by T.C. Wehner and T. Ng, 1 June 2005; design by C.T. Glenn;
send questions to T.C. Wehner; last revised on 15 December, 2009