Cucurbit Genetics Cooperative
Other Crop Genetics Cooperatives
Home About Membership Reports Gene Lists Conferences Links Search NCSU
Cucurbit Genetics Cooperative Report 20:41-44 (article 18) 1997

Discriminating Fruit from 3n and 4n Vines in Progeny of an Open Pollinated 4n x 2n Melon

Jeffrey Adelberg

Department of Horticulture, Clemson University, Clemson SC 29634

Triploid melon are not seedless (1,3). Observation that triploid fruit were firmer at full slip indicates there may be post-harvest advantages to working with triploid fruit. There were also observations made that polyploidy melon may have better tolerance to foliar diseases. It is difficult to make seed from hand pollination of tetraploid melons with diploid pollen (Adelberg and Nugent, unpublished data). The following study attempts to lay a foundation for working with mixed 3n and 4n populations in the filed, as a means to identify genotypes pairs to make triploid hybrids and conduct post-harvest evaluation of fruit.

4n x 2n progeny are categorized as tetraploid or triploid by viability of seed in progeny fruit. Fruit with less than 10% viable seed will e classified as triploid, greater than 10% are classified as tetraploid. This classification was applied, retroactively, to more common screening techniques such as fruit shape, scar size and fruit size. Other classification schemes to distinguish ploidy, such as seedling cotyledon shape, pollen morphology, and a new index termed “fatxcar” (scar size/shape) were verified or refuted using seed viability data.

Materials and Methods. Experiments were conducted on an autoploid series (2n, 3n, 4n) of B-line and ‘Planter’s Jumbo’ (P. Nugent, USDA Vegetable Lab, Charleston SC) in a completely randomized design. Experiments for heterosis and ploidy effects between ‘Planter’s Jumbo’ (PJ) and ‘Green Ice’ (GI) were conducted using GI 2n and 4n PJ, diploid hybrid PJ x GI [2n (PJ x GI)], and triploid hybrid PJ x GI [3n (4n PJ x GI)].

Tetraploid melon were bee-pollinated when surrounded by flowering diploid plants (open pollinated [OP] 4n x 2n). This seed, and seed of tetraploid parents, were germinated at 30C in vertical rolls of germination paper. After 4 days, seedlings were transplanted to a Todd 200 planter in the greenhouse. Diploid seed were directly germinated in a Todd 200 planter.

Seedlings of 4n x 2n were screened as tetraploid or triploid at seedling stage based on cotyledon shape. Broad, rounded cotyledons were assumed to be from tetraploid seedlings and narrow, mottled cotyledons were assumed to be from triploid seedlings.

Transplants were placed in the field the first week of May 1996. The field had been bedded, and a subsoil trench was laid with palletized chicken manure ( 4-2-2), on center, as the only fertilizer application. Transplants were provided with trickle irrigation at 0.76m spacing, by planting at emitter sites. Rows were 1.83m apart, and 3 subplots of 12-20 hills (depending on germination) were repeated in the field design. Field design also placed triploid or hybrids adjacent to their diploid and tetraploid parents.

Pollen was collected on June mornings and fixed in refrigerated 70% ethanol and examined at 100X for determination of ploidy, based on uniformity, shape and size of starch filled grains. Estimates were made of the number of triploid plants in (4n x 2n) populations.

Fruit were harvested at skip twice a week, July 9 through August 9. Fruit were weighed, and their circumference measured lengthwise with string (scar to scar) and equatorially (width). Fruit shape ratio was quantitatively described by dividing longitudinal and equatorial circumference. This was directly compared to an invasive method of slicing fruit and dividing internal diameter dimensions. Blossom scar was measured with a string. Soluble solid (% brix) was measured with a hand-held refractometer. Seed were removed and dried (all fruit had many seed). Fifty dried seed per fruit were imbibed on vertically rolled germination paper and percentage germination was recorded after 4 days at 30C.

An index, “fatxcar,” was calculated, where scar size was divided by fruit shape ratios. The intention was to emphasized the scar size of oblate fruit and shrink scar sized of long fruit, so that triploids and tetraploids can be separated in field evaluations using non-invasive means.

Data were tabulated per fruit. Putative triploids and tetraploids from 3n and 4n seeds were separated based on an arbitrary criterion that tetraploids germinate greater than 10% and triploid germinate less than 10%. This criterion was established prior to the analysis.

Results. Screening ploidy prior to fruit set. When 4n x 2n seedlings were separated based on cotyledon shape, 29% of the putative 3x seedlings were triploid from the germination test. This figure was appropriate for both PJ and B. For PJ and B, 19% and 35% respectively, were triploid in progeny that were screened as tetraploids. Cotyledon shape was not useful to distinguish 3n and 4n progeny at seedling stage.

Pollen morphology showed triploid plots of B to contain 27% triploid plants, and PJ contained 48% triploid plants, as compared to 29%, from the germination test. Single hill harvests would be necessary to draw a tighter correlation but were not conducted. Pollen morphology as a field screen of a mixed population prior to harvest would be logistically difficult.

Fruit of Autoploid Series (Table 1).

  1. Autotriploid fruit is intermediate in size between diploid parent (large) and tetraploids parent (small) for both genotypes.
  2. Soluble solids was not improved in either autotriploid. Effects are specific for genotype and should not be generalized for ploidy.
  3. Autotriploids take longer to mature than either diploids or tetraploids.
  4. Fruit shape ratios can be quantified by non-destructive means (circumference with string) and was used to describe ploidy differences. 3n and 4n were not distinguishable. The invasive method of internal diameters did not give any more precision in separating 3n and 4n (data not shown). Differences between shape determined by circumference and internal diameter, showed fruit was estimated to be longer by circumference. Diploids were estimated 2% longer and polyploids were estimated 8 and 9% longer, respectively. There is a 75% chance that this error is randomly applied to tetraploids and triploids (paired t-test), and therefore using circumference, as opposed to internal diameter, would not bias these distinctions.
  5. Blossom end scar is intermediate in autotriploids between small (diploid) and large (tetraploid) scar sizes. Distinction of3n and 4n, made by blossom end scar size, had 5% chance of error (paired t-test). Increasing this difference by fatxcar index, reduces the error in 3n and 4n distinction, to 3% (paired t-test). This small math trick reduced the probability of type I errors by 40% in this data set.
  6. Yield was not estimated due to the low percentage of 3n in mixed progeny rows.

Heterosis and Ploidy Effects on Fruit Qualities (Table 2).

  1. Diploid PJ were larger than diploid GI. Diploid PJ x GI were the same size as the larger parent. Tetraploid PJ is the same size as GI. The triploid hybrid PJ x GI and the autotriploid PJ were the same size as their parents. Heterosis was not observed for fruit size.
  2. Soluble solids in diploid PJ x GI is higher than either parent, a heterotic response. The triploid hybrid were the same as parents, similar to triploid inbred. Heterosis at 3n Ploidy was not observed.
  3. Diploid PJ x GI were earlier in maturity than diploid parents, a heterotic response. Triploid hybrid and inbred are likely to mature after their parents. Heterosis was not observed at higher Ploidy.
  4. Estimated shape can distinguish diploids from polyploids, regardless of background. 3n and 4n were not distinguished. Internal diameter did not make better distinctions (data not shown).
  5. Scar will only distinguish diploids from polyploids. Losing the ability to separate 3n PJ from 4n PJ was an artifact of fewer internal replications in the one dimensional analysis, compared to factorial in Table 1. This casts doubt on utility of scar character in simple field experiments. Heterosis possibly creates larger scars in diploid hybrids compared to diploid parents, which can only become a worse problem for discrimination of triploid and tetraploid.
  6. Fatxcar index does not help the situation in 5. Yield was not estimated due to the low percentage of 3n in mixed progeny rows.

Conclusions. June and July 1996 were atypically dry with two brief periods of light rain, and foliar disease was not observed on any plots. Post-harvest evaluations of triploids melon would require separating 3n and 4n fruit in the field at time of harvest. We have not found a satisfactory way to make this distinction. Two options we will follow in future work are: 1) conduct field trials on methods of hand-pollination techniques for tetraploid females with diploid parents, and 2) create new tetraploid populations, with recessive fruit characters to diploid parent (e.g., smooth skinned, white fleshed tetraploids for hybridization with netted orange diploids) for easy field discrimination of 3n and 4n.

Heterosis does not occur in all parental combinations. In this study we had parents which were heterotic as diploids, but were not heterotic at higher ploidy. The fruit of diploid hybrid PJ x GI was sometimes bicolor. Salmon (orange) is generally thought to be dominant to green. We have made previous observations in other triploid hybrids, where an orange fleshed tetraploid from a diploid hybrid of orange and green, when backcrossed to the green parent, produces green triploid progeny (2,3). This poses a dilemma when proposing the easy field discrimination of 3n and 4n by fruit characters.

Table 1. Progeny of OP audoploid 4n x 2n melon, with diploid and tetraploid parents, in Kiwi field at Musser Farm,. Oconee county, SC, 1996. Less than 10% germination on paper in an incubator at 30 C was criterion used to distinguish 4n from 3n.

Ploidy and Genotypez
Mean fruit weight (kg)
Soluble solids (% brix)
Days to harvest (post transplant)
Fruit shapey
Blossom scar (mm)
Fatxcarx (Scar/shape)
Germination (%)
2n PJ
3n PJ
4n PJ
2n B
3n B
4n B
Probability > F from Analysis of Variance
Ploidy (P)
Genotype (G)


P x G





z PJ='Planter's Jumbo', B=B-line.
y  longitudinal circumference/equitorial circumference.
x an index calculated by scr size/fruit shape ratios  

Table 2. Progeny of hybrid and polyploid hybrid melon from 4n x 2n o.p. cross, grown in Kiwifield at Musser Farm, Oconee County, SC, 1996. Less than 10% germination on paper in incubator at 30 C was criterion used to distinguish 4n from 3n.

Ploidy and Genotypez
Mean fruit weight (kg)
Soluble solids (% brix)
Days to harvest (post transplant)
Fruit shapey
Blossom scar (mm)
Fatxcarx (Scar/shape)
2n PJ
2n GI
2n (PH x GI)
3n (PJ x GI)
3n PJ
4n PJ

z PJ='Planter's Jumbo', GI='Green Ice'.
y longitudinal circumference/equitorial circumference.
x an index calculated by scar size/fruit shape ratios.
w a,b,c, denote differences between means within columns as determined by Tukey's Honestly Significant Difference test, with slpha = 0.05.  

Literature Cited

  1. Adelberg, J., P. Nugent, X. Zhang, B. Rhodes, H. Shorupska. 1995. Fertility and fruit characters of hybrid triploid melon. Breeding Sci. 45:37-43.
  2. Adelberg, J., and M. Takagaki. 1995. Observations of triploid melon under intensive greenhouse management in Japanese hydroponic production system. In: G. Lester and J. Dunlap (eds.), Proceedings of Cucurbitaceae '94: Evaluation and Enhancement of Cucurbit Germplasm, pg. 128-135. Gateway Publishing, Edinburg TX.
  3. Adelberg, J. 1993. Tetraploid melon from tissue culture and their triploid hybrids. Ph.D. dissertation. Clemson Univ., Clemson SC.
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