First field generation in Maine (2014).

On 30 May 2014, these tubers were planted contiguously in the field on Chapman Farm (lat. 46°37′15.32″ N, long. 68°6′42.02″ W), in Mapleton, ME, on a Caribou gravelly loam soil (fine-loamy, mixed, frigid Typic Haplorthod) at an elevation of 200 m above sea level. Plants were spaced 0.91 m apart in the row to allow separation at harvest by a one-row digger. Rows were also spaced 0.91 m apart. Common commercial practices regarding fertilization and pesticide application were followed. No irrigation was available. At harvest, teams of breeders and agronomists went through the field and made selections based on horticultural characteristics of the tubers, such as smoothness, lack of sprouting, and size. Tubers were rated for skin color, skin texture, shape, eye depth, overall appearance, and flesh color according to the rating scale in Table 1. Tubers from each clone selected were put into red mesh bags, labeled with the family identification, and then all bags of tubers selected from each family (i.e., female parent) were put together into a wooden box. From 14,035 single hills, 689 clones were initially selected.

Table 1.

Rating scale for tuber characteristics.

View Table

The 89 boxes (one for each family) were placed in storage. Beginning 26 Sept. 2014 and weekly thereafter, the bags were examined for tuber sprouting. Bags with sprouted tubers were discarded until only five bags remained for each family. In some cases, the families had fewer than five selections (i.e., bags) to start with, and these were saved. This was done so to initiate a recurrent selection scheme to lengthen tuber dormancy in this material (not part of this study). In all, 357 clones were saved for additional evaluation. Tubers from the saved clones were divided equally three ways: up to four tubers, where available, were sent to Florida and New Mexico, and kept in Maine.

First year field evaluations in the south (2015).

Field evaluations were conducted in the spring of 2015 at the University of Florida, Hastings Agricultural Extension Center in Hastings, FL (lat. 29°41′27.58″ N, long. 81°26′31.37″ W) at an elevation of 8 m above sea level. The soil at the field trial location is classified as sandy, siliceous hyperthermic Arenic Ochraqualf belonging to the Ellzey series. On 19 Dec. 2014, the trial area was fumigated with Telone® [1, 3-Dichloropropene (63.3%) + Chloropicrin (34.7%) C-35; Dow AgroScience; Indianapolis, IN] at a rate of 103 L·ha⁻¹ and raised rows (0.35 m height) were formed with 1.0 m spacing between rows. On 6 Feb. 2015, tubers were planted in a completely randomized design with one replication on a 0.20 m in-row spacing. Irrigation was supplied via seepage irrigation. Fertilizer practices follow the guidelines of the Vegetable Production Handbook for Florida (Zotarelli et al., 2012). Plants were vine-killed on 18 May 2015 with a single application of the chemical desiccant glufosinate ammonium (Rely® 280 Herbicide; Bayer Crop Science, Research Triangle Park, NC) applied at a rate of 1,535 mL·ha⁻¹. Tubers were mechanically harvested using a one-row digger on 1 June 2015. After harvest, tubers were immediately washed and graded into marketable and culls (e.g., mechanical injuries, rotten). A total of 12 marketable tubers were put into mesh bags and maintained a dark, cold room at 10 °C and 80% to 85% relative humidity. Tubers were then weekly evaluated for sprouting.

The New Mexico trials were conducted at New Mexico State University’s Los Lunas Agricultural Science Center (lat. 34°46′9.63″ N, long. 106°45′33.09″ W), located about 43 km south of Albuquerque, NM at an elevation of 1475 m above sea level. Although potatoes are a major crop in the northwest area of New Mexico, they are not widely grown in the central to southern part of the state. However, there is great interest in the crop, particularly if cultivars that are high performers in the warmer, southern areas can be identified, providing impetus for testing these clones in Los Lunas. In 2015, four plants per clone were planted per plot at a spacing of 0.23 m between plants. Spacing between rows was 0.91 m. Plots were furrow irrigated and fertilized in accordance with common practices for other Solanaceous crops grown in the area. Clones were planted on 2 Apr. and harvested on 6 July 2015. For harvest, the tubers were dug up with a one-row, tractor-propelled potato digger. All tubers were picked up and bagged by hand to obtain a total harvest weight for each plot. A 10-tuber sub-sample was taken for each line and placed in storage for long-term sprouting evaluation.

At both locations, tubers were evaluated for skin color, skin texture, shape, eye depth, appearance, and flesh color. Correlations were calculated among locations (ME 2014, FL 2015, NM 2015) for the tuber traits evaluated. The numbers of clones selected at two or more locations were counted; and the percent agreement, determined as the number of clones both locations agreed to save or agreed to discard, among locations was calculated.

Second field generation in Maine (2015).

Clones were planted in a completely randomized design on Chapman Farm, Mapleton, ME on 8 June 2015. Plots consisted of four plants per clone. Plants were spaced 0.23 m within the row, with 0.96 m between plots in a row, in rows spaced 0.91 m apart. Common commercial practices were once again used. Plots were harvested on 22 Sept. 2015. Following harvest, 135 clones were selected to continue evaluation; however, all clones were harvested because ME was the seed source for all locations.

Second year field evaluations in the south (2016).

The second year field evaluation was conducted in the same area in Hastings, FL. Fertilizer, fumigant, and herbicides were applied at the same rate as the previous year. The experimental area was fumigated on 21 Dec. 2015 and the trial was planted on 2 Feb. 2016. Clones were planted in a randomized complete block design with two replications. In-row plant spacing was 0.30 m and spacing between rows was 1.0 m. Plants were vine-killed on 26 Apr. 2016 and harvested on 10 May 2016. Tuber evaluation and storage conditions were identical to the 2015 season.

The trial was repeated in Los Lunas, NM in 2016; however, additional material was available, so most plots were planted with five tubers, and duplicate plots were planted for most accessions. Similar field management and harvest protocols to the previous year’s test were implemented. In 2016, clones were planted on 23 Mar. and harvested on 30 June.

Results from FL 2016, NM 2016, and ME 2015 were compiled and used to narrow the number of selections for additional evaluation down to 82 clones, based on their being selected at two or three locations.

Third year field generation in Maine (2016).

Clones were planted in a randomized complete block design with two replications on 2 June 2016. A plot consisted of five plants of a clone at the same in-row and between-row spacing as in 2015. Common commercial practices were once again used. Plots were harvested on 29 Sept. 2016. The number of tubers and total weight of tubers per plot was recorded.

Statistical analysis.

Correlations among traits across locations were calculated using the correlation procedure in SAS, version 9.2 (SAS Institute, Cary, NC). Yield data from each location were analyzed separately using the mixed models procedure in SAS. Clones were considered fixed effects and replications were considered random effects. The distribution of residuals from each location was fairly normal, but the error variances among locations were heterogeneous. Therefore, the data were not combined for further analysis. Least square means were generated for each clone and ranked within a location from highest (ranking = 1) to lowest.

First field generation in Maine (2014).

From 14,035 clones initially planted in ME 2014, 689 (4.9% of the total) were initially selected. This represents a higher percentage than normal, which may be due to the number of selectors, the large amount of variation present within a family, the lack of defined selection criteria for the intense yellow-flesh creamer market, and/or the unique nature of the germplasm. Three two-person teams (inexperienced selectors paired with experienced selectors) went through the field making selections, bagging, and labeling clones. Unlike tetraploid families, which usually result from controlled crosses within market classes, variation for all tuber characteristics was much greater in the diploid open-pollinated families. For those with selection experience limited to traditional tetraploid market classes, the variation in tuber size, color, shape, and other characteristics in this diploid population, coupled with a lack of defined selection criteria for this new marketing niche, presented a wide array of possibilities. In the U.S. Department of Agriculture–Agricultural Research Service’s (USDA-ARS) tetraploid potato breeding populations, the normal selection percentage in the first field generation ranges from 2% to 3% (Haynes et al., 2014).

Nine different skin colors were represented in the 689 original selections: 249 (36.1%) were yellow, 158 (22.9%) were red, 116 (16.8%) were tan, 78 (11.3%) were pink, 35 (5.1%) were buff, 32 (4.6%) were purple, 6 (0.9%) were orange, 4 (0.6%) were brown, and 11 (1.6%) were bicolored (combinations of red or pink and tan or yellow).

Overall skin texture rating of the 689 original selections was 7.3 and ranged from 4 to 9. Only one clone had a lightly russeted texture, five had a netted texture, and 76 were slightly netted. The rest were moderately smooth to very smooth. Because creamer potatoes are meant to be eaten either whole or halved with the skins on, smoother textures are more desirable.

Overall shape rating of the 689 original selections was 3.7 and ranged from 1 to 8. Oblong shapes predominated (ratings = 4 or 5, 40% of the clones), followed by round shapes (ratings = 1 or 2, 31% of the clones) and long shapes (ratings = 7 or 8, 6% of the clones). The rest of the clones had variable shapes: 16% of the clones were round to oblong (rating = 3) and 8% were oblong to long.

Overall eye depth rating of the 689 original selections was 5.2 and ranged from 3 to 9. Intermediate eye depth predominated (rating = 5, 49% of the clones), followed by moderately shallow (rating = 7, 24% of the clones), deep (rating = 3, 23% of the clones), and very shallow (rating = 9, 3% of the clones).

Overall tuber appearance rating of the 689 original selections was 5.2 and ranged from 1 to 9. Fair (rating = 5) appearance predominated in 59% of the clones, 24% of the clones were rated with a good appearance (rating = 7), and 1% of the clones were rated with an excellent appearance (rating = 9). The rest of the clones (16%) were rated poor or very poor in appearance. The poor or very poor ratings were somewhat surprising given the selection criteria. However, they were probably representative of the different selection pressure applied by the selectors (some of whom had never selected within the diploid population before) and the postharvest evaluator, who had more than ten years’ experience rating diploid populations.

Overall flesh color averaged 3.8 and ranged from white to dark yellow–orange: 5.7% of the clones were white fleshed, 3.8% were cream fleshed, 6.8% were light-yellow fleshed, 71.6% were yellow fleshed, and 12.2% were dark-yellow to orange fleshed. Among the 44 orange-fleshed families, flesh color ranged from white to orange; only 30 of the 449 clones selected from the orange-fleshed breeding nursery scored as orange fleshed, when they were all expected to be orange fleshed. Because foreign pollen in the orange-fleshed open-pollinated seed nursery was expected to be negligible due to the distance from the yellow-fleshed open-pollinated seed nursey, it is more likely that many of the so-called orange-fleshed parents that went into that open-pollinated seed nursery were not recessive for the Zep1 allele, and hence, those parents were either misclassified for flesh color or had high levels of zeaxanthin despite not having two copies of the Zep1 allele. Levels of zeaxanthin can vary greatly among individual clones in this population (Lu et al., 2001). High levels of zeaxanthin impart an orange color to the flesh (Brown et al., 1993).

Discarding the clones with the shortest dormancy in each family had little or no effect on the mean or range of values for any of the traits measured: skin color, skin texture, tuber shape, tuber eye depth, tuber appearance, or tuber flesh color.

All but three selections had sprouted by 10 Dec. 2014 (78 d after harvest). Data ceased to be recorded after this time because seed was packed for shipping to FL and NM. The average number of days to sprouting for all 689 selections was 20 d. Within 8 d of harvest, 182 selections had sprouted. Fifteen days after harvest, another 170 selections had sprouted. After discarding the clones with the shortest dormancy in each family, the average number of days to sprouting was only 25 d. Short dormancy is a characteristic of phu, and longer dormancy of stn (Dodds and Paxman, 1962). Thompson et al. (1980) reported that the mean length of dormancy for phu and stn in their study was 20 and 75 d after harvest, respectively. Progeny from their parents had a mean length of dormancy of 47 d, exactly midway between the two groups. They also found that additive genetic variance for tuber dormancy predominated; the narrow-sense heritability estimate for dormancy was 0.73. No dormancy (i.e., tubers sprouting at harvest) was one of the negative selection criteria used in this study.

First year field evaluations in the south (FL 2015, NM 2015).

Three hundred fifty-seven clones undergoing evaluation for the intense yellow-flesh creamer market were evaluated at two southern locations: Hastings, FL and Los Lunas, NM. These locations represented potential production sites with proximity to large Hispanic markets. Ratings in FL and NM were on a one- to four-plant plot basis, with seed tubers sourced from the ME 2014 single-plant harvest.

In general, tuber skin colors were similar across all three environments (Table 2). A common phenomenon in commercial red-skinned potato cultivars is the tendency for the red-skin color to fade to pink under higher temperature production environments. One hundred forty-four clones had anthocyanin (pink, red, purple) -colored skin. Among these 144 clones, the two southern locations had slightly higher average skin color (FL = 2.6, NM = 2.8) than ME (2.2), indicating a general tendency toward lighter skin color under higher temperature growing conditions, as expected. The correlations involving skin color among the locations were the highest correlations observed for any of the traits scored, ranging from 0.77 to 0.85.

Table 2.

Average ratings for tuber skin color, texture, shape, eye depth, overall appearance, and flesh color for 357 clones evaluated in Hastings, FL in 2015; Los Lunas, NM in 2015; and Presque Isle, ME in 2014; and the correlations among traits at two locations.

View Table

Tuber skin texture was more heavily netted in the southern states than in ME (Table 2). The development of skin russeting in normally smooth-skin varieties is known to occur (Yamaguchi et al., 1964) and results from an accumulation of periderm/skin layers in response to high soil temperatures (Ginzberg et al., 2009). Thus, the coarser texture in the southern states was not surprising. The correlations involving skin texture among locations were some of the lowest correlations observed for any of the traits scored, ranging from −0.03 to 0.11. There was less variation for skin texture in ME than in either FL or NM, as evidenced by the fact that the variance for skin texture in ME (σ² = 0.59) was less than one-third that in FL (σ² = 3.12) or NM (σ² = 2.00). This finding indicates that there is genetic variation in this germplasm for skin texture response to higher temperatures.

Tuber shape was generally round to oblong, although there were some clones with a long tuber shape (Table 2). Shape averaged across locations gave 73 round (average shape ≤ 2), 239 round to oblong (2 < average shape ≤ 5), and 45 oblong to long (average shape > 5). Several studies have found that round tuber shape is dominant over long (de Jong and Burns, 1993; de Jong and Rowe, 1972), and this finding has been verified by genetic mapping (Lindqvist-Kreuze et al., 2015; Vaneck et al., 1994) where the round gene, designated Ro, has been mapped to chromosome 10. The correlations involving tuber shape among the locations were moderate, ranging from 0.50 to 0.67. However, the variation for tuber shape was similar across all locations.

Tuber eye depth was generally intermediate to shallow, the average ranging from 5 to 7 (Table 2). Eye depth tended to be rated deeper in ME than in either FL or NM. Whether this rating is due to evaluator differences or represents an environmental component is unknown. Only one study was found in the literature that suggested that eye depth could be efficiently selected in two vastly different environments (Johansen et al., 1965); but details were lacking in the abstract, and the full report was never published. Several studies have shown that deep eye depth is dominant to shallow (Li et al., 2005; Mosquera et al., 2009). The correlations involving tuber eye depth among the locations were low to moderate, ranging from 0.12 to 0.30. In general, the diploid species phu and stn have deep eyes, so the moderate-to-shallow eyes recorded here are evidence of the success in selecting against deep eyes in the North Carolina State University (from 1966 to 1986) and USDA–ARS (from 1987 to present) breeding programs over the years.

Average tuber appearance ratings were similar across locations, ranging from 5.0 to 5.4 (Table 2). However, there was no correlation among the ratings at any of the locations. Neele et al. (1991) reported that selection on plant appearance (based on visual scoring of foliage, stolons, and tubers) was unreliable in the first two clonal generations due to significant genotype × environment interactions and high error variances. ME ratings were based on a one-hill plot, whereas FL and NM ratings were based on a maximum of four-hill plots.

Average flesh color ratings evaluated at harvest were similar across locations, ranging from 3.4 to 3.7 (Table 2). Average ratings were slightly darker in ME than in FL or NM. This finding may be due to evaluator differences or lighting differences while scoring. ME tubers were scored under fluorescent light in storage. In FL, tubers were scored in a barn under fluorescent lamp light. Genotype × environment interactions for flesh color, as measured objectively by a colorimeter, are known to be statistically significant, but of little practical significance (Haynes et al., 1996). Yellow-flesh intensity in potato tubers is positively correlated with total carotenoid content (r = 0.83) (Lu et al., 2001), so many of these selections are expected to have higher levels of carotenoids than what is found in commercial tetraploid cultivars.

The average number of days to sprouting for all clones harvested in FL was 24 d. This number is in close agreement with the ME 2014 first field generation. However, the range in FL, from 16 to 152 + days, was more than twice as great as in Maine in 2014. Discounting the clones that had not sprouted by the end of the observation period, the correlation between days to sprouting in ME 2014 and FL 2015 was 0.31 (P < 0.0001). The moderately low correlation was not surprising, because warmer temperature growing environments are known to shorten dormancy (Susnoschi, 1981). What was surprising was to find that five clones had a dormancy greater than 152 d. NM did not evaluate for days to sprouting in 2015.

FL selected 101 clones to continue evaluating, NM selected 28, and of those, only four clones were selected at all three locations.

Second field generation in Maine (2015).

Seven clones were lost due to rotting in storage between the first and second field generation in ME. A far larger loss occurred in the field in 2015 when 123 clones were lost due to a severe soft rot infestation brought on by an extremely wet growing season.

As expected, correlations of tuber skin color (r = 0.67, P < 0.001) and shape (r = 0.78, P < 0.001) between the first and second field generation in ME 2014 and 2015 were high (Table 3). Correlationsof skin texture (r = 0.22, P < 0.001), eye depth (r = 0.21, P < 0.01), and tuber appearance (r = 0.25, P < 0.001) between the two years were low. The correlation of days to sprouting between ME 2014 and ME 2015 was only moderate (r = 0.39, P < 0.0001). The average number of days to sprouting in 2014 was 29 for the 227 clones grown in common both years, whereas the average number of days to sprouting in 2015 was 51. There may be several reasons for this result: the first observation on sprouting in ME 2015 was nearly three weeks later than in 2014 due to harvesting demands; sufficient seed was available to enable observations on sprouting to continue nearly two months longer in 2015; and wet conditions or other environmental factors and/or seed size may also have affected tuber dormancy.

Table 3.

Correlations among the same traits at the same location in different years.

View Table

Of the 227 clones that survived the 2015 growing season in ME, 125 were selected from the second field generation (Table 4). Thirty-seven (16.3%) of these were selected in both FL 2015 and ME 2015. The percentages for selected clones between ME and FL were higher than an earlier study (9.7%) looking at early generation selection in tetraploid clones undergoing evaluation for more traditional markets (Haynes et al., 2012). FL and ME 2015 both agreed to discard 70 clones (30.8%), which was lower than in the previously mentioned study where 55.5% of the clones were discarded (Haynes et al., 2012). FL and ME 2015 disagreed on the fate of the remaining 120 clones (52.9%), which was higher than in the previously mentioned study where they disagreed on 34.8% of the clones. These discrepancies may reflect differences in ploidy level between the two studies: tetraploids are generally considered to be better buffered against environmental variables than diploids due to their greater allelic diversity. It is also possible, however, that the higher percentage of disagreement between ME and FL may be a result of dealing with developing different selection criteria for this new marketing niche.

Table 4.

Percent agreement among ME (in 2015), FL (in 2015), and NM (in 2015) selectors for 227 clones evaluated at each location.

View Table

In contrast, only 14 (6.2%) clones were selected in both NM and ME 2015 (Table 4). NM and ME both agreed to discard 97 clones (42.7%) and disagreed on the fate of the remaining 116 clones (51.1%). The lower percentage selected may indicate greater environmental sensitivity in this germplasm. This was also the first year that NM had grown potatoes, and so the low saved percentage may reflect learning how to grow and evaluate a new crop.

Only six clones were selected in both FL and NM 2015. FL and NM 2015 both agreed to discard 144 clones and disagreed on the fate of the remaining 77 clones.

It is interesting that the greatest agreement (both agree to save, and both agree to discard), 66.0%, was between the two southern locations (FL, NM). The least agreement at 47.1% was between FL and ME, two programs with a long history of collaboration. The differences in agreement between these locations were significant (χ² = 20.01, P < 0.001).

Second year field evaluations in the south (FL 2016, NM 2016).

Eighty-two clones were evaluated in replicated trials in the southern states in 2016. Compared with the clones in the first year of field evaluation in FL, this subset of selected clones in the second year in FL generally averaged lighter skin texture (7.3 vs. 5.2), slightly better appearance (6.0 vs. 5.3), and had longer dormancy (40 vs. 24 d). The averages for other characteristics were similar between the two years. High correlations were observed between years for skin color (r = 0.74, P < 0.0001) and tuber shape (r = 0.65, P < 0.0001); moderate-to-low correlations were observed for eye depth (r = 0.31, P < 0.001) and appearance (r = 0.24, P < 0.05); and there was a negative correlation for days to sprouting (r = −0.25, P < 0.05) (Table 3). There was no correlation for skin texture between years.

Compared with the clones in the first yield of field evaluations in NM, this subset of selected clones in the second year in NM generally averaged slightly lighter skin texture (5.6 vs. 4.8). The averages of other characteristics were similar between the two years. Moderate-to-high correlations were observed between years for skin color (r = 0.73, P < 0.0001), tuber shape (r = 0.67, P < 0.0001), and flesh color (r = 0.51, P < 0.0001) (Table 3). No correlations were observed for skin texture or appearance. NM did not evaluate eye depth in 2016.

Thus, these selections maintained their phenotypic diversity for most traits. Diversity in skin color, flesh color, and tuber shape could conceivably be advantageous from a marketing standpoint in allowing the development of multiple cultivars to either fill different niches or be combined into an appealing mixture of cultivars. For the locations that evaluated length of dormancy both years, clones had longer dormancy the second year. It is difficult to assign a single explanation for these differences because in ME the first year field generation was propagated from greenhouse-produced mini-tubers, whereas the second year field generation was propagated from field-produced whole seed. However, in FL both the first and second year field evaluations were propagated from field-produced whole seed. Thus differences in length of dormancy may also be due a carry-over effect of the seed growing environment. Such effects on other traits in potatoes have been reported (Hutchinson et al., 2014; Madec and Perennec 1959).

Third year field data in Maine (2016).

Growing conditions in ME in 2016 were nearly ideal for potatoes, with warm days, cool nights, and adequate rainfall. The average high air temperatures for June, July, and Aug. were 22.3, 25.2, and 25.0 °C, while the average low air temperatures were 10.5, 13.9, and 13.5 °C. Rainfall totaled 73.2, 148.6, and 119.8 mm, respectively.

The 2016 data from the ME harvest included not only tuber yield, but also the number of tubers per five-plant plot. There were significant differences among the clones for average number of tubers per plot, which ranged from 28 to 182. There were also significant differences among the clones for average weight of individual tubers, which ranged from 15 g to 68 g. Considerable progress has been made in increasing the size of tubers in the phu-stn population. Thompson et al. (1980) reported an average tuber weight of 2.6 g for phu and 2.0 g for stn. In contrast, the average tuber weight in ME 2016 was 33 g. Gautney and Haynes (1983) reported a mean yield per hill, in their phu-stn population, of 130.3 g. In contrast, the mean yield per hill in ME 2016 was 592 g.

There was a significant negative correlation (r = −0.62, P < 0.0001) between number of tubers and weight of tubers. A similar trend has been observed at the tetraploid level (Bhagowati et al., 2002; Gan et al., 2013; Gopal et al., 1994; Sanwal et al., 2003).

Statistical analysis of yield trial data from all locations.

Yields in 2016 were greatest in ME, followed by FL, with the lowest yields in NM. The average plot yield in ME 2016 was 2960 g, ranging from 1078 to 5153 g. The average plot yield in FL 2016 was 2383 g, ranging from 622 to 4944 g. The average plot yield in NM 2016 was 461 g, ranging from 119 to 874 g. The greater clonal yields in ME may be due either to the cooler growing season as compared with the southern locations and/or to better adaptation to the longer day growing conditions in ME as compared with the southern locations. Potato was a new crop for personnel at New Mexico’s Los Lunas Agricultural Science Center, so optimum production protocols have not yet been developed for this area of the state. The lower plot yields in NM may be due to environmental conditions not conducive to production of these potatoes, or the result of optimal inputs not yet determined for this crop in Los Lunas, NM.

There were significant differences among clones for total yield at each location in 2016. Least square mean yields were ranked from highest to lowest at each location. Four clones were ranked in the top 20 at all three locations. Eighteen clones consistently yielded in the top half of the rankings. Two clones yielded well in the south but poorly in the north, indicating that they might have some tolerance to heat. Two clones yielded poorly in the south but well in the north, indicating that they might be sensitive to heat stress. Twenty clones consistently yielded in the lower half of the rankings. Pictures of some of the better-yielding clones are shown in Fig. 1.

View Full Size

Pictures of some of the diploid creamer potatoes from the diploid hybrid Solanum tuberosum groups Phureja and Stenotomum population for the intense yellow flesh being evaluated.

Citation: HortScience horts 54, 12; 10.21273/HORTSCI13576-18

• Download Figure
• Download figure as PowerPoint slide

Thus far, attempts to introgress the intense yellow-flesh trait, and hence high carotenoid levels, from diploid potatoes to tetraploid potatoes have met with limited success (Haynes et al., 2011). Tetraploid germplasm with higher levels of carotenoids have been obtained, but they still fall far short of the concentrations that have been recorded in intense yellow- and orange-fleshed diploids.