Specialty Pumpkin: Laying the Groundwork for an Emerging Crop and Lucrative Products

Objective 1: Assess potential risks/benefits of specialty pumpkin production and barriers to acceptance.

Team: Jorge Ruiz-Menjivar and Marilyn Swisher

Activity 1: Consumer survey to understand consumer perceptions, preferences and willingness to pay for specialty pumpkin based on country of origin, certification labels other food product attributes.

First, we conducted a consumer study with a sample of 450 American consumers in February 2022. The sample, drawn from a nationally representative panel accessed via Qualtrics XM, adhered to predetermined quotas: Gender (52% female, 48% male), Age (18-34: 30%, 35-54: 32%, 55+: 38%), Race (White: 75%, Non-White: 25%), and Ethnicity (Hispanic/Latino: at least 30%). Respondents were compensated $9.26 each. Inclusion criteria specified U.S. adults aged 18 and above who had recently purchased fresh pumpkins or butternut squash for cooking purposes. The analysis employed a combination of descriptive and multivariate statistics to examine the factors influencing consumer preferences and purchase decisions among calabaza and pumpkin varieties. Results indicated that taste emerged as the most influential factor in purchase decisions among consumers of calabaza and pumpkin varieties, with a majority expressing a preference for specific taste profiles. Price sensitivity was also notable, with a significant proportion of respondents indicating that cost played a crucial role in their decision-making process. Flesh color, particularly in calabaza, was identified as another determinant, with consumers showing preferences for specific shades or intensities of color. The size of the calabaza was found to influence purchase decisions, with consumers often preferring larger sizes for various culinary uses. Ease of preparation for cooking emerged as a significant factor, with consumers favoring varieties that were perceived as easier to prepare and cook. Availability in nearby stores was highlighted as critical, influencing accessibility and purchase frequency among respondents. Conversely, factors such as the specific variety of calabaza, distinctions in organic labeling, and packaging did not significantly impact purchase decisions among the surveyed consumers. These findings suggest that while specific attributes like taste, price, flesh color, size, ease of preparation, and availability strongly influence consumer choices, other factors related to variety differentiation and specific marketing appeals may have less immediate impact on purchasing behavior. Second, we reviewed the literature to identify social, economic, and psychological factors influencing fruit and vegetable purchase decisions. Additionally, this identification led to a study where we conducted a formal bibliometric analysis of changes in eating habits and dietary patterns amid the COVID-19 pandemic. The coronavirus disease 2019 (COVID-19) pandemic altered consumers’ daily routines, including food consumption behavior. Data for this global bibliometric study were collected from the Scopus database, resulting in a sample of 497 peer-reviewed articles published between 2020 and 2022. Results showed that the impact of the COVID-19 pandemic on consumers’ dietary habits and food consumption varied across countries, largely dependent on economic conditions. Findings indicated that in developed countries with higher purchasing power, there was a reduction in food waste and an increase in local food consumption. For example, studies in these countries found increased demand for
“sustainable” food and food products purchased via online delivery platforms. This paper was presented at the 2023 Sociology of Consumption Conference (August 28th, Oviedo, Spain). Third, a subsample of 300 consumers was recruited to evaluate the effect of socio-psychological factors (i.e., food agency, cooking and food skills, cooking confidence, perceived behavioral control, subjective norms, attitude, food neophobia, consumer trust, and certification label) on the purchase intention of fresh calabaza and fresh produce. Results from the bibliometric search informed the data collection instrument for phase II of our consumer study. The instrument was developed based on existing literature on food preferences for new and native food products, as well as consumer demand for fresh produce. Additionally, we employed the Theory of Planned Behavior (TPB), including attitude, subjective norm, perceived behavioral control, and behavioral intention, to guide the development of our instrument, research questions, and hypotheses. Data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). Findings revealed several significant outcomes: socio-psychological factors played a crucial role in influencing the purchase intention of fresh calabaza and produce. Factors such as food agency, cooking and food skills, cooking confidence, perceived behavioral control, subjective norms, attitude towards calabaza and fresh produce, food neophobia, consumer trust in product quality, and certification labels all impacted consumers' decisions. Specifically, consumers with higher cooking and food skills confidence expressed stronger intentions to purchase calabaza and fresh produce. Positive attitudes towards calabaza and perceptions of its health benefits were also strongly correlated with purchase intentions. Moreover, consumers who perceived more significant control over their food choices and were influenced by positive subjective norms tended to express greater intent to buy calabaza and fresh produce. Interestingly, while certification labels influenced consumer trust in product quality, their direct impact on purchase intentions for calabaza and fresh produce was moderate. Food neophobia, or the fear of trying new foods, negatively correlated with purchase intentions, suggesting that consumers less familiar with calabaza might be less likely to purchase it unless encouraged by positive attitudes or social norms. Overall, these findings underscore the importance of consumer education and perception management in promoting calabaza and fresh produce. Strategies that enhance cooking confidence, emphasize health benefits, and build trust through effective certification may effectively increase consumer uptake of calabaza and similar fresh products in the market.

Figure: Importance of Calabaza attributes

Activity 2: Interviews with actors in the distribution system and sales points ranging from farmers’ markets to traditional commercial outlets, to identify key bottlenecks and opportunities to market in a variety of venues. 

We developed a 22-question interview instrument and obtained IRB approval (IRB 202102464) to conduct interviews with store managers in Florida. The interviews were completed in English or Spanish, as needed. We identified traits of importance to grocers as barriers or incentives, such as seasonal availability, size of the squash, color of the squash rind and flesh, in-store perishability, size of standard shipping units (e.g., bin vs. 50 lb. box), the value of the Fresh from Florida identity for calabaza, the value of country-of-origin labeling, and information about how to prepare the squash, including recipes. Our target sample included general managers, produce managers, and others who made decisions about produce stocked at international specialty markets that typically carry calabaza. This included a focus on points of sale in venues serving Latinx, Afro-American, Afro-Caribbean, and other clientele groups with a cultural heritage of consuming calabaza. We completed 40 interviews in 2023, covering the main metropolitan cities with large Latinx populations: Metropolitan Miami, Fort Lauderdale, Boca Raton, Tampa, Orlando, and Jacksonville. The interviews were transcribed and analyzed using thematic analysis. The results indicated several key findings. Shelf life was not a significant concern for grocers. The size of the squash was crucial, with a preference for larger calabazas, which were often cut into pieces if too large for easier cooking. Flesh color was important to consumers, while skin color was not. The variety of calabazas did not significantly influence purchasing decisions, and grocers successfully sold all calabazas stocked. Grocers in Orlando and Miami reported that consumers with Puerto Rican roots specifically demanded Puerto Rican-grown calabazas and were willing to pay a premium for them. Consumers did not highly demand the “Fresh from Florida” and organic labels. Store managers also noted that these labels did not significantly impact consumer purchasing decisions. The majority of grocers recognized the benefits of including calabaza in their regular stock and saw significant potential to expand the market by attracting consumers who were less familiar with calabaza. Effective strategies included mixing winter squash, sweet potato, butternut squash, and calabaza in a single bin to offer variety and attract attention, displaying calabaza alone with a product description also helped inform consumers about the squash and providing recipes that featured calabaza. Overall, our findings highlighted the positive reception and constant demand for calabaza among the Latinx and Afro-Caribbean clientele in Florida. Store managers expressed optimism about the potential to increase calabaza sales by educating consumers and offering it as part of a broader range of produce. The findings suggest that targeted marketing and consumer education could further enhance the appeal and sales of calabaza in various markets.

Activity 3: Interviews with farmers to identify key barriers, opportunities and research priorities for specialty pumpkin production. 

We assessed the interest of non-producers in cultivating calabaza and the interest of current producers in increasing their calabaza production. Our target sample included producers who serve Latinx, Afro-Caribbean, Afro-American, and other clientele groups with a cultural heritage of consuming calabaza. We conducted 30 interviews: 15 in the Southeastern United States during the summer and fall of 2023 and 15 with farmers in the Southwest and southern Puerto Rico in the spring of 2024. We developed an interview instrument and obtained IRB approval to conduct interviews with farmers in the Southeast. Interviews were conducted in English or Spanish, based on the farmer's preference. We used a multiple-case research design with a diverse sample of growers, including members of agricultural cooperatives, beginning farmers with less than four years of operation, smallholder farmers with less than 10 acres, farmers with large operations exceeding 250 acres, women-led farms, and farms with different production systems such as conventional, certified organic, and agroecology. The theory of planned behavior informed the interview protocol to capture attitudes, social norms, and perceived behavioral control affecting farmers' intentions to continue producing fruits and vegetables in the short and long term. This study received approval from the institutional review boards (UF-IRB: 202102464 / UPRM-IRB: 2023010008), and informed consent was obtained from all participants. The interviews were transcribed and analyzed using thematic analysis to identify salient barriers, needs, and sources of information. Our aggregated results indicate that the main barriers include issues related to (i) climate variability, (ii) labor, and (iii) market pricing and commercialization. Growers particularly emphasized experiencing crop damage, yield reductions, and sales losses due to climate shocks such as erratic weather patterns, drought, extreme heat, and an increase in the prevalence and severity of insect pests. Regarding labor, growers reported challenges in recruiting and retaining qualified and skilled workers. Farmers, regardless of scale or production system, expressed difficulties due to the recurrent oversupply of similar crops in the market, leading to the acceptance of lower prices. Additionally, growers highlighted issues related to the reliability, efficiency, and lack of distribution channels. Concerning sustainability in their production systems, most participants expressed a need for more information and dissemination of sustainable pest management practices.

Activity 4: Assessments of field research to identify most and least promising calabaza breeding lines currently under evaluation.

We developed a research assessment instrument and obtained IRB approval (IRB 202102464) to conduct assessments in Florida and Puerto Rico. Assessments in Puerto Rico were completed in Spanish. Participants identified their favorite and least favorite calabaza breeding lines under evaluation in this project. They also determined which traits were most important when selecting favorite and least favorite calabaza breeding lines. Our target sample included current calabaza producers, farmers interested in adding calabaza to their production systems, and technical advisors who consulted with farmers growing calabaza. We conducted two research assessments during this project: June 2023 in Florida and March 2024 in Puerto Rico, which was integrated into the interview protocol with producers. We
had planned an additional assessment in February 2023 in Puerto Rico, but we had to cancel it because faculty and staff were on strike at the University of Puerto Rico-Mayaguez. For the grower assessment conducted in Florida (June 2023), the first part of the assessment included a field observation where growers individually assessed anonymized varieties in the field, with a cut fruit from each variety displayed so farmers could see an example of the flesh and inside of the fruit. The second part of the assessment included a group discussion where farmers shared their observations. Five participants attended the event, and at the end, they selected as a group the top three varieties with promising characteristics that they would want to grow and three varieties that should not be used for further experimentation. For the assessment conducted in Puerto Rico (March 2024), we incorporated a set of discrete choices for desirable characteristics of pumpkin for farmers to consider from two perspectives: a) the calabaza they would consider ideal from their personal perspective as a grower and b) the calabazas they think consumers would want to purchase. The attributes evaluated included country of origin, price, production method, taste, flesh color, weight, shape of calabaza, and any desired enhanced traits. From a personal perspective, growers preferred calabazas originating from the Caribbean (West Indies) and were willing to pay between $1.00 and $1.50 per pound. They had no specific preference for the production method but favored a mildly sweet to nutty taste profile with an orange flesh color. The ideal weight for these growers was between 6 and 10 pounds, and they preferred a scalloped shape. Lastly, farmers expressed interest in pest-resistant varieties and those tolerant to drought. Conversely, from a general consumer perspective, including consumers in the contiguous U.S., Puerto Rican farmers believed that the ideal attributes would differ. They thought most consumers would not care about the country of origin and would prefer the cheapest calabaza, priced at less than $1.00 per pound. The production method was deemed irrelevant for consumers, who were thought to favor a mildly sweet to very sweet taste and a deep orange flesh color. The preferred weight for consumers was 3 pounds or less, making it a more personal size, and they would prefer a round or elongated shape. The findings imply that breeding and marketing strategies need to balance both grower and consumer preferences. Growers prioritize traits like taste, color, and size, which should guide breeding efforts. Conversely, consumers focus on price and ease of preparation, highlighting the need for cost-effective production and clear communication of the product's benefits. Addressing these differing preferences can enhance market acceptance and adoption of calabaza, benefiting both producers and consumers through targeted promotion and education strategies.

Activity 5: Advisory council meetings to review biological research and consult in the development of future calabaza research.

• We obtained IRB approval (IRB 202102464) and recruited eight panel members to attend annual or bi-annual meetings totally up to three hours per year. The advisory council is a permanent governing body for the project's duration that assesses overall results, discusses progress and impacts, and determines if the project plan requires adjustments. Our panel consists of four calabaza farmers from Florida, one calabaza farmer and one Extension faculty member from Puerto Rico, and one calabaza farmer and one Extension faculty member from Alabama.  We extended invitations to an Extension faculty member from Florida and two calabaza farmers from Puerto Rico and Alabama, respectively.  We hope to add those additional members in 2023/2024. We hosted two advisory council meetings: January 30, 2023 and March 18, 2024.  Seven panel members were in attendance in 2023, while three were in attendance in 2024. 

A summary of key points:

• Growers were interested in trialing some of the calabaza germplasm on their farms. These seeds will be distributed to growers in 2023.
• Growers wished to know how to protect the seedlings from bird and lizard damage. Netted shielding cages were suggested.
• Growers were interested in earning how to measure sweetness of calabaza on their farms. Information was provided to a suitable equipment to use.
• Consumer demand was highlighted as the main barrier to adopting calabaza as a new crop. Also resistant to pest and diseases was highlighted as important, especially for organic growers.
• Growers provided input as to the main components to focus on in the marketing of calabaza e.g. providing information on nutrition, how to prepare for cooking, suggested recipes, change of name from calabaza to something more recognizable e.g. Bellevue butternut.

   

  Objective 2: Evaluate pumpkin germplasm lines and cultivars for use as flesh, seeds, and as product ingredients.

  Team: Andrew MacIntosh and Amy Simonne

  Activity 1: Comparative analysis of qualitative attributes for selection of calabaza genotypes adapted to subtropical climates

  Flesh Yield:

  Flesh yield for the calabaza varies greatly among the squash genotype (Table 3-2) and are markedly distinguishable from the previously reported values on either extreme (Maynard, 2002). ‘Soler’ was much larger on average than any of the other cultivars (3.5 kg) while the Waltham Butternut cultivar was much smaller on average (0.72 kg). Generally, fruit size in Cucurbita is dependent on genotype but can also be influenced by the environment (Gaspera, 2016; Wetzel, 2019). In certain markets, smaller squash fruits are preferred by consumers due to convenience of handling, dish preparation, and less labor (Costa, 2003). Therefore, small fruit calabaza cultivars might be more desirable for a fresh market within the U.S. compared to large fruit cultivars (Carbonell, 1990).

  Fruit shape, Growth Habit, and Flesh Color:

  The calabaza genotypes varied in fruit shape, growth habit and flesh color (Table 3-3). The fruit shapes observed spanned from globe shaped (UFTP 8, UFTP 22) to round shaped (UFTP 24, ‘La Estrella’) to oblate shaped (UFTP 38, UFTP 42, ‘Soler’) and Bell-shaped (Butternut). Calabaza fruit shapes are dependent on the parent lineage (Ferriol, 2003) and could be a factor affecting consumer preference (Carbonell, 1990).  Flesh yield for the calabaza, shown in Table 2, ranges greatly from previously seen values on either extreme (Maynard, 2002). The UFTP 45 germplasm line was much larger on average than any of the other lines (3.44 kg) while the ‘Waltham Butternut’ squash was much smaller on average (0.72 kg). It should be noted that convenience is also one of the major contributing factors in purchasing power due to many variables including but not limited to saving time, reduced cost, less skill required for complex dishes, and less labor (Costa, 2003). Therefore, smaller calabaza might be better for a fresh market within the U.S. compared to larger fruit.

  The flesh color was determined using the Hunter lab L*a*b* color scale. Statistically significant differences were observed among the squash genotypes in flesh color attributes, including lightness (L*), red/green (a*), and blue/yellow (b*) and are presented in Table 3-3. Interestingly, the values observed for the Waltham Butternut cultivar are different from values of those previously reported in literature. For example, Mashitoa, (2021) reported average L*, a*, b* values of 33.22, -7.45, and 10.61 in butternut, respectively (Mashitoa, 2021). However, the findings in the current study are within the range of those found in literature for other C. moschata cultivars (Men, 2020; Provesi, 2011; Itle, 2009). Silva (2019) found that cultivars with high a* and high b* (for an intense orange color) were preferred by panelists. Using this preference trend, UFTP 8, UFTP 22, and UFTP 24 would have the highest consumer color liking. The chroma and hue angle were determined using Equations 3-2 and 3-3, respectively (Table 3-4). These additional color indicators are also within ranges seen for various calabaza germplasm lines (Wessel-Beaver, 2006; Maynard, 1994).   Chroma value (color purity) was statistically the highest for UFTP 8 with UFTP 24 being the only other germplasm line that was not statistically significantly different. This color intensity would further support these germplasm lines’ higher consumer acceptability, as multiple studies have shown that appearance is a very important quality that consumers consider when buying a food (Silva, 2019; Trejo Araya, 2009). 

  Color parameters (a*b* and Chroma) were shown to be positively correlated with Brix and hardness, while lightness saturation parameters (L* and hue angle) were positively correlated with peroxidase as shown in Table 3-6. Helyes (2006) found that the relationship for color parameters (a*b* and Chroma) could be explained through the ripeness of the fruit, as color and Brix concentration naturally change throughout the fruit development process. The ripening process could also explain the positive correlation of Chroma, a*, and b* with hardness and many other texture parameters in squash fruits reflect an increase in hardness characteristics during the ripening process (Valenta, 2022). The L* value correlates with the lightness or darkness of a sample, with a positive L* value being lighter and a negative L* value being darker. Lightness (L*) and hue angle was shown to be positively correlated with POD. This is to be expected as POD is responsible for enzymatic browning on fruit. Therefore, when the concentration of the enzyme increases, color change will be evident – mainly changing to brown (Ndiaye, 2009). It should be noted that POD was negatively correlated with a*, (+, red;-, green), which is most likely due to the ripeness of the fruit being more of a red color than a green color. With POD making the overall color of the sample more brown, this would decrease the red hue resulting in an inverse relationship between these two attributes.  

  Peroxidase Enzyme (POD):

  Among the calabaza genotypes, there was a narrow range (1.5 - 1.8 abs/min) of observed POD activity, as seen in Table 3-4. The enzyme activity observed for these calabaza genotypes were relatively high when compared to values found in literature (0.9 - 1.1 abs/min) (Zhou, 2017; Jamali, 2018). This could be due to plant stress (Jaiswal, 2012), level of fruit maturity (Sharma, 2013), and days in storage (Suo, 2022). Peroxidase enzyme, unlike polyphenol oxidase, is primarily known for its browning effect on color (Tomás-Barberán, 2001; Toivonen, 2008). This relationship is supported with the finding shown in Table 3-6, where the primary correlations present were for color-based parameters. It should be noted that POD was also negatively correlated with “springiness”, a*, and SCC/TA. Springiness is defined as the ability of a food sample to recover its original height after being compressed between the first and second compression (Hanusz, 2018). Previous research on Cucurbita maxima shows that springiness is related to the moisture and fibrousness of the squash sample, which is also strongly correlated with ripeness (Corrigan, 2006). As previously described, a* is the color range from red (+) to green (-). With a* values being an indicator for fruit ripeness with the C. moschata species, a negative correlation with POD, would support the fact the the higher the concentration present, the lower the a* value and therefore undesirable fruit ripeness. 

  Brix:

  The Brix values across the calabaza genotypes ranged between 6.2 and 12.2, with a mean of 9.9 (Table 3-4).  The observed values are within the range present in existing literature (6.3 – 15), affirming that the Brix measured is within an acceptable range (Iacuzzo, 2009; Abbas, 2020a; Jacobo-Valenzuela, 2011).  Notably, UFTP 22 had the highest Brix value (12.2), which may indicate this germplasm line has a higher perceived sweetness for consumers as °Brix is considered a reasonable measure of sugar content and the overall evaluation of fruit quality (Harrill, 1998). It should be noted that variation in Brix is primarily genetic (Akter, 2013); however, several other factors can influence this trait including growing conditions (Alam, 2001), and optically active compounds such as pectins, amino acids, fiber, and organic acids (Byrne, 1991; Li, 2021).

  Correlations between Brix and many other tested parameters are shown in Table 3-6 with positive correlations including TA, malic acid, a*, b*, chroma, hardness, cohesiveness, gumminess, chewiness, and resilience. Brix concentration in relation to organic acid concentration has been used as an indicator for fruit and vegetable ripeness for quite some time (Kuti, 1992). This would therefore support the strong correlation (0.78) between Brix and TA as well as Brix and malic acid.  

  Titratable Acidity and pH:

  Previous studies have shown that the Butternut squash contains malic, citric, fumaric, ascorbic, and gallic acid (Nawirska-Olszańska, 2014; Pevicharova, 2017; Men, 2021), however, the most prevalent acid among these was malic acid (Papanov, 2021; Zhou, 2017). The reported malic acid concentration for C. moschata ranges from 0.16-0.28 mg/100 g (Table 3-4) (Zhou, 2017). This range is within that observed (0.05-0.17 mg/100 g) in the current study. The germplasm line UFTP 22 and the cultivars Waltham Butternut and La Estrella had the three highest TA values of 0.15, 0.17, and 0.13, respectively (Table 3-4). To account for the range and lower concentrations within this study, previous research has shown that the acid concentrations are expected to be influenced by the cultivar (Kim, 2012) and growing region (Pevicharova, 2017).  These higher TA concentrations strongly negatively impact the germplasm lines SSC/TA ratio (−0.84). However, the SSC/TA ratio is not a catch-all for consumer acceptability, as it has been shown to be cultivar-dependent (Crisosto, 2003). 

   Fruit pH is an important fruit -quality parameter due to its influences on color (Gliemmo, 2009), microbial growth (Gliemmo, 2013), and change during storage (Nawirska-Olszańska, 2014). Previous studies in C. moschata showed a pH range between 5.3-7.79. The values for pH observed in the current study (5.98 - 6.58) are within the range (5.3-7.79) of those previously reported for C. moschata (Provesi, 2011; Zhou, 2017; Papanov, 2021; Jacobo-Valenzuela, 2011).  

   Previous literature suggests that titratable acidity does not directly correlate to the pH of the sample in all cases (Marsh, 2006).  Papnov et al. (2021) previously showed that pH and TA (malic acid) had no correlation (0.04) while Zinash et al. showed a strong negative correlation (-0.86) between the two measurements (Papanov, 2021; Zinash, 2013). Based on the data presented in Table 4, TA (as malic acid equivalent) and pH were strongly negatively correlated for the fruits in this study. Acids in fruits of some species, such as tomatoes, have been shown to decrease acid concentration through respiration only (Anthon, 2011). This could mean that the germplasm lines, which are currently being compared could have varying degrees of ripeness. It should be noted that TA and pH should have a negative correlation as the increase of organic acids would lead to a decrease in pH values. Therefore, the variation in concentrations and inversely proportional relationship between pH and TA is to be expected. Additionally, pH was shown to be positively correlated with the SSC/TA ratio which follows the same logical reasoning as the negative correlation between pH and TA.

  Yeast Fermentable Extract (YFE):

  The YFE for the calabaza genotypes showed an initial extract range (6.1-12.4 ± 0.01) and a final extract range (1.1-3.2 ± 0.88) yielding a fermentable extract range of 60.5 - 78.5% (Table 3-4). The concentration of YFE for C. moschata has not been previously reported. However, this measurement is important to quantify as it reflects the amount of simple (fermentable) sugars affecting the consumer's perceived sweetness (Goldfein, 2015; Maicas, 2020).  Furthermore, YFE has been shown to be a reliable indicator for fermentability of produce (Moreno, 2022). It should be taken into account that yeast strain, fermentation temperature, enzymes, and yeast nutrients have all been shown to impact fermentation (Parcunev, 2012). Interestingly, UFTP 22 had the highest Brix values (12.2) but had one of the lowest YFE% of 60.8. Comparing this to UFTP 8 which had the highest YFE% of 78.5 with a relatively high Brix value of 11.6. This is supported by the correlation between Brix and YFE (0.23) not being significant. This distinction between Brix and YFE is important to assess as it can demonstrate the difference in market value potential between these germplasm lines for application-based purposes as described in the introduction. When comparing YFE to the Brix values within a germplasm line, one can extrapolate that the UFTP 8 contains the highest concentration of simple sugars and is therefore likely to be perceived as most sweet. The opposite is true for UFTP 42 and UFTP 22.

  Texture Profile:

  The texture profile for each calabaza germplasm line was determined with average values and statistical differences for each textural attribute presented in Table 3-5. All textural measurements (hardness, adhesiveness, springiness, cohesiveness, gumminess, chewiness, and resilience) showed some statistically significant differences between germplasm lines, but this range was not wide. This is most likely due to the genotypes similarity to one another but could also be due to storage time (Ratnayake, 2004), and sampling location (Marian, 2020). Firmness/hardness had statistically significant extremes with UFTP 8 having the highest firmness/hardness value (40,323 g) and UFTP 38 having the lowest firmness/hardness value (22,027 g). Gumminess was one of the most variable textural parameters with Soler having the lowest gumminess (ease of swallowing) but not being statistically significantly different than UFTP 38, and UFTP 42. As previously described, firmness/hardness is a strong indicator of fruit ripeness (Valenta, 2022); however, variation in texture parameters is to be expected based on genetics (Toivonen, 2008; Corrigan, 2006).  

  Concentration of compositional elements varied slightly, as seen in Table 3-4, but these ranges did not seem to impact or strongly correlate with textural parameters - excluding Brix and YFE - as seen in Table 3-6. YFE is a measurement of the amount of fermentable sugars present within a sample while Brix is a measurement of the total amount of soluble solids (I.e starch, sugar, and fiber) suspended in the blended sample. Therefore, it would be expected to see some similarities between Brix and YFE correlations as YFE is part of the Brix measurement.  

  Brix was positively correlated with all textural measurements except adhesiveness and springiness which have almost no correlation at (0.01 and 00.06 respectively) (Table 3-6). Adhesiveness is the amount of force required to remove the food sample from the probe it was in contact with, better known as stickiness (Chandra, 2014). Both adhesiveness and springiness are related to the amount of moisture within a food sample (Corrigan, 2006), which could decrease the concentration of sugars and other soluble solids impacting texture.  As previously discussed, hardness and many other texture parameters are impacted during the ripening process of fruits which corresponds with an increase in Brix (Valenta, 2022), supporting the strong correlations between Brix and all other textural measurements.  

  YFE was positively correlated with hardness and gumminess as seen in Table 3-6. Gumminess is determined by multiplying the hardness of a sample by its cohesiveness (Parcunev, 2012). Cohesiveness is a food sample’s ability to resist a second deformation compared to the first deformation due to compression (Chandra, 2014). With both textural parameters being positively correlated and related to hardness, it can be reasonably assumed that as the fruit ripens the sugar concentration and hardness parameters would increase. Additionally, springiness was negatively correlated with YFE. As previously described, both adhesiveness and springiness are related to the amount of moisture within a food sample (Corrigan, 2006), which could decrease the concentration of sugars and other soluble solids within a sample. As YFE is only the measurement of fermentable sugars, it would be more impacted by the dilution of sugars than Brix, resulting in this negative correlation. 

   

  Table 3-2.  Yield table for calabaza germplasm lines used in the study.

  Cultivar	Avg. Total Fruit Weight (kg)	Avg Fruit Flesh (kg)	Flesh Yield (%)
  Butternut	0.72 d	0.54 d	75.17 c
  UFTP 8	1.62 c	1.24 c	76.55 bc
  UFTP 22	1.70 c	1.36 c	80.18 a
  UFTP 24	2.51 b	1.99 b	78. 36 abc
  UFTP 38	2.75 b	2.14 b	77.52 abc
  UFTP 42	2.58 b	2.11 b	79.88 ab
  ‘Soler’	3.50 a	2.77 a	78.96 ab
  ‘La Estrella ‘	1.83 c	1.38 c	75.21 c

  n = 10-15 squash of each cultivar were measured to determine average values. Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per column.

  Table 3-3. Color (L*, a*, b*), chroma, hue angle, pictures of vine growth and cross section of Cucurbita moschata genotypes grown at the Plant Science Research and Education Unit in Citra, Florida (2022).

  Cultivar	L*	a*	b*	Chroma	Hue Angle	Illustration (fruit and vine)
  Butternut	69.13 bc	22.37 b	70.21 c	73.93 c	72.41 cd
  UFTP 8	67.90 cd	25.23 a	77.40 ab	81.44 a	71.93 d
  UFTP 22	69.93 ab	22.72 b	74.79 b	78.21 b	73.15 cd
  UFTP 24	68.63 bc	20.10 c	77.93 a	80.52 ab	75.50 b
  UFTP 38	66.41 d	22.20 b	59.92 e	63.93 e	69.74 e
  UFTP 42	69.70 abc	20.66 b	70.21 c	73.20 c	73.62 c
  ‘Soler’	71.18 a	14.85 d	67.51 d	69.14 d	77.63 a
  ‘La Estrella’	71.17 a	19.00 c	76.06 ab	78.49 b	75.96 b

  N=5 squash were sampled per genotype, with 3 measurements taken per squash. Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per column.

   

  Table 3-4. Titratable acidity (TA) expressed as malic acid (g/100g), Brix/soluble solids content (g/L), SSC/TA ratio, peroxidase (POD) (abs/min), and yeast fermentable extract (YFE) for Cucurbita moschata germplasm lines grown at the Plant Science Research and Education Unit in Citra, Florida (2022). 

  Cultivar	Malic Acid (g/100g)	pH	Brix  (g/100 g)	SSC/TA  Ratio	POD (abs/min)	YFE  %
  Butternut	0.17 a	5.98 f	11.9 b	66.43 e	1.71 a	66.19 bc
  UFTP 8	0.09 d	6.58 a	11.6 c	127.83 bc	1.59 a	78.47 a
  UFTP 22	0.15 b	6.12 e	12.2 a	76.66 de	1.49 ab	60.80 d
  UFTP 24	0.08 e	6.16 d	10.7 e	134.29 ab	1.70 a	67.19 b
  UFTP 38	0.05 g	6.34 c	7.3 g	141.46 a	1.26 b	62.19 d
  UFTP 42	0.06 f	6.40 b	8.2 f	122.91 c	1.55 a	60.53 d
  ‘Soler’	0.07 f	6.36 c	6.2 h	87.34 d	1.76 a	67.81 b
  ‘La Estrella’	0.13 c	6.17 d	11.2 d	81.78 d	1.76 a	64.72 c

  Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per column.

   

  Table 3-5. Hardness, adhesiveness, springiness, cohesiveness, gumminess, chewiness, and resilience for each calabaza germplasms using a double compression texture test (TA.XT). Plant Science Research and Education Unit in Citra, Florida.2022.

  Cultivar	Hard.	Adhes.	Spring.	Cohes.	Gum.	Chew.	Resilience
  Butternut	33009 bc	-47.74 bc	0.52 cd	0.21 d	9237 bcd	4885 bc	0.11 bc
  UFTP 8	40323 a	-47.99 a	0.51 d	0.38 a	15792 a	8154 a	0.20 a
  UFTP 22	32841 bc	-16.58 a	0.65 a	0.32 abc	10730 bc	7104 ab	0.16 ab
  UFTP 24	35974 ab	-37.71 abc	0.50 d	0.35 ab	12725 ab	6420 ab	0.18 a
  UFTP 38	22027 d	-47.99 bc	0.55 bc	0.24 cd	5628 de	3207 cd	0.11 bc
  UFTP 42	26515 cd	-25.38 ab	0.58 b	0.25 bcd	7234 cde	4360 bcd	0.12 bc
  ‘Soler’	22541 d	-41.81 abc	0.52 cd	0.16 d	3665 e	1859 d	0.07 c
  ‘La Estrella’	36432 ab	-61.81 c	0.49 d	0.32 abc	11595 abc	5793 abc	0.16 ab

  Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per column.

  Table 3-6. Pearson correlation coefficient for fresh calabaza quality parameters

  Parameter	Brix	pH	TA%	Malic Acid	SSC/TA	YFE	L*	a*	b*	Chroma
  Brix	1.00	-0.43	0.78	0.77	-0.35	0.23	0.02	0.62	0.73	0.80
  pH	-0.43	1.00	-0.73	-0.73	0.63	0.49	-0.27	0.11	-0.10	-0.08
  TA%	0.78	-0.73	1.00	1.00	-0.83	-0.09	0.36	0.25	0.39	0.41
  Malic Acid	0.77	-0.73	1.00	1.00	-0.84	-0.09	0.36	0.24	0.38	0.41
  SSC/TA	-0.35	0.63	-0.83	-0.84	1.00	0.18	-0.71	0.25	-0.16	-0.12
  YFE	0.23	0.49	-0.09	-0.09	0.18	1.00	-0.18	0.25	0.41	0.44
  L*	0.02	-0.27	0.36	0.36	-0.71	-0.18	1.00	-0.67	0.34	0.24
  a*	0.62	0.11	0.25	0.24	0.25	0.25	-0.67	1.00	0.19	0.33
  b*	0.73	-0.10	0.39	0.38	-0.16	0.41	0.34	0.19	1.00	0.99
  Chroma	0.80	-0.08	0.41	0.41	-0.12	0.44	0.24	0.33	0.99	1.00
  Hueangle	-0.17	-0.13	0.00	0.01	-0.36	0.02	0.84	-0.83	0.39	0.25
  Hard	0.88	-0.11	0.49	0.49	-0.11	0.55	0.04	0.50	0.89	0.93
  Adhes.	0.01	-0.04	0.00	0.00	0.01	-0.42	0.06	0.11	0.06	0.07
  Spring.	0.06	-0.09	0.17	0.17	-0.14	-0.57	0.00	0.26	-0.15	-0.12
  Cohes.	0.68	0.16	0.11	0.10	0.33	0.39	-0.25	0.61	0.74	0.80
  Gum.	0.83	0.04	0.33	0.32	0.37	0.56	-0.15	0.62	0.84	0.90
  Chewi.	0.87	-0.01	0.41	0.40	0.52	0.40	-0.15	0.70	0.81	0.88
  Resilien.	0.75	0.11	0.19	0.18	0.24	0.44	-0.19	0.60	0.81	0.87
  POD	0.21	-0.28	0.34	0.34	-0.51	0.32	0.73	-0.51	0.57	0.48

  ¹ Many physiochemical attributes were abbreviated including Hue (Hue angle), Hard (harndness), Adhes. (Adhesiveness), Spring. (Springiness), Cohes. (Cohesiveness), Gum. (Gumminess), Chewi. (Chewiness), and Resilien. (Resilience).

  ² Color values (L*, a*, and b*) are denoted with an asterisk.

   

  Table 3-6.  Continued

  Parameter	Hue	Hard.	Adhes.	Spring.	Cohesiv.	Gum.	Chew.	Resilien.	POD
  Brix	-0.17	0.88	0.01	0.06	0.68	0.83	0.87	0.75	0.21
  pH	-0.13	-0.11	-0.04	-0.09	0.16	0.04	-0.01	0.11	-0.28
  TA%	0.00	0.49	0.00	0.17	0.11	0.33	0.41	0.19	0.34
  Malic Acid	0.01	0.49	0.00	0.17	0.10	0.32	0.40	0.18	0.34
  SSC/TA	-0.36	-0.11	0.01	-0.14	0.33	0.10	0.04	0.24	-0.51
  YFE	0.02	0.55	-0.42	-0.57	0.39	0.56	0.40	0.44	0.32
  L*	0.84	0.04	0.06	0.00	-0.25	-0.15	-0.15	-0.19	0.73
  a*	-0.83	0.50	0.11	0.26	0.61	0.62	0.70	0.60	-0.51
  b*	0.39	0.89	0.06	-0.15	0.74	0.84	0.81	0.81	0.57
  Chroma	0.25	0.93	0.07	-0.12	0.80	0.90	0.88	0.87	0.48
  Hue	1.00	0.03	-0.04	-0.31	-0.18	-0.12	-0.20	-0.13	0.81
  Hard	0.03	1.00	-0.19	-0.27	0.83	0.97	0.91	0.89	0.39
  Adhes.	-0.04	-0.19	1.00	0.85	0.01	-0.12	0.11	-0.01	-0.29
  Spring.	-0.31	-0.27	0.85	1.00	-0.03	-0.19	0.09	-0.07	-0.55
  Cohes.	-0.18	0.83	0.01	-0.03	1.00	0.93	0.93	0.99	-0.06
  Gum.	-0.12	0.97	-0.12	-0.19	0.93	1.00	0.96	0.96	0.19
  Chewi.	-0.20	0.91	0.11	0.09	0.93	0.96	1.00	0.96	0.05
  Resilien.	-0.13	0.89	-0.01	-0.07	0.99	0.96	0.96	1.00	0.04
  POD	0.81	0.39	-0.29	-0.55	-0.06	0.19	0.05	0.04	1.00

  ¹ Many physiochemical attributes were abbreviated including Hue (Hue angle), Hard (harndness), Adhes. (Adhesiveness), Spring. (Springiness), Cohes. (Cohesiveness), Gum. (Gumminess), Chewi. (Chewiness), and Resilien. (Resilience).

  ² Color values (L*, a*, and b*) are denoted with an asterisk.

  Activity 2: Chemical and physical properties of winter squash and their correlation with sensory attributes

  Fresh Cooked – Quality Parameters:

  The genotypes differed significantly in physiochemical attributes, though the degree of separation varied (Table 1 and Table 2). Most notably, Brix, malic acid, and pH showed more separation between genotypes than hue angle, POD, and all texture profile measurements. The Brix measurements range found within this study (7.7 - 16.03°Bx) were higher than those previously reported by Gajewski (2008), Santos Jr. (2017), and Khanobdee (2001) for cooked C. moschata (3.0-13.6), and lower than reported by Amaro in 2017 (18.4°Bx). Variations in sugar levels are likely due to the maturity of the fruit, which is synonymous with a decline in starch and an increase in simple sugars (Irving, 1997). Other factors that commonly impact Brix levels include genetics (Akter, 2013) and growing conditions (Alam, 2001). Notably UFTP 8 had the highest Brix level (16.03°Bx) of the analyzed germplasms, while ‘Soler’ and ‘La Estrella’ had the lowest (7.73°Bx and 7.83°Bx, respectively).   

  Malic acid analysis showed UFTP 22 had the highest concentration (0.06 mg/100 g), while UFTP 38 had the lowest concentration (0.02 mg/100 g). This acidity range aligns with previous literature. However, it should be noted that winter squash acid dominance is cultivar-dependent (malic or citric acid) and variation between germplasm lines may cause this variation (Nawirska-Olszanska, 2014).   

  The average color values (L*, a*, b*) of the freshly cooked microwaved samples (50.98, 10.91, and 50.96 respectively) were higher than of microwaved samples previously reported , which averaged 38.22, 4.6, and 11.3 respectively (Silva, 2019). However, other means of cooking C. moschata show a range which encompasses the values shown in this study (Silva, 2019; Gajewski, 2008; de Almeida, 2019). Regarding L* value, UFTP 22 (56.74) was significantly higher than all other germplasm lines. This shows that the pulp from this genotype was lighter than all other lines. Many of the germplasm lines displayed an orange color with low a* values (red color) and high b* values (yellow color). UFTP 22 showed the highest average a* value (19.59) and b* value (65.77) which resulted in a more intense orange pulp color. This is further supported with UFTP 22 having the highest chroma value (68.69). The germplasm lines UFTP 38, ‘Soler’, and UFTP 42 were pale yellow/orange with the lowest chroma values. Hue angle defines the hue of the color of a sample without considering the lightness. This categorization of color ranged from 73.48 - 85.14, which encompasses a range from dark orange to yellow (Pathare, 2012).   

  A double compression texture profile was determined and slight statistical variations between germplasms were observed. These statistical differences are to be expected as all samples were cooked until an internal temperature of 80°C was maintained for 3 minutes as this was considered cooked, therefore, limited variation in texture between samples is to be expected. To account for the statistical differences reported, fruit development variation (Abbas, 2020) and genetic differences (Gomes, 2020) could explain the observed variation. Additionally, the values from this study are similar to those expressed within literature for winter squash (Theanjumpol, 2022; Rinaldi, 2020).  In summary, the measured physiochemical properties of fruit from this study fall within those reported in previous literature.

  Fresh Cooked – Consumers’ Acceptability: 

  Demographic data was collected whereby 65.2% of panelists were women and 34.8% were men. This ratio aligns with findings from a prior study performed in the U.S. by Bowman (2006) which showed that 67% of meals were planned and prepared by women. This indicates that the results of this study may reflect the significance of women’s opinions on the characteristics of the genotypes, as they are frequent purchasers of food. Additionally, the majority of responses were from adults in the age range of 18 and 24 years old (43.5%), followed by those from 35 to 44 (33.9%), 25 to 34 (19.4%), and lastly over 45 (3.2%). It is noteworthy that age is a relevant factor of consumer acceptability due to dietary habits changing with developmental stages in life (Jaime et al., 2015). The panelists’ self-identified race was most often White (48.3%), followed by Asian (32.6%), then Hispanic (15.4%), and lastly African American (11.2%). Satterwaite et. al. (2010) showed that participants’ cultural habits and economic conditions influence their product consumption in their daily lives, which would therefore influence their product preference. 

  Appearance liking and color liking for the genotypes showed UFTP 38 and ‘Waltham Butternut’ as the most liked visually (Table 3). The overall liking comparison among samples, however, showed that UFTP 24 was statistically the highest performing variety. This indicates that other factors, such as taste and texture, greatly impacted the perception of a food, a finding supported by Anderson et al., (2019). When consumers rated their sweetness liking and flavor liking, UFTP 8 and UFTP 24 were statistically the most liked. In comparison, Brix values shown in Table 1 indicate UFTP 8 to be the highest, followed by UFTP 22. The discrepancy in Brix for UFTP 24 could be due to many reasons including varying concentrations of simple sugars (Godshall, 1995), and sweetness-contributing aromatic compounds (Schwieterman, 2014). Furthermore, sugar is one of the major factors affecting overall flavor perception (Godshall, 1995); understanding the composition and concentration of sugars is part of identifying consumer preferences (Beauchamp, 2011). In the case of the texture liking, there was less variation and no significant differences among the genotypes, however,  UFTP 38 and the ‘La Estrella’ were the least liked. The texture of produce is traditionally associated with sensory hardness which in turn is correlated with the presence or level of fibers (Kutle, 2021). The term ‘presence of fibers’ is frequently used, since it refers to the geometric properties of the food particles and is applied to fibrous materials (mainly materials containing hemicellulose and cellulose). These properties can be perceived during sensory analysis by chewing (Lillford, 2011) and are very important to solid vegetable foods. High fiber levels are typically undesirable in vegetable crops, since the consumer usually associates their presence with a lack of succulence in food (Laignier, 2021).  

  Fresh Cooked – Correlation: 

  Data on physicochemical characteristics and sensory ratings were analyzed by Principal Component Analysis (PCA). In this analysis, attributes that presented statistically significant differences amongst other physicochemical characteristics in a correlation matrix were used. These included individual color values (L*, a*, b*), pH, malic acid, Brix, POD, hardness, adhesiveness, cohesiveness, springiness, chewiness, resilience, gumminess, overall liking, appearance liking, color liking, sweetness liking, flavor liking, texture liking, and price. The two components of the biplot PCA model explained 37.01% and 22.06%, respectively, for the total variability (59.07%) found in the original variables (Figure 3). The distance and the angle of the variables from one another using the origin as a vector indicates the positive or negative correlation those variables have. This angular interpretation is further explained with an acute angle indicating a positive correlation, an obtuse angle being a negative correlation and a 90° degree angle showing no correlation (Bro, 2014). Furthermore, the farther a variable is from the origin the more that component is impacting the model (Abdi, 2010).  

  Attributes that had a correlation with the primary components above 0.7 were considered important (Figure 3). The first principal component showed that many sensory profile attributes (overall liking, price, sweetness liking, and flavor liking), were strongly positively correlated, while appearance and color liking showed a strong negative correlation. For the second principal component, texture parameters including hardness, gumminess, and chewiness were positively correlated and negatively correlated to texture liking. PCA analysis confirmed genotype variation in physicochemical and sensory attributes (Figure 4). ‘Waltham Butternut’, ‘La Estrella’, UFTP 8, UFTP 22, and UFTP 24 placed on the right, whereas, on the left, were ‘Soler’, UFTP 38, and UFTP 42. Germplasm lines on the right that were negatively correlated with the second principal component had high overall liking, price, sweetness liking, and flavor liking for consumer acceptability; and high intensity for Brix. For the second principal component, germplasms ‘La Estrella’, UFTP 38, and UFTP 8 were above the axis indicating that these lines expressed more hardness, gumminess, and chewiness compared to other lines. 

  With respect to fresh microwaved pumpkin germplasms, Brix was strongly positively correlated with willingness to pay (0.79). This put UFTP 24 as the most liked genotype, overall. 

  Frozen Cooked– Quality Parameters: 

  There was significant variation in the physiochemical composition among the genotypes when frozen fruit cuts were assayed (Table 4 and Table 5). Parameters which showed a great deal of statistical variation among the genotypes include Brix, malic acid concentration, and pH, whereas POD, hue angle, and texture composition showed the least variation. The UFTP 24 genotype had the lowest Brix values of 5.56 compared to UFTP 8 which had the highest at 12.53. The process of freezing involves the formation of ice crystals which damage the tissue and cells, thus resulting in softer foods (Van Buggenhout, 2006; Paciulli, 2015). This cell rupturing can cause leaching of nutrients including sugar and acids (Caliskan, 2017). However, Brix is considered a reasonable measure of sugar content and the overall evaluation of fruit quality (Harrill, 1998).

  Color values of frozen germplasm lines had L* values ranging from 53.5 – 60.5 indicating a light color for the samples. Additionally, some of the germplasm lines displayed a vibrant orange color with positive a* values (8.04 - 20.43) and b* values (49.93 - 65.72). The highest color values (L*, a*, b*) and chroma (68.88) were seen for UFTP 8 (60.50, 20.43, and 65.72) indicating a bright and highly saturated orange colored pumpkin. However, other genotypes including UFTP 22 and UFTP 24 were not statistically significantly different that UFTP 8 with regards to L* and b* values. Furthermore, the relationship between the chromatic parameters (a*, and b*) should be taken into account to interpret color. This is demonstrated with Soler and La Estrella which had the lowest a* values (8.04, and 8.48) and varying b* values (51.77, and 59.10). These a* and b* values for Soler and La Estrella resulted in the highest hue angle values of 81.5 and 82, respectively, indicating a yellow color. This is due to the difference between the a* and b* values and not just the b* value alone. The increase in chromatic values has previously been reported in salmon (Kono, 2017), cherries (Bilbao-Sainz, 2019), and papaya (Cano,1995) after freezing. Thus, freezing rate was shown to affect surface color.  

  Overall, the texture profile of the genotypes showed slight statistically significant separation amongst germplasm lines except for gumminess, and resilience (Table 5). These germplasm line differences are most likely due to several factors including starch gelatinization, pectin degradation, cell wall breakdown, and cell separation (Gallego-Castillo, 2018; van Dijk,2002). These phenomena translate into a release of water, which decreases tissue stiffness, hardness, and firmness and but increases adhesiveness, springiness, gumminess, and chewiness (Corrigan, 2006; Chandra, 2014). Freezing causes cell wall rupture, whereas just the cooking of fresh squash causes softening via cellular separation, without breakage, and leaving the structure of the cellular wall almost intact (Alvarez, 2005).  

  Frozen Cooked – Consumers Acceptability: 

  Freezing fruits is a value-added process that aids in elongating shelf life and increasing value to fruits that otherwise would not have been used or wasted. Genotypes UFTP 8, UFTP24, and ‘Waltham Butternut’ had the highest overall liking scores (Table 6). Consumer acceptability for these genotypes was further supported by high flavor liking and sweetness liking ratings. As previously noted, consumers typically associate the term “overall liking” with “liking of taste” when compared to other sensory terms such as liking of appearance, odor, and texture (Andersen, 2019; Hellwig, 2022). Additionally, American consumers usually prefer sweeter foods on average (Barthes, 2012), demonstrating the correlation between sweetness and overall liking. 

  The appearance liking preference separated one half of the genotypes from the other with UFTP 8, UFTP 22, UFTP 42 and ‘Waltham Butternut’ being amongst the most preferred. When comparing this to the color liking, only UFTP 8, UFTP 22, and UFTP 42 were most preferred. This could be due to more prevalent shrinking, leading to higher rates of discoloration, or non-uniformity on the outside of the ‘Waltham Butternut’ samples. A clear, statistically significant difference in texture was observed among the genotypes. The four most liked samples- UFTP 8, UFTP 22, UFTP 24 and ‘Waltham Butternut’-were contrasted with UFTP 38 and ‘Soler’ as the least liked. This genotype preference was seen in the willingness to pay per pound of each sample, with the most liked samples being ‘Waltham Butternut’ at $1.40 and UFTP 8 at $1.44 per pound. As previously discussed, overall liking has been shown to be highly correlated with flavor liking and sweetness liking. These attributes were both lower for UFTP 24 (5.98- flavor liking) (6.00- sweetness liking) therefore decreasing the consumer acceptability of this genotype. Additionally, price ranged more with frozen samples (0.81-1.44) ) compared to fresh samples (1.21-1.53) indicating that the effects of freezing negatively affected some genotypes more than others. 

  Frozen Cooked – Correlation: 

  Principal Component Analysis (PCA) was used to analyze physicochemical characteristics and sensory profile of frozen winter squash. Two components of the biplot PCA model explained 42.62% and 25.49% respectively, for the total variability (68.11%) found in the original variables (Figure 5). As previously stated, attributes with correlation to each axis >0.7 were considered important: sensory profile attributes (overall liking, price, sweetness liking, texture liking, and flavor liking) were strongly positively correlated to first principal component (C1), as were physicochemical characteristics (L*, a*, b*, and hardness). Visually, cohesiveness and adhesiveness were strongly negatively correlated (>-0.7) to C1 and with values at -0.68 and -0.65, respectively. This is an indication that texture in an important factor which influences the consumer perception of frozen samples. The second principal component (C2) supports the importance of texture, with texture profile characteristics (gumminess and resilience) and compositional characteristics (POD and pH) on either extreme.  

  The visualization of the PCA illustrates the components of consumer overall liking and willingness to pay (price). Flavor liking and overall liking have similar trends in correlations, with the latter being a better indicator for consumer acceptability (Andersen, 2019; Hellwig, 2022). Factors found to be strongly positively correlated with overall liking include L* and a* color values. This strong correlation (>0.7) for compositional characteristics in the frozen state is slightly different for price with L*, a*, and b* color values all being strongly positively correlated. It should also be noted that Brix no longer has a strong correlation with price, but sweetness liking is highly correlated (0.98). This could be due to other factors such as flavor overwhelming the consumers perception of the sample. As previously discussed, texture attributes were shown to impact consumer perception but were not strongly negatively correlated (>-0.7). However, for frozen samples, texture liking is strongly correlated with both overall liking (0.86) and price (0.91).  

  The C1 sample variation (Figure 6) revealed a separation between the most liked and least liked genotypes. The right side showed the most preferred germplasm lines: ‘Waltham Butternut’, UFTP 8, UFTP 22, and UFTP 24, while the left side included the lesser-liked germplasm lines: ‘Soler’, ‘La Estrella’, UFTP 38, and UFTP 42. ‘Waltham Butternut’ is strongly correlated with appearance and color liking compared to UFTP 8 which is strongly correlated with hardness, L*, and a* color values. Having the most preferred cultivars associated with factors strongly positively correlated with overall liking and price further support the use of these genotypes for the frozen American market. Conversely, UFTP 38 and ‘Soler’ were strongly correlated with cohesiveness and adhesiveness and were negatively correlated with overall liking and price. 

   Table 1. Freshly cooked compositional values including color values (L*a*b*), hue angle, chroma, titratable acidity (TA) (g/L), malic acid (g/100g), Brix/soluble solids content (g/L), peroxidase (POD) (abs/min), for Calabaza germplasms. Plant Science Research and Education Unit in Citra, Florida.2022. 

  Cultivar	L*	a*	b*	Hue Angle	Chroma	pH	TA (g/L)	Malic Acid (mg/100g)	SSC  (Brix)	POD (abs/min)
  Butternut	53.37 b	10.98 c	54.01 c	78.61 b	55.25 c	6.86 f	0.08 c	0.05 c	14.80 c	2.39 a
  UFTP 8	53.71 b	15.96 b	57.96 b	74.68 c	60.17 b	7.35 b	0.03 f	0.02 f	16.93 a	0.56 b
  UFTP 22	56.74 a	19.59 a	65.77 a	73.48 c	68.69 a	6.70 g	0.11 a	0.06 a	15.73 b	0.88 a
  UFTP 24	46.87 e	8.66 cd	47.16 d	79.63 b	47.98 d	7.10 d	0.05 e	0.03 e	14.37 d	1.11 a
  UFTP 38	47.69 ed	11.47 c	41.63 e	75.16 c	43.54 e	7.58 a	0.02 h	0.01 h	10.57 e	0.56 a
  UFTP 42	48.75 d	7.32 d	47.24 d	81.25 b	47.83 d	7.28 c	0.03 g	0.02 g	10.70 e	0.71 b
  Soler	49.47 ed	4.07 e	46.76 d	85.14 a	46.97 d	6.85 f	0.06 d	0.03 d	7.73 f	0.91 b
  La Estrella	51.20 c	9.25 cd	47.14 d	79.20 b	48.29 d	7.01 e	0.09b	0.06 b	7.83 f	1.97 a

  ¹ Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per row. n=10-15 fruits.

  Table 2. Freshly cooked texture profile (hardness, adhesiveness, springiness, cohesiveness, gumminess, chewiness, and resilience) for each Calabaza germplasms using a double compression texture test (TA.XT). Plant Science Research and Education Unit in Citra, Florida.2022.

  Cultivar	Hardness (g)	Adhesiveness (g/sec)	Springiness	Cohesiveness	Gumminess	Chewiness	Resilience
  Butternut	313.73 bc	-53.96 c	0.20 a	0.1454 a	43.93 b	8.80 ab	0.0241 bc
  UFTP 8	393.46 abc	-11.39 a	0.22 a	0.1424 ab	58.49 ab	14.63 ab	0.0372 a
  UFTP 22	361.15 abc	-18.56 ab	0.18 a	0.1151 c	42.62 b	8.14 ab	0.0260 b
  UFTP 24	315.15 bc	-25.50 b	0.21 a	0.1203 c	36.84 b	7.66 b	0.0243 bc
  UFTP 38	418.94 ab	-40.43 c	0.18 a	0.1170 c	46.88 b	8.43 ab	0.0246 bc
  UFTP 42	274.08 c	-48.43 c	0.20 a	0.1311 abc	34.68 b	6.91 b	0.0210 bc
  Soler	309.49 bc	-40.15 c	0.21 a	0.1225 abc	36.68 b	8.03 ab	0.0196 c
  La Estrella	480.82 a	-19.62 ab	0.20 a	0.1423 ab	73.38 a	17.08 a	0.0331 a

  ¹ Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per row. n=10-15 fruits. Texture characteristics n=12 and values are means ± 1 standard deviation.

   

  Table 3. Freshly cooked sensory rating using the 9-point hedonic ratings for overall liking, appearance liking, color liking, sweetness liking, flavor liking, and texture liking. A line scale for price was used ranging from $0.80 - 3.00 per pound.

  Cultivar	Overall Liking	Appearance Liking	Color Liking	Sweetness Liking	Flavor Liking	Texture Liking	Price (Per pound)
  Butternut	6.02 bc	6.98 a	7.10 ab	6.10 bc	6.10 b	6.28 ab	1.43 ab
  UFTP 8	6.37 b	6.04 c	6.27 c	6.38 ab	6.54 ab	6.08 abc	1.46 b
  UFTP 22	6.15 b	5.82 cd	6.10 c	5.93 c	6.25 b	6.07 abc	1.41 ab
  UFTP 24	6.88 a	6.02 c	6.12 c	6.63 a	6.85 a	6.44 a	1.53 a
  UFTP 38	5.37 d	7.36 a	7.17 a	5.26 d	5.43 c	5.60 c	1.18 d
  UFTP 42	6.15 b	6.45 b	6.71 b	5.78 c	6.25 b	5.96 abc	1.34 c
  Soler	5.61 cd	6.54 b	6.73 b	5.01 d	5.52 c	6.02 abc	1.21 cd
  La Estrella	5.97 bc	5.52 d	5.62 d	5.88 c	6.10 b	5.87 bc	1.31 bcd

  ¹ Letters represent two-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per row. n=10-15 fruits. Sensory panel N= 89 panelists and values are means ± 1 standard deviation.

   

  

• Figure 3. Principal component analysis for fresh microwaved sensory attributes and quality parameters

  
  

  Figure 4. Principal component analysis for fresh microwaved samples

  Table 4. Frozen cooked compositional values including color values (L*a*b*), hue angle, chroma, titratable acidity (TA) (g/L), malic acid (g/100g), Brix/soluble solids content (g/L), peroxidase (POD) (abs/min), for Calabaza germplasms. Plant Science Research and Education Unit in Citra, Florida.2022. 

  Cultivar	L*	a*	b*	Hue Angle	Chroma	pH	TA (g/L)	Malic Acid (mg/100g)	SSC  (Brix)	POD (abs/min)
  Butternut	55.82 cd	14.75 b	56.11 d	75.26 c	58.04 c	5.66 f	0.10 d	0.06 d	5.80 f	2.08 a
  UFTP 8	60.50 a	20.43 a	65.72 a	72.77 d	68.88 a	6.14 ab	0.12 b	0.07 b	12.53 a	1.07 c
  UFTP 22	57.83 abc	16.23 b	61.45 abc	75.28 c	63.59 b	6.03 c	0.06 h	0.04 h	5.86 ef	1.24 b
  UFTP 24	59.43 ab	11.13 c	63.24 ab	80.16 b	64.26 b	6.10 b	0.07 g	0.04 g	5.56 g	1.25 b
  UFTP 38	55.12 cd	13.71 b	49.93 e	74.80 c	51.83 d	5.92 d	0.09 f	0.05 f	7.56 c	1.35 b
  UFTP 42	57.11 bc	15.62 b	58.68 cd	75.20 c	60.74 bc	5.95 d	0.09 e	0.06 e	6.83 d	1.36 b
  Soler	53.49 d	8.04 d	51.77 e	81.49 ab	52.46 d	5.88 e	0.11 c	0.07 c	8.83 b	1.21 bc
  La Estrella	55.40 cd	8.48 d	59.10 cd	82.00 a	59.77 bc	6.15 a	0.16 a	0.10 a	5.96 e	1.34 b

   ¹ Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per row. n=10-15 fruits. Texture characteristics n=12 and values are means ± 1 standard deviation.

   

  Table 5. Frozen cooked texture profile (hardness, adhesiveness, springiness, cohesiveness, gumminess, chewiness, and resilience) for each Calabaza germplasms using a double compression texture test (TA.XT). Plant Science Research and Education Unit in Citra, Florida.2022.

  Cultivar	Hardness (g)	Adhesiveness (g/sec)	Springiness	Cohesiveness	Gumminess	Chewiness	Resilience
  Butternut	592.60 cd	-66.83 c	0.35 a	0.1870 bc	107.11 e	38.01 c	0.0308 e
  UFTP 8	2365.60 a	-60.62 bc	0.28 bc	0.1569 c	345.39 a	92.93 a	0.0429 dc
  UFTP 22	1127.00 bc	-51.68 ab	0.28 bc	0.2049 b	223.87 bc	61.35 b	0.0409 d
  UFTP 24	1425.90 b	-63.96 bc	0.27 bc	0.2140 b	297.53 ab	78.91 ab	0.0533 a
  UFTP 38	595.50 cd	-52.67 ab	0.30 b	0.3256 a	190.08 cd	58.62 b	0.0493 abc
  UFTP 42	638.20 cd	-46.58 a	0.27 bc	0.2204 b	128.88 de	34.09 c	0.0425 d
  Soler	427.30 d	-43.64 a	0.25 c	0.2977 a	124.62 de	31.04 c	0.0437 bcd
  La Estrella	1592.90 b	-64.84 bc	0.30 b	0.1923 bc	280.80 ab	83.07 a	0.0499 ab

  ¹ Letters represent one-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per row. n=10-15 fruits. Texture characteristics n=12 and values are means ± 1 standard deviation.

   

  Table 6. Frozen cooked sensory rating using the 9-point hedonic ratings for overall liking, appearance liking, color liking, sweetness liking, flavor liking, and texture liking. A line scale for price was used ranging from $0.80 - 3.00 per pound.

  Cultivar	Overall Liking	Appearance Liking	Color Liking	Sweetness Liking	Flavor Liking	Texture Liking	Price (Per pound)
  Butternut	6.34 a	6.72 abc	6.77 bc	6.62 a	6.40 a	6.00 a	1.40 a
  UFTP 8	6.40 a	7.11 a	7.14 ab	6.56 a	6.40 a	6.32 a	1.44 a
  UFTP 22	5.62 bc	7.11 a	7.26 a	6.00 b	5.17 bc	6.11 a	1.17 b
  UFTP 24	5.97 ab	6.41 cd	6.46 cd	6.00 b	5.98 ab	5.84 a	1.22 b
  UFTP 38	5.09 d	6.26 d	6.24 d	5.44 c	5.26 d	4.31 c	0.90 e
  UFTP 42	5.30 cd	6.84 ab	7.06 ab	5.67 bc	5.23 cd	5.18 b	1.07 c
  Soler	4.41 e	6.60 cd	6.53 cd	4.92 d	4.43 e	4.60 c	0.81 e
  La Estrella	5.02 d	4.84 e	4.76 e	5.50 c	5.08 d	5.17 b	0.98 cd

  ¹ Letters represent two-way ANOVA with Duncan MRT results and mean separation to determine significant differences at α>0.05 per row. n=10-15 fruits. Sensory panel N= 90 panelists and values are means ± 1 standard deviation.

   

•  

  
  Figure 5. Principal component analysis for fresh microwaved sensory attributes and quality parameters

• 
• Figure 6. Principal component analysis for frozen microwaved samples
  

  Activity 3: Determination of macronutrients and bioactive compounds of calabaza germplasm

  Moisture Content and Total Soluble Solids Content:

  The average moisture content of the calabaza varieties and butternut squash ranged from 76.30 to 94.4% on a wet weight basis, as shown in Table 3-2. UFTP45 was found to have the highest moisture content, while UFTP46 had the lowest observed moisture content. Samples from lines UFTP 8, UFPT22, UFTP38, UFTP48, and UFTP47 had significantly higher moisture content than samples from UFTP24 and UFTP57.  The butternut squash used as a control contained 87.6% moisture, with a standard deviation of 0.8, which did not significantly differ from the reported value for moisture content of butternut squash, 86.4% (USDA FDC).

  Total soluble solids of the calabaza varieties and butternut squash, reported in °Brix, ranged from 5.2 to 13.2 (Table 3-2). UFTP46 was found to have the highest soluble solids content among the calabaza varieties, while UFTP45 had the lowest concentration of soluble solids. This finding demonstrates the inversely proportional relationship between moisture and soluble solids concentrations within the mesocarp of the calabaza fruit and is in accordance with Loy’s (2004) claim that greater fresh weight, attributed to higher moisture, may be inversely proportional to consumer acceptability as it relates to sweetness of the edible portion of calabaza fruit. If 11% total soluble solids is to be considered the minimum threshold for consumer palatability, as determined by Loy (2004), then only one of the experimental varieties of calabaza would meet this standard, UFTP46 with 13.2% soluble solids, in addition to the butternut squash control which was found to have an average of 11.6% soluble solids. Average soluble solids were significantly higher in samples from lines UFTP8 and UFTP24 than in samples from lines UFTP22, UFTP47, and UFTP57. Along with average TSS in samples from UFTP45, average TSS in samples from lines UFTP38 and UFTP42 did not differ significantly.

  Carotenoid Concentration:

  The average concentrations of carotenoids among the calabaza varieties, expressed as β-carotene equivalents, ranged from 17.4 to 73.5 micrograms per one gram of raw calabaza mesocarp on a wet weight basis (Table 3-3), 21.0 to 278 micrograms per gram on a dry weight basis. UFTP57 had the highest concentration of β-carotene equivalent carotenoids and, therefore, also retinol activity equivalents (RAE), but concentrations did not vary significantly from lines UFTP47, UFTP42, UFTP24, or UFTP8. UFTP45 had the lowest concentrations of β-carotene equivalent carotenoids but only varied significantly from two other lines, UFTP8 and UFTP57.

  The measured concentrations of β-carotene equivalent carotenoids for each calabaza variety, as well as the butternut squash control, fell within the range of β-carotene concentrations published by the USDA’s FoodData Central for raw winter squash (butternut), specifically 13.6 to 83.8 μg β-carotene per one gram of mesocarp. Concentrations of β-carotene equivalent carotenoids determined in this study were higher than those determined by Itle and Kabelka (2009); from six different varieties of Cucurbita moschata they found β-carotene concentrations ranging from 0.2 to 15.3 micrograms per gram on a wet weight basis using high-performance liquid chromatography. This could be because Itle and Kabelka (2009) determined β-carotene concentrations in Cucurbita moschata varieties exhibiting a wide range of mesocarp colors, from light yellow to orange-red, suggesting a wide range of β-carotene concentrations as β-carotene has already been associated with “deep orange flesh colour” (Robinson and Decker-Walters 1997). Using liquid chromatography analysis of lyophilized samples from different varieties of Cucurbita moschata, Azevedo-Meleiro and Rodriguez-Amaya (2007) and Kulcyński and Gramza-Michalowska (2019) determined concentrations of β-carotene to range from 56.7 to 66.7 micrograms per gram of dry matter from two sample C. moschata varieties and 12.9 to 52.6 micrograms per gram dry matter from six sample C. moschata varieties, respectively. This discrepancy between results of these studies and those of the study presented here can be explained by the methods of analysis chosen for each. Chromatography is a very precise and accurate primary method of chemical analysis that extracts and separates individual target compounds, such as β-carotene. Spectrophotometry is less precise as it measures absorbance and transmittance values at a given wavelength for all compounds extracted but not separated (e.g., total carotenoids). Therefore, the concentrations reported in this study are likely an overestimation of β-carotene in the calabaza samples. Though chromatographic methods of analysis are more precise and accurate than spectrophotometry, the latter is more cost-effective, less labor intensive, and more rapid, maintaining its status as a valuable secondary method of analysis.

  The dietary reference intakes (DRI) for RAE, based on the estimated average requirements (EAR) for humans based on their gender and age, are 210 μg RAE/day for male and female children aged 1 through 3 years, 275 μg RAE/day for male and female children aged 4 through 8 years, 445 μg RAE/day for males and 420 μg RAE/day for females aged 9 through 13 years, 630 μg RAE/day for males and 485 μg RAE/day for females aged 14 through 18 years, 625 μg RAE/day for males and 500 μg RAE/day for females aged ≥19 years (Institute of Medicine 2006). Table 3-4 shows the percentage of the DRI for retinol activity equivalents for each group specified previously that a single serving of each experimental calabaza variety and butternut squash would provide assuming that a serving size of calabaza is equal to 100 grams and that a negligible amount of β-carotene is lost during typical cooking (Kläui and Bauernfeind 1981).

  Samples from each breeding line, as well as samples of the butternut squash, would, on average, be considered “excellent” sources of retinol activity equivalents across every life stage and regardless of gender. An “excellent” source of a nutrient is defined as that which provides ≥ 20% of the respective DRI per reference amount customarily consumed (FDA 2022). Fruits from lines UFTP8, UFTP42, and UFTP57 were among the top samples consistently providing the highest percentages of the DRI for RAE at each life stage, though the carotenoid profile is complex and β-carotene is not the only carotenoid present.

  Ascorbic Acid Concentration:

  Average ascorbic acid contents among the experimental varieties of calabaza varied significantly and ranged from 51.1 to 207 micrograms per single gram of raw mesocarp (Table 3-5). The highest concentrations of ascorbic acid were found in samples from breeding lines UFTP46 and UFTP24, while the lowest concentrations of ascorbic acid were observed in samples from breeding line UFTP45 and in samples of the butternut squash used as a control. Databases and literature values support ascorbic acid concentrations of 166 to 229 micrograms per gram of raw winter squash (butternut) on a wet weight basis (Hopp and Merrow 1963; USDA FDC; Roura et al. 2007).

  The DRIs for vitamin C (“ascorbic acid”), based on the EARs for humans at different life stages and by gender, are 13 mg/day for male and female children aged 1 through 3 years, 22 mg/day for male and female children aged 4 through 8 years, 39 mg/day for male and female children aged 9 through 13 years, 63 mg/day for males and 56 mg/day for females aged 14 through 18 years, 75 mg/day for males and 60 mg/day for females aged ≥19 years (Institute of Medicine 2006). Table 3-6 shows the percentage of the DRI for vitamin C (“ascorbic acid”) for each group specified previously that a single serving of each experimental calabaza variety and butternut squash would provide assuming that a serving size of calabaza is equal to 100 grams. It is important to note that ascorbic acid is sensitive to thermal treatments, such as those used in standard home cooking procedures. Roura et al. (2007) found that approximately 20% of the ascorbic acid content of fresh butternut squash was lost during a simulated cooking regime of 30 minutes at 95 ± 2 ºC.

  Each breeding line and butternut squash produced samples that, on average, would serve as “excellent” sources of vitamin C as a function of ascorbic acid content. A “good” source is one which provides 10 – 19% of a respective DRI per reference amount customarily consumed (FDA 2022). Samples from breeding lines UFTP22, UFTP38, UFTP45, UFTP47, and butternut squash produced samples that could be considered “good” sources of vitamin C as a function of ascorbic acid for children aged 9 through 13 years. Samples from lines UFTP8, UFTP24, UFTP42, UFTP46, and UFTP57 would serve as “excellent” sources of ascorbic acid. A single serving of calabaza from each experimental variety, and excluding butternut squash, could be considered a “good” source of ascorbic acid, with a few exceptions; samples from line UFTP45 would not be considered “good” sources of vitamin C as a function of ascorbic acid  for males aged 14 years and older, while samples from UFTP22 would not be considered “good” sources for males 19 years of age and older. Samples from lines UFTP24 and UFTP46 were most consistently found to be “excellent” sources of ascorbic acid across the life stages.

  Phenolic Compound Concentration:

  The average concentration of phenolic content, expressed as gallic acid equivalents (GAE), among the experimental varieties of calabaza and the butternut squash control ranged from 168.4 to 526.6 micrograms GAE per gram of raw mesocarp (Table 3-7). Line UFTP46 produced samples with the significantly highest average concentration followed by samples from line UFTP24, while line UFTP38 produced samples with the lowest average concentrations.

  The USDA’s FoodData Central does not provide information on the total phenolic compound concentration of calabaza or any other related winter squash, raw or otherwise. Researchers have reported phenolic compound concentrations of 159.3 ± 5.7 micrograms GAE per gram of mesocarp in “pumpkin” (no species designation) on a wet weight basis using the Folin-Ciocâlteu method (Chun et al. 2005). Other researchers have also used the Folin-Ciocâlteu method to enumerate total phenolic concentrations, in C. moschata species specifically, and suggest a wide range of 69.7 to 974.3 micrograms GAE per gram of mesocarp on a wet weight basis (Deng at al. 2013; Mokhtar et al. 2021).

  There are no DRIs for total phenolic content, gallic acid, or antioxidants in general, save vitamins and minerals which also happen to confer biologically relevant antioxidant capacity (e.g., β-carotene, vitamin C).

  Figure 3-2 shows the average concentrations of each bioactive compound analyzed in this study between each breeding line and butternut squash. Samples from lines UFTP57 and UFTP8 showed the highest concentrations of β-carotene concentrations, but these concentrations did not vary significantly from those found in samples from lines UFTP24, UFTP42, and UFTP47. Samples from lines UFTP46 and UFTP24 were found to have the highest concentrations of ascorbic acid, though concentrations in UFTP24 did not significantly differ from several other lines. Calabaza from line UFTP46 was found to have a significantly higher concentration of phenolic content (GAE) than any other experimental line and butternut squash. This study found that, among the breeding lines analyzed, UFTP24 and UFTP46 stand to confer the most nutritional benefit to human health as it relates to the bioactive and antioxidant capacities of β-carotene, ascorbic acid, and total phenolic content.

  Commercial Calabaza Cultivars:

  Following experimentation, it was revealed that UFTP45 and UFTP46 are commercially available calabaza cultivars. The UFTP45 breeding line was revealed to be the cultivar ‘Soler’, released from the Agricultural Experiment Station at the University of Puerto Rico in 2004 (Wessel-Beaver 2005). Breeding line UFTP46 was revealed to be the cultivar ‘La Estrella’, developed by breeding programs at the University of Florida and released in 2002 (Maynard et al. 2002).

  In this study, these two cultivars performed vastly differently from each other compositionally. ‘Soler’ was among the lines with the highest moisture content (94.4%) and lowest soluble solids content (5.20 °Brix), while ‘La Estrella’ was among the lines with the lowest moisture content (76.3%) and highest soluble solids content (13.2 °Brix) (Table 3-2). ‘Soler’ exhibited one of the lowest concentrations of β-carotene equivalent concentrations among the lines analyzed in this study (17.4 μg βCE/g), while ‘La Estrella’ had approximately twice as much as ‘Soler’ (36.9 μg βCE/g), though it did not rank among the highest lines for β-carotene equivalent concentrations (Table 3-3). In a comparison of ascorbic acid content ‘Soler’ was found to have some of the lowest concentrations among the lines analyzed in this study (58.3 μg/g), while ‘La Estrella’ was found to have the highest concentration of ascorbic acid (207.4 μg/g) among all lines studied (Table 3-5). Among the breeding lines ‘Soler’ was found to have some of the lowest concentrations of phenolic content (193.6 μg GAE/g), while ‘La Estrella’ had the highest concentration of phenolics among all lines analyzed (526.6 μg GAE/g) (Table 3-7). Combining the results of these analyses, ‘Soler’ is associated with a less sweet-tasting mesocarp that has less health-promoting antioxidants than the sweeter-tasting ‘La Estrella’, which also boasts greater antioxidant capacity beneficial to human health.

  Applying nutrient content claims, as defined by the Food and Drug Administration, both ‘Soler’ and ‘La Estrella’ could be considered “excellent” sources of retinol activity equivalents because a single serving (approximately 100 grams) of either would provide at least 20% of the dietary reference intake for RAE across all life stages. ‘Soler’ could be considered a “good” source of vitamin C (10-19% DRI) across most life stages, with the exception of males 14 years of age and older (Table 3-5). In contrast, ‘La Estrella’ could be considered an “excellent” source of vitamin C across all life stages (Table 3-5). As previously stated, no DRIs have been established for phenolic content so nutrient content claims cannot be applied to ‘Soler’ or ‘La  Estrella’ in that regard.

   

  Table 3-2. Average moisture content  and °Brix of calabaza samples from each breeding line and samples from butternut squash.

  Breeding Line	Moisture Contentwb (%)			°Brix
  Waltham Butternut Squash	87.6	±	0.80ab	11.6	±	0.99ab
  UFTP8	87.7	±	1.90ab	9.70	±	0.68bc
  UFTP22	91.5	±	0.58ab	7.96	±	0.57cd
  UFTP24	84.1	±	1.29b	10.2	±	0.42bc
  UFTP38	91.4	±	2.20ab	6.67	±	0.90de
  UFTP42	91.6	±	1.15ab	7.02	±	0.59de
  UFTP45	94.4	±	1.55a	5.20	±	0.52e
  UFTP46	76.3	±	1.02c	13.2	±	1.80a
  UFTP47	89.4	±	2.86ab	8.38	±	0.83cd
  UFTP57	86.5	±	4.28b	8.66	±	0.35cd

  Average ± standard deviation; n = 3, except UFTP47 (n = 8) and UFTP57 (n = 9); Averages sharing the same letter do not differ significantly from one another.

   

  Table 3-3. Average β-carotene equivalent carotenoid concentrations (βCE) and calculated retinol activity equivalents (RAE) of calabaza samples from each breeding line and samples from butternut squash.

  Breeding Line	Carotenoids (μg βCE/gwb)			RAE (μg/gwb)
  Waltham Butternut Squash	27.7	±	6.87bc	2.31	±	0.57
  UFTP8	54.9	±	13.38ab	4.58	±	1.11
  UFTP22	36.4	±	4.64bc	3.03	±	0.39
  UFTP24	44.4	±	12.78abc	3.70	±	1.07
  UFTP38	30.0	±	8.94bc	2.50	±	0.75
  UFTP42	48.4	±	4.53abc	4.04	±	0.38
  UFTP45	17.4	±	9.78c	1.45	±	0.81
  UFTP46	36.9	±	9.48bc	3.07	±	0.79
  UFTP47	35.6	±	10.85abc	2.97	±	0.90
  UFTP57	73.5	±	27.35a	6.13	±	2.28

  Average ± standard deviation; n = 3; Averages sharing the same letter do not differ significantly from one another.

  
  

  Table 3-4. Percent Dietary Reference Intake (DRI) per life stage for retinol activity equivalents (RAE) provided by 100 grams raw calabaza.

  	1 - 3 y	4 - 8 y	9 - 13 years		14 - 18 years		≥19 years
  Breeding Line			Males	Females	Males	Females	Males	Females
  Waltham Butternut Squash	110	84	52	55	37	48	37	46
  UFTP8	218	166	103	109	73	94	73	92
  UFTP22	144	110	68	72	48	63	49	61
  UFTP24	176	135	83	88	59	76	59	74
  UFTP38	119	91	56	59	40	51	40	50
  UFTP42	192	147	91	96	64	83	65	81
  UFTP45	69	53	33	34	23	30	23	29
  UFTP46	146	112	69	73	49	63	49	61
  UFTP47	141	108	67	71	47	61	47	59
  UFTP57	292	223	138	146	97	126	98	123

   

  
  

  Table 3-5. Average ascorbic acid concentrations of raw calabaza samples from each breeding line and from the control, butternut squash.

  Breeding Line	Ascorbic Acid (μg/gwb)
  Waltham Butternut Squash	51.09	±	3.92c
  UFTP8	92.96	±	11.87bc
  UFTP22	66.79	±	4.20bc
  UFTP24	138.9	±	10.74ab
  UFTP38	72.39	±	9.03bc
  UFTP42	85.98	±	18.08bc
  UFTP45	58.26	±	7.45c
  UFTP46	207.4	±	71.60a
  UFTP47	73.42	±	12.19bc
  UFTP57	96.50	±	29.73bc

  Average ± standard deviation; n = 3; Averages sharing the same letter do not differ significantly from one another.

   

   

   

   

  Table 3-6. Percent Dietary Reference Intake (DRI) per life stage for ascorbic acid (vitamin C) provided by 100 grams raw calabaza.

  				14 - 18 years		≥19 years
  Breeding Line	1 - 3 years	4 - 8 years	9 - 13 years	Males	Females	Males	Females
  Waltham Butternut Squash	39**	23**	13*	8	9	7	9
  UFTP8	72**	42**	24**	15*	15*	12*	15*
  UFTP22	51**	30**	17*	11*	11*	9	11*
  UFTP24	107**	63**	36**	22**	23**	19*	23**
  UFTP38	56**	33**	19*	11*	12*	10*	12*
  UFTP42	66**	39**	22**	14*	14*	11*	14*
  UFTP45	45**	26*	15*	9	10*	8	10*
  UFTP46	160**	94**	53**	33**	35**	28**	35**
  UFTP47	56**	33**	19*	12*	12*	10*	12*
  UFTP57	74**	44**	25**	15*	16*	13*	16*

  * denotes “good” source (10 -19% DRI); ** denotes “excellent” source (≥ 20% DRI).

  Table 3-7. Average total phenolic compound concentrations of raw calabaza samples from each breeding line and from the control, butternut squash.

  Breeding Line	Phenolics (μg GAE/gwb)
  Waltham Butternut Squash	265.83	±	65.80bc
  UFTP8	238.43	±	37.58c
  UFTP22	225.23	±	20.16c
  UFTP24	377.77	±	44.96b
  UFTP38	168.42	±	33.58c
  UFTP42	181.23	±	22.71c
  UFTP45	193.62	±	14.74c
  UFTP46	526.59	±	84.44a
  UFTP47	241.04	±	8.82c
  UFTP57	263.77	±	20.06bc

  Average ± standard deviation; n = 3; Averages sharing the same letter do not differ significantly from one another.

   

   

   

   

   

  

  Figure 3-2. Average bioactive compound concentrations for each breeding line plus butternut squash samples.

  Activity 4: Correlations between bioactive compound concentrations and tristimulus colorimetric values to approximate moisture, soluble solids, carotenoid content, ascorbic acid content, and phenolic content in calabaza mesocarp

   

  Correlations between Colorimetric Values and Bioactive Compounds of Randomly Selected Calabaza Subsamples

  Significant correlations (Pearson, r) were observed between the measured L*, a*, and b* readings taken at a single point along the equatorial circumference of each calabaza sample, as well as the respective chroma value (C*) and hue angle (h_a,b) calculations, and target compound concentrations within random subsamples of calabaza mesocarp from each fruit sample (Table 4-2). Weak correlations were observed between L* (lightness) and all target compounds analyzed, save total soluble solids content, which showed a moderately strong, positive, and statistically significant correlation with L*, r(28) = .42, p < .05. The measured colorimetric values for a* (redness-greenness) showed a strong, negative correlation with moisture content (r(64) = -.51, p < .0001) and strong, positive correlations with β-carotene equivalent concentrations (r(34) = .54, p < .001), ascorbic acid content (r(35) = .44, p < .01), and total soluble solids (r(28) = .64, p < .001). A moderately strong, positive correlation was observed between a* and total phenolic content extracted from the calabaza subsamples, but it was not statistically significant (p > .05). A statistically significant correlation was found between measured b* (yellowness-blueness) values and each target compound analyzed; moisture content showed a strong, negative correlation with b* (r(64) = -.56, p < .0001), while β-carotene equivalent concentration, ascorbic acid content, and total soluble solids showed strong, positive correlations with b* (r(34) = .56, p < .001, r(35) = .51, p < .01, r(28) = .60, p < .001, respectively). Though significant, total phenolics shared only a moderately strong, positive correlation with b* (r(28) = .41, p < .05).

  Chroma (C*) and hue angle (h_a,b) were calculated from the L*, a*, and b* measurements and showed significant correlations with the analyzed compounds extracted from the calabaza subsamples. Moisture content exhibited a strong, negative correlation with C* (r(64) = -.57, p < .0001), while β-carotene equivalent concentration, ascorbic acid content, and total soluble solids showed strong, positive correlations with b* (r(34) = .57, p < .001, r(35) = .51, p < .01, r(28) = .62, p < .001, respectively). Total phenolic content showed a positive, moderately strong correlation with C* (r(28) = .41, p < .05). Moisture content showed a moderately strong, positive correlation with h_a,b that was statistically significant (r(64) = .48, p < .0001). Beta-carotene equivalent concentrations and ascorbic acid content also demonstrated significant, moderate-strength correlations with h_a,b, though negatively correlated (r(34) = -.48, p < .001, r(35) = -.40, p < .05, respectively). Total soluble solids showed a strong, negative correlation with h_a,b (r(28) = -.64, p < .001). Finally, total phenolic content showed a negative, moderate-strength correlation with h_a,b, but it was not found to be statistically significant (p > .05).

  These results suggest that as redness of the calabaza mesocarp increases, moisture content decreases, while also indicating that β-carotene equivalent concentrations, ascorbic acid content, and total soluble solids may be increased. Similar to redness and moisture, increased yellow pigmentation in calabaza mesocarp may indicate decreased moisture content. Increased yellowness may also indicate increased β-carotene equivalent concentrations, ascorbic acid, total soluble solids, and total phenolic compounds. Increased saturation of the hue of the mesocarp may indicate decreased moisture content; that is, a dull hued mesocarp is likely to have more moisture than a clear, brightly hued mesocarp. In contrast, increased saturation of hue may indicate increased concentrations of β-carotene equivalent concentrations, ascorbic acid, total soluble solids, and total phenolic compounds. Lastly, results show that an increased hue angle may indicate increased moisture, which is in accordance with other findings of this study that correlate yellowness and redness with decreased moisture. Since increased yellowness and redness were also found to be likely indicators of increased β-carotene equivalent carotenoid concentrations, ascorbic acid, total soluble solids, and total phenolic compounds, it should hold that these target compound concentrations should be decreased when hue angle is increased, which is supported by the inverse correlation found between these compound concentrations and hue angle.

  The directions of the correlations reported by Simonne et al. (1993) and Itle and Kabelka (2009) support the directions of correlations found in this study, however a discrepancy in the strengths of correlations may be attributed to the methods of analysis used in each study and, in the case of research conducted by Simonne et al. (1993), the use of a different color space (Hunter L a b vs. CIEL*a*b* used here).

  Significant correlations were also observed between L*, a*, and b* measurements, C* and h_a,b calculations, and the phenotypic characteristics of the calabaza samples, specifically mass, widest circumference, and height measured from blossom end to stem beginning following a longitudinal cut through the middle of each fruit sample (Table 4-3). Weak, negative correlations were observed between L* and mass and L* and circumference while almost no correlation was observed between L* and fruit height. A weak, negative correlation was found between a* and mass of the sample fruits, but the association was not significant (p > .05). Equatorial circumference showed a significant, moderate-strength correlation with a*, negatively correlated (r(64) = -.34, p < .01), while fruit height also showed a significant, moderate-strength correlation with a*, positively correlated (r(64) = .24, p < .05). A significant, though weak, negative correlation was found between circumference and b* (r(64) = -.29, p < .05). A weak, negative correlation between mass and b* and a weak, positive correlation between height and b* were also observed, though neither were statistically significant (p > .05). Chroma exhibited one significant, negative correlation, though weak, with circumference (r(64) = -.28, p < .05). It also showed weak correlations with mass, negatively correlated, and height, positively correlated. Hue angle was found to have significant correlations, though only of weak to moderate strength, with circumference (r(64) = .34, p < .01) and height (r(64) = .25, p< .05).

  Moderate-strength correlations between a* and circumference and height of calabaza fruit may suggest that as fruits widen the redness of the mesocarp decreases. In contrast, redness may increase as fruit height increases. In general, though some of the correlations between color space values and calculations and phenotypic characteristics of calabaza fruit showed statistical significance, none of the correlations were particularly strong (-.50 < r  < .50) suggesting that phenotypic characteristics are not related to or good predictors of calabaza mesocarp colorimetric values.

  Significant correlations were also observed between proximate and bioactive compound concentrations extracted from random subsamples from each sample calabaza fruit (Table 4-4). Moisture content only weakly correlated with mass and height, with the latter practically not at all. A moderately strong, positive correlation was observed between moisture and circumference, and the correlation was statistically significant (r(64) = .32, p < .01). Moisture negatively correlated with target and bioactive compound concentrations within the subsamples of mesocarp; it was moderately correlated with β-carotene (r(34) = -.41, p < .05), strongly correlated with ascorbic acid (r(35) = -.84, p < .0001), strongly correlated with total soluble solids (r(28) = -.87, p < .0001), and strongly correlated with total phenolic compounds (r(28) = .88, p < .0001). Besides its correlation with moisture content, β-carotene concentration was only found to significantly correlate with ascorbic acid, and only moderately so (r(34) = .44, p < .01). Concentrations of β-carotene showed weak and negatively correlated associations with phenotypic characteristics of calabaza samples and weak and positively correlated associations with soluble solids content and total phenolics (p > .05). Like β-carotene concentrations, ascorbic acid content was only weakly and negatively associated with phenotypic characteristics of calabaza fruit; however, ascorbic acid did demonstrate a strong, positive correlation with total soluble solids content (r(28) = .65, p < .0001) and total phenolics (r(28) = .83, p < .0001). Strong, negative correlations were observed between total soluble solids and mass (r(28) = -.55, p < .01) and circumference (r(28) = -.66, p < .0001), but only a weak, positive correlation was observed between soluble solids and height (p > .05). There was a strong, positive correlation found between soluble solids and total phenolic content (r(28) = .81, p < .0001). Lastly, total phenolic content demonstrated moderate, negative correlations with mass and circumference of calabaza fruit, though neither association was statistically significant (p > .05), and almost no correlation with height of calabaza fruit.

  These results suggest that calabaza higher in moisture content may have lower concentrations of β-carotene equivalent carotenoids, ascorbic acid, soluble solids, and total phenolic compounds than calabaza fruits lower in moisture. This observed inverse relation of moisture with bioactive compounds and soluble solids concentrations within calabaza mesocarp, and to a lesser extent, the observed positive association of total soluble solids with bioactive compounds,  may be useful as a harvest index not only for fruit maturity, but also for phytonutrient density, especially with regard to total phenolic compound concentrations.

  A “fairly good” negative correlation was reported by Culpepper and Moon (1945) between mature fruit size and total solids content in Cucurbita spp. This supports the strong, negative correlation of TSS with mass and circumference observed in this study since TSS is a fraction of total solids. Results of this study are comparable to those of Kulczyński and Gramza-Michalowska (2019) who observed a significant, moderate strength, positive correlation between total carotenoid content and phenolic content (r = .44, p < .05) and phenolic content and ascorbic acid, reported as vitamin C (r = .42, p < .05)  in another cucurbit species, Cucurbita pepo. Unlike results of the present study, Kulczyński and Gramza-Michalowska (2019) found no statistically significant correlations between bioactive compounds when varieties of Cucurbita moschata were analyzed. Differences in the results of these studies may be attributed to the analytical methodology chosen; HPLC, a more precise method of analysis, was used to analyze carotenoids and vitamin C in Kulczyński and Gramza-Michalowska’s (2019) study.

  The strong, positive correlation between ascorbic acid and TPC is most likely due to the method for extraction of total phenolics used in this study; the Folin-Ciocâlteu method measures phenolic content, as well as other reducing agents found within a given substrate and, therefore, likely reflects the content of ascorbic acid. Attempts were made to mitigate this overlap by incubating samples for 48 hours, however future work may consider other FC protocols with different incubation periods better suited for phenolic extraction from calabaza, specifically.

  Correlations between Colorimetric Values and Bioactive Compounds of Non-randomized Calabaza Subsamples

  Significant correlations (Pearson, r) were observed between colorimetric readings and calculations and target bioactive compounds extracted from 2 gram, non-randomized plugs of calabaza subsamples (Table 4-5). Similar to the results reported in section 4.3.1, L* showed only weak correlations with β-carotene and ascorbic acid. In this experiment, L* also showed a weak correlation with total phenolic compound concentrations, but a moderate correlation with total soluble solids, though the correlation was not statistically significant (p > .05). Concentrations of β-carotene equivalent carotenoids and TPC concentrations showed moderately strong, positive correlations with a* colorimetric values, but were not statistically significant (p > .05), unlike correlations between a* and ascorbic acid (r(14) = .61, p < .05) and a* and TSS (r(13) = .65, p < .01) which were strong. Concentrations of β-carotene equivalent carotenoids and ascorbic acid demonstrated strong, positive correlations with b* colorimetric values (r(13) = .64, p < .05 and r(14) = .59, p < .05, respectively). A positive, moderate-strength correlation was observed between b* and TSS, while a weak but positive correlation was observed between b* and TPC. Neither correlation was found to be statistically significant (p > .05). Concentrations of β-carotene equivalent carotenoids and ascorbic acid were found to significantly correlate with calculated C* values (r(13) = .66, p < .01 and r(14) = .61, p < .05, respectively), both exhibiting a strong, positive relation. Chroma only moderately correlated with TSS and weakly correlated with TPC. Calculated h_a,b showed a moderate-strength, negative correlation with β-carotene equivalent carotenoids and TPC, but neither correlation was found to be statistically significant (p > .05). There were strong, negative correlations observed between h_a,b and ascorbic acid and TPC that were found to be statistically significant (r(14) = -.55, p < .05 and r(13) = -.67, p < .01, respectively).

  Results of this study further support that redness of calabaza mesocarp may indicate increased concentrations of ascorbic acid and total soluble solids, though results did not support redness as an indication of β-carotene equivalent carotenoid concentrations as they did in Section 4.3.1. Yellowness of calabaza mesocarp may be an indication of β-carotene equivalent carotenoid concentrations and ascorbic acid, as suggested previously (Section 4.3.1), but not of TSS or TPC. A bright, clear hue (denoted by increased C*) of calabaza mesocarp may indicate increased concentrations of β-carotene equivalent carotenoids and ascorbic acid, but there was insufficient evidence to support saturation of hue as an indication of increased TSS or TPC in this experiment. Similar to previous results, results from this experiment suggest that the farther away the color of mesocarp is from red and yellow along the a*, b* plane within the 3-dimensional CIEL*a*b* color space, the lower the concentrations of bioactive compounds, particularly ascorbic acid and TSS.

  

  Figure 4-1. Schematic illustrating triplicate colorimetric measurements (‘x’) taken from skin side of a plug of mesocarp subsample (skin removed). Opposite side of calabaza reserved for processing for subsequent moisture, soluble solids, β-carotene equivalent carotenoids, ascorbic acid, and total phenolic compound analyses.

   

  

   

  Figure 4-2. Subsample plugs of mesocarp, each measured for color (‘x’) on both sides of plug, in triplicate (skin removed) for a total of six colorimetric measures per plug. Each plug was designated for a specific analysis, either of β-carotene equivalent carotenoids, ascorbic acid, soluble solids, or total phenolic content.

  
  

  Table 4-1. Average mass, equatorial circumference, height, and color space values of (n = 3) calabaza samples from each breeding line, plus butternut squash samples; average ± standard deviation.

  Breeding Line	Mass (kg)	Equatorial Circumference (cm)	Height (cm)	L*	a*	b*
  Waltham Butternut	0.429 ± 0.12	24.39 ± 3.00	14.34 ± 1.89	71.84 ± 1.67	17.07 ± 0.72	70.31 ± 0.19
  UFTP8	1.189 ± 0.02	47.89 ± 1.51	10.07 ± 0.45	69.80 ± 0.46	14.48 ± 3.69	72.77 ± 1.80
  UFTP22	1.300 ± 0.17	45.35 ± 2.05	12.53 ± 0.80	76.19 ± 1.03	12.49 ± 1.88	72.66 ± 3.39
  UFTP24	0.763 ± 0.36	38.47 ± 0.48	10.03 ± 0.35	69.64 ± 4.38	14.25 ± 2.21	73.07 ± 2.76
  UFTP38	1.797 ± 0.08	54.77 ± 0.16	12.37 ± 0.61	70.45 ± 3.26	8.72 ± 2.91	63.93 ± 3.80
  UFTP42	1.393 ± 0.29	49.53 ± 5.01	10.90 ± 0.61	68.03 ± 2.70	7.72 ± 7.19	68.86 ± 5.56
  UFTP45	2.236 ± 0.57	64.61 ± 7.10	10.60 ± 0.62	68.79 ± 1.45	-2.70 ± 3.14	54.63 ± 2.70
  UFTP46	1.248 ± 0.19	44.56 ± 0.59	11.83 ± 0.59	74.94 ± 1.27	11.35 ± 4.64	74.69 ± 6.60
  UFTP47	2.464 ± 0.52	58.21 ± 4.65	13.13 ± 0.35	73.58 ± 6.98	2.62 ± 4.23	58.99 ± 8.84
  UFTP57	1.707 ± 0.09	49.64 ± 1.33	13.73 ± 0.57	73.91 ± 0.89	15.95 ± 3.00	79.26 ± 0.97

  
  

  Table 4-2. Pearson coefficients denoting strength and direction of linear correlations between target compounds and colorimetric values.

  Bioactive Compound	L*	a*	b*	C*	ha,b (°)
  Moisture	-.16	-.51*	-.56*	-.57*	.48*
  β-carotene	.03	.54*	.56*	.57*	-.48*
  Ascorbic Acid	.26	.44*	.51*	.51*	-.40*
  Total Soluble Solids	.42*	.64*	.60*	.62*	-.64*
  Total Phenolics	.33	.32	.41*	.41*	-.31

  * denotes statistically significant correlations.

   

  Table 4-3. Pearson coefficients denoting strength and direction of linear correlations between phenotypic characteristics of calabaza samples and colorimetric values.

  Phenotypic Characteristic	L*	a*	b*	C*	ha,b (°)
  Mass	-.13	-.15	-.18	-.15	.14
  Circumference	-.23	-.34*	-.29*	-.28*	.34*
  Height	.05	.24*	.19	.21	-.25*

  * denotes statistically significant correlations.

   

  Table 4-4.  Pearson coefficients denoting strength and direction of linear correlations between phenotypic characteristics and bioactive compound concentrations of calabaza samples.

  	Moisture	β-carotene	Ascorbic Acid	Total Soluble Solids	Total Phenolics
  Mass	.18	-.20	-.23	-.55*	-.31
  Circumference	.32*	-.19	-.19	-.66*	-.34
  Height	-.03	-.15	-.25	.17	-.06
  Moisture	-	-.41*	-.84*	-.87*	-.88*
  β-carotene	-.41*	-	.44*	.20	.13
  Ascorbic Acid	-.84*	.44*	-	.65*	.83*
  Total Soluble Solids	-.87*	.20	.65*	-	.81*
  Total Phenolics	-.88*	.13	.83*	.81*	-

  * denotes statistically significant correlations.

  
  

  Table 4-5.  Pearson coefficients denoting strength and direction of linear correlations between bioactive compounds of non-random calabaza plugs and each plug’s colorimetric values.

  Bioactive Compound	L*	a*	b*	C*	ha,b (°)
  β-carotene	.18	.44	.64*	.66*	-.36
  Ascorbic Acid	.11	.61*	.59*	.61*	-.55*
  Total Soluble Solids	.42	.65*	.47	.49	-.67*
  Total Phenolics	.19	.33	.20	.22	-.32

  * denotes statistically significant correlatio

Objective 3: Determine yield, fruit quality and disease resistance of tropical pumpkin cultivars in the Southeastern U.S. and Puerto Rico in organic and conventional cropping systems and determine phenotypic relationships among nutrition, flavor and fruit size traits in select germplasm.

Team: Geoffrey Meru, Angela Ramírez, Andre da Silva, and Carlene Chase

Activity 1: Performance of calabaza germplasm in Florida

The fruit weight varied across the germplasm evaluated and ranged from 0.85lb (JW6823 butternut) to 5.75 lb (Dickson) (Table 1). Similar results were observed for yield per plant which was highest in JW6823 butternut but least in Fortuna and UFTP25 (Table 1). However, the fruit weight and total weight/ acre was highest yield in UFTP42 followed by UFTP24, but least in Fortuna followed by WB butternut (Table 1). Fruit quality parameters varied significantly across the squash germplasm evaluated (Table 2). Flesh color was ranked highest in UFTP 1990 but least in Taina Dorada. On the other hand, the Brix index was highest in  JW6823 butternut but least in Dickson (Table 2). Fruit size also varied across the squash germplasm with the butternut cultivars having the least fruit length, fruit width, and flesh thickness. On the contrary these parameters were highest in Dickson pumpkin (Table 2).

Results from 2022 trial (FL):

Table 1: Fruit yield data for calabaza germplasm evaluated in Florida under conventional production system in 2022

Table 2: Fruit quality of the calabaza germplasm evaluated in  Florida under conventional production system in 2022

Results from 2023 trial (FL):

Table 1: Fruit yield data for calabaza germplasm evaluated in Florida under conventional production system in 2023

Table 2: Fruit quality of the calabaza germplasm evaluated in  Florida under conventional production system in 2023

Activity 2: Performance of calabaza germplasm in Puerto Rico

Results from 2022 trials (PR):

Horticultural traits evaluated like silverleaf, vigor, vine length, flowering male, and female) were statistically significant in the Spring trial. Means separation was performed using Tuckey (a= 0.05). Data on disease evaluation demonstrate that lines UFTP58 (Table 1) and Verde Luz were the only ones to show <10% of leaves symptoms for silverleaf.  Most lines under the conventional trial had silverleaf on >50-80% of the leaves (scores of 3.5 to 4). Genotypes Fortuna and Soler had higher (5) scores for silverleaf symptoms under the conventional trial. While under the organic trial genotypes UFTP58, Fortuna and Verde Luz had scores of Regarding powdery mildew (PM) evaluation, the only genotype showing PM symptoms was UFTP58 under both cropping systems. Under the horticultural evaluation lines UFTP34 and UFTP81 had excellent vigor (score of 5) under the conventional cropping system (Table 1) However, those genotypes did not have the same vigor under the organic trial. Genotypes UFTP38 and UFTP32 had the poorest vigor under the organic, while under the conventional those genotypes had scores of 4.6 and 4, respectively. Genotype UPFTP58 had the poorest vigor the conventional trial. In general, female flowering began at 38 after transplant (UFTP58), followed by UFTP4 (39 d), and the latest genotypes to have females flower was at 61 (Fortuna) days after transplanted under the conventional, while in the organic, genotype UPFT58 (34 d) was the earliest and UFPT10 and Soler was the latest (54 days). Considering male flowering under the organic started at 22 days (UFTP34) and the latest at 56 days (Fortuna). Under the conventional, the earliest male flower was in UFTP58 at 22 days after transplanting and UFPT22, UFPT26, and Fortuna (51 days) at the latest. Average male flowering regardless of the cropping system was at 44 days after transplanting. In general, most genotypes tended, under both cropping systems, to have small to medium vine lengths. Regarding yield, yield components (Table 2), the number of fruits per plot was 1 to 14 with a mean value of 8 for the conventional and 7 for the organic system. The number of marketable fruits in the plot was an average of 7 for both management systems (Table 2). The average fruit weight in the conventional system was 7.5 pounds and in the organic 6.6 pounds (Table 3), a small pumpkin and desirable size for some of the Puerto Rico vegetable markets. Most lines presented an orange color flesh with high sugar content (average of 8 Brix), average sweetness at consumption time and overall rating of 2 for both management growing systems (Table 4).  Some lines performed equally under both systems like UFPT 4, UFPT22, UFPT46 and Fortuna behave similarly in the downscale of yield, UFPT34 (C), UFTP42 and UFPT45 had the higher yield, however, this result was in one of the growing systems, non-showing consistent under both parameters. In general, a reduction of about + 6,000 kg ha^-1 is observed in the others breeding lines when comparing yield under conventional versus organic, being the organic the one that have less production (Table 5). Promising lines are also observed under organic conditions (UFTP42, UFPT45) when compared to the conventional.