Pennsylvania State University Statewide Cultivar Evaluation Program Model

: Cultivar selection is an important managerial task for commercial farmers. Numerous available options provide marketing opportunities for farmers; however, they can also make cultivar selection time-consuming. For the past 13 years, a team of Penn State Extension educators and university faculty has implemented a new model to evaluate cultivars for key vegetable crops to assist farmers. Our approach started by building relationships within the vegetable industry with farmers, seed company representatives, and farmer-driven organizations who support our program with funding and gift-in-kind donations. We also added rigor to cultivar evaluations by conducting trials of key vegetable crops over 2 years at three locations. The results were used to develop cultivar recommendations, which were disseminated through various face-to-face and distance Extension avenues. The results were also published in refereed journals. The combination of funding, Extension products, and refereed articles led to university support. A drawback to our approach was the difﬁculty implementing uniform production methods at three experimental sites. The input of a statistical consultant assisted with overcoming this drawback. This model also required the use of strategies common to successful teams. Our methods are described here for those who desire to build a similar program.


Introduction
Cultivar selection is critical to the success of commercial vegetable farmers. The Mid-Atlantic Fruit and Vegetable Convention (the Convention; Hershey, PA, USA) is a regional farmer-based meeting attracting about 2200 participants annually. Every year, a survey is administered to farmer attendees. It consists of two parts: demographic questions (e.g., location of farming activity, years of farming, size of operation, etc.) and a section that asks about farming needs and issues. This second section changes annually and is developed by Penn State Extension Educators and university faculty to help guide future Extension programming and research initiatives. To ensure robust response rates, respondents participate in a face-to-face interview conducted by volunteers. During the 3-day convention, attendees are approached and asked to participate. If they agree, the volunteer goes through the list of survey questions. After completion, the interviewee is given a sticker place on their nametag so as not to be asked to participate more than once in the survey. In 2019, the survey focused on cultivar selection.
On average, respondents rated cultivar selection as third in terms of importance from a list of vegetable management activities, behind irrigation management and crop rotation ( Table 1). Commercially available cultivars for any given crop are numerous, providing farmers with options and also leaving them with the task of determining which ones fit their markets and management style. Additionally, the market life cycle of individual cultivars is often short, and new ones are continuously added to available offerings [1]. The Convention survey also indicates that farmers gather information for cultivar selection from a variety of sources, with other farmers, seed company representatives, and university trials ranking highest on average ( Table 2). University cultivar evaluations have the advantage of providing an unbiased assessment, and 69.8% (n = 139; Table 3) of Convention survey respondents rate cultivar evaluation as important or very important to their success. Despite this, university cultivar evaluations have been in decline in the U.S. over the last 30-40 years as a result of fewer public breeding programs and declining industry support [2]. Additionally, these evaluations have been deemed as lacking rigor by university administrations [3] and peers, further shrinking support for them.  Against this backdrop, a team of Penn State Extension educators and university faculty has been conducting annual cultivar evaluations since 2008 with the overall goal of helping farmers select cultivars for key vegetable crops. Common objectives of cultivar evaluations include updating recommendations and identifying top and/or poor performing cultivars [5]. The objective of our program has been to provide farmers in Pennsylvania and the mid-Atlantic region with recommendations for top-yielding cultivars, primarily in conventional systems [6][7][8][9][10][11]; although, performance in organic systems has also been evaluated [10]. Additionally, we have measured some produce quality attributes. Here, we provide our methods for building industry and university support for cultivar evaluations.

Program Initiation and Building Industry Support
In 2007, seven Penn State Extension educators and university faculty developed the concept for this program. Initial decisions were to conduct parallel cultivar evaluations at three sites to account for geographic environmental diversity across Pennsylvania, evaluate each vegetable type for 2 years to account for year-to-year environmental variability, and attempt to run the program for 10 years (Figure 1).
Against this backdrop, a team of Penn State Extension educators and university faculty has been conducting annual cultivar evaluations since 2008 with the overall goal of helping farmers select cultivars for key vegetable crops. Common objectives of cultivar evaluations include updating recommendations and identifying top and/or poor performing cultivars [5]. The objective of our program has been to provide farmers in Pennsylvania and the mid-Atlantic region with recommendations for top-yielding cultivars, primarily in conventional systems [6][7][8][9][10][11]; although, performance in organic systems has also been evaluated [10]. Additionally, we have measured some produce quality attributes. Here, we provide our methods for building industry and university support for cultivar evaluations.

Program Initiation and Building Industry Support
In 2007, seven Penn State Extension educators and university faculty developed the concept for this program. Initial decisions were to conduct parallel cultivar evaluations at three sites to account for geographic environmental diversity across Pennsylvania, evaluate each vegetable type for 2 years to account for year-to-year environmental variability, and attempt to run the program for 10 years (Figure 1).  With this framework, we applied for competitive funding from the Pennsylvania Vegetable Growers Association (PVGA; Richfield, PA, USA), an industry association operated by Pennsylvania vegetable and small fruit farmers and having about 1000 farmer members. Since 2008, the PVGA has provided between USD 10,000 and USD 12,500 annually, totaling over USD 150,000 to date in support of our program. Among competitively funded projects, ours has received the largest amount of funding each year and is one of the longest continuous research projects. The ongoing financial support received from the PVGA is also a sign of appreciation and positive evaluation of the information generated by our program. It has remained a research priority partly because of the relationships we have built with PVGA leadership and members.
The PVGA leadership determines which crops are evaluated within the boundaries of the amount of funding provided. Funding amounts preclude crops with high labor requirements (e.g., tomato (Solanum lycopersicum L.), cucumber (Cucumis sativus L.)) and limits which parameters are measured. We have evaluated green bell peppers , and cabbage (Brassica oleracea var. capitata L.; 2020-21). The number of cultivars included in each evaluation has ranged from 16 to 34. We have focused on measuring yield and some produce quality parameters including the rind color and shape of pumpkin fruit, tip cover of sweet corn ears, head diameter of broccoli, and some postharvest physicochemical traits of muskmelon fruit. In 2018-19, we have included grafted entries with the standard cultivar tested on its own rootstock and grafted onto alternative rootstocks. Grafting has gained popularity among Pennsylvania farmers as a strategy for managing soilborne pests, diseases [12,13], and abiotic stressors [14][15][16], and for increasing yield [17,18]. The number of rootstocks commercially available for solanaceous (Solanaceae) and cucurbit (Cucurbitaceae) crops has been increasing, and determining which to use to address specific crop issues is becoming an additional challenge for which farmers need assistance.
The cultivars evaluated and the standards we have used are largely farmer-driven, at first through a formal survey [19] and more recently through informal conversations with several farmers. We have also established relationships with seed company representatives who provide recommendations for newer cultivars and often provide seed at no cost. The standard cultivars we have used are determined by PVGA members or by farmers we have built relationships with through Penn State Extension. Using farmer-selected standards provides more meaningful results and a reference point for comparison [5]. As one program team member states, "[Farmers] want to be able to see how what they are already growing compares with newer cultivars." The program is also supported by gift-in-kind donations provided by a farmer cooperator in the form of land and equipment use. Each year, one of the three sites where the evaluation is conducted is a commercial farm. Farmer cooperators vary from year to year depending on where the offices of Extension educators on the program team are located. We seek out farmer cooperators with the capabilities to handle the production system and the crop being evaluated. Crops that require raised beds, plastic mulch, and drip irrigation and those that are direct-seeded and started from transplants are matched to a farmer with those resources. For example, the sweet corn evaluation was direct-seeded; therefore, the farmer cooperator needed to select a field and prepare the soil. The Extension educator then hand-seeded replicated plots. The farmer applied pesticides as needed, and the Extension educator harvested and collected data. This minimized the amount of effort required by the farmer. Good cooperation, communication, and coordination are required.
Approaching different farmers for each crop or year also helps ensure that the project does not become a burden. Crops with a concentrated harvest allow for collaboration with farmers located farther away from the educator's office. In recent years, we have selected cooperators closer to the Extension office to help minimize time and travel costs. The largest challenge we have encountered is not having complete control of the land, equipment, labor, and timing of pest control and other management practices. However, involving a farmer cooperator helps build relationships within the farming community and increases the farmer's understanding of research as they participate in the process on their farm.

Building University Support
In the U.S., cultivar evaluations are conducted at numerous universities. In the typical model, evaluations are conducted for a single year at a single location. It has been stated that cultivar evaluations do not require "large amount[s] of intellectual input" [2], and they have been deemed not rigorous by university administration [3] and faculty. We  [20], and U.S. Department of Agriculture plant hardiness zones, average last spring frost dates, and average first fall frost dates differ at each site [7]. Our program team includes members housed in these three areas to facilitate trial management at each site.
Conducting evaluations in more than 1 year is more important than at multiple sites in a single year to measure cultivar stability [3]. In our program, evaluations are conducted for 2 years and at three sites. Multi-location trials provide opportunities to quickly evaluate cultivars over a range of conditions [21] and expand the usefulness of the data collected. 71.7% of farmer respondents to the Convention survey indicate it is important or very important to have cultivar evaluations conducted in areas like theirs (n = 138; Table 3). Our program is designed so that results from individual locations can be disseminated at the local level, allowing farmers to focus on results from the area most like theirs. We also combine results from individual locations to develop statewide recommendations. By detailing the environmental conditions and production practices employed, the results can also be extended to similar areas [5] outside of Pennsylvania. Indeed, recommendations based on results from this program are included in a multi-state vegetable production guide [4]. Conducting the experiment over 2 years and at three sites adds complexity and raises the rigor of evaluations, and the results are publishable in refereed articles. We know of no other cultivar evaluation programs in the U.S. using this model.
Having three evaluation sites also presents some challenges. Ideally, production practices are standardized for all sites such that only soil and environmental factors vary. Before the evaluations are conducted, efforts are made to use uniform protocols across sites, except for planting date, which varies depending on the climate and seasonal weather conditions at each site (see reference [7] for hardiness zones and average first and last frost dates). Despite these efforts, different methods are employed, including irrigation use and scheduling, nitrogen and other nutrients applied, pesticides used, and other production practices. Successful farmers use methods established for their farm based on experiences with their land, environment, and equipment. Honoring this experience, different methods are sometimes used at the farmer cooperator site. Additionally, of the sites, the southeastern site is the area most concentrated with vegetable farms. Local variations in production practices are sometimes used at this site. For example, staking bell peppers is common in southeastern but not in central Pennsylvania. While an effort to control experimental variation within a location is made, it has varied between locations. We work with a statistical consultant, because of this complexity, to assist with analyzing and interpreting these data collected.
Despite the differences in production systems at each site, other contributors to experimental variation between sites are minimized to allow for the site-to-site comparison of yields and quality parameters. Cultivars evaluated, including standards, are identical between sites. Plant spacing at times varies by site. Plant spacing has a large effect on crop yield. For some crops, it has been reported that higher plant densities can result in higher yield on an area basis; however, at an individual plant level, yields are lower (e.g., bell pepper [22]; broccoli [23]; watermelon (Citrullus lanatus (Thunb.) Matsum. & Naki) [24]; hydroponic tomato [25]) because of increased competition for resources between plants. For other crops, increasing plant populations can lead to lower yields (pumpkin [26]). Crop maturity criteria are also standardized between sites. Additionally, four replications are used at each site to increase statistical power.
Since production protocols differed across sites and the visual inspection of the yield data indicated a three-way interaction between location, year, and cultivar, we initially analyzed data by location for the first crop (bell peppers), as recommended by our statistical consultant. Data from the second evaluation (acorn, butternut, and buttercup/kabocha winter squash) were combined for statistical analysis in response to journal peer reviewers. With this change, the experiment now tests for the main effects of sites, cultivars, and years as well as all interactions. Environment, soils, and some production practices are confounded, but we can compare cultivars grown under different conditions. Large differences in cultivar performance often occur and may be economically important; however, due to high variation the differences are often not significant at the 5% level. Reporting probability values, even if above the 5% level, may be useful for farmers, since yield differences above this level may be economically relevant for them.
Our model of conducting evaluations for 2 years at three sites brings rigor to the experiment, which facilitates publishing results in refereed journals. Thus far, six refereed articles from this program have been published [6][7][8][9][10][11], with two more in the pipeline. Additionally, recommendations have been disseminated through various extension channels, including articles and presentations at local and regional meetings and webinars. The impact of these has been positive. As one farmer has stated about cultivar evaluation presentations at the Convention, "the best-attended sessions are often on variety [cultivar] trials." This has had positive implications in annual staff and faculty evaluations of program team members. For Extension educators, it has contributed annually to the number of scholarly presentations completed and the involvement in research projects. It has also served as a major program area important in the promotion process. For university faculty, refereed publications and external competitive funding have been important contributors to the development of a dossier for promotion and tenure, as well as for annual and post-tenure evaluations. For faculty with Extension appointments, presentations and articles disseminated have also contributed to an Extension program. However, as has been noted previously [2], cultivar evaluation can only serve as part of a promotion and/or tenure packet, with contributions focused on other topics also needed.

Other Elements of This Model
The size of our team is another factor that makes our model successful. Having a relatively large team maintains and reinforces relationships and reciprocal trust among a sizeable group of Extension educators, university faculty, farmers, and other members of the vegetable industry. Membership for our team has changed since the program began 13 years ago. Four members have retired from Pennsylvania State University and two have left to pursue other employment opportunities. Our team consisted of seven members at its inception and is currently composed of six members, three of whom are original members. New members have been a combination of new hires and current educators interested in seeing the program continue.
Our program is built on teamwork. Effective communication within the team and follow-through on action items are critically important indicators of the success of teams [27]. For our program, team members are located across the state, which facilitates management at each trial site; however, this presents a barrier to face-to-face communication. The Penn State Extension system is organized around teams supporting different agricultural industries. All team members of our program are part of the same Extension team and therefore interact regularly at team meetings and events. Early in the project, communication occurred primarily by phone or e-mail. Now, communication occurs primarily by e-mail, text messages, and e-meetings to discuss research methods, cultivar entries, and more.
Clearly establishing team member responsibilities is also important to ensure followthrough on tasks. Our team consists of a person dedicated to administrative duties (writing annual proposals, analyzing data, generating reports, etc.). Another team member communicates with farmers and seed company representatives to determine cultivar entries and standards and acquire seeds. Two team members take the lead on manuscript preparation. Another two members work to schedule presentations of results at a large annual regional farmer convention. Yet another team member identifies farmer cooperators. This arrangement uses all team members' time efficiently. Leadership in these roles has been self-determined, with members identifying and then filling needs.

Conclusions
Our program has been operating successfully for 13 years and will continue into the foreseeable future. The key to our success has been creating a new model for cultivar evaluations. This has included building long-term support from the vegetable industry and