Growing media

Growing media approved for certified organic production were used in the trials: Premium Flower 201 (M201; Morgan Composting, Sears, MI, USA); Promix MP Organik (PMPO; Premier Tech Horticulture, Quakertown, PA, USA); Seed Catapult (SCOE; Ohio Earth Food, Hartville, OH, USA); Sunshine No. 1 Natural and Organic (SUN1; Sun Gro Horticulture, Agawam, MA, USA); and Fort Light (VCFL; Vermont Compost, Montpelier, VT, USA). We intentionally chose products that contained compost (M201, SCOE, VCFL) and those that did not (PMPO and SUN1) because prior research showed significant differences in crop performance with compost and noncompost media (Maynard et al. 2024). Chemical characteristics of the media based on the saturated media extract (SME) analysis and compost analysis procedures conducted by a commercial laboratory are provided in Table 2.

Table 2. pH, electrical conductivity (EC), total nitrogen, ammonium-nitrogen, nitrate-nitrogen, and plant mineral nutrients in five growing media before planting tomatoes.

View Fullscreen

Fertilizers

Treatments included each growing medium with and without a dry fertilizer derived from composted turkey litter, feather meal, and sulfate of potash and containing 5N–0.88P–3.32K–3Ca (5–2–4; Suståne Natural Fertilizer, Cannon Falls, MN, USA). In two additional treatments, a liquid fertilizer derived from poultry manure containing 3N–0.88P–2.49K (3–2–3; Envirokure, Philadelphia, PA, USA) was applied to tomatoes growing in PMPO and VCFL. These two media were chosen for the liquid fertilizer treatments to represent a noncompost medium and a compost medium. Each fertilizer was applied at the same rate to all media. The rate was calculated to contain enough N for a 4% concentration in aboveground plant tissue on a dry weight basis at the end of transplant production. This concentration was more than adequate (Wilcox 1993), and the dry weight of 1 g/plant used in the calculation was based on the size of large transplants in a previous study (Maynard et al. 2024).

In Expt. 1 (PP1, PS1, and SW), the dry fertilizer was mixed by hand into media at a rate of 15 kg·m⁻³ [25 lb/yard³ providing (in kg·m⁻³) 0.75 N, 0.13 phosphorus (P), 0.50 potassium (K), and 0.45 calcium (Ca)] before the media was moistened and put into flats, and the liquid fertilizer was applied in a solution of 500 mg·L⁻¹ (ppm) N beginning 8 to 11 d after seeding (DAS). Application of liquid fertilizer continued approximately weekly until 3000 mL [116 fl oz, providing (in kg·m⁻³) 0.78 N, 0.23 P, 0.65 K] had been applied or the experiment was ended. A sample of 5–2–4 analyzed by a commercial laboratory (Midwest Laboratories, Omaha, NE, USA) had pH of 6.19, salt index of 16, total N level of 6.05%, ammonium-N level of 0.36%, undetectable nitrate-N, available P level of 1.6%, soluble K level of 3.93%, Ca level of 3.83%, magnesium (Mg) level of 0.73%, sodium (Na) level of 0.38%, sulfur (S) level of 2.32%, iron (Fe) level of 0.259%, undetectable boron (B), chloride (Cl) level of 0.52%, and carbon (C) level of 30.16%. A sample of the fertilizer solution that was sent to a commercial laboratory (A and L Great Lakes Laboratories, Fort Wayne, IN, USA) for analysis had a pH of 8.3, EC of 3.93 dS·m⁻¹, and the following (in mg·L⁻¹): nitrate-N, 383; ammonium-N, 44.6; P, 7.0; K, 502; Ca, 74; Mg, 24; Na, 581; S, 220; Fe, 0.07; B, 0.09; molybdenum (Mo), 0.02; silicon (Si), 5; Cl, 37; and alkalinity [calcium carbonate (CaCO₃) equivalent], 425. Preliminary results of Expt. 1 showed that fertilizer affected emergence; therefore, during Expt. 2 (PP2 and PS2), the same rates of fertilizer were used, but applications were delayed. The dry fertilizer was sprinkled on the surface of the growing media in two applications: at 7 and 10 DAS (first application) and at 21 and 26 DAS (second application) at PP2 and PS2, respectively. The liquid fertilizer was applied daily beginning 22 and 26 DAS at PP2 and PS2, respectively, until 3000 mL of the 500 ppm N solution had been applied.

Emergence and seedling performance

Flats were filled with media, seeded with tomato (Solanum lycopersicum cv. Big Beef; Johnny’s Selected Seeds, Winslow, ME, USA), and watered until water leached from holes in the bottom of the cells (hereafter referred to as to run-through). Thereafter, flats were hand-watered as needed based on visual observation to maintain moderate moisture in the growing media. At SW, flats were on germination mats set to maintain 26.7 °C (80 °F) until 90% of seeds germinated. At other locations, flats were on greenhouse benches with no supplemental heat. Emergence was recorded daily from initiation until at least 90% of seedlings had emerged.

Plant growth was measured three times during seedling production: at 14 to 17 DAS and then at 4-d to 12-d intervals. On each date, three randomly selected plants were cut at the soil level; thereafter, leaves per plant, combined fresh weight, and combined dry weight after drying at 55 to 60 °C (131 to 140 °F) were recorded. At the final growth measurement, plant height and diameter of the stem below the cotyledons were measured on the sampled plants. At SW, a fourth measurement was performed at 35 DAS.

The final plant samples from PP1 and PP2 and the second and final samples from PS2 were sent to a commercial laboratory (Brookside Laboratories, New Bremen, OH, USA) to determine the nutrient concentration following methods described for N and minerals by Miller et al. (2013).

Media measurements

The EC and/or pH of growing media were measured twice in most trials: between 9 and 12 DAS and between 24 and 36 DAS. Flats were watered to run-through, and sampling began 30 to 60 min later. At SW, values were determined in one cell per experimental unit using an ion-sensitive field-effect transistor for pH (Wireless ISFET pH probe; Spectrum Technologies, Aurora, IL, USA). At PP and PS, three cells per experimental unit were sampled using a modification of the pour-through (PT) method described by Wright (1986). Five to 15 mL (0.19–0.58 fl oz) of distilled or deionized water was poured into a cell, the leachate was collected, and EC and pH were measured. Measurements of each experimental unit were averaged to evaluate treatment effects.

Field performance

Tomatoes from one greenhouse trial at each location (PP2, PS1, and SW) were transplanted to the field. Each field trial was arranged as a continuation of the associated greenhouse trial and had a randomized complete block design with four replications and 12 treatments. An experimental unit consisted of six plants at PP and SW and four plants at PS. At PP and PS, the field trials were located in organically managed but noncertified experimental areas. At SW, the trial was in a conventionally managed area that received no synthetic inputs during the trial other than plastic mulch. No fertilizers were applied to the soil at any location before transplanting. Seedlings were transplanted by hand 0.46 m (18 inches) apart in a single row on beds covered with 1.2-m-wide (width, 4 ft) black plastic mulch [PP: 25 µm (1 mil) embossed; Trickleez Irrigation, Inc., St. Joseph, MI, USA; SW: 20 µm (0.8 mil); Ginegar Plastic, Inc., Santa Maria, CA, USA] or with 1.8-m-wide (6-ft-wide) black woven weed mat (DeWitt Company, Sikeston, MO, USA) at PS. Beds were 3 m (10 ft) on center at PP, 0.9 m (3 ft) at PS, and 1.8 m (6 ft) at SW. At PS and SW, plants were supported with a Florida weave trellis. At PP, plants were not supported or pruned. Plants were watered in by hand at transplanting and then irrigated as needed through a single line of drip tape under the mulch (Rivulis Ro-drip; Rivulis Irrigation Inc., San Diego, CA, USA). At PP and PS and at SW, 30-cm and 20-cm (12-inch and 8-inch) emitter spacing [0.9 L·h⁻¹ per emitter at 55 kPa (0.24 gal/h per emitter at 8 psi)] were used, respectively.

Vegetative growth was evaluated 14 d after transplanting (DAT) and 28 DAT at PP2, 16 DAT, 21 DAT, and 28 DAT at PS1, and 11 DAT, 23 DAT, 29 DAT, and 37 DAT at SW by counting the number of fully expanded leaves on the main stem, measuring stem diameter just below the cotyledonary node, and measuring height from soil to the main stem growing point for each plant. On the final growth measurement date at PP2 and PS1, the number of buds, open flowers, fruit set, and aborted buds or flowers on the first main stem cluster, and fresh and dry weights were determined for three randomly selected plants per plot. Additionally, at PP2, the total number of clusters with open flowers or fruit were determined for three randomly selected plants per plot.

At SW, ripe fruit was harvested twice weekly from 8 Jul through 5 Aug, and the number and weight were recorded. Early yield and fruit number were calculated by summing the first three harvests. Average weight per fruit was calculated over the entire harvest period.

Data analyses

Maximum emergence, days to 50% emergence, and emergence uniformity were estimated for each experimental unit by fitting the sigmoid equation:

where y = % emergence, x = days after seeding, a = maximum emergence, b = days to 50% emergence, and c is related to the steepness of the sigmoid curve; a smaller value indicated a steeper curve and, therefore, more uniform emergence. The nonlinear platform of JMP Pro software (version 14, 15, or 16; SAS Institute Inc., Cary NC, USA) was used to estimate parameters.

Seedling fresh and dry weights for each experimental unit were used to estimate parameters a1, b1, and c1 for the following equation:

where x1 and x2 are orthogonalized values of x (days after seeding) and x² and y is fresh weight + 0.01 or dry weight + 0.001; the fractional amounts were added to allow inclusion of aboveground weights of 0 on day 0. Parameter a1 represents the intercept, parameter b1 represents the slope or overall relative growth rate, and parameter c1 represents the departure from a constant relative growth rate over the course of the experiment, with negative values indicating slowing of the relative growth rate. The fit model standard least squares platform of JMP Pro software (version 14, 15, or 16) was used to estimate parameters.

An analysis of variance (ANOVA) was performed separately for each trial to evaluate effects of treatments on responses. Fixed effects included replication and treatment. Preplanned single degrees of freedom contrasts were constructed to evaluate the main effects of each fertilizer averaged across media, of the media type (compost or noncompost) averaged across 5–2–4 fertilizer treatments, and of the interaction between each fertilizer × media type. Contrasts within compost media were used to compare M201 and VCFL to SCOE and M201 to VCFL and to evaluate the interaction of these media effects with the effects of 5–2–4 fertilizer. M201 and VCFL were combined because their chemical characteristics are more similar than those of SCOE (Table 2). When interactions were significant at P < 0.05, the effect of fertilizer within the media type was evaluated using preplanned contrasts and Fisher’s least significant difference. Effects were considered significant when P < 0.05. Statistical results of contrasts are provided in Supplemental Tables 1 to 11. At PS1, PMPO with 3–2–3 was inadvertently omitted from the experiment; therefore, preplanned contrasts to evaluate liquid fertilizer effects were valid only for VCFL. At PS2, SUN1 without 5–2–4 was accidentally omitted from block 1, and SCOE with 5–2–4 was accidentally omitted from block 2. To account for this, the analysis relied on least square means and adjusted least significant difference values for comparisons with treatments represented in only three blocks. For EC, pH, and plant tissue analyses, separate ANOVAs were used for each measurement date. Analyses were performed using the fit model standard least squares platform of JMP Pro software (version 15 or 16). The EC and growth measurements were log-transformed and counts were square root-transformed when needed to improve equality of variances.

Growing media EC

Interpretation of EC measurements depends on the crop, production system, and plant growth stage. This discussion relied on guidelines developed with systems that use synthetic fertilizer to produce plugs, bedding plants, and potted plants for sale (Warncke 2015). Precise conversions are not available for EC methods used in this study, but values published for EC measured using the PT method provide an approximation (Cavins et al. 2000). Interpretations included the following: EC <0.75 (SME) or 1.0 (PT), very low, not enough nutrients for rapid growth; EC = 0.76 to 2.0 dS·m⁻¹ (SME) or 1.0 to 2.6 dS·m⁻¹ (PT), low, acceptable, or suitable for seedlings; EC = 2.0 to 3.5 dS·m⁻¹ (SME) or 2.6 to 4.6 dS·m⁻¹ (PT), optimum or normal for established plants; EC >3.5 (SME) or 4.6 (PT), high and may reduce growth; EC >5 (SME) or 6.5 (PT), very high, likely to reduce growth, and may lead to salt injury (Cavins et al. 2000; Warncke 2015).

The initial EC of M201 and that of VCFL were high enough to limit growth, both SCOE and SUN1 had an EC acceptable for seedling growth, and PMPO did not have enough nutrients for growth (Table 2). Additionally, M201 and VCFL had high levels of sodium, which would have contributed to the EC without providing any nutritional benefit. As expected, EC increased with fertilization (Table 3, Supplemental Table 1). Incorporating 5–2–4 fertilizer resulted in EC levels at 11 DAS that could reduce growth, even in media without high initial EC (PP1). A single top-dressing of 5–2–4 at the half rate frequently raised EC to potentially harmful levels in media with high initial EC (M201 and VCFL at PP2 and PS2). In media with a very low initial EC (PMPO), top-dressing at the half-rate led to EC suitable for seedlings; in media with low initial EC (SCOE and SUN1), top-dressing led to EC suitable for seedlings or established plants at 9 to 12 DAS. Early application of 3–2–3 fertilizer (PP1) brought media with very low initial EC (PMPO) into the normal range for established plants by 11 DAS. In media with high initial EC (VCFL), the early application of 3–2–3 fertilizer led to very high EC likely to reduce growth.

Table 3. Electrical conductivity (EC; dS·m⁻¹) of five growing media with and without supplemental fertilizer 9 to 12 and 28 to 36 d after seeding (DAS) in three trialsⁱ.

View Fullscreen

The increase in EC caused by fertilizer remained significant through measurements performed at 28 to 36 DAS (Table 3, Supplemental Table 1). With 5–2–4 fertilizer either incorporated or top-dressed, media generally had EC in the range suitable for seedlings, but it was low for established plants. With 3–2–3 fertilizer applied late in production (PP2 and PS2), media EC was in the normal range for established plants. With 3–2–3 applied beginning early in production (PP1), EC in PMPO was in the range suitable for seedlings; in VCFL, it was in the range for established plants. Without fertilizer, EC of noncompost media and that of the compost media SCOE were typically too low for growth, and both EC of M201 and that of VCFL were in the range suitable for seedlings. The results indicated that whether media contain compost or not and whether additional fertilizer is added or not, maintaining EC at levels appropriate for growth through the end of tomato transplant production is unlikely without frequent applications of readily available nutrients.

Previous studies have documented similar results with solid organic fertilizers. Li and Mattson (2019) performed a study of cucumbers grown in peat/perlite medium and reported a rapid increase in growing media EC followed by a decline when N derived from composted turkey litter, feather meal, and bone meal was incorporated at 0.57 kg·m⁻³ (0.96 lb/yard³). In that study, a vermicompost-based fertilizer applied at the same N rate also increased EC compared with the unfertilized media, but to a lesser extent. Without fertilizer additions, media EC can decrease by 50% within 10 d with regular overhead watering (Taylor et al. 2016).

Growing media pH

Initial media pH was within the recommended range (5.0–6.8) (A and L Great Lakes Laboratories, 2002) for M201, SUN1, and VCFL, low for PMPO, and high for SCOE (Table 2). The 5–2–4 fertilizer increased media pH measured 9 to 12 DAS only when it was incorporated into media (PP1) (Fig. 1A, Supplemental Table 2). By the end of the production cycle (24–36 DAS), the use of 5–2–4 before or after seeding reduced media pH by averages of 0.48, 0.39, and 0.43 points at SW, PP2, and PS2, respectively, but it did not affect pH at PP1 (Fig. 1C). At both times, changes in pH were larger in noncompost media and not always significant in compost media. The reduction in pH brought all media except PMPO closer to, but not always within, the desired range (Table 4); PMPO started with a low pH and, in some cases, 5–2–4 reduced the final pH below the desired range. A transient pH increase followed by a decrease to below the pH of unfertilized media after incorporation of a similar product were also reported by Li and Mattson (2019).

View Figure

• Download as Powerpoint
• Download Figure

Table 4. pH of five growing media with and without supplemental fertilizer 9 to 12 and 25 to 36 d after seeding (DAS) in four trials.

View Fullscreen

Liquid 3–2–3 applied before the early pH measurement did not change media pH measured 3 d later (PP1) (Fig. 1B, Supplemental Table 2). By the end of production, 3–2–3 frequently increased pH, but the effect was not consistent (Fig. 1D). The largest increases occurred at PP1, with 2.17 points in PMPO, and the next largest occurred at PP2, with 0.75 points averaged across PMPO and VCFL. The final pH in PMPO fertilized with 3–2–3 was within the desired ranged, but pH in VCFL was high (Table 4). The pH increase is, in part, explained by the alkalinity of the fertilizer solution: 425 ppm CaCO₃ equivalent when prepared using irrigation water with alkalinity of 155 ppm CaCO₃ equivalent. Therefore, it would be advisable to use 3–2–3 in systems where pH is monitored and can be readily adjusted. When that is not possible, using it in media with a low initial pH, such as PMPO, may be preferable.

Media pH changes are influenced by many factors, including the initial pH, buffering capacity of the medium, alkalinity of irrigation water, nitrogen mineralization and nitrification, temperature and moisture, plant nutrient uptake, and microbial activity (Argo and Biernbaum 1997). Taylor et al. (2016) found that compost and peat at the same initial pH have a similar buffering capacity on a dry weight basis. The smaller changes observed in compost media in this trial are probably caused by higher bulk density of those media (data not shown): a flat filled with compost media would have greater weight of media and, therefore, more buffering capacity than a flat filled with peat media. In a number of cases in this work, media pH exceeded 6.8, which is the upper limit of the recommended range, regardless of whether fertilizer was added. This occurred less often at PS2, where irrigation water was acidified to reduce alkalinity. We did not identify widespread negative effects on plants caused by media pH, perhaps because of the relatively short production cycle.

Seedling emergence

Emergence curves differed among treatments (Fig. 2). Across all growing media, when solid fertilizer was mixed into growing media before seeding (PP1, PS1, and SW) the time to 50% emergence was delayed by 1 to 2 d, emergence was less uniform in two of the trials, and maximum emergence at SW was reduced (Fig. 3, Table 5, Supplemental Table 3). Solid or liquid fertilizer applied after emergence had begun (PP2 and PS2) rarely affected emergence parameters. Media containing compost significantly delayed the time to 50% emergence by approximately 1 d in four trials and reduced emergence uniformity in the two trials at PP compared with media without compost.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Table 5. Estimated maximum emergence (a, %), days to 50% emergence (b), and emergence uniformity (c) derived from emergence data for tomato grown in five growing media with and without supplemental fertilizer in five trials.

View Fullscreen

In some trials, the effect of incorporating solid fertilizer was larger in compost media than in noncompost media or differed among compost media (Fig. 3, Table 5, Supplemental Tables 3 and 4). At SW, in noncompost media, maximum emergence with and without incorporated 5–2–4 differed by less than 1% on average; however, in compost media, incorporating 5–2–4 reduced emergence by an average of 6.7%. At PP1, for noncompost media, incorporating 5–2–4 delayed emergence by less than half a day; however, in compost containing media, the delays were approximately 1 d for SCOE, 1 to 2 d for VCFL, and 3 d for M201. At both PP1 and SW, incorporating 5–2–4 reduced emergence uniformity (increased parameter c) more in SCOE than in M201 and VCFL.

When the response to fertilizer did not depend on media type (i.e., when the fertilizer × media type interaction was not significant at P < 0.05), the main effects of media were of interest (Fig. 3, Supplemental Table 3). For the two noncompost media, emergence parameters did not usually differ. For compost media, emergence required one-half to 1 d longer in M201 and VCFL than in SCOE in two trials (PS1 and PP2) and was less uniform in M201 and VCFL than in SCOE in one trial (PP2). Emergence was faster in M201 than in VCFL in one trial (PS2), and it was more uniform in one trial (PS2) and less uniform in a second trial (PS1).

These results are consistent with those of our previous work that compared emergence in these media and illustrated the relationship between growing media EC and emergence parameters in which higher EC was associated with delayed and less uniform emergence (Maynard et al. 2024). In this trial, media EC measured 9 to 12 DAS was significantly higher when 5–2–4 was applied before seeding, and this was consistent with the delay in emergence and reduced uniformity and maximum emergence in some trials. Notably, negative effects of 5–2–4 occurred in all media, not only in the two compost media with high initial EC; however, the negative effects were often larger in compost media.

Li and Mattson (2019) also reported seedling injury consistent with high EC when they incorporated a product containing 8N–1.8P–3.3K from the same manufacturer into noncompost organic growing media at rates of 2.37 to 7.12 kg·m⁻³ (4 to 12 lb/yard³). In that trial, EC measured using the PT method increased from less than 2 dS·m⁻¹ to greater than 4 dS·m⁻¹ at 20 DAS with the high rate of fertilizer; cucumber seedling survival was reduced and root rot was observed. The manufacturer has suggested that an 8N–0.9P–3.3K formulation with lower EC than that of the 5–2–4 formulation is preferred by organic growers, with a recommended incorporation rate of 3.6 to 8.9 kg·m⁻³ (6–15 lb/yard³) (Holden C, personal communication, 10 Mar 2020). The results reported here support the use of a lower rate of 5–2–4 when incorporated to avoid negative effects on emergence, particularly in media with an initially higher EC.

Seedling growth

Growth of tomatoes differed among treatments. Results were generally similar for fresh and dry weights, height, stem diameter, and leaf counts; therefore, only dry weight data are presented (Fig. 4). The relative growth rate (b1), its change over time (c1), and the final dry weight are most helpful for understanding the plant response (Fig. 5, Supplemental Tables 5 and 6).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

When 5–2–4 was incorporated into compost media before planting (trials PP1, PS1, and SW), the results were inconsistent. In most cases, tomatoes grew more slowly and produced seedlings with lower dry weight than that of those in unfertilized media (Fig. 5, Table 6, Supplemental Tables 5–7); however, sometimes they grew more quickly (e.g. in VCFL at PP1 and SCOE at SW) or there was no difference in overall growth rate caused by 5–2–4 (e.g., M201 at PP1 and SW). When 5–2–4 was top-dressed onto compost media (trials PP2 and PS2), tomatoes usually grew more quickly than those without fertilizer, had a smaller decrease in relative growth rate, and produced up to twice the dry weight; however, in some instances, there was no significant effect of 5–2–4 (Fig. 5, Table 6, Supplemental Tables 5–7). Among the compost media, significant effects of topdressing 5–2–4 were observed least often in SCOE. Fertilizing VCFL with 3–2–3 increased plant dry weight up to 75%, but the effect was not consistent when 3–2–3 was first applied early in production (Fig. 5, Table 6). These inconsistent effects of fertilizer in compost media could be attributable to confounding by negative fertilizer effects on emergence timing in the case of 5–2–4 and variation in the availability of nutrients contained in the media caused by temperature, moisture, and leaching fraction variations among the trials.

Table 6. Final seedling dry weight for tomato grown in five growing media with and without supplemental fertilizer in five trials.

View Fullscreen

In noncompost media, either incorporating or top-dressing 5–2–4 significantly increased the overall relative growth rate, reduced the decline in relative growth rate, and resulted in tomato seedlings with final dry weight 1.8- to 5.3-times greater than that of those without 5–2–4 (Fig. 5, Table 6, Supplemental Table 5). In contrast to compost media, any negative effects of 5–2–4 fertilizer on emergence that may have reduced growth were counteracted by the positive response to nutrients. Growth improved significantly more in SUN1 than in PMPO for at least one response in all trials except PS1 (Table 6, Supplemental Tables 5–7). SUN1 fertilized with 5–2–4 consistently produced plants as large as or larger than those grown in unfertilized compost media. In contrast, PMPO with 5–2–4 typically produced plants smaller than or the same size as plants grown in unfertilized compost media. It is possible that the low pH of PMPO observed during the initial media test and also at the end of production (Tables 2 and 4) limited mineralization of N in the fertilizer or limited plant nutrient uptake.

Fertilizing with 3–2–3 in PMPO increased the overall relative growth rate, reduced the decline in relative growth rate only when applied late in production, and increased final dry weight by 1.5- to 7-times (Fig. 5, Table 6, Supplemental Table 5). When 3–2–3 was first applied within 12 DAS (trials PP1 and SW), the effects were larger, and the final plant dry weight was the same as or greater than that of plants grown in compost media without fertilizer. When 3–2–3 was first applied later in the growing period (trials PP2 and PS2), plants were smaller than those grown in unfertilized compost media.

Treatments that consistently produced plants with the greatest final dry weight were VCFL + 3–2–3 applied at any time, VCFL + 5–2–4 top-dressed, PMPO + 3–2–3 applied early, and SUN1 + 5–2–4 incorporated.

Seedling nutrient concentrations

Both 5–2–4 and 3–2–3 significantly increased N and K concentrations (Fig. 6, Supplemental Table 8). According to Wilcox (1993), tomato seedlings are deficient in N or K if the aboveground concentration is less than 2.8% or 3%, respectively. Plants in both compost and noncompost media were often deficient in N by the end of the production cycle whether they received 5–2–4 fertilizer (Fig. 6A), usually deficient in K without 5–2–4, and deficient in K with 5–2–4 at PP2 (Fig. 6C). When the liquid 3–2–3 fertilizer was first applied early, the final tomato N concentration was insufficient, with an average of 1.85% (PP1), and the K concentration was below the adequate level (4%), but not deficient. When 3–2–3 was first applied several weeks after seeding, the final tomato N and K concentrations were sufficient, with average N values of 3.64% and 4.46%, and average K values of 4.50% and 5.61%, for PP2 and PS2, respectively. Earlier in production, at 24 DAS (PS2), all treatments had N and K concentrations above adequate levels, except for unfertilized noncompost media, which were deficient in both N and K, and unfertilized M201, which was deficient in N.

View Figure

• Download as Powerpoint
• Download Figure

Fertilizer effects on tomato P concentrations at the end of production differed for 5–2–4 and 3–2–3 and differed for 5–2–4 incorporated into media or top-dressed, with increases in some cases and decreases in others (Fig. 6B, Supplemental Table 8). However, P concentrations were above the adequate level of 0.35% (Wilcox 1993) in nearly all cases. The two instances when P was deficient (<0.28%) occurred in noncompost media with fertilizer were as follows: PMPO with 3–2–3 at PP1 and SUN1 with 5–2–4 at PP2. A lower P tissue concentration with fertilization is probably explained by dilution of P in the plant tissue attributable to an increase in plant dry weight caused by the fertilizer.

The N and K deficiencies corresponded to final media EC less than 2.6 dS·m⁻¹ in nearly all cases. When the N concentration was adequate (≥3.5%), final EC was always more than 2.6 dS·m⁻¹; however, in several cases, when the K concentration was adequate (≥4.0%), final EC was less than 2.6 dS·m⁻¹. This supports the suggestion (Cavins et al. 2000) that EC levels less than 2.6 dS·m⁻¹ measured using PT indicate nutrient levels that are inadequate for established plants and emphasizes the unsuitability of the EC measurement alone for determining which nutrients are needed.

These results confirm previous findings that plants grown in these media are likely to be deficient in N after 4 to 5 weeks of growth without additional fertilizer (Maynard et al. 2024), similar to the results of studies of organic transplant production, as reviewed by Pascual et al. (2018) and Burnett et al. (2016). Plants are also likely to be deficient or close to deficient in K at that time. In noncompost media, this deficiency was more extreme and likely occurs much earlier than that in compost media. Results from PS2 show that deficiency can develop rapidly in the final weeks of the production cycle between 24 and 33 DAS. Tissue P concentrations are likely to be adequate regardless of whether supplemental fertilizer is applied.

The use of fertilizers following the protocols in this trial will improve tissue N and K concentrations at the end of the transplant production period, but these protocols are not optimized. The N concentration will likely be insufficient unless 3–2–3 fertilizer is applied frequently toward the end of production. The K concentration is likely to be higher than the 3.0% reported as deficient by Wilcox (1993) and fall within the range (2.6%–4.5%) reported for other trials in which potassium fertilization was intended to be adequate (Hartz et al. 2002; Lee et al. 2017; Widders and Garton 1992). When 3–2–3 is applied frequently toward the end of production, tissue K concentration may increase well above adequate levels. Creating media with lower initial K levels and/or using fertilizers with lower K may still provide adequate K while reducing the likelihood of harmful EC levels in media.

The tissue tests indicate that the N component of fertilizers was most important for improving plant growth, and that K was second most important. The fact that N was deficient by the end of the production cycle emphasizes the importance of understanding N release from organic media and fertilizers and N losses from growing media to effectively manage nutrient supply for seedlings. In this experiment, the total N applied in fertilizer treatments was calculated as sufficient to supply all N required by seedlings; however, clearly, adequate amounts were not available to plants in all fertilizer treatments, even in compost media that contained additional organic N.

Field performance

Size differences between unfertilized and fertilized transplants existed 4 weeks after transplanting (WAT) (Fig. 7, Supplemental Table 9). In noncompost media, both fertilizers significantly increased tomato plant dry weight (PS1 and PP2) and height (SW); increases ranged from 36% to 75% for dry weight and from 57% to 84% for height. In compost media, increases in size with fertilizer were smaller (average of 9%–19% for dry weight and 6%–18% in height) and not always significant. Stem diameter and leaf number showed similar trends (data not shown).

View Figure

• Download as Powerpoint
• Download Figure

The ranking of plant size 4 WAT generally corresponds to the ranking in size at the end of transplant production, but there is one notable exception. At PP2, plants grown in PMPO with 3–2–3 first applied 3 weeks after seeding (WAS) had dry weight 4 WAT larger than that of plants from unfertilized PMPO or SUN1 and equivalent to that of all other treatments (Fig. 7C), even though these plants were smaller than any except unfertilized PMPO or SUN1 at the end of transplant production (Table 6). The rapid growth after transplanting required for this result seems likely related to the relatively high tissue N concentration at the end of seedling production when 3–2–3 was applied 3 WAS (PP2) (Fig. 6A). Research of conventional tomato production has documented the value of applying higher rates of fertilizer soon before transplanting to increase levels of readily available N in the seedling that can be used for vegetative growth during transplant establishment (Garton and Widders 1990; Widders and Garton 1992).

Fertilizers applied during seedling production influenced the early yield and fruit number produced by tomatoes at SW (Table 7). In noncompost media, 5–2–4 led to more than 4.5-times and 6.7-times the early weight and number, respectively, produced without 5–2–4, and 3–2–3 led to 9.4- and 16.5-times the early weight and number, respectively, compared with the unfertilized treatment (Table 7, Supplemental Tables 10 and 11). In compost media, neither 5–2–4 nor 3–2–3 significantly increased the early weight of fruit; however, in M201 only, 5–2–4 nearly doubled the early number. It is worth nothing that the coefficient of variation was high (40%); therefore, practically significant differences were not statistically significant.

Table 7. Fruit set on first main stem cluster, flower clusters with open or past-bloom flowers, and yield for tomato grown as seedlings in five growing media with and without supplemental fertilizer.

View Fullscreen

Fertilizer effects on total production were more variable. In one compost media, M201, 5–2–4 significantly reduced total yield by 28%, but the 15% and 9% yield reductions in VCFL with 5–2–4 and 3–2–3, respectively, and the 7% yield increase in SCOE with 5–2–4 were not statistically significant (Table 7). Neither fertilizer significantly influenced total fruit number in compost media. In noncompost media (PMPO and SUN1), 5–2–4 did not significantly influence total yield, but it increased total fruit number by 29% (Table 7, Supplemental Table 11). In PMPO, 3–2–3 increased total yield by 30% and fruit number by 49% (Table 7). Across all media, average fruit weight decreased significantly by 12% to 13% with the use of either 5–2–4 or 3–2–3 fertilizer. In this trial, total yield was evaluated after 1 month of harvest. It is possible that if harvest had been extended beyond 1 month, then the results for total yield would have differed from those reported here.

At PP2 and PS1, the effect on yield was not measured directly, but the effects on number and stage of reproductive structures 4 WAT provide information on potential yield effects (Table 7, Supplemental Tables 10 and 11). The results show that, if any effect is observed, then the fertilizer use increases the number of fruit set on the first cluster and the number of clusters with open blooms or developing fruit by 4 WAT, and the effect is largest in noncompost media. This supports the finding at SW of increased early yield and fruit number in fertilized noncompost media.

These yield outcomes are consistent with the work of Vavrina et al. (1998), who described the effects of increasing N application during transplant production on timing and quantity of tomato yield in conventional production. Of three studies of fresh market cultivars, two reported increased early yield and no effect on total yield with higher N application, and one reported increases in both early and total yields (Masson et al. 1991; Melton and Dufault, 1991; Weston and Zandstra 1989; all cited in Vavrina et al. 1998).

Differences between compost and noncompost media were clear in this trial, but differences among compost media are also important. Initial EC of SCOE was lower than that for VCFL and M201, in part as a result of lower nitrate-N, K, Ca, Mg, and Na (Table 2). This finding is likely the reason why incorporated 5–2–4 did not negatively affect emergence, growth, and yield in SCOE to the extent it did in M201 and VCFL. Unlike M201 and VCFL, SCOE contains sand, which could dilute compost salts and alter porosity and water-holding characteristics. In future work, it would be helpful to include a wider variety of compost media to identify categories that respond similarly to management practices.

The results clarified that nutrient limitation is likely in organic tomato transplant production in flats whether compost or noncompost media are used, and it might not be overcome even with application of fertilizer containing enough total N for plant needs. Of the fertilization practices used in this trial, the liquid 3–2–3 applied late resulted in sufficient tissue N and K concentrations at the end of the production period more reliably than treatments with incorporated or top-dressed 5–2–4 or earlier application of 3–2–3. That treatment also consistently led to final media EC measurements above the level suggested for established plants. This is probably because most of the N in 3–2–3 was in a readily available form, and fertilizer was applied more frequently and closer to the time of tissue and media sampling, likely reducing the percentage lost to leaching.

The experiment was not designed to identify optimal materials, rates, or timing of fertilizer application, but it illustrates the importance of tailoring these to initial media characteristics and provides information about the magnitude of effects that can be expected. If more specific information is not available, then categorizing media based on whether they contain compost provides a basis for differential guidelines. For example, in noncompost media, fertilizer is needed soon after emergence to support growth. Although incorporating additional fertilizer before seeding poses some risk to emergence in noncompost media, it still increases growth and yield. To avoid harming emergence, it is better to delay fertilizer application until after emergence. If liquid fertilizer is used in noncompost media, then early application will produce larger transplants than delaying application for 3 weeks, but plants that receive later application may perform as well in the field. This result should be confirmed, but it suggests that timing of liquid fertilizer application to noncompost media could be used to manage organic transplant size without detriment to field performance.

In contrast, in compost media, additional nutrients are probably beneficial, but they may not be essential to increase early or total yield. Incorporating additional fertilizer in compost media before seeding is not advised because it will likely inhibit emergence and possibly growth. Top-dressing or applying liquid fertilizer after emergence eliminates that risk, but caution is necessary because it could cause EC to exceed desirable levels early in production. Because compost media will likely provide adequate nutrients until the rapid increase in size that begins approximately 21 DAS, early application of nutrients is not critical. Delaying liquid nutrient application until that time will likely support growth at least as well as earlier fertilization and also improve tissue nutrient status at the end of transplant production.