Introduction

Tomato (Solanum lycopersicum L.) is one of the most cultivated vegetables in the world. In 2022, 49,196 ha were harvested in Mexico, with a production of 3.46 million tons, of which 13.3 % were of the round type (Servicio de Información Agroalimentaria y Pesquera [SIAP], 2023). International demand for tomato requires increased production and quality fruits, which underscores the need to develop new and better varieties (Carrillo-Rodríguez et al., 2013). However, breeding of this species in Mexico is mainly carried out by transnational seed-producing companies and, to a lesser extent, by some public agencies (Hernández-Ibáñez et al., 2017).

Despite the great socioeconomic importance of tomato cultivation in Mexico, there are only 22 varieties registered in the country’s National Catalog of Plant Varieties (CNVV). This highlights the need to develop national and public breeding programs that take advantage of the potential of native cultivated, semi-cultivated and wild tomatoes. To do so, it is necessary to: 1) collect and evaluate germplasm that allows expanding the genetic base of the crop (Carrillo-Rodríguez et al., 2013), 2) generate varieties with fruits that respond to the demands of producers and consumers in terms of flavor, color, aroma, texture, size, shape, shelf life, and yield, 3) evaluate the performance of improved lines and agricultural practices that maximize their yields in different environments according to the target market (Frasca et al., 2014), and 4) register varieties in the CNVV.

Some research in Mexico has used commercial varieties of round or ball-type tomatoes to evaluate the effects of different practices on yield, fruit quality and plant development. These studies have analyzed the impact of the application of copper nanoparticles and grafting (Peralta-Manjarrez et al., 2023), the contributions of nitrogen and other nutrients through compost tea (Ochoa-Martínez et al., 2009), variations in the concentration of Steiner's nutrient solution in mixtures of organic substrates based on coconut fiber and humus (Valenzuela-López et al., 2014), substrates based on sand and vermicompost as a nutrient source (Moreno-Reséndez et al., 2008; Rodríguez-Dimas et al., 2007), different levels of fruit thinning (Gaytán-Ruelas et al., 2016), and the yield and fruit quality potential of commercial hybrids in low-tech greenhouses in desert climates (Grijalva-Contreras et al., 2011). However, the literature on the breeding and development of Mexican round tomato varieties with indeterminate growth habit remains limited.

Considering the above, the objective of this work was to study the agronomic performance and fruit quality of 29 advanced F₅ lines of round tomatoes with indeterminate growth habit in order to identify outstanding materials for possible commercial use and as sources of germplasm for breeding programs.

Materials and methods

Plant material

The study included 29 advanced inbred lines of round tomatoes with indeterminate growth habit (IGH) and the commercial hybrid Caimán (Enza Zaden®) as a control. The 29 lines came from the Program for the Conservation and Improvement of Tomato Genetic Resources of the Colegio de Postgraduados. These lines were obtained from a cycle of mass selection, followed by four cycles of self-pollination and pedigree selection from a collection carried out in 2015 in Tehuacán, Puebla, Mexico. This was done as part of the National Tomato Network of the National System of Plant Genetic Resources for Food and Agriculture (SINAREFI). The source populations and the study lines are part of the Tomato Breeding Program and are under the custody of the Postgraduate Program in Genetic Resources and Productivity-Genetics of the Colegio de Postgraduados. The genotypes were studied under greenhouse and hydroponic conditions in Montecillo, Texcoco, State of Mexico (19º 30’ N and 98º 53’ W, at 2,250 m a. s. l.).

Conduction of experiments

Sowing was carried out on May 29, 2020 and March 18, 2021 in 200-cavity polystyrene trays with peat substrate (Kekkilä Professional®, Kekkilä). One week after seedling emergence, irrigation was started with Steiner's (1984) nutrient solution at a concentration of 25 %, which increased to 50 % after transplanting and to 100 % from flowering onwards. During the entire growing cycle, the pH of the nutrient solution was maintained between 5.5 and 6.0.

Transplanting was carried out 35 days after sowing (das), in 2020 and 2021. One plant was placed per 40 × 40 cm black polyethylene bag, with volcanic sand (red tezontle) as a substrate and a density of 4.44 plants∙m^-2. A randomized complete block design with three replications was used in each trial. Each experimental unit consisted of 10 plants managed to one stem. Pruning and tutoring were performed weekly, and at 150 das the plants were topped to limit their growth to 10 clusters.

Fungicides and pesticides were applied preventively, according to the doses and times recommended by the manufacturers: Captan (Captan®, Arysta), Imidacloprid (Confidor®, Bayer), Flonicamid (BeLeaf®, FMC), Copper oxychloride (Cupravit®, Bayer), Boscalid + Pyraclostrobin (Cabrio C®, BASF), Lambdacialotrina + Chlorantraniliprole (Ampligo®, Syngenta) and Azoxystrobin + Difenoconazole (Amistar Gold®, Syngenta). During flowering, the plants were gently shaken every third day, at eleven o'clock, by vibrating the wires supporting them to enhance pollination.

Variables evaluated

Initially, two variables were recorded: 1) days to first cluster flowering (DTF), counted from sowing to anthesis of the first flower of the first cluster, and 2) days from sowing to ripeness (DTR), recorded when one fruit from the first cluster of each plant reached grade 2 ripeness (breaker stage) according to the United States Department of Agriculture (USDA, 1975) color grading scale for fresh tomatoes.

At 155 das, in two red fruits (grade six on the USDA ripening scale) from the third cluster of each plant (20 fruits per experimental unit), the following were obtained: the average length (FL, mm) and equatorial diameter (ED, mm) with a digital millimeter vernier (Caldi-6MP, Truper®, Mexico), weight (FW, g) with a digital balance (SP2001, Ohaus®, USA), firmness (FF, N) with a texturometer (GY-1, Sundoo Instruments®, China) and a plunger of 3. 5 mm diameter, and total soluble solids of tomato juice (TSS, °Brix) with a digital refractometer (PAL-1, Atago®, Japan).

Harvests were made at 170, 190 and 210 das, and the total number of fruits (TNF) and the total fruit weight (TFW, g) of each plant were recorded with a digital balance (Torrey®, Mexico).

Statistical analysis

Variable averages were obtained and a combined analysis of variance was performed with the sources of variation: lines, year, block nesting in years and Lines × Years interaction. A Tukey's comparison of means (P ≤ 0.05) was made and a correlation analysis was performed to determine associations between pairs of variables. SAS® statistical package v.9.3 (SAS Institute, 2011) was used for the statistical analyses.

Results and discussion

Analysis of variance

The combined analysis of variance detected significant statistical differences (P ≤ 0.01) among lines for all variables (Table 1), indicating the existence of phenotypic and genetic differences within the germplasm. This condition is necessary in breeding programs for the species. There were also significant differences between years for all variables, except for FL and TNF. In the Lines × Years interaction, some genotypes had contrasting behaviors for DTF, in response to the evaluation environments.

Table 1. Mean squares of the combined analysis of variance for agronomic and fruit quality traits of 29 advanced lines (F₅) and the commercial hybrid Caimán of round tomatoes with indeterminate growth habit evaluated in Montecillo, State of Mexico, Mexico, in 2020 and 2021.

Comparison of lines

Phenological development variables

Earliness is a phenological trait that conditions the early onset of reproduction and progress towards fruiting (Burbano-Erazo et al., 2020). The evaluated materials took from 59.7 to 68.7 das to flower in the first cluster (Table 2). Lines 21093 and 21094 were the earliest, outperforming 21129, 21085, 21084, 21134, 21131, 21104, 21130, 21081, 21136, 21106, 21127, Caimán, 21133 and 21126 in a range of 4.8 to 9.0 das. The DTF obtained coincide with the 61 to 81.4 das observed by Rodriguez-Dimas et al. (2007) in the Big Beef and Red Chief varieties.

Table 2. Agronomic and fruit trait averages of 29 advanced lines (F₅) and the commercial hybrid Caimán of round tomatoes with indeterminate growth. Averages for 2020 and 2021.

Regarding DTR, variations were observed in the genotypes from 128 to 145 das. Lines 21087, 21105 and 21099 were earlier in this trait than 21085, 21129, 21106, 21098, 21097, 21134, 21136, Caimán and 21133, with differences of 8.5 to 17 days. The DTF and DTR were similar to the 58-65 das and 128 das recorded by Porres et al. (2015) in the Dominique, Beverly, Criollo and Nemonetta varieties (round tomatoes), when evaluating their responses to different fertigation periods in greenhouses in Sololá, Guatemala.

Fruit quality

Firmness, color, aroma, TSS and pH in tomato fruits are determining attributes of postharvest quality and influence consumer preferences (Bonilla-Barrientos et al., 2018). In this regard, the FF of the 30 genotypes ranged from 6.7 to 10.2 N (7.0 to 10.8 kg∙cm^-2) (Table 2). The lines with the firmest fruits were 21102, 21094, Caimán, 21083, 21103, 21095, 21104, 21133, 21084 and 21096, while the one with the lowest firmness was 21106, with differences from 2.8 to 3.5 N. The firmness observed is consistent with that reported by Morales-Ruiz et al. (2021) (7.0 to 11.0 kg∙cm^-2), who determined the physicochemical and postharvest behavior of fruits of the Maxcesa, Komett and Merlissen varieties (IGH round type), grown in greenhouses in Tehuacán, Puebla.

A high TSS concentration is associated with greater nutrient absorption and transport capacity, and is, among the quality variables of tomato fruits, the most important for nutritional and processing purposes (Marouelli et al., 2012). In the lines evaluated, TSS content ranged from 3.78 to 5.12 °Brix. Lines 21086 and 21097 were significantly superior to 21083, 21136, 21080 and 21087, with a difference of 1.05 to 1.34 °Brix (Table 2). Martínez-Rodríguez et al. (2017) point out that the TSS content in tomatoes must be greater than 4.0 °Brix for them to be considered of sufficient quality for fresh consumption. In this sense, 24 genotypes produced fruits with TSS above 4.0 °Brix, so they could be considered suitable for fresh consumption. However, the TSS in the fruits of Caimán, 21081, 21083, 21136, 21080 and 21087 were lower than this value.

The observed TSS are consistent with values reported in previous studies: 5.5 and 5.3 °Brix in the André and Adela hybrids (Moreno-Reséndez et al., 2008); 3.71 °Brix in Bosky, Romina and PX01636262 (Ochoa-Martínez et al., 2009); 4.53 °Brix in the Piranha hybrid (Peralta-Manjarrez et al., 2023), and 4.8 to 5.0 °Brix in Big Beef and Red Chief fruits (Rodríguez-Dimas et al., 2007).

Yield components

Some important traits for selection in breeding programs are yield, size, average weight and number of fruits per plant (Burbano-Erazo et al., 2020). Fruit length ranged from 53.4 to 73.2 mm (Table 2). Lines 21105, 21103, 21133, 21087 and 21093 had longer fruits than 21098, 21086, 21126, 21129, 21127, 21097, 21130, 21131, 21084 and 21081. Caimán fruits were only statistically longer than those from 21086 and 21098. In Coahuila, Mexico, Peralta-Manjarrez et al. (2023) evaluated the fruit quality of the Piranha variety (round fruit and plants with IGH) under greenhouse and hydroponic conditions, and observed that the control treatment plants produced fruits with lengths of 51.66 mm.

Equatorial and polar diameters determine the size of tomato fruits; in addition, ED is one of the main quality indicators for marketing (Secretaría de Comercio y Fomento Industrial [SCFI], 1998). In the evaluated genotypes, EDs from 61.4 to 87.1 mm were observed, and only the fruits of 21106 and Caimán (87.1 and 81.0 mm) were wider than those of 21096, 21086, 21134, 21129, 21080 and 21098, with differences of 13.4 to 25.7 mm (Table 2). Mexican Standard NMX-FF-031-1997-SCFI establishes the classification of round tomatoes by size according to their ED: small (from 54 to 58 mm), medium-sized (from 54 to 64 mm), large (from 63 to 71 mm), and extra-large (from 70 mm and up) (SCFI, 1998). According to this criterion, line 21098 produced medium-sized fruits (61.4 mm), 14 lines developed large fruits and 18 lines would be classified as extra-large. Moreno-Reséndez et al. (2008) observed that the fruits of the Adela and André varieties had diameters of 66.7 to 69.1 mm, values similar to those reached by some genotypes in the present study. On the other hand, Ochoa-Martínez et al. (2009) reported an ED of 76 mm in the Bosky, Romina and PX01636262 hybrids, while Rodríguez-Dimas et al. (2007) recorded an ED of 73 to 79 mm in the Big Beef and Red Chief varieties, which are very close to the ED of the lines 21106, Caimán, 21087, 21133 and 21085 (76.5 to 87.1 mm).

Regarding FW, genotypes ranged from 124.8 to 338.7 g. Line 21106 (338.7 g) and Caimán (285.6 g) outperformed 21098, 21129, 21080, 21081, 21097, 21086, 21134, 21131, 21083, 21096, 21099, 21127, 21102, 21084 and 21126, with differences of 94.7 to 214.0 g (Table 2). The results partially coincide with the FWs reported by Moreno-Reséndez et al. (2008) in the André and Adela hybrids (218.26 and 177.48 g, respectively), in plants grown in sand under greenhouse conditions and irrigated with Hoagland's nutrient solution. Likewise, Ochoa-Martínez et al. (2009) recorded weights of 223.0 g per fruit in the Bosky, Romina and PX01636262 hybrids, when applying nutrient solution. Rodriguez-Dimas et al. (2007) had fruit weights of 184.8 to 214.0 g with the Big Beef and Red Chief varieties. In the present work, 23 genotypes developed tomatoes weighing more than 177.48 g, and only 21106, Caimán, 21087, 21133, 21085, 21103 and 21105 had fruits weighing more than 223 g.

The TNF per plant in the evaluated materials ranged from 18 to 43 in 10 clusters. Lines 21081, 21098 and 21080 produced 41 to 43 fruits, outperforming 21106, Caimán, 21136, 21133, 21085 and 21086, which had 18 to 30.6 fruits per plant. These results agree with those reported by Rodríguez-Dimas et al. (2007), who obtained 32 to 34 fruits per plant with the Big Beef and Red Chief varieties (round type with IGH) grown in sand with inorganic fertilizers in greenhouses in Torreón, Coahuila. Likewise, these authors point out that by managing the plants at eight clusters and a planting density of 4.2 plants∙m^-2 they achieved yields of 24.7 to 28.0 kg∙m^-2, values close to those of the present study.

The TFW of the 30 genotypes evaluated ranged from 4249.0 to 6828.0 g (18.87 to 30.32 kg∙m^-2 or 188.7 to 303.2 t∙ha^-1) (Table 2). Line 21136 had the highest yield (6827.7 g), followed by 21102 and 21105 (with 6334.4 and 6280.0 g, respectively). These lines were statistically superior to the commercial control Caimán and to lines 21127, 21130, 21098, 21129, 21126, 21106, 21321 and 21080, with differences of 1,737.0 to 2,579.0 g per plant. In Coahuila, Mexico, Ochoa-Martínez et al. (2009) evaluated the effect of compost tea on the yield and quality of round tomato fruits of the Bosky, Romina and PX01636262 varieties under greenhouse conditions, and observed that the control treatment (nutrient solution in sand substrate) had the highest average yield (21.84 kg∙m^-2). Grijalva-Contreras et al. (2011) reported yields of 20.1 to 31.1 kg∙m^-2 when evaluating the performance of 10 round tomato hybrids in low-tech greenhouses in Sonora, Mexico, for two years. These yields are close to the range of 22.1 and 30.3 kg∙m^-2 achieved by 24 genotypes in the present study.

In 2022, the average greenhouse yield of round tomatoes in Mexico was 187.68 t∙ha^-1. The yields of the 30 genotypes evaluated (188.7 to 303.2 t∙ha^-1) are close to those reported by SIAP (2023) for the states of Coahuila, Guanajuato, Zacatecas, and San Luis Potosí (164.35 to 315.6 t∙ha^-1), which were higher than those of Michoacán, Baja California Sur, Nayarit, Sinaloa, Sonora, Colima and Jalisco (77.2 to 141.9 t∙ha^-1), and lower than those of the State of Mexico, Chihuahua, Querétaro and Aguascalientes (378.1 to 517.0 t∙ha^-1). It is important to note that in several of these states more sophisticated greenhouses and production systems are used, with production cycles longer than 210 das, resulting in higher yields.

Pearson correlations

In genetic improvement programs, breeders evaluate the variability of multiple traits in individuals of a population and the relationships between these traits (Acquaah, 2007). Therefore, it is important to know the magnitude and nature of the correlations between the traits of interest, since the selection of one trait can influence the expression of another, depending on the genetic correlation between the two (Souza et al., 2012). In this study, significant positive linear correlations (P ≤ 0.01) were observed between TFW with ED, FW, FL, FF and TNF (28 to 66 %) (Table 3). In this sense, the positive values of the phenotypic and genotypic correlations between yield, number, average fruit weight and fruit wall thickness should be considered as primary traits in tomato breeding (Souza et al., 2012).

Table 3. Pearson correlation coefficients of 10 agronomic and fruit quality traits of 29 lines (F₅) and the commercial control Caimán of the round tomato type with indeterminate growth grown in Montecillo, Texcoco, State of Mexico, in 2020 and 2021.

There were also significant correlations between DTF with ED and DTR (35 and 83 %, respectively), FW with ED (93 %) and FL (71 %), and FL with ED (63 %). In this regard, López et al. (2015) point out that to increase tomato fruit yield and size, genotypes with larger fruit dimensions (length and width) should be selected. In this study, no linear correlations were observed between FF and TSS with quality traits such as FW, ED or FL, suggesting the possibility of improving these qualities independently, without affecting the other variables.

On the other hand, a significant (P ≤ 0.01) negative correlation (-31 %) was found between TNF and DTR, indicating that the longer the ripening time of a genotype, the lower its fruit production. This result is similar to the -33 % reported by Liu et al. (2021) when evaluating combining abilities, heterosis and heritabilities of 55 tomato genotypes.

Monge-Pérez and Loría-Coto (2019) calculated Pearson correlations between fruit quality variables of 17 fat-type tomato varieties grown under greenhouse and hydroponic conditions in Costa Rica. As in the present work, these authors found no significant associations between TSS with FF, FW (-0.29) or DTR (-0.39), nor between FW with FF (0.39) or DTR, FF-DTR (0.06) or TFW-DTR. Unlike the present results, these authors also did not observe significant correlations between TSS and TFW, nor between TFW with FF and FW.

On the other hand, Gaytán-Ruelas et al. (2016), when evaluating the effect of fruit thinning on the quality and yield of six round tomato varieties under greenhouse conditions, observed significant positive correlations between FW-ED (0.93**), FW-TFW (0.52**) and ED-TFW (0.52**). The first correlation had a value similar to that found in the present work, while the other two almost doubled the coefficients.

Bdr et al. (2020) reported a correlation between ED-FL (64 %) similar to that of the present study, when producing lycopene-rich IGH tomato hybrids, although this was not significant. The correlation coefficients obtained are close to those found by Souza et al. (2012) in tomato genotypes with IGH, with correlations of 0.94* for TFW-TNF, 0.73* for ED-FW, 0.53* for TFW-FW and 0.645 (not significant) for FL-TFW. These authors point out that the high genetic correlation between two traits can favor gain by indirect selection, which is beneficial in economically relevant traits with low heritability or that are difficult to measure.

Conclusions

The F₅ lines of round tomatoes with indeterminate growth habit showed significant differences for days to flowering and ripening, length, diameter, weight, firmness, soluble solids content, and total number and weight of fruits. These differences allowed us to identify outstanding round tomato lines in earliness, yield and fruit quality variables, which make them suitable for fresh consumption. Of the genotypes evaluated, 28 showed yields statistically equivalent to those of the control, and line 21136 outperformed the control by 34.9 % in total fruit weight. This line has the potential to be used as a commercial variety or as a source of germplasm in breeding programs for the species, except for total soluble solids content.