Heritability of Morpho-Agronomic Traits in Cocona (Solanum sessiliflorum Dunal) and Efficiency of Early Visual Selection for Fruit Yield

Abstract

Cocona (Solanum sessiliflorum Dunal) is an underutilized Amazonian fruit species with significant food, nutritional, and economic potential, but its genetic improvement remains limited. This study aimed to estimate the heritability of 13 morpho-agronomic traits in two F₂ populations, assess the efficiency of early visual selection, and identify traits associated with fruit yield. Approximately 250 plants from each population were grown in the Central Amazon. One week before the first harvest, plants were visually screened for yield potential, and the selected individuals were further evaluated for fruit traits. Broad-sense heritability (h²) was significant for most traits; the highest values were for number of flowers per plant (h² = 0.88), petiole length (h² = 0.87), collar diameter (h² = 0.71), canopy diameter (h² = 0.58), and fruit length (h² = 0.55). Early visual selection achieved ~65% efficiency. Fruit yield was correlated strongly and positively with the number of fruits per plant. These results indicate that phenotypic selection is effective for improving key plant and fruit traits in cocona. Early visual selection can be used to identify high-yielding individuals and the number of fruits per plant can be used as a complementary criterion to enhance selection accuracy for fruit yield.

1. Introduction

Cocona (also known as cubiu, maná, or peach tomato) is an underutilized Amazonian species belonging to the Solanaceae family. Its fruits are morphologically similar to tomatoes but covered by urticating trichomes that are easily removed at harvest. Fruit peel color ranges from yellow to purple, while the mesocarp and placenta are light yellow. Fruits are acidic, especially in the placenta, and contain many small seeds (≈500–2000 seeds per fruit; ≈1200 seeds g⁻¹) [1]. Seed germinative vigor can be maintained for up to one year when stored in hermetic containers at 12 °C in darkness.

Plants reach 1 to 2 m in height with canopy diameters of 1.5 to 2.0 m and tolerate pruning. The production cycle lasts 9 to 10 months, with the first three months dedicated to seedling preparation. Harvesting begins in the fifth month after field transplantation. In Spodosol soils, fruit yield ranges from 25 to 140 t ha⁻¹ depending on fertilization management [1]. Its reproduction is predominantly sexual, although propagation via cuttings is also possible. Cocona is self-compatible [1], although some studies indicate that cross-pollination can occur [2,3], suggesting that the species is still undergoing domestication [4].

In Brazil, Amazonian communities traditionally use cocona in fish-based dishes, whereas in Peru it is consumed as a soft drink using placenta or cooked fruits. However, several studies have shown that it can be utilized to produce various other products, such as jams and jellies [5], nectars [6,7], cocona-cashew nectar (Anacardium occidentale) [8], cocona-quinoa nectar (Chenopodium quinoa) [9] and cocona-pineapple juice [10]. Additional studies indicate medicinal potential, including reductions in blood glucose [11], cholesterol and triglycerides [7,12,13], hepatic, renal, and pancreatic protective effects [14]; cytoprotective activity with inhibition of cancer cell growth [15]; and anxiolytic effects [11].

Despite this potential, only one variety, SRN9 from Peru, is officially listed in the global registry of plant varieties [16]. In southeastern Brazil, cv. Santa Luzia and cv. Thais are commercially cultivated [8]. In addition, 101 conserved accessions have been registered in Peru [17] and nine in Brazil [18]. These numbers indicate that the genetic improvement of cocona could substantially enhance its commercial value. Nevertheless, progress in this area is hampered by a scarcity of data concerning the heritability of key morpho-agronomic characteristics and the efficacy of selection methodologies.

Previous studies on cocona have explored the inheritance of specific traits. The presence of thorns on stems and leaves is controlled by a recessive allele, while maternal effects influence fruit mass and shape [10]. The broad-sense heritability (h²) of fruit mass has been estimated to be 0.89 [1]. In contrast, in tomato, a related species, heritability estimates have been reported for a wide range of morpho-agronomic traits [19,20,21], supporting the development of numerous cultivars. These results underscore the need to quantify the heritability of cocona traits to support more efficient breeding strategies.

Visual selection can be a promising strategy for improving fruit yield in cocona. This approach allows for rapid phenotypic screening of large F₂ populations or segregation population and has proven effective in other crops, such as maize [22], soybean [23], and Urochloa ruziziensis [24]. However, studies in rice [25] and potato [26] have indicated limitations, possibly due to variation in trait expression or genotype × environment interactions. In crops with multiple harvests, such as cocona, tomato, and sweet pepper (Capsicum annuum L.), early visual selection can predict the fruit yield throughout the production cycle. Therefore, assessing its efficiency in cocona may offer valuable guidance for optimizing breeding strategies and enhancing genetic gain.

Another bottleneck is the lack of detailed information on the relationships among various fruit traits, which could facilitate indirect selection for enhanced productivity or quality. For instance, in tomatoes, a small peduncle scar diameter and a large fruit diameter are linked to greater post-harvest resistance [27], while fruit weight is associated with the number of locules [28].

This study aimed to address these gaps by estimating the heritability of key morpho-agronomic traits in cocona, evaluating the efficacy of early visual selection for fruit yield, and identifying traits that correlate with productivity. Such a comprehensive approach is expected to contribute significantly to the development of more robust and efficient breeding programs for cocona, potentially leading to the release of superior varieties.

To achieve these objectives, two F₂ populations were assessed to quantify the heritability of various morpho-agronomic traits, evaluate the effectiveness of early visual selection for fruit yield, and identify traits associated with fruit yield.

2. Materials and Methods

2.1. Plant Material

The plant material consisted of two F₂ families (n = 250 each) and a control line, CUB-4 (n = 13) (Figure 1). CUB-4 is maintained in the germplasm bank of the Instituto Nacional de Pesquisas da Amazônia (INPA), while the two F₂ families were derived from two F₁ plants identified within CUB-4. These F₁ plants originated from spontaneous outcrosses between CUB-4 and two distinct, unidentified lines from the same collection (Figure 1). All nine cocona lines involved are considered autogamous, having maintained phenotypic stability since 2005 [18]. Occasional natural outcrosses were readily detectable due to atypical fruit morphology, as observed in the F₁ plants.

2.2. Seedling Production

The seeds were sown in 128-cell polystyrene trays filled with a commercial sterile substrate (Tropstrato HT, Vida Verde, Mogi Mirim, SP, Brazil) and then placed on benches in the greenhouse. Irrigation was automated, with sprinkling occurring twice a day for 10 min each time. After two months, the seedlings were transplanted into 180 mL polyethylene cups containing a substrate composed of chicken manure, soil, and organic compost in a 1:2:1 ratio. After one month, the seedlings reached a height of 10 to 15 cm, developed at least four leaves, and were ready for field transplantation. All procedures were conducted at the Horticultural Experimental Station/INPA (2°59′48″ S, 60°01′20″ W, 51 m above sea level), Manaus, AM, Brazil.

2.3. Field Experiment

The field experiment was conducted from November 2021 to April 2022 in the floodplain of the Solimões River at the Ariaú Experimental Station (3°15′17″ S, 60°14′47″ W, 21 m above sea level) of INPA, Iranduba, AM, Brazil (Figure 2). The soil was classified as Humic Glei soil. The climate is humid equatorial, type “Af” according to the Köppen classification. The monthly rainfall ranged from 188 to 320 mm [29]. The mean temperature was 27 °C, with minimum and maximum temperatures of 22 °C and 32 °C, respectively.

The seedlings were transplanted following the family block design, placing them in holes 15 cm in diameter and 5 cm deep, with a spacing of 1.5 × 1.0 m between rows and plants, respectively. In the experiment, no organic or chemical fertilizers, soil amendments, or pesticides were used. Cultivation consisted of weekly weeding with hoes and gasoline-powered brush cutters, as well as pruning the lateral branches and leaves.

The following traits were recorded: plant height (cm), collar diameter (cm), leaf length (cm), leaf width (cm), petiole length (cm), number of tillers, number of fruits, number of flowers, plant vigor (classified on a scale from 1 = healthy to 5 = dead), fruit hairiness (classified on a scale from 1 = smooth to 3 = hairy), length of the largest fruit (cm), diameter of the largest fruit (cm), and length-to-diameter (L/D) ratio (Figure 3).

2.4. Plant Selection

The early visual selection of the plants was based on the apparent quantity of ripe and immature fruits at 113 days after transplanting (DAT). The selected plants were marked using 2 m rods (Figure 3B). A total of 45 plants from Pop 1 and 80 from Pop 2 were selected for further evaluation. Subsequently, five harvests were performed on 11, 22 and 31 March, 8 April and 3 May 2022. At each harvest, the following ten fruit traits were assessed for each plant: fruit yield (t ha⁻¹) [6666.7 × total mass (t)], number, mass (g), length (cm), diameter (cm), length-to-diameter (L/D) ratio, locule number, pericarp thickness (mm), total soluble solids content (°Brix), and pH.

The number of fruits per plant was recorded, and five medium-sized fruits were selected from each plant for length and diameter measurements. The locule number, pericarp thickness, soluble solids content (°Brix), and pH were assessed in three of these fruits. All fruits were counted per plant and weighed using an electronic balance. Pericarp thickness was measured with a digital caliper, soluble solids content was determined with a handheld refractometer, and pH was measured using a benchtop pH meter.

2.5. Statistical Analysis

Descriptive statistics, including means, standard deviations, and variances, were calculated for Pop 1, Pop 2, and the control line CUB-4 based on the remaining individuals: Pop 1 (n = 203), Pop 2 (n = 202), and CUB-4 (n = 7). Subsequently, Student’s t-tests were performed to compare the means of the F₂ populations with the CUB-4 control line.

The genetic variance (${\sigma }_g^2$) of the traits was estimated as follows: ${\sigma }_g^2={\sigma }_p^2-{\sigma }_e^2$, where ${ \sigma }_p^2$ = phenotypic variance, and ${\sigma }_e^2$ = environmental variance estimated as variance of the control line CUB-4. A chi-square (χ²) test was conducted to assess whether the genetic variance differed significantly from zero. These analyses were performed using Statdisk 13.0.1 [30]. Broad-sense heritability (h²) was calculated as the ratio of genetic variance to phenotypic variance: h² = ${\sigma }_g^2/{\sigma }_p^2$

Pearson correlation coefficients among fruit traits (p < 0.05) were calculated using Statistica 12 (StatSoft Inc., Tulsa, OK, USA) and a correlation heatmap was generated using JMP 14 software (SAS Institute Inc., Cary, NC, USA).

3. Results

3.1. Heritability Estimates

Few studies have been conducted on the estimation of the heritability of cocona traits. This study provides values that will be useful to guide the breeding of this species. Significant heritability values for Pop 1 ranged from 0.14 to 1.0, while for Pop 2, they ranged from 0.24 to 1.0 (

Table 1. Variances of F₂ families (Pop 1 and Pop 2), variances of the control CUB-4, genetic variances (${\sigma }_g^2$), and broad-sense heritabilities ($h^2$) for morpho-agronomic traits of cocona (Solanum sessiliflorum Dunal).

Trait	P1 (n = 203)	P2 (n = 202)	CUB-4 (n = 7)	${\mathit{\sigma }}_{\mathit{g}}^2$ (Pop 1) §	${\mathit{\sigma }}_{\mathit{g}}^2$ (Pop 2) §	${\mathit{h}}^2$ (Pop 1)	${\mathit{h}}^2$ (Pop 2)
Plant height (cm)	225.39	415.88	139.95	85.44 **	275.93 **	0.38 **	0.66 **
Collar diameter (cm)	0.18	0.28	0.08	0.10 **	0.20 **	0.55 **	0.71 **
Leaf length (cm)	55.04	67.36	31.81	23.23 **	35.55 **	0.42 **	0.52 **
Leaf width (cm)	32.82	39.93	105.81	−72.99	−65.88	0.00	0.00
Petiole length (cm)	5.67	13.18	1.62	4.05 **	11.56 **	0.71 **	0.87 **
N° tillers	0.70	0.85	1.00	−0.30	−0.15	0.00	0.00
Canopy diameter (cm)	382.82	431.82	179.67	203.15 **	252.15 **	0.53 **	0.58 **
N° fruits per plant	69.92	85.27	59.70	10.22 **	25.57 **	0.14 **	0.30 **
N° flowers per plant	574.26	570.80	67.81	506.45 **	502.99 **	0.88 **	0.88 **
Plant vigor †	0.64	2.06	4.07	−3.43	−2.01	0.00	0.00
Fruit hairiness ††	0.25	0.28	<0.01	0.25 **	0.28 **	1.00 **	1.00 **
Fruit length—L (cm)	1.85	1.10	0.83	1.02 **	0.27 **	0.55 **	0.24 **
Fruit diameter—D (cm)	1.72	1.25	0.88	0.84 **	0.37 **	0.48 **	0.29 **
L/D ratio	0.0108	0.0128	0.0107	0.021 **	0.0002 **	0.38 **	0.27 **

Legend: D = diameter. L = length. ^† Notes (1 = healthy, 5 = dead). ^†† Notes (1 = smooth to 3 = hairy). ^§ Genetic variance. ** Significantly different from zero based on χ² test (p < 0.01).

).

Pop 1 exhibited higher heritability values than Pop 2 for fruit-related traits, such as length, diameter, and L/D ratio. However, Pop 2 showed greater heritability than Pop 1 for plant traits such as plant height, collar diameter, petiole length, and canopy diameter. These findings suggest that Pop 1 is more suitable for selecting fruit traits, whereas Pop 2 is better suited for selecting traits related to plant architecture.

These contrasting heritability estimates between Pop 1 and Pop 2 were also reflected in the phenotypic variability observed for fruit size and shape (Figure 4 and Figure 5). In Pop 1, fruits ranged from rounded to very elongated and from small to large. In contrast, Pop 2 displayed limited variability in fruit shape, but considerable variation in fruit size.

The phenotypic contrast between CUB-4 and the unknown parental lines was evaluated by t-tests comparing CUB-4 with the F₂ populations. Significant differences were observed between CUB-4 and the F₂ population (Pop 1), as well as between CUB-4 and F₂ population (Pop 2) for six key traits (

Table 2. Means and standard deviations of morphological characteristics of the F₂ population (Pop 1), F₂ population (Pop 2), and control CUB-4 of cocona (Solanum sessiliflorum Dunal). Iranduba, Amazonas.

Trait	Pop 1 (n = 203)	Pop 2 (n = 202)	CUB-4 (C) (n = 7)	Δ Pop 1-C §	Δ Pop 2-C §
Plant height (cm)	79.51 (15.01) †††	97.28 (20.39)	61.57 (11.83)	17.94 **	35.71 **
Collar diameter (cm)	2.50 (0.43)	2.60 (0.53)	2.20 (0.28)	0.30 **	0.40 **
Leaf length (cm)	51.84 (7.24)	55.41(8.21)	46.86 (5.64)	4.98 **	8.55 **
Leaf width (cm)	38.59 (5.73)	40.72 (6.32)	39.14 (10.29)	−0.55	1.58
Petiole length (cm)	9.16 (2.38)	10.90 (3.63)	8.43 (1.27)	0.73	2.47 **
N° tillers	1.92 (0.84)	1.61 (0.92)	2.00 (1.00)	−0.08	−0.39
Canopy diameter (cm)	117.15 (19.57)	126.69 (20.78)	116.00 (13.40)	1.15	10.69
N° fruits per plant	12.05 (8.36)	12.35 (9.23)	12.71 (9.41)	−0.66	−0.36
N° flowers per plant	60.45 (23.96)	62.79 (23.89)	48.86 (8.23)	11.59 **	13.93 **
Plant vigor †	1.37 (0.80)	1.80 (1.43)	2.45 (2.02)	−1.08	−0.65
Fruit hairiness ††	2.64 (0.50)	2.56 (0.53)	3.00 (0.00)	−0.36	−0.44 **
Fruit length—L (cm)	4.81 (1.36)	4.58 (1.05)	3.23 (0.91)	1.58 **	1.35 **
Fruit diameter—D (cm)	4.41 (1.31)	4.70 (1.12)	3.04 (0.94)	1.37 **	1.66 **
L/D ratio	1.11 (0.10)	0.98 (0.11)	1.08 (0.10)	0.03	−0.10 **

^† Notes (1 = healthy to 5 = dead). ^†† Notes (1 = smooth to 3 = hairy). ^††† Standard deviation in parentheses. ^§ Difference between the family mean and the control line, CUB-4. ** Significant based on Student t-test (p < 0.01).

). These results indicate a genetic difference between CUB-4 and the unknown parental lines, particularly in plant architecture, fruit shape, and fruit size characteristics.

3.2. Early Visual Selection for Fruit Yield

In Population 1 (Pop 1), early visual selection identified 45 promising plants, with 29 plants (64.4%) exhibiting significantly higher fruit yield than the CUB-4 parental line (

Table 3. Mean values of fruit traits for 45 plants selected by early visual selection from the F₂ (Pop 1) family of cocona (Solanum sessiliflorum Dunal). Iranduba. Amazonas. 2021–2022.

Plant	Code	Fruit Yield (t ha−1)	Fruit Number per Plant	Fruit Mass (g)	Fruit Length (L) (cm)	Fruit Diameter (D) (cm)	L/D Ratio	Locule Number	Pericarp Thickness (mm)	Total Soluble Solids (Brix)	pH
1	P1-198	12.99	14	139.21	6.25	7.25	0.86	4.67	0.73	5.10	3.11
2	P1-113	11.61	18	96.76	6.63	6.15	1.08	4.67	0.71	4.90	3.73
3	P1-18	11.37	29	58.79	5.30	4.96	1.07	4.67	0.40	5.07	5.00
4	P1-15	10.57	29	54.66	4.83	4.43	1.09	4.00	0.44	4.73	2.88
5	P1-215	9.63	16	90.31	6.25	4.75	1.32	4.00	0.80	4.10	2.94
6	P1-98	9.60	23	62.61	5.80	5.53	1.05	4.00	0.47	4.93	5.50
7	P1-195	8.63	20	64.75	6.28	4.70	1.34	5.00	0.43	5.10	6.30
8	P1-79	8.13	19	64.21	4.20	3.50	1.20	4.00	0.47	4.77	3.22
9	P1-17	7.78	7	166.67	4.25	4.43	0.96	4.67	0.77	5.03	5.70
10	P1-16	7.43	13	85.77	6.40	5.65	1.13	3.00	0.65	4.95	3.16
11	P1-214	7.33	18	61.11	4.24	6.24	0.68	4.00	0.56	5.03	4.17
12	P1-95	7.27	7	155.71	8.40	7.50	1.12	5.00	1.14	3.95	3.06
13	P1-196	7.10	15	71.00	6.00	4.80	1.25	4.00	0.60	5.00	3.26
14	P1-219	6.87	14	73.57	5.90	5.40	1.09	4.00	0.50	4.97	5.70
15	P1-267	6.87	15	68.67	4.60	4.00	1.15	4.00	0.60	4.70	2.98
16	P1-252	6.83	18	56.94	5.63	5.23	1.08	4.00	0.47	5.03	3.87
17	P1-197	6.72	22	45.85	6.28	4.70	1.34	4.67	0.60	5.33	4.70
18	P1-13	6.43	16	60.31	6.00	6.37	0.94	4.00	0.67	4.97	5.43
19	P1-206	6.09	14	65.31	5.28	5.12	1.03	4.67	0.53	5.00	4.00
20	P1-178	6.05	9	100.91	5.33	4.77	1.12	4.67	0.57	4.77	2.90
21	P1-199	5.81	8	108.87	6.56	6.28	1.04	4.00	0.63	5.03	4.06
22	P1-168	5.80	22	39.55	7.33	7.50	0.98	4.00	0.33	5.03	4.53
23	P1-181	5.40	14	57.86	5.66	4.54	1.25	4.00	0.40	4.43	3.75
24	P1-220	5.38	7	115.29	6.23	6.30	0.99	4.00	0.60	4.77	5.70
25	P1-190	5.30	14	56.79	5.27	4.67	1.13	4.00	0.53	5.03	4.57
26	P1-214	5.00	8	93.75	7.50	4.70	1.60	-	-	-	-
27	P1-228	4.21	8	78.92	6.35	5.50	1.15	4.00	0.50	4.60	4.11
28	P1-173	3.93	5	118.00	5.20	5.50	0.95	4.67	0.53	5.00	5.20
29	P1-275	3.73	8	70.00	5.50	4.65	1.18	4.00	0.40	5.10	4.90
30	P1-32	2.95	12	36.82	4.46	4.52	0.99	4.00	0.49	5.00	-
31	P1-175	2.66	10	39.91	7.17	7.33	0.98	4.00	0.33	5.10	5.73
32	P1-274	2.57	2	192.50	6.85	7.70	0.89	5.00	1.15	5.05	5.75
33	P1-97	1.97	4	73.75	4.35	4.15	1.05	4.00	-	-	-
34	P1-176	1.50	3	75.00	5.20	5.50	0.95	-	-	-	-
35	P1-62	1.37	3	68.33	5.57	5.13	1.08	4.00	0.43	5.40	-
36	P1-255	1.03	2	77.50	4.80	5.00	0.96	4.00	0.60	5.10	5.30
37	P1-145	0.90	1	135.00	5.50	5.80	0.95	4.00	0.80	6.00	5.00
38	P1-133	0.87	1	130.00	7.20	5.60	1.29	5.00	0.80	4.10	4.90
39	P1-254	0.80	2	60.00	3.07	3.17	0.97	4.67	0.33	4.10	5.00
40	P1-34	0.70	3	35.00	5.20	5.70	0.91	6.00	0.60	4.10	-
41	P1-36	0.70	2	52.50	5.20	4.60	1.13	4.00	0.85	5.70	-
42	P1-177	0.70	1	105.00	6.90	5.50	1.25	4.00	0.70	5.10	2.93
43	P1-158	0.60	2	45.00	4.50	4.00	1.13	4.00	0.45	5.10	5.00
44	P1-61	0.33	1	50.00	5.20	4.60	1.13	4.00	0.40	4.90	-
45	P1-101	0.33	1	50.00	5.00	4.50	1.11	4.00	0.90	5.00	5.20
	Mean	5.11	10.67	80.19	5.68	5.29	1.09	4.26	0.59	4.91	4.41
	s2	12.08	61.91	1308.18	1.09	1.11	0.03	0.24	0.04	0.16	1.08
	CUB-4	3.60	5.23	35.70	3.98	4.06	0.98	4.00	0.36	5.10	4.41

). A similar pattern was observed in Population 2 (Pop 2), where visual screening of 80 plants yielded 52 high-performing genotypes (65%) surpassing CUB-4’s productivity (

Table 4. Mean values of fruit traits for 80 plants selected by early visual selection from the F₂ (Pop 2) family of cocona (Solanum sessiliflorum Dunal). Iranduba. Amazonas. 2021–2022.

Plant	Code	Fruit Yield (t ha−1)	Fruit Number per Plant	Fruit Mass (g)	Fruit Length (L) (cm)	Fruit Diameter (D) (cm)	L/D Ratio	Locule Number	Pericarp Thickness (mm)	Total Soluble Solids (Brix)	pH
1	P2-27	15.62	29	80.80	5.54	5.38	1.03	4.00	0.53	5.23	3.19
2	P2-117	14.87	22	101.36	6.14	6.04	1.02	4.00	0.60	4.77	4.10
3	P2-177	14.63	14	156.79	6.40	6.60	0.97	6.00	0.75	4.10	3.15
4	P2-4	13.53	18	112.78	5.32	5.24	1.02	4.67	0.63	5.03	5.45
5	P2-87	12.50	28	66.96	5.60	5.86	0.96	4.00	0.77	5.10	3.38
6	P2-115	12.23	31	59.19	5.28	4.90	1.08	4.00	0.53	4.77	3.59
7	P2-173	12.13	20	91.00	5.93	5.68	1.04	4.00	0.80	4.50	3.03
8	P2-206	12.12	27	67.36	5.36	5.44	0.99	4.00	0.61	5.07	-
9	P2-204	11.80	29	61.03	5.30	5.36	0.99	4.00	0.47	5.00	5.70
10	P2-198	11.53	34	50.85	4.64	4.74	0.98	4.00	0.33	5.10	4.44
11	P2-269	11.50	17	101.47	6.68	5.83	1.15	4.00	0.77	3.77	5.32
12	P2-29	11.13	17	98.24	5.80	6.00	0.97	4.50	0.75	4.60	3.12
13	P2-84	11.07	18	92.22	5.94	5.66	1.05	4.00	0.60	4.90	5.67
14	P2-175	10.77	19	85.00	6.20	5.15	1.20	4.67	0.69	4.97	3.25
15	P2-33	10.50	34	46.32	4.60	4.96	0.93	4.00	0.40	5.40	4.90
16	P2-256	10.37	20	77.75	5.72	5.56	1.03	4.00	0.63	4.77	5.60
17	P2-163	10.20	17	90.00	5.65	5.10	1.11	4.00	0.81	5.00	-
18	P2-151	9.87	13	113.85	6.10	6.13	1.00	4.67	0.67	4.93	4.14
19	P2-153	9.87	30	49.33	4.70	4.64	1.01	4.00	0.48	5.10	-
20	P2-236	9.70	13	111.92	6.63	6.38	1.04	4.00	0.66	4.80	3.37
21	P2-202	9.66	16	90.56	5.80	7.03	0.83	4.00	0.81	5.05	4.70
22	P2-234	9.53	16	89.38	5.58	5.58	1.00	4.00	0.53	5.10	4.70
23	P2-34	9.30	13	107.31	5.86	6.12	0.96	4.00	0.77	3.77	3.68
24	P2-23	9.23	12	115.42	6.07	6.23	0.97	5.00	0.73	4.43	3.00
25	P2-122	9.09	20	68.21	5.46	5.50	0.99	4.67	0.70	4.90	4.97
26	P2-54	8.70	22	59.32	5.00	5.30	0.94	4.33	0.53	5.10	4.20
27	P2-28	8.37	18	69.72	5.35	5.38	1.00	4.00	0.57	4.43	4.90
28	P2-268	8.37	14	89.64	5.68	5.92	0.96	4.67	0.63	4.43	5.47
29	P2-105	8.33	15	83.33	5.83	5.80	1.01	4.00	0.53	5.00	4.34
30	P2-220	7.93	16	74.38	5.70	5.85	0.97	4.00	0.63	4.43	3.70
31	P2-5	7.80	15	78.00	5.50	5.47	1.01	5.00	0.60	5.43	2.98
32	P2-6	7.63	25	45.80	4.40	4.75	0.93	4.00	0.49	5.00	5.47
33	P2-79	7.53	20	56.50	4.77	4.90	0.97	4.00	0.47	4.97	-
34	P2-190	7.51	15	75.13	4.24	4.12	1.03	4.00	0.63	5.13	5.00
35	P2-162	7.47	19	58.95	5.13	4.58	1.12	4.00	0.52	4.93	6.67
36	P2-138	7.30	22	49.77	4.40	4.60	0.96	4.00	0.50	4.90	2.98
37	P2-217	7.17	13	82.69	5.46	5.04	1.08	4.67	0.71	4.97	2.82
38	P2-170	6.96	12	87.00	6.10	6.24	0.98	5.67	0.83	5.10	3.60
39	P2-92	6.93	17	61.18	5.08	5.42	0.94	4.33	0.37	3.03	4.50
40	P2-93	6.93	9	115.56	6.35	6.63	0.96	5.00	0.73	5.10	5.26
41	P2-180	6.87	14	73.59	5.20	5.25	0.99	5.50	0.50	4.90	-
42	P2-25	6.87	15	68.67	4.04	4.10	0.99	5.00	0.51	5.17	4.56
43	P2-114	6.87	11	93.64	5.63	6.13	0.92	4.00	0.83	3.77	6.30
44	P2-187	6.80	23	44.35	4.20	4.40	0.95	4.00	0.40	4.97	5.30
45	P2-148	6.59	11	89.88	3.17	3.37	0.94	4.00	0.60	4.43	3.02
46	P2-161	6.00	20	45.00	3.85	4.10	0.94	4.00	0.51	4.95	-
47	P2-179	5.30	13	61.15	4.86	5.00	0.97	4.67	0.53	5.20	6.05
48	P2-288	5.13	5	154.00	6.75	5.90	1.14	5.33	0.80	4.77	2.97
49	P2-52	5.06	12	63.25	5.46	5.38	1.01	5.33	0.47	5.13	-
50	P2-8	5.03	14	53.93	4.86	4.72	1.03	5.00	0.47	5.10	4.90
51	P2-121	4.53	6	113.33	5.50	6.93	0.79	4.00	0.70	5.00	6.03
52	P2-86	3.73	6	93.33	5.60	5.86	0.96	4.00	0.63	4.77	4.30
53	P2-82	3.57	8	66.88	5.83	6.00	0.97	4.00	0.50	4.10	3.78
54	P2-19	3.10	7	66.43	5.24	5.16	1.02	4.00	0.47	5.10	4.87
55	P2-287	3.03	3	151.67	6.67	6.33	1.05	4.67	0.87	4.87	5.67
56	P2-189	2.90	5	87.00	5.20	5.40	0.96	4.67	0.57	5.20	6.30
57	P2-2	2.73	4	102.50	5.10	5.23	0.97	6.00	0.50	4.77	4.90
58	P2-144	2.67	9	44.44	4.50	4.45	1.01	4.00	0.45	5.50	3.03
59	P2-61	2.50	7	53.57	4.64	4.72	0.98	4.67	0.50	5.10	4.37
60	P-26	2.47	4	92.50	5.10	5.50	0.93	5.33	0.57	4.77	4.90
61	P2-123	2.40	3	120.00	6.00	5.93	1.01	4.00	0.57	4.93	5.00
62	P2-21	2.30	4	86.25	5.30	5.58	0.95	5.33	0.57	5.67	4.80
63	P2-55	2.13	2	160.00	6.15	6.70	0.92	4.67	0.67	4.77	-
64	P2-50	1.97	4	73.75	4.88	4.95	0.98	4.00	0.50	-	-
65	P2-20	1.87	6	46.67	4.84	4.64	1.04	6.33	0.47	5.13	4.80
66	P2-239	1.87	2	140.00	4.00	4.57	0.88	4.00	0.60	4.10	-
67	P2-145	1.80	2	135.00	5.85	6.35	0.92	4.00	0.67	5.10	4.93
68	P2-66	1.57	3	78.33	5.27	5.53	0.95	4.00	0.57	4.77	4.95
69	P2-56	1.40	2	105.00	5.80	6.25	0.93	4.67	0.80	4.43	-
70	P2-57	1.33	2	100.00	5.75	5.00	1.15	4.00	0.65	5.00	-
71	P2-67	1.13	2	85.00	5.15	5.30	0.97	4.00	0.50	4.77	4.85
72	P2-188	1.10	1	165.00	10.80	11.50	0.94	-	-	-	-
73	P2-60	1.00	3	50.00	4.33	4.40	0.98	4.00	0.50	5.13	-
74	P2-141	0.97	2	72.50	4.65	4.75	0.98	4.00	0.55	5.10	4.90
75	P2-149	0.80	1	120.00	5.60	5.70	0.98	4.00	0.70	4.20	4.90
76	P2-7	0.73	1	110.00	4.70	4.50	1.04	4.00	0.60	5.20	4.90
77	P2-74	0.57	1	85.00	5.10	4.70	1.09	4.00	0.70	4.10	4.95
78	P2-119	0.47	1	70.26	4.20	4.70	0.89	-	-	-	-
79	P2-285	0.47	1	70.00	5.10	4.50	1.13	4.00	0.50	6.00	-
80	P2-72	0.33	1	50.00	5.30	5.10	1.04	4.00	0.40	5.10	4.85
	Mean	6.45	12.75	85.17	5.39	5.45	0.99	4.37	0.60	4.85	4.52
	s2	17.16	81.10	865.14	0.87	0.99	0.00	0.33	0.02	0.21	0.97
	CUB-4	3.60	5.23	35.70	3.98	4.06	0.98	4.00	0.36	5.10	4.41

). Both selection efficiencies significantly exceeded the theoretical random selection threshold of 50% (χ² test, p < 0.05), demonstrating that early visual phenotyping is an effective strategy for yield improvement in this breeding program.

3.3. Correlations Among Fruit Traits

In selected Pop 1 plants, fruit yield exhibited a strong positive correlation with fruit number (r = 0.86, p < 0.010; Figure 6). The efficiency of selection based on fruit number was similar to that observed for fruit yield, with 29 out of 45 plants (64.4%) producing more fruits than the CUB-4 control line. Similarly, in Pop 2, fruit yield was significantly correlated with fruit number (r = 0.84, p < 0.010; Figure 6), with 56 out of 80 plants (70.0%) exceeding CUB-4 in fruit number. These high selection efficiencies (64.4% and 70.0%, respectively) suggest that visual selection for fruit yield indirectly favors genotypes with increased fruit number, even when selection is primarily focused on overall yield performance.

The correlations between fruit mass and fruit length were r = 0.30 (p = 0.068) in Pop 1 and r = 0.66 (p = 0.001) in Pop 2 (Figure 6). The correlations between fruit mass and fruit diameter were r = 0.41 (p = 0.010) and r = 0.62 (p = 0.001) in Pop 1 and Pop 2, respectively (Figure 6). In contrast, the correlations between fruit mass and L/D ratio were r = −0.23 (p = 0.176) and r = 0.05 (p = 0.704) in Pop 1 and Pop 2, respectively. Therefore, selection for fruit mass is likely to include fruit length and diameter to some extent, but not fruit shape.

In the selected plants, pericarp thickness showed a significant correlation with fruit mass in both populations: Pop 1 (r = 0.76, p < 0.01) and Pop 2 (r = 0.70, p < 0.01). However, no significant correlation was observed between pericarp thickness and fruit yield (r = −0.05 in Pop 1 and r = 0.19 in Pop 2). These results suggest that visual selection for fruit yield is independent of pericarp thickness.

4. Discussion

4.1. Heritability

This study evaluated the heritability of several plant and fruit traits, the efficiency of early visual selection for fruit yield, and estimated correlations among fruit traits in cocona. Among the plant traits, heritability estimates ranged from 0.42 to 0.71 for characteristics such as plant height, collar diameter, leaf length, petiole length, and canopy diameter. For fruit traits, heritability values ranged from 0.27 to 0.88 for number of flowers per plant, fruit length, fruit diameter, and the L/D ratio. Early visual selection for fruit yield was effective. In Pop 1, 29 of the 45 selected plants had high yields, and in Pop 2, 52 of the 80 selected plants were productive. This represents ~65% selection efficiency, exceeding the 50% efficiency expected under random selection. Additionally, strong positive correlations were observed between fruit yield and number of fruits per plant, as well as between fruit mass and pericarp thickness.

Research on heritability in cocona remains scarce. Pahlen [1] reported high heritability for fruit mass (h² = 0.89) and noted that fruit size and shape were consistent across successive generations, although heritability values were not specified. In the present study, the heritability of fruit mass was not evaluated, but broad-sense heritability estimates for fruit length, fruit diameter, and the L/D ratio ranged from 0.27 to 0.55. Although Pahlen [1] did not detail the method used to estimate heritability, it is possible to infer an approximate value of 0.86 based on an ANOVA performed under a randomized complete block design using soil type as a blocking factor. This suggests that fruit mass can be selected more efficiently than size or shape, and that the latter traits should be selected independently.

The moderate heritability observed for fruit size and shape traits suggests polygenic control. In tomato, a related species, at least seven genes have been identified as regulators of these traits, influencing processes such as hormonal regulation, locule number, endoreduplication, and other factors [20]. Conversely, Salick [10] proposed maternal inheritance for these traits, based on crosses among seven parental lines.

This apparent discrepancy can be explained by the phenotypic maternal effect, in which maternal nuclear genes or the maternal environment predominantly influence the offspring’s phenotype [31], with effects persisting for up to two generations [32]. Both the present study and Salick [4] found that the F₁ progeny displayed fruit size and shape similar to those of the maternal genotype, confirming the presence of a phenotypic maternal effect. However, the present study also observed a high degree of segregation of these traits in the F₂ generation, indicating that the maternal effect does not extend beyond F₁. This supports the recommendation to start selection for size and shape from F₂ generation.

The moderate heritability observed for fruit size indicates a significant environmental influence. Pahlen [1] demonstrated that high soil fertility levels can increase fruit mass. This suggests that proper fertilization and pH correction are necessary to ensure accurate selection for this trait.

4.2. Early Visual Selection

Early visual selection also reduces the time and labor needed to evaluate fruit yield and handle seeds. In the present study, a selection efficiency of 65% was observed when CUB-4 was used as the control. At the time of visual selection, selected plants bore several fruits; however, only one to four fruits per plant were harvested in some plants (Table 3 and Table 4) due to fruit damage caused by Alternaria sp. and plant mortality resulting from Sclerotium rolfsii infection [8]. Plant losses were estimated at 18%, as 250 plants were originally planted, but only 203 and 202 were evaluated in Pop 1 and Pop 2, respectively. Therefore, preventive fungicide application (6–8 times per cycle, as suggested by Terrones [8]) may improve selection accuracy.

4.3. Analysis of Correlations Among Fruit Traits

Strong correlations among key agronomic traits facilitate genotype selection. In this study, fruit yield showed a strong positive correlation with the number of fruits per plant (Figure 6). This strong association was reflected in the efficiency of selection for the number of fruits per plant (64 and 70%), which was similar to fruit yield (65%). These results indicate that evaluating the number of fruits per plant before the harvest can be used to improve visual selection of fruit yield. Correlations between fruit mass and fruit length; fruit mass and fruit diameter; and fruit mass and L/D ratio in both populations were zero to moderate. It indicates that the selection for these traits should be evaluated and selected independently.

Pericarp thickness is a trait that determines the potential uses of the fruit. Greater thickness is desirable for industrial processing [8], such as the production of nectars, jellies, canned fruits [8,19], and powdered products. Even though it is not directly linked to yield, it correlates with fruit mass. Therefore, it is advisable to evaluate pericarp thickness in plants selected for high yield, especially if the goal is to develop cultivars for industrial use.

4.4. Implications, Limitations and Further Studies

Based on these results, we propose several strategies for breeding cocona. The F₂ population should consist of at least 200 healthy plants to enable effective selection. Moderate selection pressure should be used for plant height, collar diameter, and number of flowers per plant, while leaf length, canopy diameter, number of fruits per plant, and fruit size and shape should be selected more gently. Early visual selection for fruit yield can be improved by preventing disease and evaluating the number of fruits per plant. Fruit mass should be selected separately from fruit yield and shape. Selection for size and shape should start from the F₂ generation.

These strategies may not perform similarly in non-flooded soils, where harvesting can extend up to five months [1] and where fertilization and soil acidity corrections are applied. In contrast, cultivation in flooded soils usually occurs from August to February, resulting in a harvest window of only two months. Additionally, these soils are naturally rich in nutrients and organic matter.

In this study, fruit hairiness heritability was h² = 1.00, indicating complete genetic fixation of this trait within both parental lines. The absence of known glabrous genotypes in the species suggests a narrow genetic base for this trait in cultivated cocona. Studies in tomato breeding have shown that when Solanum lycopersicum cv. MT is pollinated with Solanum galapagense (a hairy-fruited species), the resulting fruits are also hairy. This implies potential paternal genetic control of the trait [23]. Thus, generating glabrous-fruited cocona may require hybridization with tomato species, though such an approach demands reproductive compatibility and the evaluation of trait stability in segregating generations. To elucidate the genetic architecture of fruit hairiness in cocona, future research should integrate molecular marker analysis, gene expression profiling, and histological characterization of fruit trichomes.

Finally, given the contrasting conditions between floodplain and non-floodplain soils, similar investigations should be conducted in Amazonian Oxisols and Spodosols to guide the improvement of cocona for these environments.

5. Conclusions

Most morpho-agronomic traits evaluated in cocona showed moderate-to-high broad-sense heritability, supporting the effectiveness of phenotypic selection. Early visual selection in the F₂ generation was 65% efficient, underscoring its potential as a rapid and low-cost strategy for identifying high-yielding genotypes. The number of fruits per plant is a reliable complementary criterion for selecting high-yielding genotypes.

Fruit mass, size, shape, and pericarp thickness should be assessed independently in subsequent selection stages to meet processing quality standards. The observed complete fixation of fruit hairiness (h² = 1.00) highlights a narrow genetic base and supports the need for molecular characterization and interspecific hybridization to broaden genetic variability.

Future studies should validate these findings under non-floodplain conditions, particularly in Amazonian Oxisols and Spodosols, to guide the development of superior cultivars adapted to diverse environments.