Efficiency of Stratification on Yield and Popping Expansion of Popcorn in the Context of Mass Selection

Abstract

The current popcorn production in Mexico is insufficient to meet the demand for this grain, since there are no genotypes suitable for production within the country. The native popcorn varieties do not meet market standards; therefore, it is necessary to carry out genetic improvement on yield and popping expansion. The objective of this study was to estimate the response to selection on yield and popping expansion of popcorn under two modalities of mass selection. We used a population of popcorn from the first selection cycle (C1) of the cross between a commercial popcorn of the North American Yellow Pearl race and a native population of the Mexican race Palomero Toluqueño. An additional cycle of mass selection (C2) was carried out with stratification (MSS) and without stratification (MNSS) followed by a field trial that included the different cycles of selection. The parents of the initial cross and two controls, under a complete randomized block experimental design with 10 replications in two localities, and the genetic gain per cycle was calculated. Genotypes C2_MSS, C2_MNSS, C1_MSS and the control Palomero Ixtenco showed the highest average grain yield with values of 4.9 t ha⁻¹. For popping expansion, the Jack Superior control showed the best flake volume with 34.7 cm³ g⁻¹, in contrast to the native popcorn Criollo Plaza with a value as low as 7.2 cm³ g⁻¹. Chapingo was the locality with the highest grain yield, averaging 5.08 t ha⁻¹, while in San Salvador Atenco, the yield was 2.78 t ha⁻¹. Genetic gains were found with a popping expansion of 12.50% with stratification and 11.42% without stratification. For grain yield, a gain of 6.82% was obtained using stratification and 0.74% without stratification. Stratified visual mass selection is an efficient method for genetic advancement in popcorn yield, while popping mass selection is efficient regardless of stratification.

1. Introduction

The diet of Mexicans is mainly based on corn (Zea mays L.), in very varied foods. In the popcorn sector, Mexico has a production deficit of popcorn. SIAP [1] reported a production volume of just 246 tons, so import is used to meet demand; thus, Mexico has to import approximately 80 thousand tons of popcorn annually from countries like the United States and Argentina.

Mexico has native popcorn varieties, such as the Palomero Toluqueño race, which are grown for food purposes, but not in the form of popcorn, due to its low expansion. Santacruz-Varela [2] in a test of popping, obtained expansion calibers of 1.62 cm³ g⁻¹ in a collection of Mexican popcorns conserved at the US National Plant Germplasm System, which contrasts with the 38.87 cm³ g⁻¹ shown as an average by commercial North American popcorns. In order to meet the demand for popcorn in Mexico, without resorting to imports, national popcorn breeding programs need to be promoted to increase the yields and volume of popcorn varieties in this country.

Maize breeding methods have made it possible to develop varieties that meet the needs of man through hybridization and selection. The effectiveness of selection is based on the increase in the frequency of favorable alleles [3]. Simple mass selection (with no stratification), which is carried out by the farmer, has little genetic advancement, mainly due to the high environmental variability within the plot. Stratification of the whole plot into small sub-plots, as proposed by Gardner more than six decades ago [4], seeks to reduce variation caused by environmental effects. This procedure involves more activities than simple selection. Several authors have obtained average gains in grain yield in maize, around 6.5% per selection cycle [5,6].

Authors such as [7,8,9] obtained estimates of response to selection for different methods of selection, based on known genetic variances of a population of Brazilian popcorn. In the case of simple mass selection, they reported gains of 2.23% in yield and 11.77% in popping expansion, while with stratified visual mass selection gains of 2.33% in yield and 12.13% in popping expansion volume were obtained. In the United States of America, emphasis has been placed on producing hybrids, mainly three-way hybrids, to take advantage of each of the three components of the heterotic pattern identified for yellow popcorn: Amber Pearl, South American and Supergold, with which they have reached expansion volumes greater than 50 cm³ g⁻¹.

In Mexico, researchers [10] developed the V460P popcorn variety from a broad-based population from the United States of America. This variety was obtained through 11 cycles of visual stratified mass selection, where the selection criteria were grain yield and popping expansion. The variety had a grain yield potential of up to 5 t ha⁻¹ under irrigation, with an expansion volume of 30 cm³ g⁻¹; however, this volume is still below the popping expansion of the imported popcorn, the same situation as for the recently released three-way hybrid H-301 P with a popping expansion around 24 cm³ g⁻¹, below the minimum of at least 40 cm³ g⁻¹ required for the market.

The objective of this study was to evaluate the efficiency of the stratified visual mass selection method in comparison to simple mass selection, on a broad-genetic base popcorn population for grain yield and popping expansion, in order to optimize genetic gains in the process of generating new varieties.

2. Materials and Methods

2.1. Genetic Material

The population used comes from the first selection cycle (C1) of stratified mass selection of a composite of outstanding individuals of the F2 cross between an advanced generation of the commercial three-way hybrid Iowa Pop 12 (IP) with a high popping expansion (around 40 cm³ g⁻¹), classified as a North American Yellow Pearl popcorn [11] and the landrace Criollo Plaza (CP) belonging to the Mexican maize race Palomero Toluqueño, which is well-adapted to the local environment.

2.2. Site for Selection and Agronomic Management

The planting was carried out manually on 4 June 2021, at Montecillo, State of Mexico located at 19°28′ North latitude and 98°54′ West longitude, at 2240 masl. Two seeds were deposited every 50 cm and conducted under irrigation conditions. Two manual weedings were carried out just before stage V5 [12] and at the final stage of cultivation. The soil of the site is a chromic vertisol type with a clay texture. The fertilization formula used was 140N-60P-00K, applying 60% of the N and all the P at the planting and the rest of the N after the second weeding. The crop was conducted under irrigation during the 2021 Spring–Summer cycle. The flood irrigation system was used, applying one irrigation immediately after sowing and three auxiliary irrigations during crop development, as a complement to rainfall. An average annual temperature of 16.1 °C and an average annual precipitation of 586 mm were registered in 2021.

2.3. Procedures for Selection

The whole selection plot consisted of 70 rows 100 m long and 0.8 m wide. Five rows from each North and South edge as well as 5 m at the East and West ends were left as borders. The remaining 60 rows were divided into two equal sections for applying a second cycle of mass selection under two variants as follows:

2.3.1. Mass Selection with Stratification (MSS)

The section of 60 rows of 50 m long was subdivided into 24 sub-plots of 10 rows of 9 m long each accommodated in four blocks of six sub-plots with 1-m separation. Each sub-plot held a maximum of 380 plants (340 useful plants since edge plants and non-competent plants were not considered). The goal was set for selecting the 50 best plants (around 15%) based on visual estimation of yield from individuals within sub-plots based on ear size and general appearance [13], for a large total of 1200 selected individuals (Figure 1).

2.3.2. Mass Selection with No Stratification (MNSS)

The section of land devoted to mass selection with no stratification was considered as a single unit (Plot 25, Figure 1) in which the same selection pressure (≈15%) was applied. By extracting the best 1200 individuals for grain yield, as explained in the previous paragraph, based on visual selection from the total individuals of the unit.

2.3.3. Selection for Popping Expansion

Popping tests were applied in the laboratory individually to the kernels of plants selected in the field in both MSS and MNSS. Using rehydrated samples with a moisture content in the range of 13.5 to 15.5%. This was obtained by conditioning the seed samples in a 5 °C controlled environment chamber with 78% relative humidity for six weeks. The popping tests were performed on 30 g samples of grain in hot air-based popcorn maker machines (Hamilton Beach, Model 73400, Bienne, Switzerland) until the grains stopped bursting (approximately 1:30 min). The expansion volume was determined in a 2000 mL graduated cylinder, with a diameter of 8.89 cm, as proposed by [14], by dividing the volume obtained by the initial weight of the sample, to express its value in cm³ g⁻¹.

Based on the data from the popping tests of 15% of the plants from the visual selection in the field, the best third was selected for popping expansion. This results in a combined selection pressure (field-laboratory) of 5%. The remaining seed from the selected plants was used to generate the balanced composites corresponding to mass selection with stratification (C2_MSS) and without stratification (C2_MNSS).

2.4. Performance Evaluation of Selection Cycles

Performance tests included progeny materials IP and CP, cycle 0 (C0, recombinant composite of outstanding S₁ lines derived from the original cross between IP and CP) and subsequent selection cycles C1_MSS, C2_MSS and C2_MNSS. Palomero Ixtenco (PI) acquired in a maize fair in the state of Tlaxcala and Jack Superior (JS), a commercial North American Pearl popcorn imported from the US were included as controls.

2.4.1. Field Trials

Trials were established during the Spring–Summer cycle of 2022 at two localities of the highlands of Central Mexico. The first one in Chapingo, Texcoco, State of Mexico (19°29′ North latitude and 98°53′ West longitude, at an altitude of 2247 m, with an average annual temperature of 15.2 °C and average annual precipitation of 637 mm) [15]. The second locality was San Salvador Atenco, State of Mexico (19°32′ North latitude and 98°56′ West longitude, at an altitude of 2250 m, with an average annual temperature of 15.1 °C, average annual precipitation of 605 mm) [15] and a chromic vertisol soil.

The experimental design used was randomized complete blocks with 10 replications in each locality. The experimental unit consisted of three rows 5 m long and 0.8 m wide, where two seeds were deposited every 0.5 m, to obtain a population density of 50,000 plants ha⁻¹.

Planting was carried out on 29 May and 13 June 2022 in Chapingo and San Salvador Atenco, respectively. The crop was conducted under irrigation conditions in both localities. Fertilization in Chapingo was with the formula 180N-85P-15K. Applying 50% of the N and all the P and K at planting and the rest of the N in the second weeding. In San Salvador Atenco the formula was 140N-00P-00K, applying the complete formula at the second weeding. Weed control was carried out with herbicides Atrazine and Acetochlor in a pre-emergent manner, with doses of 2 kg and 2 L per ha, respectively. Chemical control of the locust (Schistocerca piceifrons Walker) pest was carried out by applying the insecticide Karate Zeon^® (Lambdacihalothrin) at a dose of 300 mL ha⁻¹ at the V5 phenological stage.

Of the three rows in each experimental unit, two were used for the evaluation of grain yield. In the remaining row, plant-to-plant crosses were performed to obtain seeds free from external pollen to perform the popping tests. Thus, avoiding eventual xenia effects on popping expansion [16]. The grain yield (GY) was calculated by taking the total weight of ears from each experimental unit, discounting the corncob percentage (corncob weight/ear weight) estimated from a 5-ear sample, and involving an additional adjustment to 14% moisture, the result was reported in t ha⁻¹. Other traits registered were the number of grains in 10 g (NG), obtained by weighing 10 g of grain and counting the number of kernels present there; the number of ears per plant (NEP) was obtained by dividing the total number of ears by the total number of plants present in the experimental unit.

2.4.2. Popping Tests

The popping tests were performed on grains with moisture content between 13.5 and 15.5%, following the same process of popping and equipment as described in the previous section. The expansion volume (EV) was obtained in cm³ g⁻¹ of grain. Other traits registered in the lab were the flake color (COL) with a visual scale of 0-1, where 0 corresponds to a white flake and 1 to a creamy one. Flake shape (FS), using a scale of 1–5 as proposed by [14], where 1 corresponds to a round flake (“mushroom”) and 5 to a flake with pronounced peaks (“butterfly”). Pericarp retention (PR) was rated on a scale of 1–5, where 1 corresponds to a flake with high retention of pericarp (81–100%), while 5 corresponds to flakes with little retention (10–20%), according to [14]. Unpopped grain weight (UPGW) was measured in g with an analytical scale and expansion volume per hectare (VHA) was created by multiplying GY × EV and the result was reported in m³ ha⁻¹.

2.4.3. Estimation of Response to Selection

The response to selection (ΔG) per cycle was estimated based on the linear regression coefficient of the phenotypic means on the number of selection cycles. The regression coefficient was expressed as a percentage of the average performance of the original population (C0) [17]. The response to the selection was estimated only for GY and EV.

2.5. Statistical Analysis

A combined analysis of variance across localities and comparison of means (Tukey, p ≤ 0.05) were performed using the statistical package SAS OnDemand for Academics (version 4.3) [18]. The response to the selection was calculated from a simple linear regression coefficient on the number of recurrent selection cycles.

3. Results

3.1. Analysis of Variance

The analysis of variance detected significant differences (p ≤ 0.05) caused by the effect of localities (LOC) for all variables (

Table 1. Sources of variation and mean squares of evaluated characters from the yield and popping trials in San Salvador Atenco and Chapingo, Mexico, 2022.

	Sources of Variation
	LOC		GEN		REP/LOC		GEN × LOC		ERROR	CV (%)
Traits	Mean squares
GY	155.63	**	43.97	**	0.38	NS	3.39	**	0.38	15.58
NG	6981.25	**	4607.85	**	84.21	**	34.49	NS	64.61	11.6
NEP	7.69	**	1.41	**	0.05	NS	0.04	NS	0.06	16.94
COL	0.04	NS	0.99	**	0.02	NS	0.02	NS	0.01	36.64
FS	14.26	**	8.67	**	0.45	NS	1.17	**	0.37	14.77
PR	3.87	**	9.16	**	0.90	**	1.65	**	0.42	16.13
UPGW	35.23	**	317.71	**	3.91	NS	14.17	**	2.66	43.54
EV	30.19	NS	1186.94	**	15.81	NS	75.54	**	12.87	14.80
VHA	58,587.60	**	36,595.48	**	606.63	NS	2555.32	**	516.21	24.34

NS: non-significant, ** Significant at p ≤ 0.01, LOC: locality, GEN: genotype, REP/LOC: replications nested into localities, GY: grain yield, NG: number of grains in 10 g, NEP: number of ears per plant, COL: color, FS: flake shape, RP: pericarp retention, UPGW: unpopped grain weight, EV: expansion volume, VHA: volume of flakes per hectare, CV: coefficient of variation.

), except for COL and EV; that is, the last two characteristics are little influenced by the environment.

Genotypes (GEN) significantly influenced all variables, while in the GEN × LOC interaction differences were found (p ≤ 0.05) for GY and EV; that is, the rate of change in the performance of the genotypes was different when moving from one environment to another.

Coefficients of variation were below 17% for GY, NG, NEP, FS, PR, and EV, while in the rest of the variables, they were above 24%. The COL and UPGW were the highest, with 36.7% and 43.5% respectively. This is because, in COL, the evaluation is visual and accuracy can be affected by factors such as luminosity.

3.2. Effect of Genotypes

Comparisons of genotype means indicated differences (p ≤ 0.05) in all the evaluated variables (

Table 2. Comparison of mean values of eight genotypes in grain yield and popping expansion in San Salvador Atenco and Chapingo, Mexico, 2022.

Traits	Genotype
	C2MSS		C2MNSS		C1MSS		C0		IAP12		CP		PI		JS		THSD (0.05)
GY (t ha−1)	5.23	a §	4.93	a	4.90	a	4.52	b	2.31	d	3.88	c	5.03	a	0.41	e	0.64
NG	65.32	c	66.50	c	62.94	c	62.55	c	80.10	b	52.90	d	62.50	c	103.94	a	7.97
NEP	1.49	a	1.55	a	1.55	a	1.38	b	1.56	a	1.06	c	1.66	a	0.88	c	0.23
COL (0-1)	0.27	b	0.27	b	0.28	b	0.22	b	0.66	a	0.00	c	0.22	b	0.67	a	0.11
FS (1-5)	4.55	a	4.55	a	4.10	b	3.75	c	4.67	a	2.85	d	3.90	c	4.94	a	0.61
PR (1-5)	4.40	a	4.45	a	3.85	b	3.90	a	4.44	a	2.45	c	4.10	a	4.50	a	0.64
UPGW (g)	2.31	b	2.05	b	2.38	b	2.63	b	1.37	b	13.55	a	2.53	b	2.82	b	1.61
EV (cm3 g−1)	28.27	b	28.00	b	25.13	b	21.77	c	27.85	b	7.23	d	23.53	c	34.68	a	3.58
VHA (m3 ha−1)	143.85	a	138.48	a	124.12	a	94.49	c	62.01	d	27.69	e	117.99	b	13.79	e	24.62

^§ Means with the same letter within rows do not differ statistically (Tukey, p ≤ 0.05), C2_MSS: Cycle 2 mass stratified selection, C2_MNSS: Cycle 2 mass non-stratified selection, C1_MSS: Cycle 1 mass stratified selection, C0: Cycle 0, IAP12: Iowa Pop 12, CP: Criollo Plaza, PI: Palomero Ixtenco, JS: Jack Superior, GY: grain yield, NG: number of grains in 10 g, NEP: number of ear per plant, COL: color, FS: flake shape, PR: pericarp retention, UPGW: unpopped grain weight, EV: expansion volume, VHA: Flakes volume per hectare. THSD: Tukey honest significant difference.

). The Ixtenco Palomero and the C2_MSS, C2_MNSS and C1_MSS selection cycles recorded grain yield of more than 4.8 t ha⁻¹, ranking within the first group of significance (a), while the lowest grain yield was Jack Superior without reaching 0.5 t ha⁻¹.

Regarding the number of grains per 10 g, the Jack Superior and Iowa Pop 12 genotypes had the highest values with 104 and 80, respectively, while Criollo Plaza had the lowest value with 53.

Ixtenco Popcorn, Iowa Pop 12 and the C1_MSS, C2_MNSS and C2_MSS selection cycles showed to be prolific, obtaining more than 1.4 ears per plant. Criollo Plaza and Jack Superior recorded values of 1.06 and 0.88 ears per plant, respectively.

The genotypes Jack Superior and Iowa Pop 12 had the highest COL values, presenting a cream-like color, while Criollo Plaza had the lowest value, being their popcorn practically white. The genotypes Jack Superior, Iowa Pop 12 and the selection cycles C2_MSS and C2_MNSS presented FS from bilateral to butterfly, while in Criollo Plaza the flake was from semi-mushroom type towards unilateral. For pericarp retention and unpopped grain weight, seven of the eight genotypes evaluated had values below 40% and 3.0 g (10%), respectively. As for the landrace Criollo Plaza, pericarp retention in their flakes ranged from 61 to 80%, with an unpopped grain weight of 13.55 g (45%).

Regarding expansion volume, Jack Superior, a commercial genotype in the USA, showed the highest value with 34.68 cm³ g⁻¹. The lowest value was Criollo Plaza, with 7.23 cm³ g⁻¹.

The C1_MSS, C2_MNSS and C2_MSS selection cycles obtained VHAs larger than 120 m³ ha⁻¹, ranking within the first group of significance (a). The genotypes with lower volumes of popcorn per hectare were Criollo Plaza and Jack Superior, with 27.7 and 13.8 m³ ha⁻¹, respectively.

3.3. Effect of Localities

Chapingo, Mexico was the site with the agroclimatic conditions that produced the highest grain yields (

Table 3. Comparison of means from two locations in the tests of yield and popping characteristics in San Salvador Atenco and Chapingo, Mexico, 2022.

Traits	Localities
	Chapingo		San Salvador Atenco		THSD (0.05)
GY (t ha−1)	5.08	a §	2.78	b	0.21
NG	62.52	b	76.09	a	2.56
NEP	1.62	a	1.17	b	0.07
COL (0-1)	0.33	a	0.30	a	0.04
FS (1-5)	3.83	b	4.46	a	0.19
PR (1-5)	3.83	b	4.16	a	0.20
UPGW (g)	3.30	b	4.18	a	0.52
EV (cm3 g−1)	23.64	a	24.85	a	1.25
VHA (m3 ha−1)	118.69	a	68.03	b	7.79

^§ Means with the same letter within rows do not differ statistically (Tukey, p ≤ 0.05), C2_MSS: Cycle 2 mass stratified selection, C2_MNSS: Cycle 2 mass non-stratified selection, C1_MSS: Cycle 1 mass stratified selection, C0: Cycle 0, IAP12: Iowa Pop 12, CP: Criollo Plaza, PI: Palomero Ixtenco, JS: Jack Superior, GY: grain yield, NG: number of grains in 10 g, NEP: number of ears per plant, COL: color, FS: flake shape, PR: pericarp retention, UPGW: unpopped grain weight, EV: expansion volume, VHA: Flakes volume per hectare. THSD: Tukey honest significant difference.

), as well as larger grains, more prolific plants, higher volume of popcorn per hectare and lower loss for unpopped grains than those obtained in San Salvador Atenco, State of Mexico. In Chapingo the flake tended to take bilateral form and not butterfly, as in the site of San Salvador Atenco. Both locations produced similar results (p ≤ 0.05) in terms of expansion volume and color of the flakes.

3.4. Response to Selection

The genetic gain for GY was 8.32% and 6.82% for C1_MSS and C2_MSS cycles, respectively, and 0.74% for C2_MNSS (

Table 4. Means and genetic gains obtained in the C0, C1_MSS, C2_MNSS and C2_MSS cycles of recurrent mass selection in popcorn for grain yield (GY) and expansion volume (EV) in San Salvador Atenco and Chapingo, Mexico. 2022.

Cycle	GY (t ha−1)	ΔGGY (%)	EV (cm3 g−1)	ΔGEV (%)
C0	4.52	-	21.77	-
C1MSS	4.80	8.32	25.13	15.43
C2MNSS	4.93	0.74	28.00	11.42
C2MSS	5.23	6.82	28.27	12.50

C0: Cycle 0, C1_MSS: Cycle 1 mass stratified selection, C2_SS: Cycle 2 mass non-stratified selection, C2_MSS: Cycle 2 mass stratified selection, GY: grain yield, Δ_GGY: Genetic gain for grain yield, EV: Expansion volume, Δ_GEV: Genetic gain for expansion volume.

). Based on the initial average and using MSS, this gain is equivalent to an increase of 0.376 and 0.334 t ha⁻¹ in cycles C1_MSS and C2_MSS, respectively, while with C2_MNSS gains of 0.036 t ha⁻¹ were obtained. The genetic advancement for EV was 15.43%, 12.50% and 11.42% for the C1_MSS, C2_MSS and C2_MNSS cycles, respectively. Starting from the initial mean and using the MSS method, this gain is equivalent to an increase of 3.36 and 3.14 cm³ g⁻¹, while with simple selection (MNSS) its equivalent is 2.87 cm³ g⁻¹.

The linear regression analysis showed the evolution of genetic gains in successive selection cycles (Figure 2 and Figure 3). It was observed that the MSS method to obtain the C1_MSS and C2_MSS cycles resulted in acceptable gains for both grain yield and expansion volume, which increased by considerable amounts.

4. Discussion

The average phenotypic expression of a population is the result of a number of factors involved in the development of individuals. These factors can be grouped into two main categories, genetic and environmental, and often a significant contribution of interaction between the two is expressed. The extent of involvement of the above factors depends largely on the genetic nature and heritability of the different traits under assessment [3].

In the present study of popcorn populations, judging by the magnitude of the mean squares in Table 1, most of the traits evaluated showed statistical significance for both the environmental and genetic factors. It is clear that there is a greater influence of the environment than that of the varieties factor in most traits, especially in grain yield, number of grains per 10 g, number of ears per plant, shape of flakes and volume of flakes per hectare, as the latter is heavily influenced by the environmental-dependent grain yield [3]. Such traits as color of the flakes, pericarp retention, unpopped-grain weight and expansion volume are highly dependent on the genetic components [19].

Two traits are especially important in popcorn, grain yield per unit area and the expansion capacity per unit weight. The grain represents the initial amount of raw material to produce the snack. The expansion directly influences the profit of the vendors, as the snack is commonly sold by volume.

In popcorn as in field corn, grain yield represents an extremely complex, multigenic trait, highly influenced by the environment, as shown in Table 1. In this context, for a popcorn variety to show a high yield, it must possess favorable genes and proper adaptation to an agroecological region. Even when high-yielding popcorn is grown outside of its adaptation space it fails to produce satisfactorily. As reported by [20], where the commercial non-adapted popcorn Valle Verde grown in central Mexico did not have favorable results for grain yield to such a degree that it could not be quantified.

Genotype–environment interaction plays an important role in crop production, and it is reflected in the stability of genotypes through different production environments [21]. For this reason, popcorn varieties developed in the United States of America show excellent yields under the environmental conditions of the US corn belt. When established outside those conditions, yield is severely affected, as seen in the yield of the genotypes Iowa Pop 12 and Jack Superior (Table 2). This constitutes a significant barrier to the direct adaption of varieties from other latitudes in Mexico’s agricultural area. It might be possible to use them as gene donors of favorable genes into Mexican native popcorn landraces that already have adapted to local environmental conditions. Results support this approach as populations that received North American contribution with posterior mass selection (i.e., C1 _MSS, C2_MSS, C2_MNSS) show important levels of grain yield (Table 2).

The significant differences in popping expansion between genotypes can be attributed to the high variability of genetic nature. Ref. [22] mentioned that the expansion volume is a polygenic trait. Genes contributed to the hardness of the endosperm and pericarp thickness in a different fashion between selection cycles, which by its magnitude suggests that this complexity can be efficiently exploited by applying some method of genetic improvement.

The expansion volume represents the most important quality-related characteristic of popcorn and that which distinguishes it from other types of maize [20,22]. When it comes to commercial popcorn with years of genetic improvement, as is the case of the North American Yellow Pearl Popcorn race [2], its superiority is evident compared to popcorn which has not received improvement in expansion volume [20].

Expansion volume is usually negatively correlated with grain yield. According to [19], breeding methods that utilize genetic variation for expansion volume and grain yield simultaneously are likely to be more successful in terms of grain yield and some other flake quality traits. In order to reconcile this issue, [23] proposed the expanded popcorn volume per hectare as a “super trait” that is positively correlated with both grain yield and popping expansion; thus, the selection of this trait could be useful to obtain simultaneous gains in both traits.

In our study, populations that received mass selection (C2_MSS, C2_MNSS, C1_MSS) were outstanding in popping value per hectare, while the native popcorn Criollo Plaza and the North American Jack Superior statistically obtained the lowest values (Table 2). The first one is due to its reduced capacity for popping, and the second is due to a reduced yield because of its lack of adaptation. Nevertheless, they far exceed the values reported by [20], who evaluated accessions of the Mexican popcorn race Palomero Toluqueño and obtained an average popcorn volume per ha of 8.86 m³ ha⁻¹.

When it comes to quality, it is well known that the physical and biochemical characteristics of popcorn kernels vary depending on genetics, the growing environment and the agronomic practices used [24]. Ref. [25] considered that the phenotypic traits of flake characteristics such as taste, texture, color and shape can be determinants of popcorn quality. Quality includes physical attributes such as shape, density, size, hardness and grain pericarp thickness, as well as compositional attributes, such as the levels of zein protein and types of fatty acids in grain [26]. All these traits are influenced by environmental conditions. The proportion of unpopped kernels is also considered as a component of quality. Refs. [27,28] considered that most unpopped kernels are not inherently unpoppable, but that not all meet the minimum thermodynamic requirements for popping during heating, or do not meet individually the optimum moisture content for producing a burst.

Commonly, breeding programs in popcorn focus on the development of cultivars that have good agronomic characteristics, with high yields and expansion volumes [29]. Both traits are quantitative in nature and of special interest for this crop [30]. The success of the improvement methods depends on the selected parents [31] and the heritability of the traits to be improved [32]. Grain yield has a heritability of 18.7% [3] and the expansion volume as high as 72%, determined by four genomic regions [25]. In addition to the above, the genetic variances of the population are a key element that, together with the method applied, determines the magnitude of the response to selection.

In order to increase grain yield and popping expansion in popcorn, it is possible to apply different breeding methods. For example, [33], through full-sib families in three selection cycles and estimating a fourth cycle through linear regression, reported genetic gains in grain yield of 0.264 t ha⁻¹ and 2.01 cm³ g⁻¹ in expansion volume per selection cycle.

Mass selection is a relatively easy and quick method to execute, since it can be carried out in one agricultural cycle and does not require controlled pollinations, compared to other methodologies where it is necessary to form families, evaluate them and recombine them; however, genetic advances may be smaller. Based on their results on genetic gain in maize [34], mentioned that mass selection by prolificacy is efficient as an indirect selection index to increase grain yield when carried out in a favorable environment which maximizes the expression of prolificacy. It is possible to include this character in the selection criteria, given that prolificacy is a highly expressed attribute in popcorn germplasm.

The difference in the response to selection for grain yield between MSS and MNSS (simple, non-stratified) selection lies in the heritability of the character. In the stratified selection plot, a lower environmental variance is expected within selection sub-plots, so the phenotypic value of the individuals is mostly caused by their genetic value, which is heritable. The opposite occurs in the whole selection plot (non-stratified), where a higher environmental variance is expected [4]. There is a risk of selecting individuals with desirable phenotypes, which are mostly generated by favorable environmental factors, which are not heritable. Regarding the similarity in the response to selection for expansion volume obtained with MSS and MNSS (simple selection), this is due to the high heritability of the character, which although it is a quantitative trait is determined by few pairs of genes and is little influenced by environmental factors [35]. This allows the selection of individuals with high expansion volumes, with greater certainty that they will be inherited.

The results obtained in our experiments with non-stratified selection, which was the method used to produce C2_MNSS, were satisfactory for expansion volume, but not for grain yield for the reasons explained above. Based on the progress made with mass-stratified selection, it is expected that with two additional cycles of selection, the resulting population will present acceptable grain yields and expansion volumes. This will allow the release of the first Mexican variety of yellow popcorn that reaches expansion volumes that meet the market standard (>40 cm³ g⁻¹) by taking advantage of additive variance. Subsequently, it might be possible for the generation of inbred lines and the exploitation of dominance effects through hybridization to increase grain yield.

In the future emphasis should be placed on increasing the yield of popcorn varieties, as popcorn breeding is lagging behind field corn and currently yields are significantly lower. Another aspect that needs to be improved is the nutritional quality of the flakes, since it is a type of maize almost exclusively for humans. A third aspect to improve is the obtaining of varieties that pop with a lower amount of calorific energy, to make the process more friendly with the environment.

5. Conclusions

Stratified visual mass selection is far more efficient than non-stratified mass selection for genetic advancement in grain yield. Both methods are effective for genetic advancement in expansion volume in popcorn. The Mexican native popcorn produces an acceptable yield but low levels of popping expansion. Popcorn from the US corn belt possesses a high expansion volume, but under the Mexican highland conditions, the yield is low due to a lack of adaptation. The results show that it is possible to combine the favorable attributes of both groups of popcorn into new populations that excel in both aspects.