ISSN e:2007-4034 / ISSN print: 2007-4034

English | Español

     

 
 
 
 
 
 
 
 

Vol. 28, issue 2 Mayo/Agosto - 2022

ISSN: ppub: 1027-152X epub: 2007-4034

Scientific article

Morphological characterization of seed-donor Creole avocado trees from three areas in Colombia

http://dx.doi.org/10.5154/r.rchsh.2021.06.010

0000-0003-2450-0627 López-Galé, Yeison 1 * ; 0000-0002-5679-3885 Murcia-Riaño, Nubia 1 ; 0000-0002-6512-7307 Romero-Barrera, Yajaira 2 ; 0000-0002-8145-1764 Martínez, Mauricio Fernando 1

  • 1Corporación Colombiana de Investigación Agropecuaria - AGROSAVIA, Centro de Investigación Palmira. Diagonal a la intersección de la Carrera 36ª, calle 23, Palmira, Valle de Cauca, C. P. 763531, COLOMBIA.
  • 2Corporación Colombiana de Investigación Agropecuaria - AGROSAVIA, Centro de Investigación Tibaitatá. Vía a Mosquera km 14, Bogotá, C. P. 250047, COLOMBIA.

Corresponding author: ylopezg@agrosavia.co, tel. (+57) 310 424 19 24.

Received: May 25, 2021; Accepted: January 05, 2022

This is an open-access article distributed under the terms of the Creative Commons Attribution License view the permissions of this license

Abstract

Avocado plant propagation in Colombia is done by grafting commercial materials, such as 'Hass', 'Lorena', 'Choquette', 'Fuerte', 'Reed', and 'Trinidad', among others, onto rootstocks produced from sexual seed from Creole and 'Hass' trees. These seeds can be obtained in local markets or in avocado-producing areas located in contrasting agroecological regions without well-defined selection criteria. This work aimed to characterize the morphological variability of 80 seed-donor avocado (Persea americana Mill.) trees for rootstock production in three producing areas of Colombia. Thirty-nine morphological descriptors were evaluated for characterization, including qualitative and quantitative characters of plants, leaves, fruits and seeds. Multiple factorial analysis detected high morphological variability in fruit and seed characteristics in the three avocado-producing zones, while cluster analysis allowed the identification of three fully formed groups, so it was not possible to discriminate groups of trees by area of origin. With the results obtained, it was possible to determine a wide phenotypic divergence in the analyzed seed-donor trees, where 80 % of them presented outstanding physical characteristics of seeds for rootstock production.

KeywordsPersea americana Mill.; West Indian avocado; genetic variability; IPGRI descriptors

Introduction

Avocado (Persea americana Mill.) is a fruit species of Mesoamerican origin belonging to the family Lauraceae. This species exhibits high adaptive plasticity to different environments as a result of evolutionary and ecological processes (Barrientos-Priego, Muñoz-Pérez, Reyes-Alemán, Borys, & Martínez-Damián, 2007; Galindo-Tovar, Ogata-Aguilar, & Arzate-Fernández, 2008). To date, three subspecies or horticultural races have been recognized: the Mexican race (P. americana var. drymifolia), Guatemalan race (P. americana var. guatemalensis) and West Indian race (P. americana var. americana). Genetic diversity among these races has allowed, through hybridization and anthropic selection processes, the development of modern commercial varieties, which have adapted to specific areas with high yields (Knight & Campbell, 1999).

Avocados are naturally distributed from central Mexico to Peru, and are currently grown in more than 50 countries with tropical and subtropical climates around the world. Mexico is the leading producer (2,029,886 t·year-1), while the Dominican Republic is the country with the highest yields per unit area (43.7 t·ha-1) (FAOSTAT, 2017). In Colombia, production is estimated at 544,933 t·year-1 with an average yield of 9.77 t·ha-1, giving it the fifth highest production volume worldwide, with the departments of Antioquia, Bolívar, Caldas, Huila, Quindío, Risaralda, Santander, Tolima and Valle del Cauca contributing more than 80 % of the national production (Agronet, 2019).

Traditionally, the propagation of avocado plants for crops in Colombia is done by grafting commercial materials ('Hass', 'Lorena', 'Choquette', 'Fuerte', 'Reed', and 'Trinidad', among others) onto rootstocks produced from sexual seed (ungrafted) from Creole and 'Hass' trees (Bernal-Estrada & Díaz-Díez, 2020). Seeds are obtained from local markets, which generally do not know the origin of the material, or from producing areas located in contrasting agroecological regions without any clear selection criteria.

Easy-to-observe morphological characteristics with high discriminant action have been used to determine morphological variety in Creole avocado populations in different regions of the world (Acosta-Díaz, Hernández-Torres, & Almeyda-León, 2012; Acosta-Díaz, Hernández-Torres, & Almeyda-Leó, 2013; Gutiérrez-Díez et al., 2009; López-Guzmán et al., 2012; López-Guzmán et al., 2015; Montes-Hernández, de la Torre-Vizcaíno, Heredia-García, Hernández-Martínez, & Camarena-Hernández, 2017; Sánchez-Pérez, 1999). In Colombia, very little information is available on the diversity of Creole avocados and their effect on productive and phytosanitary behavior when used as rootstocks in commercial crops.

In studies by Cañas-Gutierrez, Galindo-López, Arango-Isaza, and Saldamando-Benjumea (2015) and Cañas-Gutiérrez, Arango-Isaza, and Saldamando-Benjumea (2019), the high morphological and molecular variability in seed-donor Creole avocados from different regions of Colombia is evident; in addition, they relate their genetic compatibility with commercial cultivars such as 'Hass', 'Fuerte' and 'Reed', when used as rootstocks. Burbano-Figueroa (2019) mentions some West Indian ecotypes for the Montes de María area (Bolívar and Sucre), identified by local producers according to external and internal characteristics of the fruit, and are known as 'Cebo', 'Leche' and 'Manteco'.

There has been little research on the identification of Creole rootstocks with tolerance or resistance to phytosanitary problems. Among these, the selections made by Rodriguez-Henao, Caicedo-Arana, Enriquez-Valencia, and Muñoz-Florez (2017) in the collection of the National Avocado Germplasm Bank located in the Palmira Research Center of AGROSAVIA (Valle del Cauca) stand out, mainly focused on the identification of materials with resistance to root rot caused by Phytophthora cinnamomi Rands.

Determining the morphological variability existing in the different avocado seed producing regions for rootstock production in Colombia would allow identifying selection attributes for the development of propagation and breeding programs, as well as identifying high diversity refuges that contribute to the implementation of conservation plans. In this sense, the objective of this work was to characterize the phenotypic variability of "Creole" avocado trees as seed donors for rootstock production in three areas of Colombia. This was done with the purpose of expanding knowledge to other regions of the country, which lack information, and providing a basis for the selection of seed-producing trees that meet the needs of nursery operators.

Materials and methods

The work was carried out during the period from March 2015 to November 2016 in three Creole avocado producing areas to obtain rootstocks in Colombia: Norcasia in the department of Caldas (NOR) (05° 36’ North latitude and 74° 52’ West longitude, at 550 m a. s. l.), Alvarado in the department of Tolima (ALV) (04° 37’ North latitude and 075° 01’ West longitude, at 1,250 m a. s. l.) and San José del Palmar in the department of Chocó (SJP) (04° 55’ North latitude and 76° 12’ West longitude, at 920 m a. s. l.).

In each study area, producing trees were selected considering the requirements established by the Colombian Agricultural Institute (ICA, 2009 and 2015), in standards 3180 of 2009 and 3168 of 2015 (modified by resolution: 0780006 of 2020). These standards describe the guidelines for the selection of seed-donor trees for rootstock production.

Each tree was georeferenced, marked and characterized by evaluating 39 morphological characteristics of plant, leaves, fruits and seeds according to the standardized descriptors published for avocado by the International Plant Genetic Resources Institute (IPGRI, 1995) (Table 1). Trees were selected during the main harvest periods to obtain information on fruits and seeds.

Table 1. Qualitative and quantitative descriptors used for the morphological characterization of seed-donor Creole avocado trees in three producing areas of Colombia.

Structure Characteristic Abbreviation
Tree Vigor TV
Shape TS
Branch distribution TBD
Height (m) TH
Trunk surface TTS
Trunk circumference at 30 cm above ground level (cm) TTC
Leaf Shape LS
Base shape LBS
Apex shape LAS
Margin LM
Groove on leaf petiole GLP
Color LC
Anise smell LAS
Length (cm) LL
Leaf area (cm2) LA
Fruit Shape FS
Weight (g) FW
Length (cm) FL
Diameter (cm) FD
Apex position FAP
Apex shape FAS
Pedicel position FPP
Skin color FSC
Skin surface FSS
Skin pliability FSP
Adherence of skin to flesh ASFF
Skin thickness (mm) FST
Skin weight (g) FSW
Flesh weight (g) FFW
Flesh color FFC
Flesh texture FFT
Seed Shape SS
Weight (g) SW
Length (cm) SL
Diameter (cm) SD
Cotyledon surface SCS
Cotyledon color SCC
Seed position SP
Free space of the seed cavity FSSC

The sample size used per tree was 15 fruits (with a healthy appearance at physiological maturity), 15 seeds and 5 leaves (fully developed and randomly collected). The morphological characterization of the trees was carried out in situ, while the characterization of leaves, fruits and seeds was done in the AGROSAVIA phytopathology laboratory. The morphometric variables of length and diameter were measured with a Mitutoyo digital caliper (±0.01 mm), and the weight variables with a Mettler Toledo balance (±0.01 g).

Data matrices were constructed with the data obtained and analyzed with descriptive and multivariate statistics. A mixed data matrix was created that included quantitative and qualitative morphological characteristics, on which a multiple factor analysis (MFA) was applied. This statistical method analyzes data tables, where individuals are described by groups of quantitative and qualitative variables, allowing their simultaneous analysis (Escofier & Pages, 1994; Abdi, Williams, & Valentin, 2013).

The MFA was performed in two stages: 1) principal component analysis (PCA) for quantitative variables and multiple correspondence analysis (MCA) for qualitative variables (ordinal or nominal variables separately), retaining in each procedure the highest eigenvalue, and 2) global PCA with the weights of the original variables obtained in the previous step (Escofier & Pages, 1994). As a result of this analysis, a linear combination of the first component of the PCA and the first dimension of the MCA was obtained, with which the scores for each individual were obtained.

With the scores obtained in the global PCA of the MFA, a cluster analysis was carried out using the squared Euclidean distance and Ward's minimum variance clustering method. The number of groups in the cluster was determined with Hotelling's pseudo t2 statistic. To determine the correct allocation of seed-donor trees, discriminant analyses were performed for each group. All analyses were conducted with the SAS ver. 9.4 statistical package (SAS Institute Inc., 2013).

Results and discussion

In the three areas evaluated, the avocado production system was characterized by using low production technology, being associated with other plant species of agricultural interest and having relicts of natural forests. Due to these characteristics, they can be considered as harvesting agroforestry production systems, which is consistent with the descriptions made by Burbano-Figueroa (2019) for the avocado-producing region of Montes de María located in the departments of Bolívar and Sucre (Colombia).

A total of 80 seed-donor trees that met the technical requirements established by the ICA (2009 and 2015) were characterized. Of these, 31 trees were selected in the producing area of Norcasia (Caldas), 28 trees in San José del Palmar (Chocó) and 21 trees in Alvarado (Tolima). These trees were characterized by being vigorous with a good phytosanitary appearance (pests and diseases), highly productive with uniform fruit and having had at least three consecutive production periods.

Two periods of fruit production or harvest were identified in the three areas. The first was between the months of March to June, as the main harvest, and the second period was between the months of October to December, with smaller productions known locally as "traviesa” or “mitaca".

The trees in the production area of Norcasia (Caldas) were characterized by variable size, reaching crown heights of up to 15 m and a maximum trunk circumference of 230 cm (Table 2); in addition, most of them were obovate in shape and had ascendant and irregular growth habits. The fruits had an average weight of 459.1 ± 99.9 g, and the fruit length/width ratio was 1.81, with a predominance of narrowly obovate, pyriform and clavate fruits, with a rounded or flattened apex, and asymmetrical pedicel. The skin surface was intermediate, green to dark green, and the fruit flesh was creamy green.

Table 2. Descriptive values of quantitative morphological variables of seed-donor Creole avocado trees in three producing areas in Colombia.

Variable Norcasia Alvarado San José del Palmar
Average SD Maximum Minimum CV Average SD Maximum Minimum CV Average SD Maximum Minimum CV
TH (m) 10.4 2.2 15.0 6.4 21.1 11.6 1.6 15.0 8.0 14.0 14.5 3.0 22.0 10.0 21.0
TTC (cm) 121.7 29.9 230.0 62.0 24.6 81.5 18.8 126.0 42.0 23.1 128.9 23.9 182.0 88.0 23.6
FS (g) 459.1 99.9 832.2 261.3 21.8 427.2 87.4 725.3 256.5 20.5 542.8 127.5 976.9 270.4 23.5
FL (cm) 15.6 2.9 29.2 10.5 18.3 16.0 2.6 24.8 9.5 16.0 15.8 2.8 23.1 10.1 17.8
FD (cm) 8.6 0.8 11.5 6.2 9.7 8.3 0.7 10.5 6.6 8.0 8.9 0.8 11.2 6.7 8.6
FST (mm) 1.2 0.3 2.8 0.5 23.9 1.0 0.2 1.9 0.5 22.1 1.1 0.3 2.2 0.4 28.0
FSW (g) 49.2 15.4 107.6 22.3 31.4 41.8 13.1 95.1 15.0 31.3 47.0 16.0 130.6 25.0 34.1
FPP (g) 302.0 84.6 587.2 136.8 28.0 267.3 70.7 486.5 33.5 26.5 364.6 103.7 761.1 34.5 28.4
SW (g) 65.4 15.8 126.9 29.7 24.2 87.1 21.3 156.3 32.5 24.4 107.3 35.6 220.6 21.1 33.2
SL (cm) 5.6 0.7 10.0 3.0 12.0 6.8 1.0 10.2 4.5 14.8 7.3 1.1 10.8 4.7 15.4
SD (cm) 4.6 0.5 6.2 3.3 10.1 5.0 0.5 7.2 3.8 10.8 5.3 0.7 9.0 3.0 14.1
LL (cm) 21.9 4.2 36.2 13.5 19.2 20.4 3.2 28.8 13.5 15.6 24.2 4.7 36.2 14.4 19.4
LA (cm2) 178.6 66.5 512.1 75.5 23.2 154.8 41.0 286.4 75.5 21.5 225.3 81.1 512.1 97.9 26.0
SD = standard deviation; CV = coefficient of variation.

Seeds in Norcasia had an average weight of 65.4 g, representing 14.2 % of total fruit weight. In this area, seeds were recorded as broadly ovate, cordiform, ellipsoid, with a flattened base and a conical and ovate apex. The cotyledon surface was intermediate rough, and creamy yellow or whitish. The position of the seed in the fruit was at the center or to one side of the fruit, and usually the testa or coat was strongly attached to the seed. The free space of the seed cavity was apical, and to a lesser extent absent.

In San José del Palmar (Chocó), trees with heights greater than 20 m and a trunk circumference of 128.9 cm were recorded. Most of the trees were vigorous, columnar in shape, with axial or irregular branch distribution. The fruits were characterized by being large, with an average weight of 542.8 g and a maximum weight of 976.9 g. The fruit length/width ratio was 1.77, and fruits were pyriform, ellipsoid, narrowly obovate, obovate, clavate or rhomboidal. Fruit apex and pedicel position were frequently asymmetrical. The surface of the fruit skin was intermediate or smooth, which varied from green to light green. The texture of the flesh was generally watery, and ivory or light green in color.

Seeds from the San José del Palmar producing area were large, with an average weight of 107.3 ± 35.6 g, representing 19.7 % of total fruit weight. The most frequent seed shapes were ovate, broadly ovate and a flattened base with conical apex. The cotyledon surface was mostly intermediate rough. The location of the seed in the fruit was central, and the free space of the seed cavity was apical.

In the Alvarado (Tolima) producing area, the trees were of variable size, with maximum crown heights of 15 m and an average trunk circumference of 81.5 cm. Most of them had a columnar or pyramidal crown, with an axial branch distribution habit. Fruits were characterized by being of medium size ( x - = 427.2 ± 87.4 g), and the fruit length/width ratio was 1.92. The most frequent fruit shape was pyriform, followed by narrowly obovate, clavate, ellipsoid, and obovate. Fruits had a rounded or asymmetrical apex, and the pedicel position was asymmetrical. The skin surface was intermediate, and light green to yellow. The texture of the flesh was buttery or pastose, with a good taste, and a creamy green to deep yellow color.

The average weight of seeds in Alvarado was 87.1 ± 21.3 g, representing 20.3 % of total fruit weight. In this area, seeds were predominantly broadly ovate, with a flattened base and a conical and ovate apex. They have cotyledons with an intermediate rough surface and detachable testa, centrally positioned and to one side of the fruit; it can be ivory, cream yellow or cream pink. The free space of the seed cavity was apical.

The quantitative morphological characters evaluated showed a wide range of variation in the three producing areas. In 58.3 % of the descriptors considered, coefficients of variation (CV) greater than 20 % were recorded (Table 2). According to Hidalgo (2003), attributes with a higher percentage may indicate high variability. The characteristics with the greatest variation in the three areas were FW, FPP, SW, FSW and FST.

According to the simple correlation analysis presented in Table 3, 16.6 % of the associations between quantitative variables were significant (P ≤ 0.05). Therefore, coefficients with significant values >0.40 were considered to represent associations with natural patterns of variation (Rojas, 2003). Among the variables with the highest correlation coefficients were: FW with FPP (r = 0.93), followed by SW with SD (r = 0.91), LA with LL (r = 0.88) and SW with SL (r = 0.70).

Table 3. Simple correlation matrix between 13 quantitative variables of 80 Creole avocado trees evaluated in three producing areas in Colombia.

TH TTC FW FL FD FST FSW FPP SW SL SD LL LA
TH 1.0000
TTC 0.5004* 1.0000
FW 0.3548 0.3010 1.0000
FL 0.1144 0.1161 0.5655 1.0000
FD 0.2458 0.1510 0.6834 -0.012 1.0000
FST -0.1282 -0.031 -0.207 -0.1406 -0.0233 1.0000
FSW 0.0205 0.1330 0.4571 0.5550 0.2611 0.4002 1.0000
FFW 0.3281 0.3392 0.9384 0.5465 0.5746 -0.3238 0.3419 1.0000
SW 0.4594 0.0722 0.3957 0.0410 0.2937 -0.0054 0.0546 0.1777 1.0000
SL 0.4437 0.0593 0.4900 0.4791 0.1095 -0.1212 0.1277 0.3794 0.7039 1.0000
SD 0.3790 0.0166 0.2415 -0.2252 0.3666 0.0459 -0.0577 0.0252 0.9114 0.4187 1.0000
LL 0.2339 0.0590 0.1152 -0.0809 0.0594 -0.1298 -0.0877 0.0926 0.3251 0.2418 0.2708 1.0000
LA 0.3445 0.0848 0.1149 -0.1539 0.0740 -0.1410 -0.1638 0.1073 0.4249 0.2526 0.3996 0.8889 1.0000
*Values in bold are statistically significant (P ≤ 0.05).

The morphological characteristics observed in the three producing areas are consistent with the descriptions made for fruits and seeds of avocados of the West Indian race (Bernal-Estrada & Díaz-Díez, 2020; Cañas-Gutiérrez et al., 2015; Sánchez-Pérez, 1999), and conform to the size and shape characteristics indicated by López-Guzmán et al. (2012), which are appreciated in the consumer market. FW ranged from 256.5 to 976.9 g, and the most common shapes were pyriform, obovate and clavate. Seed weights ranged from 29.7 to 220.6 g, and the predominant shapes were obovate and those with a flattened base and a conical apex.

SW was variable in the three areas, being higher in San José del Palmar (Chocó) (CV = 33.2 %), followed by Alvarado (Tolima) (CV = 24.4 %) and Norcasia (Caldas) (CV = 24.2 %). SD and the SL were characters of moderate variability, although positively correlated with SW. Seeds from the three study areas can be categorized as large, if compared to that cited by Acosta-Díaz et al. (2012) ( x - = 35.3 g), Acosta-Díaz et al. (2013) ( x - = 36.0 g) and Gutiérrez-Díez et al. (2009) (max. weight = 69.76 g) for Creole avocados from the state of Nuevo León (Mexico), and by Montes-Hernández et al. (2017) in Mexican avocado (P. americana var. drymifolia) germplasm from the Bajío Experimental Field in the state of Guanajuato, Mexico ( x - = 38.1 g).

The PCA generated in the MFA allowed determining the relationship between the quantitative variables evaluated and the avocado samples from the three producing areas. According to Table 4, and according to the Kaiser-Gutman criterion (Palacio, Apodaca, & Crisci, 2020), the first four components with eigenvalues greater than 1 were able to explain 92.2 % of the total variability. Table 5 shows the greatest contribution of variability explained by each variable that makes up the four principal components, where fruit and seed characteristics have the greatest weight in describing the level of variation in the three populations.

Table 4. Eigenvalues and proportion of variance explained in the principal component analysis of 80 seed-donor Creole avocado trees in Colombia.

Principal component Eigenvalue Difference Absolute proportion Cumulative proportion
1 3.62856192 1.56235761 0.4032 0.4032
2 2.06620431 0.66351465 0.2296 0.6328
3 1.40268966 0.19727081 0.1559 0.7886
4 1.20541885 0.86349048 0.1339 0.9225

Table 5. Eigenvectors of the original variables evaluated for the first four principal components of 80 seed-donor Creole avocado trees in Colombia.

Variable PC1 (40.3 %) PC2 (22.9 %) PC3 (15.5 %) PC4 (13.3 %)
TH 0.224401 0.233133 -0.312307 -0.102453
TTC 0.179247 0.098079 -0.182195 0.401073
FW 0.494885 -0.150025 -0.064279 -0.170517
FL 0.302770 -0.406090 0.024404 0.457855
FD 0.331871 0.051069 0.045659 -0.652333
FST -0.087553 0.045309 0.780370 -0.035549
FSW 0.245279 -0.314061 0.567957 0.027194
FFW 0.432256 -0.262600 -0.208337 -0.188831
SW 0.330509 0.508617 0.075274 0.181772
SL 0.365683 0.184871 -0.053928 0.511808
SD 0.235897 0.589458 0.098269 -0.055720
LL 0.199100 -0.156483 -0.259610 -0.145663
LA 0.223181 -0.409901 -0.232747 -0.078180
*The values in bold are those with the highest contribution in each component.

According to the MCA developed in the MFA for the qualitative variables, the first two dimensions were identified as those with the greatest contribution to the observed variability (41.04 %). Dimension 1 contributed 24.5 % of the variability, and dimension 2, 16.4 %. The characteristics with the highest polymorphism with respect to the highest inertia values were LS (0.2094), FS (0.1572), SS (0.0810) and SCC (0.0829) (Table 6). The characteristics TS, TBD, TTS, SP, FSSC, FSP, GLP, LAS and LAS had low polymorphism, so they were not included in the MCA.

Table 6. Partial contributions by each dimension with respect to the inertia value of the qualitative variables evaluated in 80 seed-donor Creole avocado trees in Colombia.

Variable Dim1 (24.5 %) Dim2 (16.4 %) Inertia
TV 0.0388 0.0029 0.0486
FS 0.5500 0.0002 0.1572
FAP 0.0033 0.0048 0.0304
FAS 0.0008 0.0318 0.0237
FPP 0.0108 0.0008 0.0223
FSC 0.0075 0.0714 0.0282
FSP 0.0002 0.0078 0.0310
ASFF 0.0107 0.0384 0.0530
FFC 0.0066 0.0005 0.0677
FSC 0.0040 0.0018 0.0160
FFT 0.0009 0.0996 0.0715
SS 0.0024 0.1419 0.0810
SCC 0.0303 0.1518 0.0829
SCC 0.0000 0.0210 0.0442
LS 0.8193 0.0162 0.2094
LBS 0.0138 0.0034 0.0193
LC 0.0004 0.0057 0.0087
*Values in bold are those with the highest contribution.

According to the descriptive and multivariate analyses, the characteristics that presented the greatest morphological variability were those related to fruits and seeds, which confirms that these traits constitute the most important morphological criteria to explain the phenotypic variability in the three avocado populations. López-Guzmán et al. (2012) reported similar results for Creole avocados from the state of Nayarit, and Gutiérrez-Díez et al. (2009) and Acosta-Díaz et al. (2012) for Creole avocados of the Mexican race in the state of Nuevo León (Mexico).

Avocado is a highly heterozygous cross-fertilizing species that produces fruits with monoembryonic seeds, so its progeny can be highly variable (Bernal-Estrada & Díaz-Díez, 2020; Alberti, Brogio, da Silva, Cantuarias-Avilés, & Fassio, 2018). In this sense, it is to be expected that, in the three agroecological regions evaluated, where repopulation processes occur spontaneously with little selection interference by producers, there is high morphological variability, especially in fruit and seed characteristics, which are those with the greatest natural variability (IPGRI, 1995).

Based on the global PCA scores, the cluster analysis obtained in the MFA (Escofier & Pages, 1994) allowed us to identify, at a cut-off height of 0.08 units (semi-partial R2), three groups fully formed from quantitative attributes and similar qualitative ones, which indicates that it was not possible to discriminate groups of trees by areas of origin (Figure 1). According to the discriminant analysis, no reclassification of trees between groups was observed, so it is assumed that the groupings formed are correct.

Figure 1. Dendrogram of quantitative and qualitative morphological data generated by global principal component analysis of 80 seed-donor Creole avocado trees in Colombia. NOR = Norcasia; ALV = Alvarado; SJP = San José del Palmar.

Group I was made up of 20 trees representing 25 % of the population, 14 of which belong to the producing center of San José del Palmar (SJP), four to Norcasia (NOR) and two to Alvarado (ALV). This group was characterized by having trees with higher FW, FPP, FD, SW, SL, SD, LL and LA. In addition, they are trees with elongated fruits (clavate), flesh with a pastose texture (doughy), skin with intermediate adherence to the flesh, ivory-colored seeds and narrowly obovate leaves (Table 7).

Table 7. Quantitative and qualitative characteristics determined by group (clusters) for 80 seed-donor Creole avocado trees in Colombia.

Descriptor Group I Group II Group III
Quantitative variables
FW (g) 577.3 458.6 404.8
FL (cm) 17.3 15.3 14.7
FD (cm) 8.9 8.6 8.1
FST (mm) 1.1 1.2 1.6
FSW (g) 51.2 45.4 42.9
FPP (g) 383.1 298.6 261.6
SW (g) 108.1 81.1 66.1
SL (cm) 7.5 6.2 5.7
SD (cm) 5.2 4.9 4.6
Qualitative variables
TV Strong Strong Intermediate
FS Clavate Pyriform Obovate
ASFF Intermediate Light Light
SS Broadly ovate Flattened base with conical apex Broadly ovate
SCC Ivory (white 155A) Cream (yellow 158B) Cream (yellow 158B)
LS Narrowly obovate Oval Oblong-lanceolate

Group II consisted of 44 trees (55 % population) with different centers of origin (NOR = 19 trees, ALV = 15 trees and SJP = 10 trees). This group included trees with intermediate quantitative characteristics (leaves, fruits and seeds), where the most determinant qualitative characters were trees with pyriform fruits, seeds with a flattened base and conical apex, and trees with oval leaves. In contrast, group III was made up of the smallest number of trees (16, 20 % population) (NOR = nine trees, ALV = five trees and SJP = two trees) and with the lowest quantitative characteristics, with the FST variable being the character with the highest weight. This group consisted of trees of intermediate vigor, with obovate fruits and oblong-lanceolate leaves.

Alberti et al. (2018) and Ramírez-Gill (2016) recommend that the selection of seed-donor trees be carried out based on vigor, productivity, adaptation to local climatic conditions and compatibility with the grafted variety. However, in Colombia, seed size and weight are important attributes for avocado rootstock production. This is due to the fact that nursery operators prefer large seeds for plant multiplication, as they consider that these induce better vigor and generate grafting diameters in less time, as confirmed by studies conducted by Adjei, Banful, and Idun (2011) and Ndoro, Anjichi, Letting, and Were (2018) in other regions of the world.

Seeds obtained from trees in the three evaluated areas, belonging to groups I and II (formed in the cladogram and corresponding to 80 % of the analyzed population), produced large and medium-sized seeds, which meet the physical characteristics required by nursery operators for rootstock production in nurseries in Colombia. However, when selecting seed-donor trees, additional aspects of the seed such as physiological, sanitary and genetic factors should be considered in order to determine their quality, affinity and productivity with commercial markets of interest for Colombia (Reyes-Herrera et al., 2020).

Conclusions

The high phenotypic divergence in fruits and seeds found in the three avocado populations implies a broad genetic base, in which each agroecological niche can harbor a high variability of shapes and sizes. This explains why it was not possible to identify characteristic variation patterns for each production center.

Eighty percent of the avocado trees identified as seed donors in the three producing areas presented outstanding physical characteristics for the production process of planting material (rootstocks), since in them it is possible to guarantee seeds of good weight and size for nursery activities.

References

Abdi, H., Williams, L., & Valentin, D. (2013). Multiple factor analysis: principal component analysis for multitable and multiblock data sets. Wiley Interdisciplinary reviews: computational statistics, 5(2), 149-179. doi: 10.1002/wics.1246

Acosta-Díaz, E., Hernández-Torres, I., & Almeyda-León, I. H. (2012). Evaluación de aguacates criollos en Nuevo León, México: región Sur. Revista Mexicana de Ciencias Agrícolas, 3(2), 245-257. doi: 10.29312/remexca.v3i2.1460

Acosta-Díaz, E., Hernández-Torres, I., & Almeyda-León, I. H. (2013). Evaluación de aguacates criollos en Nuevo León, México: región norte. Revista Mexicana de Ciencias Agrícolas , 4(4), 531-542. doi: 10.29312/remexca.v4i4.1184

Adjei, P., Banful, B., & Idun, I. (2011). Seed size and seed cut-length effects on germination behavior and seedling growth of avocado (Persea americana). International Journal of Agricultural Research, 6(3), 299-305. doi: 10.3923/ijar.2011.299.305

Agronet (2019). Sistema de Estadísticas Agropecuarias (SEA). Colombia: Agronet. Retrieved from http://www.agronet.gov.co/www/htm3b/ReportesAjax/VerReporte.aspx

Alberti, M., Brogio, B., da Silva, S., Cantuarias-Avilés, T., & Fassio, C. (2018). Avances en la propagación del aguacate. Revista Brasileira de Fruticultura, Jaboticabal, 40(6), 1-18. doi: 10.1590/0100-29452018782

Barrientos-Priego, A., Muñoz-Pérez, R., Reyes-Alemán, J., Borys, M., & Martínez-Damián, M. (2007). Taxonomía, cultivares y portainjertos. In: Téliz, D., & Mora, A. (Eds.), El aguacate y su manejo integrado (pp. 31-62). México, D. F.: Mundi Prensa.

Bernal-Estrada, J. A., & Díaz-Díez, C. A. (2020). Actualización tecnológica y buenas prácticas agrícolas (BPA) en el cultivo del aguacate. Mosquera, Colombia: Corporación Colombiana de Investigación Agropecuaria - AGROSAVIA. doi: 10.21930/agrosavia.manual.7403831

Burbano-Figueroa, O. (2019). West Indian avocado agroforestry systems in Montes de María (Colombia): a conceptual model of the production system.Revista Chapingo Serie Horticultura, 25(2), 75-102. doi: 10.5154/r.rchsh.2018.09.018

Cañas-Gutierrez, G., Galindo-López, L., Arango-Isaza, R., & Saldamando-Benjumea, C. (2015). Diversidad genética de cultivares de aguacate (Persea americana) en Antioquia, Colombia. Agronomía Mesoamericana, 26(1), 129-143. doi: 10.15517/am.v26i1.16936

Cañas-Gutiérrez, G., Arango-Isaza, R., & Saldamando-Benjumea, C. I. (2019). Microsatellites revealed genetic diversity and population structure in Colombian avocado (Persea americana Mill.) germplasm collection and its natural populations. Journal of Plant Breeding and Crop Science, 11(4), 106-109. doi: 10.5897/jpbcs2018.0792

Escofier, B., & Pages, J. (1994). Multiple factor analysis (AFMULT package). Computational statistics & data analysis, 18(1), 121-140. doi: 10.1016/0167-9473(94)90135-X

FAOSTAT (2017). Estadísticas de producción mundial de aguacate del año 2016. Retrieved from https://www.fao.org/faostat/es/#home

Galindo-Tovar, M., Ogata-Aguilar, N., & Arzate-Fernández, A. (2008). Some aspects of avocado (Persea americana Mill.) diversity and domestication in Mesoamerica. Genetic Resources Crop Evolution, 55, 441-450. doi: 10.1007/s10722-007-9250-5

Gutiérrez-Díez, A., Martínez-de la Cerda, J., García-Zambrano, E., Iracheta-Donjuan, L., Ocampo-Morales, J., & Cerda-Hurtado, I. (2009). Estudio de diversidad genética del aguacate nativo en Nuevo León, México. Revista Fitotecnia Mexicana, 32(1), 9-18. Retrieved from https://www.redalyc.org/articulo.oa?id=61011105002

Hidalgo, R. (2003). Variabilidad genética y caracterización de especies vegetales. In: Franco, T. L., & Hidalgo, R. (Eds), Análisis estadístico de datos de caracterización morfológica de recursos fitogenéticos (pp. 2-26). Colombia: Instituto Internacional de Recursos Fitogenéticos (IPGRI). Retrieved from http://www.bioversityinternational.org/fileadmin/_migrated/uploads/tx_news/An%C3%A1lisis_estad%C3%ADstico_de_datos_de_caracterizaci%C3%B3n_morfol%C3%B3gica_de_recursos_fitogen%C3%A9ticos_894.pdf

Instituto Colombiano Agropecuario (ICA). (2009). Resolución ICA 3180: Por medio de la cual se establecen los requisitos y procedimientos para la producción y distribución de material de propagación de frutales en el territorio nacional y se dictan otras disposiciones. Colombia. Retrieved from https://www.ica.gov.co/getattachment/Normatividad/Normas-Ica/Resoluciones-Oficinas-Nacionales/2020/2020R78006/2009R3180.pdf.aspx?lang=es-CO

Instituto Colombiano Agropecuario (ICA). (2015). Resolución ICA 3168: Por medio de la cual se reglamenta y controla la producción, importación y exportación de semillas producto del mejoramiento genético para la comercialización y siembra en el país, así como el registro de las unidades de evaluación agronómica y/o unidades de investigación en fitomejoramiento y se dictan otras disposiciones. Colombia. Retrieved from https://www.ica.gov.co/normatividad/normas-ica/resoluciones-oficinas-nacionales/2015/2015r3168.aspx

International Plant Genetic Resources Institute (IPGRI). (1995). Descriptores para aguacate (Persea spp.). Roma, Italia: International Plant Genetic Resources Institute. Retrieved from https://www.bioversityinternational.org/e-library/publications/detail/descriptors-for-avocado-persea-sppaguacate-persea-spp/

Knight, R., & Campbell, C. (1999). Ecological adaptation and the evolution of modern avocado cultivars. Revista Chapingo Serie Horticultura , 5, 49-54. doi: 10.5154/r.rchsh.1999.04.034

López-Guzmán, G., Medina-Torres, R., Guillen-Andrade, H., Ramírez-Guerrero, L. G., Aguilar-Castillo, J., & Valdivia, M. G. (2012). Características fenotípicas de hoja y fruto en selecciones de aguacate criollo de clima subtropical en el estado de Nayarit. Revista Fuente, 4, 56-62. Retrieved from http://dspace.uan.mx:8080/jspui/handle/123456789/801

López-Guzmán, G., Medina-Torres, R., Guillen-Andrade, H., Ramírez-Guerrero, L., Juárez- López, P., & Ruelas-Hernández, P. (2015). Caracterización morfológica en genotipos nativos de aguacate (Persea americana Mill.) de clima tropical en Nayarit, México. Revista Mexicana de Ciencias Agrícolas , 11, 2157-2167. doi: 10.29312/remexca.v0i11.785

Montes-Hernández, S., de la Torre-Vizcaíno, J., Heredia-García, E., Hernández-Martínez, M., & Camarena-Hernández, M. (2017). Caracterización morfológica de germoplasma de aguacate mexicano (Persea americana var. drymifolia, Lauraceae). Interciencia, 42(3), 175-180. https://www.interciencia.net/wp-content/uploads/2017/08/175-180-5976-MONTES-42_3.pdf

Ndoro, L., Anjichi, V., Letting, F., & Were, J. (2018). Effect of seed size on germination and seedling performance on grafted avocado. African Journal of Education, Science and Technology, 4(4), 95-100. Retrieved from https://www.researchgate.net/publication/331286694_Effect_of_Seed_Size_on_Germination_and_Seedling_Performance_on_Grafted_Avocado

Palacio, F., Apodaca, M., & Crisci, J. (2020). Análisis multivariado para datos biológicos: teoría y su aplicación utilizando el lenguaje R. Argentina, Buenos Aires: Fundación de Historia Natural Félix de Azara. Retrieved from https://fundacionazara.org.ar/analisis-multivariado-para-datos-biologicos/

Ramírez-Gil, J. (2016). Tratamientos pregerminativos y masa de la semilla como estrategia para mejorar la producción de plantas de aguacate raza guatemalteca. Cultivos Tropicales, 37(4), 115-125. doi: 10.13140/RG.2.2.31149.28648

Reyes-Herrera, P., Muñoz-Baena, L., Velásquez-Zapata, V., Patiño, L., Delgado-Paz, O., Díaz-Diez, C., Navas-Arboleda, A., & Cortés, A. (2020). Inheritance of rootstock effects in avocado (Persea americana Mill.) cv. Hass. Frontiers in Plant Science, 11, 555071. doi: 10.3389/fpls.2020.555071

Rodriguez-Henao, E., Caicedo-Arana, A., Enriquez-Valencia, A., & Muñoz-Florez, J. E. (2017). Evaluation of tolerance to Phytophthora cinnamomi Rands in avocado (Persea americana Mill.) germplasm. Acta Agronómica, 66(1), 128-134. doi: 10.15446/acag.v66n1.50705

Rojas, W. (2003). Análisis de la variabilidad genética en quinua. In: Franco, T. L., & Hidalgo, R. (Eds), Análisis estadístico de datos de caracterización morfológica de recursos fitogenéticos (pp. 27-39). Colombia: Instituto Internacional de Recursos Fitogenéticos (IPGRI) . Retrieved from http://www.bioversityinternational.org/fileadmin/_migrated/uploads/tx_news/An%C3%A1lisis_estad%C3%ADstico_de_datos_de_caracterizaci%C3%B3n_morfol%C3%B3gica_de_recursos_fitogen%C3%A9ticos_894.pdf

Sánchez-Pérez, J. (1999). Recursos genéticos de aguacate (Persea americana Mill.) y especies afines en México. Revista Chapingo Serie Horticultura , 5, 7-18. doi: 10.5154/r.rchsh.1999.01.010

SAS Institute Inc. (2013). Statistical Analysis System: User’s guide , ver. 9.4. Cary, N. Y., USA: Author. Retrieved from https://www.sas.com/en_us/home.html

Figures:

Figure 1. Dendrogram of quantitative and qualitative morphological data generated by global principal component analysis of 80 seed-donor Creole avocado trees in Colombia. NOR = Norcasia; ALV = Alvarado; SJP = San José del Palmar.

Tables:

Table 1. Qualitative and quantitative descriptors used for the morphological characterization of seed-donor Creole avocado trees in three producing areas of Colombia.
Structure Characteristic Abbreviation
Tree Vigor TV
Shape TS
Branch distribution TBD
Height (m) TH
Trunk surface TTS
Trunk circumference at 30 cm above ground level (cm) TTC
Leaf Shape LS
Base shape LBS
Apex shape LAS
Margin LM
Groove on leaf petiole GLP
Color LC
Anise smell LAS
Length (cm) LL
Leaf area (cm2) LA
Fruit Shape FS
Weight (g) FW
Length (cm) FL
Diameter (cm) FD
Apex position FAP
Apex shape FAS
Pedicel position FPP
Skin color FSC
Skin surface FSS
Skin pliability FSP
Adherence of skin to flesh ASFF
Skin thickness (mm) FST
Skin weight (g) FSW
Flesh weight (g) FFW
Flesh color FFC
Flesh texture FFT
Seed Shape SS
Weight (g) SW
Length (cm) SL
Diameter (cm) SD
Cotyledon surface SCS
Cotyledon color SCC
Seed position SP
Free space of the seed cavity FSSC
Table 2. Descriptive values of quantitative morphological variables of seed-donor Creole avocado trees in three producing areas in Colombia.
Variable Norcasia Alvarado San José del Palmar
Average SD Maximum Minimum CV Average SD Maximum Minimum CV Average SD Maximum Minimum CV
TH (m) 10.4 2.2 15.0 6.4 21.1 11.6 1.6 15.0 8.0 14.0 14.5 3.0 22.0 10.0 21.0
TTC (cm) 121.7 29.9 230.0 62.0 24.6 81.5 18.8 126.0 42.0 23.1 128.9 23.9 182.0 88.0 23.6
FS (g) 459.1 99.9 832.2 261.3 21.8 427.2 87.4 725.3 256.5 20.5 542.8 127.5 976.9 270.4 23.5
FL (cm) 15.6 2.9 29.2 10.5 18.3 16.0 2.6 24.8 9.5 16.0 15.8 2.8 23.1 10.1 17.8
FD (cm) 8.6 0.8 11.5 6.2 9.7 8.3 0.7 10.5 6.6 8.0 8.9 0.8 11.2 6.7 8.6
FST (mm) 1.2 0.3 2.8 0.5 23.9 1.0 0.2 1.9 0.5 22.1 1.1 0.3 2.2 0.4 28.0
FSW (g) 49.2 15.4 107.6 22.3 31.4 41.8 13.1 95.1 15.0 31.3 47.0 16.0 130.6 25.0 34.1
FPP (g) 302.0 84.6 587.2 136.8 28.0 267.3 70.7 486.5 33.5 26.5 364.6 103.7 761.1 34.5 28.4
SW (g) 65.4 15.8 126.9 29.7 24.2 87.1 21.3 156.3 32.5 24.4 107.3 35.6 220.6 21.1 33.2
SL (cm) 5.6 0.7 10.0 3.0 12.0 6.8 1.0 10.2 4.5 14.8 7.3 1.1 10.8 4.7 15.4
SD (cm) 4.6 0.5 6.2 3.3 10.1 5.0 0.5 7.2 3.8 10.8 5.3 0.7 9.0 3.0 14.1
LL (cm) 21.9 4.2 36.2 13.5 19.2 20.4 3.2 28.8 13.5 15.6 24.2 4.7 36.2 14.4 19.4
LA (cm2) 178.6 66.5 512.1 75.5 23.2 154.8 41.0 286.4 75.5 21.5 225.3 81.1 512.1 97.9 26.0
SD = standard deviation; CV = coefficient of variation.
Table 3. Simple correlation matrix between 13 quantitative variables of 80 Creole avocado trees evaluated in three producing areas in Colombia.
TH TTC FW FL FD FST FSW FPP SW SL SD LL LA
TH 1.0000
TTC 0.5004* 1.0000
FW 0.3548 0.3010 1.0000
FL 0.1144 0.1161 0.5655 1.0000
FD 0.2458 0.1510 0.6834 -0.012 1.0000
FST -0.1282 -0.031 -0.207 -0.1406 -0.0233 1.0000
FSW 0.0205 0.1330 0.4571 0.5550 0.2611 0.4002 1.0000
FFW 0.3281 0.3392 0.9384 0.5465 0.5746 -0.3238 0.3419 1.0000
SW 0.4594 0.0722 0.3957 0.0410 0.2937 -0.0054 0.0546 0.1777 1.0000
SL 0.4437 0.0593 0.4900 0.4791 0.1095 -0.1212 0.1277 0.3794 0.7039 1.0000
SD 0.3790 0.0166 0.2415 -0.2252 0.3666 0.0459 -0.0577 0.0252 0.9114 0.4187 1.0000
LL 0.2339 0.0590 0.1152 -0.0809 0.0594 -0.1298 -0.0877 0.0926 0.3251 0.2418 0.2708 1.0000
LA 0.3445 0.0848 0.1149 -0.1539 0.0740 -0.1410 -0.1638 0.1073 0.4249 0.2526 0.3996 0.8889 1.0000
*Values in bold are statistically significant (P ≤ 0.05).
Table 4. Eigenvalues and proportion of variance explained in the principal component analysis of 80 seed-donor Creole avocado trees in Colombia.
Principal component Eigenvalue Difference Absolute proportion Cumulative proportion
1 3.62856192 1.56235761 0.4032 0.4032
2 2.06620431 0.66351465 0.2296 0.6328
3 1.40268966 0.19727081 0.1559 0.7886
4 1.20541885 0.86349048 0.1339 0.9225
Table 5. Eigenvectors of the original variables evaluated for the first four principal components of 80 seed-donor Creole avocado trees in Colombia.
Variable PC1 (40.3 %) PC2 (22.9 %) PC3 (15.5 %) PC4 (13.3 %)
TH 0.224401 0.233133 -0.312307 -0.102453
TTC 0.179247 0.098079 -0.182195 0.401073
FW 0.494885 -0.150025 -0.064279 -0.170517
FL 0.302770 -0.406090 0.024404 0.457855
FD 0.331871 0.051069 0.045659 -0.652333
FST -0.087553 0.045309 0.780370 -0.035549
FSW 0.245279 -0.314061 0.567957 0.027194
FFW 0.432256 -0.262600 -0.208337 -0.188831
SW 0.330509 0.508617 0.075274 0.181772
SL 0.365683 0.184871 -0.053928 0.511808
SD 0.235897 0.589458 0.098269 -0.055720
LL 0.199100 -0.156483 -0.259610 -0.145663
LA 0.223181 -0.409901 -0.232747 -0.078180
*The values in bold are those with the highest contribution in each component.
Table 6. Partial contributions by each dimension with respect to the inertia value of the qualitative variables evaluated in 80 seed-donor Creole avocado trees in Colombia.
Variable Dim1 (24.5 %) Dim2 (16.4 %) Inertia
TV 0.0388 0.0029 0.0486
FS 0.5500 0.0002 0.1572
FAP 0.0033 0.0048 0.0304
FAS 0.0008 0.0318 0.0237
FPP 0.0108 0.0008 0.0223
FSC 0.0075 0.0714 0.0282
FSP 0.0002 0.0078 0.0310
ASFF 0.0107 0.0384 0.0530
FFC 0.0066 0.0005 0.0677
FSC 0.0040 0.0018 0.0160
FFT 0.0009 0.0996 0.0715
SS 0.0024 0.1419 0.0810
SCC 0.0303 0.1518 0.0829
SCC 0.0000 0.0210 0.0442
LS 0.8193 0.0162 0.2094
LBS 0.0138 0.0034 0.0193
LC 0.0004 0.0057 0.0087
*Values in bold are those with the highest contribution.
Table 7. Quantitative and qualitative characteristics determined by group (clusters) for 80 seed-donor Creole avocado trees in Colombia.
Descriptor Group I Group II Group III
Quantitative variables
FW (g) 577.3 458.6 404.8
FL (cm) 17.3 15.3 14.7
FD (cm) 8.9 8.6 8.1
FST (mm) 1.1 1.2 1.6
FSW (g) 51.2 45.4 42.9
FPP (g) 383.1 298.6 261.6
SW (g) 108.1 81.1 66.1
SL (cm) 7.5 6.2 5.7
SD (cm) 5.2 4.9 4.6
Qualitative variables
TV Strong Strong Intermediate
FS Clavate Pyriform Obovate
ASFF Intermediate Light Light
SS Broadly ovate Flattened base with conical apex Broadly ovate
SCC Ivory (white 155A) Cream (yellow 158B) Cream (yellow 158B)
LS Narrowly obovate Oval Oblong-lanceolate