- Email: [email protected]
Morphometry of the organs of cherimoya (Annona cherimola Mill.) and analysis of fruit parameters for the characterization of cultivars, and Mexican germplasm selections
Scientia Horticulturae 107 (2006) 337–346 www.elsevier.com/locate/scihorti
Morphometry of the organs of cherimoya (Annona cherimola Mill.) and analysis of fruit parameters for the characterization of cultivars, and Mexican germplasm selections ´ ngel c, A.F. Barrientos-Priego c J. Andre´s-Agustı´n a, F. Gonza´lez-Andre´s b,*, R. Nieto-A b
a CRUCO-Universidad Auto´noma Chapingo, Morelia, Michoaca´n, Mexico Departamento de Ingenierı´a Agraria, ESTIA, Universidad de Leo´n, Avda de Portugal 41, 24071 Leo´n, Spain c Departamento de Fitotecnia, Universidad Auto´noma Chapingo, Chapingo, Estado de Me´xico, Mexico
Received 15 November 2004; received in revised form 22 September 2005; accepted 16 November 2005
Abstract Three commercially available cherimoya cultivars (Campas, Burtons and White), one cultivar recently obtained and registered in Mexico (Corte´s-II-31) and seven Mexican germplasm selections (S-196, S-256, S-260, S-266, S-9651, S-Selene and S-Carapan) were characterized by using the morphometric traits of various organs of the adult plant, together with several agronomical and chemical characteristics. The objectives were: (i) to seek an alternative approach to the definition of cherimoya cultivars through multivariate analysis, using commercial varieties and Mexican germplasm selections and (ii) to elucidate the grouping of cultivars and selections obtained by multivariate analysis on the basis of their origin and geographical distribution. Plant material was collected in 2002 from adult plants 3 years after grafting. Twenty-one morphometric characteristics (seven of leaves, nine of flowers, two of fruits and three of seeds), plus five fruit characteristics of agronomical importance and three chemical parameters of the fruit were selected for characterizing accessions. The intra-accession variability recorded for the traits selected made them suitable for identifying cultivars. All of the traits but one were capable of showing up differences between accessions at a significance level of 0.001. Principal Component Analysis (PCA) showed that the traits which yielded the maximum separation between accessions were: leaf blade form factor, angle of the fifth vein of leaves, upper angle of leaves, area of the cross-section of petals, sepal maximum projected area, weight of fruits, total soluble solids in fruits, resistance of the skin of fruits to a penetrometer and width-to-length ratio of the maximum projected area of seeds. Consequently, all of these characteristics may be of interest as descriptors for cherimoya varieties. Furthermore, four consistent groups of accessions were defined by the three-dimensional plot obtained through projecting the accessions onto the first three principal components and the tree-diagram, or dendrogram, obtained through Cluster Analysis (CA). Two of these groups were made up of accessions with a well-defined common origin. The first consisted of selections S-196, S-256, S-260 and S-266 obtained from Coatepec Harinas in Mexico State, while the second comprised selections S-Selene and S-Carapan from Pure´pecha native Indian communities in Michoaca´n State. For the other two groups the origin of the accessions forming them is not fully known. Hence, further studies, based on molecular markers, might be carried out in order to ascertain if these accessions are genetically related. # 2005 Elsevier B.V. All rights reserved. Keywords: Cherimoya germplasm; Multivariate analysis; Principal component analysis; Cluster analysis; Diversity analysis; Descriptors; Phenotypic characterization; Cultivar identification; Genetic resources for fruit
1. Introduction The cherimoya (Annona cherimola Mill.) is a subtropical fruit tree, indigenous to Andean South America (Lee and Ellstrand, 1987), most probably originating in Ecuador and the neighbouring part of Peru (DeCandolle cited by Popenoe
* Corresponding author. Tel.: +34 987291833; fax: +34 987291810. E-mail address: [email protected] (F. Gonza´lez-Andre´s). 0304-4238/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.scienta.2005.11.003
(1921)). Nevertheless, the presence of the fruit in Mexico and Central America has led botanists to assume that it might also be indigenous to the latter countries (Popenoe, 1921). As for most of cultivated fruit–tree species, the identification of cherimoya cultivars is convoluted. Sometimes more than one genotype has been accidentally given the same cultivar name. Conversely, some cultivars have several synonymous names. The problem of identifying them is worsened by variation within cultivars, even when their name represents a single genotype (Ellstrand, 1997). In addition, the parentage of most
338
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
cultivars is unknown and their phylogenetic relationships are not clear (Rahman et al., 1998). Moreover, the diversity of cherimoya germplasm, estimated from genetic parameters, is high in comparison with the average diversity of cultivated plants (Pascual et al., 1993b). The germplasm diversity in a cultivated crop depends on the pollination system, the method of propagation and the process of domestication. The mating system of cherimoya exhibits the phenomenon of protogynous dichogamy, a condition in which the receptivity of stigmas occurs prior to the release of pollen from a given flower. This increases diversity within natural populations because of cross mating (Ellstrand, 1997; George, 1984). A very large number of cherimoya cultivars trace their origin to open pollination populations (Richardson and Anderson, 1996) and have thereafter been sexually and asexually propagated. Furthermore, very few breeding programmes have been carried out on the cherimoya (Pascual et al., 1993b; Andre´s-Agustı´n, 1997), with the development of cultivars based chiefly on selection. One particular problem is that classification of cherimoya cultivars has traditionally been based on characteristics of the surface of the fruit. These are in accordance with five botanical varieties, smooth, fingerprint, tuberculate, mamillate and umbonate, recorded long ago (Irazazabal, 1985), together with other pomological characteristics (Thomson, 1970). However, the surface and shape may vary substantially from fruit to fruit even on the same tree (Ellstrand and Lee, 1987). Thus, classification of cherimoya cultivars on the basis of skin type has been difficult (Perfectti et al., 1993; Rahman et al., 1998). In the light of this, there is a need for a reliable classification of cherimoya cultivars. Several approaches have been tried in recent years. These have involved isoenzymatic (Ellstrand and Lee, 1987; Perfectti et al., 1993; Pascual et al., 1993a,b; Perfectti and Pascual, 1998) or DNA-based molecular characterization: AFLP (Rahman et al., 1998), PCR-RFLP (Rahman et al., 1997), and RAPD (Ronning et al., 1995). Isoenzymatic and molecular characterization have the advantage that they are expressed in young material, making cultivar identification easy and reliable (Ellstrand and Lee, 1987). However, morphological characters are the most evident features. Hence, they are the basis for the description and identification of cultivars (Perrier, 1998), with crop descriptors mainly based upon morphology (i.e. IPGRI, 2004). In spite of the increasing interest in biochemical and molecular techniques for characterizing and identifying cultivars (Christie, 2001), it remains crucial to have adequate morphological traits. Morphological features have well-known problems for germplasm characterization. When used for assessing variability, they may lead to systematic errors. Several factors explain this. Chief among them are the polygenic inheritance of morphological features, their generally unknown hereditability (Perrier, 1998) and the influence of domestication on some of the morphological and agronomical parameters, which is a generalized phenomenon in cultivated species (Casas et al., 1999). In order to overcome these problems, morphological traits must fulfil certain requirements in order to be adequate for characterization (Perrier, 1998; Gonza´lez-Andre´s, 2001). In principle, the set of traits must be as extensive as possible and
drawn from various different organs of the plant. Other major considerations are (i) objectivity: qualitative characteristics need to have unambiguous and objective alternative expressions, whereas, quantitative characters derived from measurements are more objective in themselves; (ii) consistency: it is preferable for traits to be related to organs either unaffected or only slightly affected by selection and the environment. For metrical (or measurement-based) traits, ratios between measurements remain stabler than absolute measurements. In the present work morphometric characteristics of several organs of the adult plant (leaf, flower, fruit and seed) have been used, together with several agronomical and chemical traits of fruits, in order to characterize four commercial cherimoya cultivars, one of them obtained and registered in Mexico, and seven Mexican selections. The objectives were (i) to look for an alternative approach to the definition of cherimoya cultivars through multivariate analysis using commercial varieties and Mexican germplasm selections, and (ii) to explain the grouping of cultivars and selections obtained from multivariate analysis on the basis of their origin and geographical distribution. 2. Materials and methods The cherimoya cultivars and selections under investigation (Table 1) were three foreign cultivars widely cultivated in Mexico, one cultivar obtained and registered in Mexico in 2003 and seven germplasm selections developed in this same country. Mexican germplasm came from landraces. The plant material, leaves, flowers, fruits and seeds, was obtained in part from open-air experimental plots with a randomized complete block design at the Fundacio´n Salvador Sa´nchez Colı´n CICTAMEX, S.C., in Coatepec Harinas in Mexico State. This is located at latitude 188460 3800 N, longitude 998460 3800 W and at a height of 2240 m above sea level. It has an average annual rainfall of 1100 mm and an average temperature of 16 8C. Other locations yielding material were Jujacato, Tingambato and Carapan in Michoaca´n State, Mexico, lying at latitudes 198260 –198510 N, longitudes 1018490 –1028220 W and heights of 1900–2198 m above sea level. The average rainfall for these sites was between 1300 and 1800 mm yearly and their average temperature 19.9 8C. Plant material was collected in 2002 from adult plants 3 years after grafting. Ten different plants were selected per accession and for each specimen one single leaf and flower, plus three fruits, were selected on the basis of the criteria given below. For leaves, a healthy and fully developed organ located between the eighth and ninth node was collected. This was scanned and the digital photograph used for morphometric measurements. One healthy fully developed flower was randomly selected per plant. It was dissected into the following parts: petals, receptacle, pedicle and sepals. The polar projection (maximum projected area) of each of these pieces, plus the cross-section of the petal, were scanned and the digital photograph used for measurements. For fruits, three healthy well-developed fruits were collected per specimen at the ripening stage, from three different places on the plant: upper, intermediate and lower. After collection they were weighed,
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
339
Table 1 Selections and cultivars of cherimoya (Annona cherimola Mill.) studied Accession
Typea
County of registration for cultivars and reference for selections
Origin
Localization of the experimental plot in Mexico
Fruits surface characteristics
Campas Burtons White S-196 S-256 S-260 S-266 S-9651 Corte´s II 31 S-Selene S-Carapan
CC CC CC GS GS GS GS GS CC GS GS
Spain New Zealand USA Nicola´s-Cruz et Nicola´s-Cruz et Nicola´s-Cruz et Nicola´s-Cruz et Andre´s-Agustı´n Mexico Andre´s-Agustı´n Andre´s-Agustı´n
Unknown Unknown Unknown Mexico: Coatepec Harinas, Mexico Mexico: Coatepec Harinas, Mexico Mexico: Coatepec Harinas, Mexico Mexico: Coatepec Harinas, Mexico Mexico: Tingambato, Michoacan Mexico: Tancı´taro, Michoacan Mexico: Tingambato, Michoacan Mexico: Carapan, Michoacan
CICTAMEXb CICTAMEX CICTAMEX CICTAMEX CICTAMEX CICTAMEX CICTAMEX CICTAMEX Jujacato, Michoacan Tingambato, Michoacan Carapan, Michoacan
Umbonate Umbonate Fingerprint-tuberculate Mamillate Smooth Mamillate Umbonate Umbonate Mamillate Fingerprint Fingerprint
a b
al. (1996) al. (1996) al. (1996) al. (1996) (1997) (1997) (1997)
CC: Commercial cultivar; GS: germplasm selection. Fundacio´n Salvador Sa´nchez Colı´n-CICTAMEX. Coatepec Harinas, Mexico.
morphometric measurements were performed with a vernier calliper and skin resistance was estimated with a multipurpose penetrometer. Skin, pulp and seeds were then separated, the two former were weighed and the seeds were counted. Total soluble solids in the juice were estimated in terms of Brix degrees with a refractometer and the pH of the pulp was measured with a pHmeter. To determine the malic acid content (milliequivalent per
100 g), juice was extracted, appropriately diluted with distilled water, then titrated with 0.1 N NaOH solution, phenolphthalein being used as an indicator. Twenty seeds were randomly selected from each accession. They were scanned and the digital photograph used for measurements. The traits scored (Table 2 and Fig. 1) were selected in accordance with the general recommendations of Sneath and Sokal (1973), Perrier
Table 2 List of characters evaluated Organ
Variable
Abbreviation 2
Leaf
Blade area (cm ) Blade form factora Petiole length (cm) Fifth vein angle (8) Upper angle (8) Width length1 of the blade Number of secondary veins in the right side of the blade
LA LFF LPL LFVA LUA LW/L LNV
Flower
Petal maximum projected area (cm2) Width length1 of the petal maximum projected area Area of the petal cross-section (cm2) Form factor of the petal cross-sectiona Receptacle maximum projected area (cm2) Form factor of the receptacle maximum projected areaa Pedicele length (mm) Pedicele diameter (mm) Sepal maximum projected area (cm2)
PA PW/L PCS PCSFF RA RFF PdL PdD SpA
Fruit
Length of the polar axis (cm) Length of the polar axis (length of the ecuatorial axis)1 Fruit weight (g) Skin weight (g) Pulp percentage as: pulp weight (fruit weight)1 100 (%) Seed index: number of seeds (fruit weight)1 100 Skin resistance to penetrometer (kg cm2) Total soluble solids (8Brix) pH of the pulp Malic acid content (meq 100 g1)
FPA FPA/EA FW FSW FPP FSI FSR FTSS FpH FMA
Seed
Width length1 of the seed maximum projected area Maximum projected area form factora Weight of a single seed obtained from weight of 100 seeds (g)
SW/L SFF SW
a
[(4 p area) (perimeter)2].
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
340
Fig. 1. Characters measured in leaves, flowers and seeds of cherimoya (Annona cherimola Mill.). Character codes according to Table 2.
(1998) and Gonza´lez-Andre´s (2001) concerning independence, objectivity and consistency. All the measurements in digital photographs were obtained with the image analysis software Image Tool 3.00 (UTHSCSA, 2000). Prior to the multivariate analysis of results, some preliminary statistical analyses were performed. In order to estimate diversity within accessions, the coefficient of variation (CV, as a percentage), defined as the standard deviation mean value1 100, was calculated for each trait in each accession. Moreover, ANOVA was carried out for each character, so as to detect which traits showed up significant differences between accessions. For the multivariate analysis, a data matrix was prepared with mean values (Table 3) for every accession. Two different multivariate analyses were carried out: Principal Component Analysis (PCA) and Cluster Analysis (CA). The data matrix was standardized by subtracting the mean value of each trait and dividing the result by the standard deviation. For the PCA, a similarity matrix among characteristics was calculated based on the Pearson product-moment correlation. From this correlation matrix eigenvalue and eigenvector matrices were obtained, the accessions then being projected onto the first three principal components. For CA, the similarity matrix among accessions was calculated using the average taxonomic distance. The unweighted pair-group method arithmetic average (UPGMA) was used for clustering. The numerical analyses were carried out with the assistance of the software package NTSYS-pc version 2.1 (Rholf, 2000). 3. Results Table 3 shows the average values obtained after measurement of the parameters listed above in ten leaves and flowers,
30 fruits and 20 seeds per accession. The mean coefficient of variation (CV) for each characteristic (the mean of the CV for each of the 11 accessions studied) ranged from 2.5% for the pH of the fruit pulp (FpH) to 25.3% for the ratio between the length of the polar axis and length of the equatorial axis in fruits (FPA/ EA). The resistance of the skin of fruits to a penetrometer (FSR) showed a CV of 20.1%, the seed index (FSI) scored 17.3%, while the remainder of the traits showed values under 15%. Twenty-one out of the 29 traits showed a CV below 12% and 14 of less than 10%. The mean CV for each accession (the mean of the CVs for each of the 29 characteristics measured) ranged from 8.9% to 14.6%. Selections S-196, S-256, S-260 S-256, SSelene and S-Carapan showed the lowest intra-accession variability, with a mean CVof under 10%. Conversely, selection S-9651 and the commercial cultivars showed the highest values for mean CV. For all traits other than the form factor of the maximum projected area of the flower receptacle (RA), there were significant differences between accessions at a significance level of 0.001 (ANOVA analysis, data not shown). The PCA (Tables 4 and 5 and Fig. 2) showed that the first three components accounted for 25.33%, 22.97% and 16.59% of the variance respectively, their cumulative variance being 64.91%. On the basis of the eigenvector values for traits along the first three components (Table 4), the three attributes responsible for maximum separation along the first component (with values in parentheses) were the maximum projected area of the sepal, SpA (0.905), total soluble solids in fruit, FTSS (0.878), and weight of fruits, FW (0.705). Along the second component they were the width-to-length ratio for seeds, SW/L (0.793), the area of the petal cross-section, PCS (0.764), and the angle of the fifth vein of the leaf, LFVA (0.759), while along the third component they were the upper angle of leaf, LUA (0.808),
Table 3 Mean values for characters scored in leaves flowers, fruits and seeds of cherimoya (Annona cherimola Mill.)
PW/L
PCS
PCSFF
RA
RFF
PdL
PdD
SpA
Seeds Fruits
FSW
FPP
FSI
FSR
FTSS
FpH
FMA
SW/L
SFF
SW
0.76 0.68 0.68 0.70 0.74 0.73 0.78 0.71 0.74
1.77 2.13 1.66 1.49 1.76 1.38 1.51 2.06 1.91
67.60 55.86 64.27 66.53 68.97 63.88 66.60 59.13 65.12
92.98 85.19 87.02 99.62 105.25 94.92 105.58 83.65 99.87
0.59 0.58 0.66 0.66 0.69 0.62 0.70 0.64 0.59
13 13 14 14 13 15 13 13 14
2.51 1.05 1.58 1.55 1.25 2.21 1.68 1.35 2.02
0.29 0.29 0.25 0.29 0.33 0.25 0.28 0.30 0.25
23.5 8.1 11.1 17.6 14.6 21.7 16.7 10.8 19.3
0.54 0.69 0.58 0.67 0.69 0.66 0.69 0.69 0.69
0.24 0.17 0.21 0.23 0.21 0.22 0.26 0.20 0.22
0.63 0.70 0.79 0.76 0.73 0.73 0.65 0.77 0.66
1.38 1.00 1.32 1.02 1.13 0.92 1.08 1.06 1.60
0.25 0.19 0.22 0.26 0.21 0.24 0.22 0.23 0.27
0.65 0.35 0.42 0.37 0.34 0.45 0.39 0.31 0.51
14.5 12.5 12.5 11.7 12.5 11.2 14.0 12.6 14.3
1.21 1.20 1.21 1.10 1.17 1.04 1.19 1.29 1.38
1002.50 664.40 672.10 614.60 648.00 634.60 823.00 508.30 695.30
120.0 120.0 94.3 140.0 109.0 144.0 140.0 81.3 63.9
85.41 76.95 80.63 68.27 73.38 71.40 74.22 75.74 87.24
4.04 7.97 3.94 8.64 12.56 9.45 13.99 14.04 5.05
3.15 3.34 4.67 7.04 6.11 3.51 3.44 3.80 4.20
21.82 12.19 18.20 18.70 18.00 17.15 24.76 18.10 17.67
5.11 4.82 4.76 5.18 4.78 5.12 4.78 4.68 5.28
0.20 0.20 0.19 0.20 0.20 0.20 0.20 0.21 0.19
0.62 0.54 0.63 0.60 0.56 0.53 0.54 0.54 0.62
0.69 0.68 0.71 0.67 0.58 0.65 0.67 0.68 0.71
0.43 0.37 0.49 0.81 0.58 0.49 0.48 0.45 0.53 174.87 154.17
0.77 0.78
2.53 1.98
62.12 54.50
99.68 98.27
0.59 0.60
15 14
1.45 1.63
0.24 0.28
9.3 10.8
0.69 0.75
0.33 0.35
0.77 0.79
0.92 1.20
0.28 0.27
0.44 0.54
12.5 12.3
1.07 1.13
807.40 700.10
71.1 80.7
87.11 83.00
7.2 9.39
0.53 0.99
19.20 26.40
4.92 4.88
0.25 0.34
0.54 0.54
0.60 0.65
0.47 0.45
Character codes according to Table 2.
341
Fig. 3. Dendrogram obtained after cluster analysis of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.) (UPGMA method).
leaf blade form factor, LFF (0.709), and resistance of the fruit skin to a penetrometer, FSR (0.619). The three-dimensional grouping obtained after projection of the accessions onto the first three principal components (Fig. 2) defined four groups. The first was composed of cultivar Campas and cultivar CortesII-31, the second group, of the germplasm selections S-196, S256, S-260 and S-266. The third comprised the Burtons cultivar together with selection S-9653 and the fourth and last was made up of selections S-Selene and S-Carapan. The White cultivar showed an intermediate position between the second and the third group. On the plane defined by the first two principal components, White was closer to the second group. However, it differed from the rest of the accessions in this group on the basis of the third axis, consisting mainly of the upper angle of the leaf and the leaf blade form factor traits. The dendrogram obtained after CA (Fig. 3) defined four groups of accessions at a dissimilarity level of 1.23. These groups were consistent with the groups defined in the PCA analysis and White was included in the third group. Table 5 shows the mean values for every character in each of the four groups. Group 1 (Campas and Cortes-II-31) had the highest values for maximum projected area of petals (PA), area of the petal cross-section (PCS), pedicle length (PdL), maximum projected area of sepals (SpA), weight of fruits (FW), pulp percentage (FPP), length of the polar axis of fruits (FPA), the
Fig. 2. Scatter diagram obtained from the first three principal components after Principal Component Analysis (PCA) of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.).
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
128.19 153.11 191.83 104.80 103.46 88.10 115.41 117.64 148.96
Campas Burtons White S-196 S-256 S-260 S-266 S-9651 Corte´s II 31 S-Selene S-Carapan
PA
FW Sepal
FPA/EA Pedicel
FPA Receptacle Petal LNV LW/L LUA FFVA LPL LFF LA
Flowers Leaves Accession
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
342
Table 4 Eigenvector values on the first three axes after Principal Component Analysis (PCA) of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.) data Organ
Character
Code
1st Component
2nd Component
3rd Component
Leaf
Blade area Blade form factora Petiole length Fifth vein angle Upper angle Width length1 of the blade Number of veins in the right side of the blade
LA LFF LPL FFVA LUA LW/L LNV
0.475 0.493 0.263 0.013 0.088 0.588 0.168
0.432 0.244 0.713 0.759 0.083 0.338 0.392
0.564 0.709 0.359 0.368 0.808 0.330 0.200
Flower
Petal maximum projected area Width length1 of the petal maximum projected area Area of the petal cross-section Form factor of the petal cross-sectiona Receptacle maximum projected area Form factor of the receptacle maximum projected area Pedicele length Pedicele diameter Sepal maximum projected area
PA PW/L PCS PCSFF RA RFF PdL PdD SpA
0.664 0.552 0.307 0.314 0.477 0.351 0.640 0.609 0.905
0.495 0.198 0.764 0.634 0.619 0.628 0.427 0.298 0.139
0.320 0.078 0.452 0.251 0.563 0.082 0.297 0.434 0.194
Fruit
Length of the polar axis Length of the polar axis (length of the ecuatorial axis)1 Fruit weight Skin weight Pulp percentage as: pulp weight (fruit weight)1 100 Seed index: number of seeds (fruit weight)1 100 Skin resistance to penetrometer Total soluble solids pH of the pulp Malic acid content
FPA FPA/EA FW FSW FPP FSI FSR FTSS FpH FMA
0.543 0.613 0.705 0.686 0.446 0.687 0.392 0.878 0.242 0.477
0.715 0.391 0.208 0.199 0.521 0.199 0.145 0.283 0.318 0.419
0.052 0.125 0.260 0.374 0.383 0.374 0.619 0.272 0.585 0.268
Seed
Width length1 of the seed maximum projected area Maximum projected area form factora Weight of a single seed obtained from weight of 100 seeds
SW/L SFF SW
0.268 0.464 0.312
0.793 0.613 0.438
0.310 0.333 0.551
a
[(4 p area) (perimeter)2].
ratio between the length of the polar axis and length of the equatorial axis in fruits (FPA/EA), fruit pulp pH, the width-tolength ratio of the maximum projected area of seeds (SW/L) and the form factor of the maximum projected area of seeds (SFF). In contrast, Group 1 presented the lowest values for the width-to-length ratio of the leaf blade (LW/L), the form factor of the petal cross-section (PCSFF), the form factor of the maximum projected area of receptacles (RFF), the seed index (FSI) and malic acid content (FMA). Group 2 (selections S-266, S-256, S-260 and S-196) showed the highest values for the following characteristics: angle of the fifth vein of the leaf (LFVA), upper angle of leaf (LUA), width-to-length ratio of the leaf blade (LW/L), width-to-length ratio of the maximum projected area of petals (PW/L), seed index (FSI), fruit skin weight (FSW), fruit skin resistance to penetrometer (FSR) and seed weight (SW). The traits that yielded the lowest values for Group 2 were: leaf blade area (LA), petiole length (LPL), pedicle length (PdL), fruit pulp percentage (FPP) and length of the polar axis (FPA). Group 3 (the Burtons and White cultivars and selection S-9651) did not show maximum values for any of the characteristics, but conversely showed the lowest values for leaf blade form factor (LFF), upper angle of leaf (LUA), maximum projected area of petals (PA), area of the petal crosssection (PCS), maximum projected area of receptacles (RA), pedicle diameter (PdD), maximum projected area of sepals
(SpA), fruit weight (FW), total soluble solids in fruit pulp (FTSS), pulp pH and seed weight (SW). Group 4 (selections SSelene and S-Carapan) showed the highest values for leaf blade area (LA), leaf blade form factor (LFF), petiole length (LPL), number of veins on the right-hand side of the leaf blade (LNV), form factor of the petal cross-section (PCSFF), maximum projected area of receptacles (RA), form factor of the maximum projected area of receptacles (RFF), pedicle diameter (PdD), total soluble solids in fruit pulp (FTSS) and malic acid content (FMA). Conversely it showed the lowest values for the angle of the fifth vein of the leaf (LFVA), width-to-length ratio of the maximum projected area of petals (PW/L), fruit skin weight (FSW), fruit skin resistance to penetrometer (FSR), the ratio between the length of the polar axis and the length of the equatorial axis of fruits (FPA/EA), the width-to-length ratio of the maximum projected area of seeds (SW/L) and the maximum projected area form factor of seeds (SFF). 4. Discussion The grouping obtained through multivariate analysis is not consistent with the long-standing classification of cultivars into five botanical varieties based on characteristics of the fruit surface: smooth, fingerprint, tuberculate, mamillate and umbonate (Popenoe, 1921). In the scatter diagram obtained
0.48 0.59 0.44 0.46 0.70 0.64 0.69 0.63 0.62 0.56 0.57 0.54 0.19 0.20 0.20 0.29 5.20 4.97 4.75 4.90 19.75 19.65 16.16 22.80 3.68 5.03 3.94 0.76 4.55 11.16 8.65 8.30 86.33 71.82 77.77 85.06 92.17 133.26 98.44 75.91 848.90 680.05 614.93 753.75 1.30 1.13 1.23 1.10 14.37 12.35 12.53 12.38 0.58 0.39 0.36 0.49 0.26 0.23 0.21 0.27 1.49 1.04 1.13 1.06 0.65 0.72 0.75 0.78 0.23 0.23 0.19 0.34 0.62 0.68 0.65 0.72 21.40 17.66 10.02 10.06 0.27 0.29 0.28 0.26 2.26 1.67 1.32 1.54 13.33 13.67 13.33 14.50 0.59 0.67 0.63 0.60
Character codes according to Table 2.
96.42 101.34 85.29 98.98 66.36 66.49 59.75 58.31 1.84 1.54 1.95 2.26 138.58 102.94 154.20 164.52 1 2 3 4
0.75 0.74 0.69 0.78
SW/L FMA FpH FTSS FSR FSI FPP FSW FW FPA/EA FPA SpA PdD PdL RFF RA PCSFF PCS PW/L PA LNV LW/L LUA FFVA LPL LFF LA Group according to Fig. 3
Table 5 Mean values for every character, in each one of the groups of accessions defined after multivariate analysis (Figs. 2 and 3) of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.)
SFF
SW
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
343
after PCA analysis (Fig. 2) and in the dendrogram (Fig. 3), Group 1 included one accession with umbonate skin and another with mamillate skin, while Group 2 had two mamillate accessions, one umbonate and one smooth. Group 3 had two umbonate and one fingerprint-tuberculate, whereas Group 4 had two accessions with fingerprint skin. The grouping of accessions obtained by other authors after isoenzymatic (Perfectti et al., 1993) or DNA-based molecular characterization (Rahman et al., 1998) was similarly not consistent with botanical varieties. A classification of cultivars based on surface characteristics of the fruit has been extensively criticized (Irazazabal, 1985; Ellstrand and Lee, 1987; Perfectti et al., 1993; Rahman et al., 1998), because it is not natural and because of the lack of consistency of these characteristics within a cultivar or even a single plant. Currently, this approach has been rejected (Scheldeman et al., 1999) in favour of a classification based on multiple genetic and morphological traits, as proposed in this work. The characteristics chosen here fulfil the requirements recommended by Perrier (1998) and Gonza´lez-Andre´s (2001). Firstly, they relate to several different organs of the plants: seven to leaves, nine to flowers, 10 to fruits and three to seeds. They are objective, since all of them are quantitative and are determined by measurement. In respect of their consistency, two points should be considered. The first is stability when subjected to selection processes or environmental influences. More than half of the morphometric characteristics measured relate to flowers and seeds. In general, seeds (Eames, 1961) and flowers (Eames, 1961; Sherry and Lord, 1996) are the stablest organs within a given taxon, for example, in comparison with leaves. This has been specifically reported for the cherimoya (Perfectti and Camacho, 1999). The second important aspect about consistency is intra-accession variability. In spite of the great genetic variability noted in the cherimoya (Pascual et al., 1993b), bringing with it considerable variation within cultivars (Ellstrand, 1997), the intra-accession variability of the traits studied makes them suitable for cultivar identification. A possible exception would be of the trait of the ratio between the length of the polar axis and the equatorial axis in fruits, as this showed a mean variation coefficient of more than 25%. From a general point of view coefficients of variation of 12% or lower are acceptable in plant organ characterization. If this coefficient is higher, then it would be advisable to increase the sample size. According to Hidalgo (2003) a sample size of 25 should be enough to keep experimental error below 10% when the coefficient of variation is 25%. In this work, the sample size for fruits, the organs with the highest coefficient of variation for most of the traits, was 30. On the basis of the ANOVA results, 28 of the characteristics considered, in other words all except for the form factor of the maximum projected area of the flower receptacle, may be useful for identifying cultivars. However, PCA showed that the traits, which yielded the maximum separation between the accessions studied were three relating to leaves, two to flowers, three to fruits and one to seeds. The leaf traits were leaf blade form factor, angle of the fifth vein of leaves and upper angle of leaves. For flowers, the characteristics were area of the
344
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
cross-section of petals and maximum projected area of sepals. In the case of fruits, the traits were weight of fruits, total soluble solids in fruits and resistance of the skin of fruits to a penetrometer. The seed trait was the width-to-length ratio of the maximum projected area of seeds. Of these more discriminatory characteristics, leaf blade area and total soluble solids of fruit pulp are the two most likely to be affected by the environment. Cultivation of the material being characterized in homogeneous conditions of soil and climate would reduce this problem. Skin resistance to a penetrometer is the most variable trait within accessions (with a mean coefficient of variation of 20.1%). In order to use this characteristic for identifying cultivars, the number of fruits tested should be at least 16, in order to keep the error of the sample below 10% (Hidalgo, 2003). For the remainder of the characters, the coefficients of variation observed were at most 14.8% (weight of fruits), all the others being less. Eight out of the 28 traits in which ANOVA detected significant differences between accessions are included in the list of characteristics in the ‘‘Guidelines for the Conduct of Tests for Distinctness, Uniformity and Stability’’ for the cherimoya issued by the International Union for the Protection of New Varieties of Plants or UPOV (UPOV, 2003). The inclusion of the other characteristics into the UPOV’s guidelines, especially those classified as stronger discriminants in the PCA analysis, might be proposed. The characteristics in the UPOV guide are treated from a qualitative viewpoint, while in the present work they have been treated from a quantitative viewpoint. However, most of the UPOV traits are easily transformable into quantitative ones. Utilization of an inexpensive and easy-to-use modern image-analysis system, for instance Image Tool, would simplify the data collection process. The grouping of accessions emerging from the work reported here (Figs. 2 and 3) was based on a large number of characteristics from several different organs. Except for the well-known cultivars Campas, Burtons and White, the accessions were local selections and a new Mexican cultivar, which were being studied in depth for the first time. Pascual et al. (1993a) and Perfectti et al. (1993) analyzed isoenzymes from the cultivars Campas and White. The isoenzymatic results showed a great distance between the two varieties, which is consistent with the morphology-based results being reported in this paper. Group 2 consisted of selections S-196, S-256, S-260 and S-266. These are outstanding genotypes selected from a single segregated population of 420 individuals located at the Fundacio´n Salvador Sa´nchez Colı´n CICTAMEX, S.C., situated at Coatepec Harinas, in the State of Mexico (Nicola´s-Cruz et al., 1996). This population was grown from the seeds of fruits collected in the area surrounding Coatepec Harinas. These four genotypes were selected on the basis of local farmers’ criteria concerning fruit size and productivity (Rubı´-Arriaga et al., 1992; Nicola´s-Cruz et al., 1996). The main fruit characteristics of this group contrasting with the other groups (Table 5) are a high seed index, low pulp content and a relatively resistant skin. The first two of these characteristics are not much liked by consumers.
Group 3 consisted of the cultivars Burtons and White and selection S-9651. This latter was a garden tree grown from seed. It was picked out in Tingambato, a community located in the Michoaca´n State of Mexico. White is a cultivar bred and registered in the United States of America, while Burtons is a New Zealand cultivar. One possible hypothesis to explain this grouping would be that the three accessions have some common origin. The grounds for proposing such a hypothesis would be that the cherimoya was introduced into the United States from Mexico by Judge Robert B. Ord in 1871, with some of the finest selections, such as White, being sent to New Zealand and other countries (Schroeder, 1997). However, S9651 has the same geographical origin as S-Selene, but S-9651 is clearly different from S-Selene. It would be possible to use microsatellite data (Sefc et al., 2000) to discover whether the morphological, agronomical and chemical similarities of Burtons, White and S-9651 are due to a genetic relationship. If that were so, two alternative hypotheses could be considered. The first would be that White and Burtons came from Mexico and thus had the same origin as selection S-9651. The alternative hypothesis would be that S-9651 was introduced into Mexico from White or Burtons material. From the point of view of traits of the fruits, this group of accessions showed the lowest fruit pulp pH and total soluble solid content. Moreover, they have small flowers, which was reported by Richardson and Anderson (1996) as hindering hand pollination of the Burtons cultivar. The selections S-Selene and S-Carapan, both in Group 4, come from the Pure´pecha ethnic communities of Tingambato and Carapan in Michoaca´n State, these places being located at a distance of 30 km one from the other. Both selections are garden trees. On the basis of their phenotypical similarity and the strong cultural links between the two communities, the material could be related. From point of view of fruit characteristics, these two selections showed high total soluble solids and malic acid content, which is a positive feature for taste, and they are highly appreciated by consumers. However, they have the disadvantage of having the least resistant skin on their fruits, which leads to difficulties with handling after harvesting. The cultivar Cortes-II-31 comes from Michoaca´n State, like S-Selene and S-Carapan, but multivariate analysis located it in Group 1, alongside the Campas cultivar, grown in Spain. The separation between S-Selene and S-Carapan, on the one hand, and Corte´s-II-31, on the other, that can be observed in the scatter diagram from PCA is based on the second and third principal components. The characteristics responsible for the maximum separation along these components were the widthto-length ratio of the maximum projected area of seeds, the petal cross-section, the angle of the fifth vein of the leaf, the upper angle of the leaf, the leaf blade form factor and the resistance to a penetrometer of the skin of the fruit. Differences are very clear in petal cross-section, in skin resistance to the penetrometer and in the width-to-length ratio of the maximum projected area of seeds, these being much greater in CortesII-31 than in S-Selene and S-Carapan. Although the three accessions come from Michoaca´n state, S-Selene and
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
S-Carapan are from the Pure´pecha indigenous communities, whereas, the cultivar Cortes-II-31 comes from Tancı´taro, which is not an indigenous community but mixed race, with more outside influences. With respect to the similarity between Corte´s-II-31 and Campas, since the parentage of Campas is unknown, molecular markers might help ascertain whether there is a relationship between the two genotypes. Both cultivars share the fruit traits of high pulp percentage and pH content, with a low malic acid content. In conclusion, 20 of the 21 morphometric characteristics chosen from several different organs of the plants, together with the five fruit traits of agronomical importance and the three chemical parameters for fruits, were all found useful for classification of cherimoya germplasm. The characteristics yielding the maximum separation between accessions were three relating to leaves, two to flowers, three to fruits and one to seeds. Consequently, all these traits might be of interest as cherimoya descriptors, so that their inclusion in UPOV might be considered. Moreover, four consistent groups of accessions were defined. Two of them were made up of accessions with a well-defined common origin. The origin of the accessions in the other two groups is unknown, hence further DNA-based studies might be carried out in the future in order to ascertain the reasons for this grouping. Acknowledgements This work was financially supported by the Servicio Nacional de Inspeccio´n y Certificacio´n de Semillas (SNICS), the Consejo Nacional de Ciencia y Tecnologı´a (CONACYT), the Centro de Investigacio´n y Desarrollo del Estado de Michoaca´n (CIDEM) and the Direccio´n General de Investigacio´n y Posgrado of the Universidad Auto´noma Chapingo in Mexico. References Andre´s-Agustı´n, J., 1997. El cultivo del chirimoyo (Annona cherimola MIll.) en el Estado de Michoaca´n, Me´xico. Proc. Am. Soc. Trop. Hort. 41, 152– 161. Casas, A., Caballero, J., Valiente-Banuet, A., Soriano, J.A., Da´vila, P., 1999. Morphometrical variation and the process of domestication of Snetocereus stellatus (Cactaceae) in Central Me´xico. Am. J. Bot. 86 (4), 522–523. Christie, T.C., 2001. Use of molecular techniqus in the Distinctness, Uniformity and Stability (DUS) testing of Brassicas: UPOV Working Group on Biochemical and Molecular Techniques. In: Scottish Agricultural Science Agency Scientific Review 1997-2000. Scottish Agricultural Science Agency 6.2, 72–73. Eames, A.J., 1961. Morphology of the Angiosperms, vol. I. McGraw-Hill, New York. Ellstrand, N.C., 1997. Cultivars characteristics. In: California Cherimoya Association (Eds.), The Cherimoya Handbook. Universitiy of California, Department of Botany and Plant Sciences, Los Angeles, CA, (Chapter 3), pp. 2–9. Ellstrand, N.C., Lee, J.M., 1987. Cultivar identification of cherimoya (Annona cherimola Mill.) using isozymes markers. Sci. Horti. 32, 25–31. George, A.P., 1984. Annonaceae. In: Page, P.E. (Ed.), Tropical Tree Fruits for Australia. Queensland Department of Primary Industries, Brisbane, Australia, p. 3541. Gonza´lez-Andre´s, F., 2001. La Caracterizacio´n Morfolo´gica. In: Gonza´lezAndre´s, F., Pita-Villamil, J.M. (Eds.), Conservacio´n y Caracterizacio´n de
345
Recursos Fitogene´ticos. Publicaciones I.N.E.A. Valladolid, Espan˜a, pp. 199–217. Hidalgo, R., 2003. Variabilidad gene´tica y caracterizacio´n de especies vegetales. In Ana´lisis estadı´stico de datos de caracterizacio´n morfolo´gica en Recursos Fitogene´ticos. IPGRI, 2004. Descriptors lists. International Plant Genetic Resources Institute, Rome. Irazazabal, P., 1985. Descripcio´n pomolo´gica de una coleccio´n de cultivares de chirimoya (Annona cherimola Mill.) Estacio´n Experimental ‘‘La Mayora’’, Ma´laga. Lee, J.M., Ellstrand, N.C., 1987. Inheritance and linkage of isozymes in the cherimoya. J. Hered. 78, 83–387. Nicola´s-Cruz, M., Rubı´-Arriaga, M., Cruz-Castillo, J.G., 1996. Exploracio´n y coleccio´n de fenotipos sobresalientes de chirimoya. In: Proceedings of the Memoria 1996 on Fundacio´n Salvador Sa´nchez Colı´n-CICTAMEX, Coatepec, Harinas, Edo. de Me´xico, pp. 59–65. Pascual, L., Perfectti, F., Gutie´rrez, M., Vargas, A.M., 1993a. Characterising isozymes of Spanish chirimoya cultivars. HortScience 28 (8), 845– 847. Pascual, L., Perfectti, F., Garcı´a, F.J., Vargas, A., Gutierrez, M., 1993b. Identificacio´n de cultivares de chirimoyo mediante ana´lisis de isoenzimas. Actas del II Congreso Ibe´rico de Ciencias Hortı´colas. Coleccio´n Monografı´as INIA 85, 858–863. Perfectti, F., Camacho, J.P.M., 1999. Analysis of genotypic differences in developmental stability in Annona cherimola. Evolution 53 (5), 1396– 1405. Perfectti, F., Pascual, L., 1998. Characterization of cherimoya germplasm by isozyme markers. Fruit Var. J. 52 (1), 53–62. Perfectti, F., Pascual, L., Garcı´a, F.J., Vargas, A., Gutie´rrez, M., 1993. Relacio´n y agrupamiento de cultivares de chirimoyo en base al ana´lisis de marcadores gene´ticos. Actas del II Congreso Ibe´rico de Ciencias Hortı´colas. Coleccio´n Monografı´as INIA 85, 864–871. Perrier, X., 1998. Analyse de la diversite´ ge´ne´tique: Mesures de dissimilarite´ et repre´sentations arbore´es. The`se doctorat. Universite´ de Montpellier II. Montpellier Popenoe, W., 1921. The native home of the Cherimoya. J. Hered. 12 (7), 336– 339. Rahman, M.S.M., Yamada, M., Yoshida, M., 1997. Relationship of Annona species as revealed by PCR-RFLP analysis. Breed. Sci. 47 (4), 335– 339. Rahman, S.M., Shimada, T., Yamamoto, T., Yonemoto, J.Y., Yoshida, M., 1998. Genetic diversity of cherimoya cultivars revealed by amplified fragment length polymorphism (AFLP) analysis. Breed. Sci. 48, 5–10. Rholf, F.J., 2000. NTSYSpc. Numerical Taxonomy and Multivariate Analysys System. Version 2. 1. Exeter Software, New York. Richardson, A., Anderson, C.P.A., 1996. Hand pollination effects on the set and development of cherimoya (Annona cherimola) fruit in a humid climate. Sci. Horti. 65, 273–281. Ronning, C.M., Schnell, R.J., Gazit, S., 1995. Using randomly amplified polymorphic DNA (RAPD) markers to identify Annona cultivars. J. Am. Soc. Horti. Sci. 120 (5), 726–729. Rubı´-Arriaga, M., de la Cruz, T.E., Barrientos, P.F., 1992. Mejoramiento gene´tico de la chirimoya a trave´s de seleccio´n en poblaciones de segregantes. In: Proceedings of the Memoria Fundacio´n Salvador Sa´nchez Colı´n-CICTAMEX, S.C. Coatepec Harinas, Edo. de Me´xico, pp. 127– 134. Scheldeman, X., Uren˜a, V., Van Damme, P., 1999. Coleccio´n y caracterizacio´n de chirimoya (Annona cherimola Mill.) en la Provincia de Loja, Sur de Ecuador. Acta Horti. 497, 153–172. Schroeder, C.A., 1997. Botany. In: California Cherimoya Association (Eds.), The Cherimoya Handbook. University of California, Department of Botany and Plant Sciences, Los Angeles, CA, (Chapter 2), pp. 2–10. Sefc, K.M., Lopes, M.S., Lefort, F., Botta, R., Roubelakis-Angelakis, K.A., Ibanez, J., Pejic, I., Wagner, H.W., Glossl, J., Steinkellner, H., 2000. Microsatellite variability in grapevine cultivars from different European regions and evaluation of assignment testing to assess the geog– raphic origin of cultivars. Theor. Appl. Genet. 100 (3/4), 498– 505.
346
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
Sherry, R.A., Lord, E.M., 1996. Developmental stability in flowers of Clarkia tembloriensis (Onagraceae) as related to population outcrossing rates and heterozygosity. Evolution 50, 80–91. Sneath, P.H.A., Sokal, R.R., 1973. Numerical Taxonomy. The Principles and Practice of Numerical Classification. Freeman, San Francisco. Thomson, P.H., 1970. The cherimoya in California. Calif. Rare Fruit Growers Yearbook 2, 20–34.
UPOV, 2003. Cherimoya (Annona cherimola Mill.). Guidelines for the conduct of tests for Distinctness, Uniformity and Stability. International Union for the Protection of New Varieties of Plants, Geneva. UTHSCSA, University of Texas Health Science Center and San Antonio, 2000. Image Tool. IT Version 2.0. University of Texas, San Antonio, TX., http:// macorb.uthscsa.edu/dig/download.html.
Morphometry of the organs of cherimoya (Annona cherimola Mill.) and analysis of fruit parameters for the characterization of cultivars, and Mexican germplasm selections ´ ngel c, A.F. Barrientos-Priego c J. Andre´s-Agustı´n a, F. Gonza´lez-Andre´s b,*, R. Nieto-A b
a CRUCO-Universidad Auto´noma Chapingo, Morelia, Michoaca´n, Mexico Departamento de Ingenierı´a Agraria, ESTIA, Universidad de Leo´n, Avda de Portugal 41, 24071 Leo´n, Spain c Departamento de Fitotecnia, Universidad Auto´noma Chapingo, Chapingo, Estado de Me´xico, Mexico
Received 15 November 2004; received in revised form 22 September 2005; accepted 16 November 2005
Abstract Three commercially available cherimoya cultivars (Campas, Burtons and White), one cultivar recently obtained and registered in Mexico (Corte´s-II-31) and seven Mexican germplasm selections (S-196, S-256, S-260, S-266, S-9651, S-Selene and S-Carapan) were characterized by using the morphometric traits of various organs of the adult plant, together with several agronomical and chemical characteristics. The objectives were: (i) to seek an alternative approach to the definition of cherimoya cultivars through multivariate analysis, using commercial varieties and Mexican germplasm selections and (ii) to elucidate the grouping of cultivars and selections obtained by multivariate analysis on the basis of their origin and geographical distribution. Plant material was collected in 2002 from adult plants 3 years after grafting. Twenty-one morphometric characteristics (seven of leaves, nine of flowers, two of fruits and three of seeds), plus five fruit characteristics of agronomical importance and three chemical parameters of the fruit were selected for characterizing accessions. The intra-accession variability recorded for the traits selected made them suitable for identifying cultivars. All of the traits but one were capable of showing up differences between accessions at a significance level of 0.001. Principal Component Analysis (PCA) showed that the traits which yielded the maximum separation between accessions were: leaf blade form factor, angle of the fifth vein of leaves, upper angle of leaves, area of the cross-section of petals, sepal maximum projected area, weight of fruits, total soluble solids in fruits, resistance of the skin of fruits to a penetrometer and width-to-length ratio of the maximum projected area of seeds. Consequently, all of these characteristics may be of interest as descriptors for cherimoya varieties. Furthermore, four consistent groups of accessions were defined by the three-dimensional plot obtained through projecting the accessions onto the first three principal components and the tree-diagram, or dendrogram, obtained through Cluster Analysis (CA). Two of these groups were made up of accessions with a well-defined common origin. The first consisted of selections S-196, S-256, S-260 and S-266 obtained from Coatepec Harinas in Mexico State, while the second comprised selections S-Selene and S-Carapan from Pure´pecha native Indian communities in Michoaca´n State. For the other two groups the origin of the accessions forming them is not fully known. Hence, further studies, based on molecular markers, might be carried out in order to ascertain if these accessions are genetically related. # 2005 Elsevier B.V. All rights reserved. Keywords: Cherimoya germplasm; Multivariate analysis; Principal component analysis; Cluster analysis; Diversity analysis; Descriptors; Phenotypic characterization; Cultivar identification; Genetic resources for fruit
1. Introduction The cherimoya (Annona cherimola Mill.) is a subtropical fruit tree, indigenous to Andean South America (Lee and Ellstrand, 1987), most probably originating in Ecuador and the neighbouring part of Peru (DeCandolle cited by Popenoe
* Corresponding author. Tel.: +34 987291833; fax: +34 987291810. E-mail address: [email protected] (F. Gonza´lez-Andre´s). 0304-4238/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.scienta.2005.11.003
(1921)). Nevertheless, the presence of the fruit in Mexico and Central America has led botanists to assume that it might also be indigenous to the latter countries (Popenoe, 1921). As for most of cultivated fruit–tree species, the identification of cherimoya cultivars is convoluted. Sometimes more than one genotype has been accidentally given the same cultivar name. Conversely, some cultivars have several synonymous names. The problem of identifying them is worsened by variation within cultivars, even when their name represents a single genotype (Ellstrand, 1997). In addition, the parentage of most
338
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
cultivars is unknown and their phylogenetic relationships are not clear (Rahman et al., 1998). Moreover, the diversity of cherimoya germplasm, estimated from genetic parameters, is high in comparison with the average diversity of cultivated plants (Pascual et al., 1993b). The germplasm diversity in a cultivated crop depends on the pollination system, the method of propagation and the process of domestication. The mating system of cherimoya exhibits the phenomenon of protogynous dichogamy, a condition in which the receptivity of stigmas occurs prior to the release of pollen from a given flower. This increases diversity within natural populations because of cross mating (Ellstrand, 1997; George, 1984). A very large number of cherimoya cultivars trace their origin to open pollination populations (Richardson and Anderson, 1996) and have thereafter been sexually and asexually propagated. Furthermore, very few breeding programmes have been carried out on the cherimoya (Pascual et al., 1993b; Andre´s-Agustı´n, 1997), with the development of cultivars based chiefly on selection. One particular problem is that classification of cherimoya cultivars has traditionally been based on characteristics of the surface of the fruit. These are in accordance with five botanical varieties, smooth, fingerprint, tuberculate, mamillate and umbonate, recorded long ago (Irazazabal, 1985), together with other pomological characteristics (Thomson, 1970). However, the surface and shape may vary substantially from fruit to fruit even on the same tree (Ellstrand and Lee, 1987). Thus, classification of cherimoya cultivars on the basis of skin type has been difficult (Perfectti et al., 1993; Rahman et al., 1998). In the light of this, there is a need for a reliable classification of cherimoya cultivars. Several approaches have been tried in recent years. These have involved isoenzymatic (Ellstrand and Lee, 1987; Perfectti et al., 1993; Pascual et al., 1993a,b; Perfectti and Pascual, 1998) or DNA-based molecular characterization: AFLP (Rahman et al., 1998), PCR-RFLP (Rahman et al., 1997), and RAPD (Ronning et al., 1995). Isoenzymatic and molecular characterization have the advantage that they are expressed in young material, making cultivar identification easy and reliable (Ellstrand and Lee, 1987). However, morphological characters are the most evident features. Hence, they are the basis for the description and identification of cultivars (Perrier, 1998), with crop descriptors mainly based upon morphology (i.e. IPGRI, 2004). In spite of the increasing interest in biochemical and molecular techniques for characterizing and identifying cultivars (Christie, 2001), it remains crucial to have adequate morphological traits. Morphological features have well-known problems for germplasm characterization. When used for assessing variability, they may lead to systematic errors. Several factors explain this. Chief among them are the polygenic inheritance of morphological features, their generally unknown hereditability (Perrier, 1998) and the influence of domestication on some of the morphological and agronomical parameters, which is a generalized phenomenon in cultivated species (Casas et al., 1999). In order to overcome these problems, morphological traits must fulfil certain requirements in order to be adequate for characterization (Perrier, 1998; Gonza´lez-Andre´s, 2001). In principle, the set of traits must be as extensive as possible and
drawn from various different organs of the plant. Other major considerations are (i) objectivity: qualitative characteristics need to have unambiguous and objective alternative expressions, whereas, quantitative characters derived from measurements are more objective in themselves; (ii) consistency: it is preferable for traits to be related to organs either unaffected or only slightly affected by selection and the environment. For metrical (or measurement-based) traits, ratios between measurements remain stabler than absolute measurements. In the present work morphometric characteristics of several organs of the adult plant (leaf, flower, fruit and seed) have been used, together with several agronomical and chemical traits of fruits, in order to characterize four commercial cherimoya cultivars, one of them obtained and registered in Mexico, and seven Mexican selections. The objectives were (i) to look for an alternative approach to the definition of cherimoya cultivars through multivariate analysis using commercial varieties and Mexican germplasm selections, and (ii) to explain the grouping of cultivars and selections obtained from multivariate analysis on the basis of their origin and geographical distribution. 2. Materials and methods The cherimoya cultivars and selections under investigation (Table 1) were three foreign cultivars widely cultivated in Mexico, one cultivar obtained and registered in Mexico in 2003 and seven germplasm selections developed in this same country. Mexican germplasm came from landraces. The plant material, leaves, flowers, fruits and seeds, was obtained in part from open-air experimental plots with a randomized complete block design at the Fundacio´n Salvador Sa´nchez Colı´n CICTAMEX, S.C., in Coatepec Harinas in Mexico State. This is located at latitude 188460 3800 N, longitude 998460 3800 W and at a height of 2240 m above sea level. It has an average annual rainfall of 1100 mm and an average temperature of 16 8C. Other locations yielding material were Jujacato, Tingambato and Carapan in Michoaca´n State, Mexico, lying at latitudes 198260 –198510 N, longitudes 1018490 –1028220 W and heights of 1900–2198 m above sea level. The average rainfall for these sites was between 1300 and 1800 mm yearly and their average temperature 19.9 8C. Plant material was collected in 2002 from adult plants 3 years after grafting. Ten different plants were selected per accession and for each specimen one single leaf and flower, plus three fruits, were selected on the basis of the criteria given below. For leaves, a healthy and fully developed organ located between the eighth and ninth node was collected. This was scanned and the digital photograph used for morphometric measurements. One healthy fully developed flower was randomly selected per plant. It was dissected into the following parts: petals, receptacle, pedicle and sepals. The polar projection (maximum projected area) of each of these pieces, plus the cross-section of the petal, were scanned and the digital photograph used for measurements. For fruits, three healthy well-developed fruits were collected per specimen at the ripening stage, from three different places on the plant: upper, intermediate and lower. After collection they were weighed,
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
339
Table 1 Selections and cultivars of cherimoya (Annona cherimola Mill.) studied Accession
Typea
County of registration for cultivars and reference for selections
Origin
Localization of the experimental plot in Mexico
Fruits surface characteristics
Campas Burtons White S-196 S-256 S-260 S-266 S-9651 Corte´s II 31 S-Selene S-Carapan
CC CC CC GS GS GS GS GS CC GS GS
Spain New Zealand USA Nicola´s-Cruz et Nicola´s-Cruz et Nicola´s-Cruz et Nicola´s-Cruz et Andre´s-Agustı´n Mexico Andre´s-Agustı´n Andre´s-Agustı´n
Unknown Unknown Unknown Mexico: Coatepec Harinas, Mexico Mexico: Coatepec Harinas, Mexico Mexico: Coatepec Harinas, Mexico Mexico: Coatepec Harinas, Mexico Mexico: Tingambato, Michoacan Mexico: Tancı´taro, Michoacan Mexico: Tingambato, Michoacan Mexico: Carapan, Michoacan
CICTAMEXb CICTAMEX CICTAMEX CICTAMEX CICTAMEX CICTAMEX CICTAMEX CICTAMEX Jujacato, Michoacan Tingambato, Michoacan Carapan, Michoacan
Umbonate Umbonate Fingerprint-tuberculate Mamillate Smooth Mamillate Umbonate Umbonate Mamillate Fingerprint Fingerprint
a b
al. (1996) al. (1996) al. (1996) al. (1996) (1997) (1997) (1997)
CC: Commercial cultivar; GS: germplasm selection. Fundacio´n Salvador Sa´nchez Colı´n-CICTAMEX. Coatepec Harinas, Mexico.
morphometric measurements were performed with a vernier calliper and skin resistance was estimated with a multipurpose penetrometer. Skin, pulp and seeds were then separated, the two former were weighed and the seeds were counted. Total soluble solids in the juice were estimated in terms of Brix degrees with a refractometer and the pH of the pulp was measured with a pHmeter. To determine the malic acid content (milliequivalent per
100 g), juice was extracted, appropriately diluted with distilled water, then titrated with 0.1 N NaOH solution, phenolphthalein being used as an indicator. Twenty seeds were randomly selected from each accession. They were scanned and the digital photograph used for measurements. The traits scored (Table 2 and Fig. 1) were selected in accordance with the general recommendations of Sneath and Sokal (1973), Perrier
Table 2 List of characters evaluated Organ
Variable
Abbreviation 2
Leaf
Blade area (cm ) Blade form factora Petiole length (cm) Fifth vein angle (8) Upper angle (8) Width length1 of the blade Number of secondary veins in the right side of the blade
LA LFF LPL LFVA LUA LW/L LNV
Flower
Petal maximum projected area (cm2) Width length1 of the petal maximum projected area Area of the petal cross-section (cm2) Form factor of the petal cross-sectiona Receptacle maximum projected area (cm2) Form factor of the receptacle maximum projected areaa Pedicele length (mm) Pedicele diameter (mm) Sepal maximum projected area (cm2)
PA PW/L PCS PCSFF RA RFF PdL PdD SpA
Fruit
Length of the polar axis (cm) Length of the polar axis (length of the ecuatorial axis)1 Fruit weight (g) Skin weight (g) Pulp percentage as: pulp weight (fruit weight)1 100 (%) Seed index: number of seeds (fruit weight)1 100 Skin resistance to penetrometer (kg cm2) Total soluble solids (8Brix) pH of the pulp Malic acid content (meq 100 g1)
FPA FPA/EA FW FSW FPP FSI FSR FTSS FpH FMA
Seed
Width length1 of the seed maximum projected area Maximum projected area form factora Weight of a single seed obtained from weight of 100 seeds (g)
SW/L SFF SW
a
[(4 p area) (perimeter)2].
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
340
Fig. 1. Characters measured in leaves, flowers and seeds of cherimoya (Annona cherimola Mill.). Character codes according to Table 2.
(1998) and Gonza´lez-Andre´s (2001) concerning independence, objectivity and consistency. All the measurements in digital photographs were obtained with the image analysis software Image Tool 3.00 (UTHSCSA, 2000). Prior to the multivariate analysis of results, some preliminary statistical analyses were performed. In order to estimate diversity within accessions, the coefficient of variation (CV, as a percentage), defined as the standard deviation mean value1 100, was calculated for each trait in each accession. Moreover, ANOVA was carried out for each character, so as to detect which traits showed up significant differences between accessions. For the multivariate analysis, a data matrix was prepared with mean values (Table 3) for every accession. Two different multivariate analyses were carried out: Principal Component Analysis (PCA) and Cluster Analysis (CA). The data matrix was standardized by subtracting the mean value of each trait and dividing the result by the standard deviation. For the PCA, a similarity matrix among characteristics was calculated based on the Pearson product-moment correlation. From this correlation matrix eigenvalue and eigenvector matrices were obtained, the accessions then being projected onto the first three principal components. For CA, the similarity matrix among accessions was calculated using the average taxonomic distance. The unweighted pair-group method arithmetic average (UPGMA) was used for clustering. The numerical analyses were carried out with the assistance of the software package NTSYS-pc version 2.1 (Rholf, 2000). 3. Results Table 3 shows the average values obtained after measurement of the parameters listed above in ten leaves and flowers,
30 fruits and 20 seeds per accession. The mean coefficient of variation (CV) for each characteristic (the mean of the CV for each of the 11 accessions studied) ranged from 2.5% for the pH of the fruit pulp (FpH) to 25.3% for the ratio between the length of the polar axis and length of the equatorial axis in fruits (FPA/ EA). The resistance of the skin of fruits to a penetrometer (FSR) showed a CV of 20.1%, the seed index (FSI) scored 17.3%, while the remainder of the traits showed values under 15%. Twenty-one out of the 29 traits showed a CV below 12% and 14 of less than 10%. The mean CV for each accession (the mean of the CVs for each of the 29 characteristics measured) ranged from 8.9% to 14.6%. Selections S-196, S-256, S-260 S-256, SSelene and S-Carapan showed the lowest intra-accession variability, with a mean CVof under 10%. Conversely, selection S-9651 and the commercial cultivars showed the highest values for mean CV. For all traits other than the form factor of the maximum projected area of the flower receptacle (RA), there were significant differences between accessions at a significance level of 0.001 (ANOVA analysis, data not shown). The PCA (Tables 4 and 5 and Fig. 2) showed that the first three components accounted for 25.33%, 22.97% and 16.59% of the variance respectively, their cumulative variance being 64.91%. On the basis of the eigenvector values for traits along the first three components (Table 4), the three attributes responsible for maximum separation along the first component (with values in parentheses) were the maximum projected area of the sepal, SpA (0.905), total soluble solids in fruit, FTSS (0.878), and weight of fruits, FW (0.705). Along the second component they were the width-to-length ratio for seeds, SW/L (0.793), the area of the petal cross-section, PCS (0.764), and the angle of the fifth vein of the leaf, LFVA (0.759), while along the third component they were the upper angle of leaf, LUA (0.808),
Table 3 Mean values for characters scored in leaves flowers, fruits and seeds of cherimoya (Annona cherimola Mill.)
PW/L
PCS
PCSFF
RA
RFF
PdL
PdD
SpA
Seeds Fruits
FSW
FPP
FSI
FSR
FTSS
FpH
FMA
SW/L
SFF
SW
0.76 0.68 0.68 0.70 0.74 0.73 0.78 0.71 0.74
1.77 2.13 1.66 1.49 1.76 1.38 1.51 2.06 1.91
67.60 55.86 64.27 66.53 68.97 63.88 66.60 59.13 65.12
92.98 85.19 87.02 99.62 105.25 94.92 105.58 83.65 99.87
0.59 0.58 0.66 0.66 0.69 0.62 0.70 0.64 0.59
13 13 14 14 13 15 13 13 14
2.51 1.05 1.58 1.55 1.25 2.21 1.68 1.35 2.02
0.29 0.29 0.25 0.29 0.33 0.25 0.28 0.30 0.25
23.5 8.1 11.1 17.6 14.6 21.7 16.7 10.8 19.3
0.54 0.69 0.58 0.67 0.69 0.66 0.69 0.69 0.69
0.24 0.17 0.21 0.23 0.21 0.22 0.26 0.20 0.22
0.63 0.70 0.79 0.76 0.73 0.73 0.65 0.77 0.66
1.38 1.00 1.32 1.02 1.13 0.92 1.08 1.06 1.60
0.25 0.19 0.22 0.26 0.21 0.24 0.22 0.23 0.27
0.65 0.35 0.42 0.37 0.34 0.45 0.39 0.31 0.51
14.5 12.5 12.5 11.7 12.5 11.2 14.0 12.6 14.3
1.21 1.20 1.21 1.10 1.17 1.04 1.19 1.29 1.38
1002.50 664.40 672.10 614.60 648.00 634.60 823.00 508.30 695.30
120.0 120.0 94.3 140.0 109.0 144.0 140.0 81.3 63.9
85.41 76.95 80.63 68.27 73.38 71.40 74.22 75.74 87.24
4.04 7.97 3.94 8.64 12.56 9.45 13.99 14.04 5.05
3.15 3.34 4.67 7.04 6.11 3.51 3.44 3.80 4.20
21.82 12.19 18.20 18.70 18.00 17.15 24.76 18.10 17.67
5.11 4.82 4.76 5.18 4.78 5.12 4.78 4.68 5.28
0.20 0.20 0.19 0.20 0.20 0.20 0.20 0.21 0.19
0.62 0.54 0.63 0.60 0.56 0.53 0.54 0.54 0.62
0.69 0.68 0.71 0.67 0.58 0.65 0.67 0.68 0.71
0.43 0.37 0.49 0.81 0.58 0.49 0.48 0.45 0.53 174.87 154.17
0.77 0.78
2.53 1.98
62.12 54.50
99.68 98.27
0.59 0.60
15 14
1.45 1.63
0.24 0.28
9.3 10.8
0.69 0.75
0.33 0.35
0.77 0.79
0.92 1.20
0.28 0.27
0.44 0.54
12.5 12.3
1.07 1.13
807.40 700.10
71.1 80.7
87.11 83.00
7.2 9.39
0.53 0.99
19.20 26.40
4.92 4.88
0.25 0.34
0.54 0.54
0.60 0.65
0.47 0.45
Character codes according to Table 2.
341
Fig. 3. Dendrogram obtained after cluster analysis of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.) (UPGMA method).
leaf blade form factor, LFF (0.709), and resistance of the fruit skin to a penetrometer, FSR (0.619). The three-dimensional grouping obtained after projection of the accessions onto the first three principal components (Fig. 2) defined four groups. The first was composed of cultivar Campas and cultivar CortesII-31, the second group, of the germplasm selections S-196, S256, S-260 and S-266. The third comprised the Burtons cultivar together with selection S-9653 and the fourth and last was made up of selections S-Selene and S-Carapan. The White cultivar showed an intermediate position between the second and the third group. On the plane defined by the first two principal components, White was closer to the second group. However, it differed from the rest of the accessions in this group on the basis of the third axis, consisting mainly of the upper angle of the leaf and the leaf blade form factor traits. The dendrogram obtained after CA (Fig. 3) defined four groups of accessions at a dissimilarity level of 1.23. These groups were consistent with the groups defined in the PCA analysis and White was included in the third group. Table 5 shows the mean values for every character in each of the four groups. Group 1 (Campas and Cortes-II-31) had the highest values for maximum projected area of petals (PA), area of the petal cross-section (PCS), pedicle length (PdL), maximum projected area of sepals (SpA), weight of fruits (FW), pulp percentage (FPP), length of the polar axis of fruits (FPA), the
Fig. 2. Scatter diagram obtained from the first three principal components after Principal Component Analysis (PCA) of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.).
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
128.19 153.11 191.83 104.80 103.46 88.10 115.41 117.64 148.96
Campas Burtons White S-196 S-256 S-260 S-266 S-9651 Corte´s II 31 S-Selene S-Carapan
PA
FW Sepal
FPA/EA Pedicel
FPA Receptacle Petal LNV LW/L LUA FFVA LPL LFF LA
Flowers Leaves Accession
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
342
Table 4 Eigenvector values on the first three axes after Principal Component Analysis (PCA) of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.) data Organ
Character
Code
1st Component
2nd Component
3rd Component
Leaf
Blade area Blade form factora Petiole length Fifth vein angle Upper angle Width length1 of the blade Number of veins in the right side of the blade
LA LFF LPL FFVA LUA LW/L LNV
0.475 0.493 0.263 0.013 0.088 0.588 0.168
0.432 0.244 0.713 0.759 0.083 0.338 0.392
0.564 0.709 0.359 0.368 0.808 0.330 0.200
Flower
Petal maximum projected area Width length1 of the petal maximum projected area Area of the petal cross-section Form factor of the petal cross-sectiona Receptacle maximum projected area Form factor of the receptacle maximum projected area Pedicele length Pedicele diameter Sepal maximum projected area
PA PW/L PCS PCSFF RA RFF PdL PdD SpA
0.664 0.552 0.307 0.314 0.477 0.351 0.640 0.609 0.905
0.495 0.198 0.764 0.634 0.619 0.628 0.427 0.298 0.139
0.320 0.078 0.452 0.251 0.563 0.082 0.297 0.434 0.194
Fruit
Length of the polar axis Length of the polar axis (length of the ecuatorial axis)1 Fruit weight Skin weight Pulp percentage as: pulp weight (fruit weight)1 100 Seed index: number of seeds (fruit weight)1 100 Skin resistance to penetrometer Total soluble solids pH of the pulp Malic acid content
FPA FPA/EA FW FSW FPP FSI FSR FTSS FpH FMA
0.543 0.613 0.705 0.686 0.446 0.687 0.392 0.878 0.242 0.477
0.715 0.391 0.208 0.199 0.521 0.199 0.145 0.283 0.318 0.419
0.052 0.125 0.260 0.374 0.383 0.374 0.619 0.272 0.585 0.268
Seed
Width length1 of the seed maximum projected area Maximum projected area form factora Weight of a single seed obtained from weight of 100 seeds
SW/L SFF SW
0.268 0.464 0.312
0.793 0.613 0.438
0.310 0.333 0.551
a
[(4 p area) (perimeter)2].
ratio between the length of the polar axis and length of the equatorial axis in fruits (FPA/EA), fruit pulp pH, the width-tolength ratio of the maximum projected area of seeds (SW/L) and the form factor of the maximum projected area of seeds (SFF). In contrast, Group 1 presented the lowest values for the width-to-length ratio of the leaf blade (LW/L), the form factor of the petal cross-section (PCSFF), the form factor of the maximum projected area of receptacles (RFF), the seed index (FSI) and malic acid content (FMA). Group 2 (selections S-266, S-256, S-260 and S-196) showed the highest values for the following characteristics: angle of the fifth vein of the leaf (LFVA), upper angle of leaf (LUA), width-to-length ratio of the leaf blade (LW/L), width-to-length ratio of the maximum projected area of petals (PW/L), seed index (FSI), fruit skin weight (FSW), fruit skin resistance to penetrometer (FSR) and seed weight (SW). The traits that yielded the lowest values for Group 2 were: leaf blade area (LA), petiole length (LPL), pedicle length (PdL), fruit pulp percentage (FPP) and length of the polar axis (FPA). Group 3 (the Burtons and White cultivars and selection S-9651) did not show maximum values for any of the characteristics, but conversely showed the lowest values for leaf blade form factor (LFF), upper angle of leaf (LUA), maximum projected area of petals (PA), area of the petal crosssection (PCS), maximum projected area of receptacles (RA), pedicle diameter (PdD), maximum projected area of sepals
(SpA), fruit weight (FW), total soluble solids in fruit pulp (FTSS), pulp pH and seed weight (SW). Group 4 (selections SSelene and S-Carapan) showed the highest values for leaf blade area (LA), leaf blade form factor (LFF), petiole length (LPL), number of veins on the right-hand side of the leaf blade (LNV), form factor of the petal cross-section (PCSFF), maximum projected area of receptacles (RA), form factor of the maximum projected area of receptacles (RFF), pedicle diameter (PdD), total soluble solids in fruit pulp (FTSS) and malic acid content (FMA). Conversely it showed the lowest values for the angle of the fifth vein of the leaf (LFVA), width-to-length ratio of the maximum projected area of petals (PW/L), fruit skin weight (FSW), fruit skin resistance to penetrometer (FSR), the ratio between the length of the polar axis and the length of the equatorial axis of fruits (FPA/EA), the width-to-length ratio of the maximum projected area of seeds (SW/L) and the maximum projected area form factor of seeds (SFF). 4. Discussion The grouping obtained through multivariate analysis is not consistent with the long-standing classification of cultivars into five botanical varieties based on characteristics of the fruit surface: smooth, fingerprint, tuberculate, mamillate and umbonate (Popenoe, 1921). In the scatter diagram obtained
0.48 0.59 0.44 0.46 0.70 0.64 0.69 0.63 0.62 0.56 0.57 0.54 0.19 0.20 0.20 0.29 5.20 4.97 4.75 4.90 19.75 19.65 16.16 22.80 3.68 5.03 3.94 0.76 4.55 11.16 8.65 8.30 86.33 71.82 77.77 85.06 92.17 133.26 98.44 75.91 848.90 680.05 614.93 753.75 1.30 1.13 1.23 1.10 14.37 12.35 12.53 12.38 0.58 0.39 0.36 0.49 0.26 0.23 0.21 0.27 1.49 1.04 1.13 1.06 0.65 0.72 0.75 0.78 0.23 0.23 0.19 0.34 0.62 0.68 0.65 0.72 21.40 17.66 10.02 10.06 0.27 0.29 0.28 0.26 2.26 1.67 1.32 1.54 13.33 13.67 13.33 14.50 0.59 0.67 0.63 0.60
Character codes according to Table 2.
96.42 101.34 85.29 98.98 66.36 66.49 59.75 58.31 1.84 1.54 1.95 2.26 138.58 102.94 154.20 164.52 1 2 3 4
0.75 0.74 0.69 0.78
SW/L FMA FpH FTSS FSR FSI FPP FSW FW FPA/EA FPA SpA PdD PdL RFF RA PCSFF PCS PW/L PA LNV LW/L LUA FFVA LPL LFF LA Group according to Fig. 3
Table 5 Mean values for every character, in each one of the groups of accessions defined after multivariate analysis (Figs. 2 and 3) of leaves, flowers, fruits and seeds of cherimoya (Annona cherimola Mill.)
SFF
SW
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
343
after PCA analysis (Fig. 2) and in the dendrogram (Fig. 3), Group 1 included one accession with umbonate skin and another with mamillate skin, while Group 2 had two mamillate accessions, one umbonate and one smooth. Group 3 had two umbonate and one fingerprint-tuberculate, whereas Group 4 had two accessions with fingerprint skin. The grouping of accessions obtained by other authors after isoenzymatic (Perfectti et al., 1993) or DNA-based molecular characterization (Rahman et al., 1998) was similarly not consistent with botanical varieties. A classification of cultivars based on surface characteristics of the fruit has been extensively criticized (Irazazabal, 1985; Ellstrand and Lee, 1987; Perfectti et al., 1993; Rahman et al., 1998), because it is not natural and because of the lack of consistency of these characteristics within a cultivar or even a single plant. Currently, this approach has been rejected (Scheldeman et al., 1999) in favour of a classification based on multiple genetic and morphological traits, as proposed in this work. The characteristics chosen here fulfil the requirements recommended by Perrier (1998) and Gonza´lez-Andre´s (2001). Firstly, they relate to several different organs of the plants: seven to leaves, nine to flowers, 10 to fruits and three to seeds. They are objective, since all of them are quantitative and are determined by measurement. In respect of their consistency, two points should be considered. The first is stability when subjected to selection processes or environmental influences. More than half of the morphometric characteristics measured relate to flowers and seeds. In general, seeds (Eames, 1961) and flowers (Eames, 1961; Sherry and Lord, 1996) are the stablest organs within a given taxon, for example, in comparison with leaves. This has been specifically reported for the cherimoya (Perfectti and Camacho, 1999). The second important aspect about consistency is intra-accession variability. In spite of the great genetic variability noted in the cherimoya (Pascual et al., 1993b), bringing with it considerable variation within cultivars (Ellstrand, 1997), the intra-accession variability of the traits studied makes them suitable for cultivar identification. A possible exception would be of the trait of the ratio between the length of the polar axis and the equatorial axis in fruits, as this showed a mean variation coefficient of more than 25%. From a general point of view coefficients of variation of 12% or lower are acceptable in plant organ characterization. If this coefficient is higher, then it would be advisable to increase the sample size. According to Hidalgo (2003) a sample size of 25 should be enough to keep experimental error below 10% when the coefficient of variation is 25%. In this work, the sample size for fruits, the organs with the highest coefficient of variation for most of the traits, was 30. On the basis of the ANOVA results, 28 of the characteristics considered, in other words all except for the form factor of the maximum projected area of the flower receptacle, may be useful for identifying cultivars. However, PCA showed that the traits, which yielded the maximum separation between the accessions studied were three relating to leaves, two to flowers, three to fruits and one to seeds. The leaf traits were leaf blade form factor, angle of the fifth vein of leaves and upper angle of leaves. For flowers, the characteristics were area of the
344
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
cross-section of petals and maximum projected area of sepals. In the case of fruits, the traits were weight of fruits, total soluble solids in fruits and resistance of the skin of fruits to a penetrometer. The seed trait was the width-to-length ratio of the maximum projected area of seeds. Of these more discriminatory characteristics, leaf blade area and total soluble solids of fruit pulp are the two most likely to be affected by the environment. Cultivation of the material being characterized in homogeneous conditions of soil and climate would reduce this problem. Skin resistance to a penetrometer is the most variable trait within accessions (with a mean coefficient of variation of 20.1%). In order to use this characteristic for identifying cultivars, the number of fruits tested should be at least 16, in order to keep the error of the sample below 10% (Hidalgo, 2003). For the remainder of the characters, the coefficients of variation observed were at most 14.8% (weight of fruits), all the others being less. Eight out of the 28 traits in which ANOVA detected significant differences between accessions are included in the list of characteristics in the ‘‘Guidelines for the Conduct of Tests for Distinctness, Uniformity and Stability’’ for the cherimoya issued by the International Union for the Protection of New Varieties of Plants or UPOV (UPOV, 2003). The inclusion of the other characteristics into the UPOV’s guidelines, especially those classified as stronger discriminants in the PCA analysis, might be proposed. The characteristics in the UPOV guide are treated from a qualitative viewpoint, while in the present work they have been treated from a quantitative viewpoint. However, most of the UPOV traits are easily transformable into quantitative ones. Utilization of an inexpensive and easy-to-use modern image-analysis system, for instance Image Tool, would simplify the data collection process. The grouping of accessions emerging from the work reported here (Figs. 2 and 3) was based on a large number of characteristics from several different organs. Except for the well-known cultivars Campas, Burtons and White, the accessions were local selections and a new Mexican cultivar, which were being studied in depth for the first time. Pascual et al. (1993a) and Perfectti et al. (1993) analyzed isoenzymes from the cultivars Campas and White. The isoenzymatic results showed a great distance between the two varieties, which is consistent with the morphology-based results being reported in this paper. Group 2 consisted of selections S-196, S-256, S-260 and S-266. These are outstanding genotypes selected from a single segregated population of 420 individuals located at the Fundacio´n Salvador Sa´nchez Colı´n CICTAMEX, S.C., situated at Coatepec Harinas, in the State of Mexico (Nicola´s-Cruz et al., 1996). This population was grown from the seeds of fruits collected in the area surrounding Coatepec Harinas. These four genotypes were selected on the basis of local farmers’ criteria concerning fruit size and productivity (Rubı´-Arriaga et al., 1992; Nicola´s-Cruz et al., 1996). The main fruit characteristics of this group contrasting with the other groups (Table 5) are a high seed index, low pulp content and a relatively resistant skin. The first two of these characteristics are not much liked by consumers.
Group 3 consisted of the cultivars Burtons and White and selection S-9651. This latter was a garden tree grown from seed. It was picked out in Tingambato, a community located in the Michoaca´n State of Mexico. White is a cultivar bred and registered in the United States of America, while Burtons is a New Zealand cultivar. One possible hypothesis to explain this grouping would be that the three accessions have some common origin. The grounds for proposing such a hypothesis would be that the cherimoya was introduced into the United States from Mexico by Judge Robert B. Ord in 1871, with some of the finest selections, such as White, being sent to New Zealand and other countries (Schroeder, 1997). However, S9651 has the same geographical origin as S-Selene, but S-9651 is clearly different from S-Selene. It would be possible to use microsatellite data (Sefc et al., 2000) to discover whether the morphological, agronomical and chemical similarities of Burtons, White and S-9651 are due to a genetic relationship. If that were so, two alternative hypotheses could be considered. The first would be that White and Burtons came from Mexico and thus had the same origin as selection S-9651. The alternative hypothesis would be that S-9651 was introduced into Mexico from White or Burtons material. From the point of view of traits of the fruits, this group of accessions showed the lowest fruit pulp pH and total soluble solid content. Moreover, they have small flowers, which was reported by Richardson and Anderson (1996) as hindering hand pollination of the Burtons cultivar. The selections S-Selene and S-Carapan, both in Group 4, come from the Pure´pecha ethnic communities of Tingambato and Carapan in Michoaca´n State, these places being located at a distance of 30 km one from the other. Both selections are garden trees. On the basis of their phenotypical similarity and the strong cultural links between the two communities, the material could be related. From point of view of fruit characteristics, these two selections showed high total soluble solids and malic acid content, which is a positive feature for taste, and they are highly appreciated by consumers. However, they have the disadvantage of having the least resistant skin on their fruits, which leads to difficulties with handling after harvesting. The cultivar Cortes-II-31 comes from Michoaca´n State, like S-Selene and S-Carapan, but multivariate analysis located it in Group 1, alongside the Campas cultivar, grown in Spain. The separation between S-Selene and S-Carapan, on the one hand, and Corte´s-II-31, on the other, that can be observed in the scatter diagram from PCA is based on the second and third principal components. The characteristics responsible for the maximum separation along these components were the widthto-length ratio of the maximum projected area of seeds, the petal cross-section, the angle of the fifth vein of the leaf, the upper angle of the leaf, the leaf blade form factor and the resistance to a penetrometer of the skin of the fruit. Differences are very clear in petal cross-section, in skin resistance to the penetrometer and in the width-to-length ratio of the maximum projected area of seeds, these being much greater in CortesII-31 than in S-Selene and S-Carapan. Although the three accessions come from Michoaca´n state, S-Selene and
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
S-Carapan are from the Pure´pecha indigenous communities, whereas, the cultivar Cortes-II-31 comes from Tancı´taro, which is not an indigenous community but mixed race, with more outside influences. With respect to the similarity between Corte´s-II-31 and Campas, since the parentage of Campas is unknown, molecular markers might help ascertain whether there is a relationship between the two genotypes. Both cultivars share the fruit traits of high pulp percentage and pH content, with a low malic acid content. In conclusion, 20 of the 21 morphometric characteristics chosen from several different organs of the plants, together with the five fruit traits of agronomical importance and the three chemical parameters for fruits, were all found useful for classification of cherimoya germplasm. The characteristics yielding the maximum separation between accessions were three relating to leaves, two to flowers, three to fruits and one to seeds. Consequently, all these traits might be of interest as cherimoya descriptors, so that their inclusion in UPOV might be considered. Moreover, four consistent groups of accessions were defined. Two of them were made up of accessions with a well-defined common origin. The origin of the accessions in the other two groups is unknown, hence further DNA-based studies might be carried out in the future in order to ascertain the reasons for this grouping. Acknowledgements This work was financially supported by the Servicio Nacional de Inspeccio´n y Certificacio´n de Semillas (SNICS), the Consejo Nacional de Ciencia y Tecnologı´a (CONACYT), the Centro de Investigacio´n y Desarrollo del Estado de Michoaca´n (CIDEM) and the Direccio´n General de Investigacio´n y Posgrado of the Universidad Auto´noma Chapingo in Mexico. References Andre´s-Agustı´n, J., 1997. El cultivo del chirimoyo (Annona cherimola MIll.) en el Estado de Michoaca´n, Me´xico. Proc. Am. Soc. Trop. Hort. 41, 152– 161. Casas, A., Caballero, J., Valiente-Banuet, A., Soriano, J.A., Da´vila, P., 1999. Morphometrical variation and the process of domestication of Snetocereus stellatus (Cactaceae) in Central Me´xico. Am. J. Bot. 86 (4), 522–523. Christie, T.C., 2001. Use of molecular techniqus in the Distinctness, Uniformity and Stability (DUS) testing of Brassicas: UPOV Working Group on Biochemical and Molecular Techniques. In: Scottish Agricultural Science Agency Scientific Review 1997-2000. Scottish Agricultural Science Agency 6.2, 72–73. Eames, A.J., 1961. Morphology of the Angiosperms, vol. I. McGraw-Hill, New York. Ellstrand, N.C., 1997. Cultivars characteristics. In: California Cherimoya Association (Eds.), The Cherimoya Handbook. Universitiy of California, Department of Botany and Plant Sciences, Los Angeles, CA, (Chapter 3), pp. 2–9. Ellstrand, N.C., Lee, J.M., 1987. Cultivar identification of cherimoya (Annona cherimola Mill.) using isozymes markers. Sci. Horti. 32, 25–31. George, A.P., 1984. Annonaceae. In: Page, P.E. (Ed.), Tropical Tree Fruits for Australia. Queensland Department of Primary Industries, Brisbane, Australia, p. 3541. Gonza´lez-Andre´s, F., 2001. La Caracterizacio´n Morfolo´gica. In: Gonza´lezAndre´s, F., Pita-Villamil, J.M. (Eds.), Conservacio´n y Caracterizacio´n de
345
Recursos Fitogene´ticos. Publicaciones I.N.E.A. Valladolid, Espan˜a, pp. 199–217. Hidalgo, R., 2003. Variabilidad gene´tica y caracterizacio´n de especies vegetales. In Ana´lisis estadı´stico de datos de caracterizacio´n morfolo´gica en Recursos Fitogene´ticos. IPGRI, 2004. Descriptors lists. International Plant Genetic Resources Institute, Rome. Irazazabal, P., 1985. Descripcio´n pomolo´gica de una coleccio´n de cultivares de chirimoya (Annona cherimola Mill.) Estacio´n Experimental ‘‘La Mayora’’, Ma´laga. Lee, J.M., Ellstrand, N.C., 1987. Inheritance and linkage of isozymes in the cherimoya. J. Hered. 78, 83–387. Nicola´s-Cruz, M., Rubı´-Arriaga, M., Cruz-Castillo, J.G., 1996. Exploracio´n y coleccio´n de fenotipos sobresalientes de chirimoya. In: Proceedings of the Memoria 1996 on Fundacio´n Salvador Sa´nchez Colı´n-CICTAMEX, Coatepec, Harinas, Edo. de Me´xico, pp. 59–65. Pascual, L., Perfectti, F., Gutie´rrez, M., Vargas, A.M., 1993a. Characterising isozymes of Spanish chirimoya cultivars. HortScience 28 (8), 845– 847. Pascual, L., Perfectti, F., Garcı´a, F.J., Vargas, A., Gutierrez, M., 1993b. Identificacio´n de cultivares de chirimoyo mediante ana´lisis de isoenzimas. Actas del II Congreso Ibe´rico de Ciencias Hortı´colas. Coleccio´n Monografı´as INIA 85, 858–863. Perfectti, F., Camacho, J.P.M., 1999. Analysis of genotypic differences in developmental stability in Annona cherimola. Evolution 53 (5), 1396– 1405. Perfectti, F., Pascual, L., 1998. Characterization of cherimoya germplasm by isozyme markers. Fruit Var. J. 52 (1), 53–62. Perfectti, F., Pascual, L., Garcı´a, F.J., Vargas, A., Gutie´rrez, M., 1993. Relacio´n y agrupamiento de cultivares de chirimoyo en base al ana´lisis de marcadores gene´ticos. Actas del II Congreso Ibe´rico de Ciencias Hortı´colas. Coleccio´n Monografı´as INIA 85, 864–871. Perrier, X., 1998. Analyse de la diversite´ ge´ne´tique: Mesures de dissimilarite´ et repre´sentations arbore´es. The`se doctorat. Universite´ de Montpellier II. Montpellier Popenoe, W., 1921. The native home of the Cherimoya. J. Hered. 12 (7), 336– 339. Rahman, M.S.M., Yamada, M., Yoshida, M., 1997. Relationship of Annona species as revealed by PCR-RFLP analysis. Breed. Sci. 47 (4), 335– 339. Rahman, S.M., Shimada, T., Yamamoto, T., Yonemoto, J.Y., Yoshida, M., 1998. Genetic diversity of cherimoya cultivars revealed by amplified fragment length polymorphism (AFLP) analysis. Breed. Sci. 48, 5–10. Rholf, F.J., 2000. NTSYSpc. Numerical Taxonomy and Multivariate Analysys System. Version 2. 1. Exeter Software, New York. Richardson, A., Anderson, C.P.A., 1996. Hand pollination effects on the set and development of cherimoya (Annona cherimola) fruit in a humid climate. Sci. Horti. 65, 273–281. Ronning, C.M., Schnell, R.J., Gazit, S., 1995. Using randomly amplified polymorphic DNA (RAPD) markers to identify Annona cultivars. J. Am. Soc. Horti. Sci. 120 (5), 726–729. Rubı´-Arriaga, M., de la Cruz, T.E., Barrientos, P.F., 1992. Mejoramiento gene´tico de la chirimoya a trave´s de seleccio´n en poblaciones de segregantes. In: Proceedings of the Memoria Fundacio´n Salvador Sa´nchez Colı´n-CICTAMEX, S.C. Coatepec Harinas, Edo. de Me´xico, pp. 127– 134. Scheldeman, X., Uren˜a, V., Van Damme, P., 1999. Coleccio´n y caracterizacio´n de chirimoya (Annona cherimola Mill.) en la Provincia de Loja, Sur de Ecuador. Acta Horti. 497, 153–172. Schroeder, C.A., 1997. Botany. In: California Cherimoya Association (Eds.), The Cherimoya Handbook. University of California, Department of Botany and Plant Sciences, Los Angeles, CA, (Chapter 2), pp. 2–10. Sefc, K.M., Lopes, M.S., Lefort, F., Botta, R., Roubelakis-Angelakis, K.A., Ibanez, J., Pejic, I., Wagner, H.W., Glossl, J., Steinkellner, H., 2000. Microsatellite variability in grapevine cultivars from different European regions and evaluation of assignment testing to assess the geog– raphic origin of cultivars. Theor. Appl. Genet. 100 (3/4), 498– 505.
346
J. Andre´s-Agustı´n et al. / Scientia Horticulturae 107 (2006) 337–346
Sherry, R.A., Lord, E.M., 1996. Developmental stability in flowers of Clarkia tembloriensis (Onagraceae) as related to population outcrossing rates and heterozygosity. Evolution 50, 80–91. Sneath, P.H.A., Sokal, R.R., 1973. Numerical Taxonomy. The Principles and Practice of Numerical Classification. Freeman, San Francisco. Thomson, P.H., 1970. The cherimoya in California. Calif. Rare Fruit Growers Yearbook 2, 20–34.
UPOV, 2003. Cherimoya (Annona cherimola Mill.). Guidelines for the conduct of tests for Distinctness, Uniformity and Stability. International Union for the Protection of New Varieties of Plants, Geneva. UTHSCSA, University of Texas Health Science Center and San Antonio, 2000. Image Tool. IT Version 2.0. University of Texas, San Antonio, TX., http:// macorb.uthscsa.edu/dig/download.html.