Stem and leaf anatomical and physio logical characteristics of ‘Colín V33’ avocado seedlings

The anatomical and physiological structure of the ‘Colín V­33’ avoca­ do stem and leaf is described from samples from plants obtained from seed in order to identify genotypes and early selection parameters in a rootstock improvement program for avocado. Eighty­nine plants of 12 months of age were used, where a total of 25 anatomical variables of the stem, leaf, and phys­ iological of leaf were evaluated. A cluster analysis was conducted that generat­ ed a hierarchical dendrogram that suggested six groups of plants. Furthermore, from the 25 variables, eight were selected as discriminant when performing a canonical discriminant analysis, the variables that most discriminated for the first canonical component were: stem diameter and density of xylem vessels, for the second: thickness of the stem epidermis, temperature of the stem leaf and stomata length, while for the third: thickness of the cambium, transpira­ tion rate, and stomatal conductance. The genotypes showed a great variation between the groups, the characteristics of these indicated that the genotypes of Group 4 showed some that could be related to small or dwarf plants (smaller stem diameter, high density of xylem vessels, a higher rate of transpiration and stomatal conductance). In contrast to the genotypes of Group 3 which present­ ed opposite characteristics in the previous variables, being able to associate with vigorous plants. The anatomical traits of the stem showed to be highly related to the behavior of the avocado plants. Associating genotypes with phys­ iological and anatomical variables in leaf and stem can have great value for the selection of rootstocks at an early stage of development.


Introduction
Fruit trees used for the establishment of plantations are formed by a variety/rootstock combination, where the variety provides the productive part and the rootstock provide the root system of the tree (Dolgun et al., 2009). The rootstocks provide a simplified management of the orchard, increase the productivity, ensure the survival of the trees, control the vigor conferring dwarfism and allowing the use of high planting densities and the water balance of the plant (Solari et al., 2006), influ ence the qualitative and nutritional attributes of the fruits (Remorini et al., 2008), also influence the scion development, as well as its adaptation to different types of soil, water stress conditions and other cli matic conditions (Giorgi et al., 2005), as well as salt stress tolerance (Massai et al., 2004).
Selection of the rootstock and the cultivar or vari eties to be exploited is an economically important decision, since the degree of productivity and quality of production will depend on the variety/rootstock combinations (Pinochet, 2010), these combinations are in function of the degree of affinity in terms of vascular connections between the graft and the root stock, because vascular regeneration is a complex process that includes the differentiation of xylem and phloem (Aloni et al., 2010).
Avocado production in Mexico is based on the use of seedderived rootstocks which generate large trees that, over time, complicates the agronomic management. For this reason, one of the priorities in avocado research is to find dwarf cultivars that are highly productive, with good fruit quality and resis tant to pests and diseases in order to establish high density plantations and reduce production costs . The use of rootstocks in intensive avocado cultivation is of great importance and rootstocks for this crop have been reported to influence the size of the tree and its productivity (Bergh and Whitsell, 1962). Potentially dwarfing root stocks such as 'Wilg' and 'Colin V33' have been eval uated against 'Duke 7', for their ability to limit the growth of 'Hass' (Roe et al., 1995).
'Colín V33' avocado is a Mexican selection that has low vigor and an expanded growth habit (Roe et al., 1995), considered as a possible dwarf rootstock since it has been used as an interstock, with reported reduction of 43% of the tree height on 'Fuerte'  and 'Hass' (BarrientosVillaseñor et al., 1999). On the other hand, given the genetic variability for the height of the seedlings of this cultivar (Rubí Arriaga, 1988), it makes it a valuable material for the improvement of dwarf avocado rootstocks. In this regard, Barrientos Priego et al. (1992) have found that the cultivar seedlings produce some dwarfing individuals when used as rootstocks.
Efforts have been made mainly with 'Colín V33' for the selection of dwarf rootstocks that meet the needs of the new production systems, where stoma ta density has been proposed as a possible preselec tion index (BarrientosPérez and SánchezColin, 1982;BarrientosPriego and SánchezColín, 1987). Other important features to be investigated are the ones that contribute to the hydric status of the plant and its productivity, in which a series of physio logical, anatomical and morphological characteristics of the stem can be measured. At the physiological level, gas exchange can be studied (Vilagrosa et al., 2010) and to determine the affinity between vari ety/rootstock combinations, the anatomical structure of xylem can be studied (Sory et al., 2010;Leal Fernández et al., 2013).
The anatomical characteristics of the dimensions of vessel elements and the proportions of xylem and phloem, in stems, are important to be able to define the amount of water that can be transported through them, since as the tissues become larger and the presence of smaller diameter of the vessels (higher pressure for water movement), the amount of water transported will be greater, therefore, there will be a better adaptation of the plants to low humidity con ditions (Vasconcellos and Castle, 1994;Reyes Santamaría et al., 2002). ReyesSantamaría et al. (2002) found vessel ele ments with smaller diameters, for the avocado geno types that have less vulnerability to drought. The stomatal density (ED) and the thickness of the epi dermis are characteristics that may be related to drought tolerance (Baas, 1982), as they are the most exposed anatomical characteristics of the plant, which represent the last link in the transpiration tor rent towards the atmosphere (Faust, 1989).
The use of avocado seedling rootstocks of a local type called "Criollo" is very common but rarely stud ied from their anatomical and physiological charac teristics on stem and leaf, variables that can be useful as a preliminary study to understand their possible role when grafted. The objective of this research was to describe the anatomical and physiological struc ture of the stem and leaf of plants derived from avo cado seed of 'Colín V33', to select individuals at the seedling level with distinctive characteristics.

Materials and Methods
The research was carried out in a greenhouse of the Experimental Field of the Chapingo Autonomous University, in Chapingo, State of Mexico,located at 19° 29'25.7'' and 98° 52'24.5'' with an altitude of 2240 meters above sea level.

Plant material
As plant material, 89 avocado seedling plants of 'Colín V33' were used, that were donated by the Germplasm Bank of the Salvador Sánchez Colín FoundationCICTAMEX, S.C. located in Coatepec Harinas, State of Mexico. The seeds were established in a black bag of caliber 600 of 26 cm x 35 cm with perforations in the first lower third; using soil, perlite and compost as a substrate (3:1:1; v:v:v), watered twice a week, and located in a glass greenhouse with oscillating temperatures between 35 ± 4°C.

Leaf gas exchange variables
To each of the plants the eleventh leaf, fully expanded and healthy, counted from the base towards the apex of the plant was selected as recom mended by BarrientosPriego et al. (2003). Once marked, the variables CO 2 assimilation rate (A), tran spiration rate (E), leaf temperature, internal CO 2 con centration and, stomatal conductance was evaluated. A punctual measurement was made between 11:00 and 13:30 hours a day with an infrared gas analyzer (model CI340, CID BioScience). This was done dur ing three days of the month of September of the year 2017 and the measurement was taken in the same plant order every day to avoid more variation in the data taken for each plant. The average of the three values obtained was used in the analysis. The water use efficiency index (WUE) was calculated based on the variables CO 2 assimilation (A) and tran spiration rate (T), using the formula WUE = A/T.

Leaf anatomical variables
After the measurements of gas exchange vari ables, on the same leaf previously marked, an impression of the underside of the middle part of the lamina using silicone for dental impressions (Exactoden) was taken. By applying transparent nail varnish on the (negative) impression, the positive impression was obtained, which was placed on a slide and fixed with a coverslip.
In the positive impressions of each sample, the image area was calculated with an object microme ter, stomatal density (SD) per mm 2 and epidermal cell density (ECD) per mm 2 were determined. These vari ables were evaluated in five fields [400x, Ayala Arreola et al. (2010)] in a Motic B3 Professional Series microscope, with the adaptation of a Moticam 480 camera with a 16 mm adapter. In addition, 10 stomata were measured in each sample. Stomata counts, epidermal cells and stomata length measure ment were performed with the help of ImageJ 1.52a image analyzer.
With all this information the stomatal index (EI) was calculated, which is equal to:

Stem anatomical variables
Transverse stem samples were obtained from plants of approximately 1 cm each, which were fixed in 96 % ethanol:100 % glacial acetic acid (2:1; v:v) and processed in an automatic tissue exchanger (Tissuematon Fisher) with 2ethoxyethanol (cello solve) and xylene, then transfer to paraffin (55°C) staying 72 hours inside a stove. The paraffin pyramid was made according to Sass (1968) and in a rotary microtome (American Optical, model 820), trans verse cuts were made with a thickness of 10 µm. The cut sections were stained for 30 min at room temper ature in a mixture of equal volumes of 0.1% aqueous solutions of safranin and fast green, then washed in distilled water for 5 minutes and washed in 2 changes of absolute alcohol for 23 min (Bryan, 1955). The stained sections then mounted on slides with coverslips by means of Haupt adhesive and 10 % formalin (Sass, 1968).
Fifteen fields were observed in each preparation per replication (three repetitions). For which the 4x, 10x and 40x lenses were used as appropriate, in a Motic B3 Professional Series microscope, with the adaptation of a Moticam 480 camera with a 16 mm adapter and digital images were obtained. The area of the images was calculated with a slide micrometer and then the cell layers of the tissues and the dimen sions of xylem vessel elements were measured with the help of the ImageJ 1.52a image analyzer.
The tissues evaluated were epidermis, parenchy ma, phloem fibers, phloem, cambium, xylem, and pith. The thickness of each layer and the total diame ter of the stem was measured. For the xylem dimen sions number of vessel elements per area (vessel density), vessel element area, major axis of vessel element, minor axis of vessel element, cell wall thick ness of two contiguous vessel elements, roundness index (RI) of vessel and Feret diameter were obtained.

Statistical analysis
A cluster analysis was performed using Ward's minimum variance agglomeration method to gener ate a hierarchical dendrogram (NúñezColín and EscobedoLópez, 2011). The resulted dendrogram was divided according to Hotelling's pseudostatisti cal t 2 (Johnson, 1998). A canonical discriminant analy sis (CDA) with the Mahalanobis distance was per formed to determine the most discriminating vari ables with the greatest importance that describe the groups (NúñezColín and EscobedoLópez, 2014). All statistical analysis were performed with the SAS V.9.2 statistical package (SAS Institute, 2009).

Results and Discussion
In order to explore the homogeneity within each variable evaluated and based on the coefficients of variation (CV), it was determined that the physiologi cal variables of leaf: CO 2 assimilation rate, stomatal conductance and water use efficiency index (WUE), were the characteristics with the greatest variation (Table 1), presenting CV of 60.84%, 57.65%, and 62.5%, respectively. These characteristics being less stable within the plants studied (more heteroge neous). On the contrary, all the anatomical variables of stem and leaf showed lower coefficients of varia tion that were below 28%, so they could be consid ered more homogeneous but still show contrasting features ( Fig. 1 and 2).

Cluster analysis
Based on Hotelling's t 2 pseudostatistic (Fig. 3), the grouping by Ward's method showed the identifica tion of six groups of plants (Fig. 4). Group 1 com posed of 21 plants, Group 2 by 9, Group 3 with 13, Group 4 with 13, Group 5 with 16 and Group 6 with 17 plants (Table 2).
In the obtained dendrogram a partition of six groups was used approximately at a cutoff distance R 2 semipartial close to 0.050 (Fig. 3). The greater dis tance corresponds to Group 4 confirming that are dif ferent from the rest, that was also reinforced with  , 1994;ReyesSantamaría et al., 2002;NúñezColín et al., 2006). The CC 2 represented 28.90% of the total varia tion and was associated with: epidermis thickness, leaf temperature, and stoma length (Table 3 and 4). The anatomical stem variable (epidermis thickness) present in CC 2 is associated with drought resistance (Baas, 1982) and it is possible to relate it as acquired adaptation, to tolerate prevailing environmental con ditions of the greenhouse (38°C). On the other hand, the size of stomata is one of the anatomical variables of the leaf that could be sensitive to the change in environmental conditions (Hetherington and  Woodward, 2003). Consequently, these characteris tics can be considered as representative of the adap tation to the environment of the plant, as a constant feature of taxonomic value, that the relationship between these two anatomical variables (epidermis thickness and stoma length) have with the variable air temperature. The CC 3 was associated with the variables thick ness of the cambium, transpiration rate, and stom atal conductance. Stomatal conductance refers to the control exerted by stomata on the rate of transpira tion, representing the ability of a water molecule to diffuse through the leaf per unit of time. A minor stomatal conductance may be the expanation for lower CO 2 assimilation and at the same time a lower transpiration rate (Sholefield et al., 1980;Barrientos Villaseñor et al., 1999).
Stomata size is considered a key physiological variable on stomatal conductance (Holland and Richardson, 2009), which is why the stomatic conduc tance use can be considered as a rapid test in geno type selection, as an indirect measure of stomatic density, since this can also be a way to perform a genotype separation towards dwarf types as pro posed by BarrientosPérez and SánchezColín (1987).
The F significance test of the Mahalanobis dis tance (Table 5) indicated that there are highly signifi cant differences between the groups of plants at a level of P<0.0001.
To have a better and clearer visualization (distrib ution) of the groups of plants, the graphic represen tation was made in the first factorial plane with the first two canonical components (Fig. 5).
The graphic representation of the groups in the first factorial plane showed that the Group 4 is the only one that is isolated from the rest and was very different from the other five groups, with most of its plants in the lower left quadrant (Fig. 5). Groups 1, 2 and 6 are probably clustered because they are plants with anatomical and/or physiological similarities according to the variables evaluated in this study (Fig. 5). These groups are those that have dis persion in both planes, so they are considered the most heterogeneous groups within the plants studied.
Groups 3 and 5 were slightly more separated from the rest of the groups, but between them, there is some closeness (less dispersion), which could be inferred as similar groups according to the behavior of the variables evaluated (Fig. 5).
WUE= Water use efficiency index. Table 3 Values of the canonical discriminant analysis of stem and leaf anatomical variables and physiological leaf variables of 89 'Colín V33' avocado seedlings In the pattern of dispersion of the plants in a threedimensional graph formed by the three canoni cal components (CC), it was observed that group 4 is completely separated from the rest and is located towards the most negative values of CC 1 and CC 3 (Fig. 6), it was also observed that groups 3, 5 and 6 were not so separated from the rest of groups (less dispersed), such as 4, and if they could be easily dif ferentiated in the graph located towards positive val ues of CC 1 and CC 2; on the other hand, the most dispersed groups were 1 and 2, mixing with each other, indicating that they probably have similar characteristics.
Group 1 is formed by plants with the greatest epi dermis thickness and the greatest stoma length but the lowest stomatal conductance of all other groups (Fig. 6). The thickness of the epidermis is a character istic that may be related to drought tolerance accord ing to Baas (1982), as it is the most exposed anatomi cal characteristic of the plant along with the stomatic density, representing the last link of the transpiration stream into the atmosphere (Faust, 1989).
The plants of group 2 were characterized by hav ing the cambium with less thickness, high average values of stem diameter and stomata length.
Group 3 was characterized by having plants with greater stem diameter, a lower density of vessel ele ments and shorter stomata length, low average val ues in transpiration and stomatal conductance. This group of plants is likely to be more vulnerable to physiological problems such as drought as reported by ReyesSantamaría et al. (2002).
The plants in group 4 had the lowest values for stem diameter, but the highest density of xylem ves sel elements, also the highest rate of transpiration and the highest stomatal conductance, this being the group that separated completely from the rest (Fig.  4). The results could indicate that the plants belong  ing to this group probably use more water and can be more adapted to condition of more availability, because by decreasing the stomatal opening and closing it contributes to reducing the loss of water due to evapotranspiration (NúñezColín et al., 2006). Previous results (Sholefield et al., 1980;Barrientos Villaseñor et al., 1999) indicated that there is a posi tive correlation between the variables CO 2 assimila tion, stomatal conductance and transpiration rate, thus an increase in stomatal conductance promotes the increase in the transpiration rate and the photo synthetic rate (DamiánNava et al., 2009). Group 5 was represented by plants with the low est thickness of the epidermis, the lowest leaf tem perature, low stomatal conductance values, and tran spiration. Group 6 ( Fig. 5) was characterized by hav ing the greater cambium thickness, high average val ues in the variables stem diameter, epidermis thick ness, leaf temperature, stomata length, but low den sities values of vessel elements, transpiration, and stomatal conductance.
Groups 1 and 6 were closest to each other, according to the Mahalanobis distances (Table 4), while the most distant groups were 3 and 4 (Fig. 5).
For any type of plant, the conducting tissue of the stem, the size of the xylem vessel elements, the per centage of the xylem and phloem, and the relation ship between xylem and phloem, are anatomical fea tures that define the transport capacity of water in plants. It has been observed in some tree species, that as the percentage of vascular tissues increases and the diameter of the vessels are smaller, the amount of water transported is higher, and this may be an indicator of greater adaptation from plants to low humidity conditions in the soil (Vasconcellos and Castle, 1994;ReyesSantamaría et al., 2002). For the case of avocado rootstocks, it has been found that stems of 'Duke 7' (1248.7 µm 2 ) had narrowed xylem element cells diameters than 'Toro Caynon' (1536.1 µm 2 ), were 'Duke 7' showed higher daily sap flow (2.8 kg day 1 ) compared to 'Toro Caynon' (2.0 kg day 1) on ungrafted plants and in the case of grafted plants it also increased the sap flow when 'Duke 7' was used (Fassio et al., 2009).
Several authors compared normal (vigorous) trees with dwarf trees of citrus, olive, mango, and Copaifera langsdorfi found that dwarf trees are char acterized by narrow (small) vessel elements and a higher density of xylem vessel elements (Saeed et al., 2010;El Said et al., 2013;Rashedy et al., 2014;Longui et al., 2014, respectively) which has also been found in avocado (ReyesSantamaría et al., 2002). The results found showed that the plants of Group 4 are the ones that had the highest densities of vessels and with this, probably less vessel element diameter, increasing transport and conferring a low probability of suffering cavitation and embolism (NúñezColín et al., 2006). On the other hand, Goncalves et al. (2007) and Tombesi et al. (2010) indicated that in cherry and peach dwarf rootstocks, respectively, the diameter of xylem vessel elements is smaller. It has been found in peach that large diameters of xylem vessel elements are found in the more invigorating rootstocks (Bruckner and DeJong, 2014). For all the above, it is inferred that genotypes belonging to Group 4 are small size plants or with dwarfing characteristics, this statement is reinforced by studies that showed that intermediate and small size plants frequently have a smaller stem diameter, such as the case of the Group 4 plants that had the lowest average stem diameter values (Fig. 7). These results show a similar trend to the study  in trees of 'Colín V33' where the trunks in dwarf trees had a circumference and average diame ter smaller than the tall trees. avocado seedlings in the six groups generated according to canonical discriminant analysis (ADC) and Ward clu ster, derived from anatomical and physiological characte ristics of stem and leaf.
The classification results show that 97.8 % of origi nally grouped individuals were correctly categorized (Table 6), which indicated that the six groups based on this are consistent, showing the stability of belonging to each group. Where only two individuals were atypical of their group. These results give cer tainty to the analysis performed and the congruence of the groups obtained.

Conclusions
The variables stem diameter, the density of vessel elements, thickness of the epidermis, leaf tempera ture, stomata length, thickness of the cambium layer, transpiration rate and stomatal conductance, dis criminated the groups of 'Colín V33' avocado seedling plants correctly.
The main correlations obtained were a positive correlation between the density of xylem vessel ele ments with the rate of transpiration and with stom atal conductance. These correlations indicated a lin ear relationship of these variables and this last can be used as an index of preselection to discriminate the vessel element density in seedlings.
A higher density of elements of xylem vessel ele ments and smaller stem diameter appear to be indi cators of dwarf plants in avocado, so these character istics could allow the selection of genotypes with behavior of this type of growth.
Group 4 genotypes presented anatomical and physiological characteristics probably associated with dwarf types unlike Group 3, which were plants with typical anatomical and physiological characteristics of vigorous plants.
To demonstrate the use of the different contrast ing groups it is required further studies with grafted plants to determine their potential according to their rootstock characteristics.