Genetic Divergence among Camu-Camu Plant Populations Based on the Initial Characteristics of the Plants

The objective in the present work was to evaluate the genetic diversity among 15 indigenous populations of camu-camu plants, identifying important characteristics in the evaluation of genetic divergence, based on the initial characteristics of the seedlings. Seeds extracted from fruits deriving from fifteen indigenous populations of camu-camu were collected. The experimental design was entirely random, with fifteen treatments (populations), and fifteen repetitions (each sub-sample), considering 30 seeds per subsample as an experimental unit. At 40 days after sowing the following were evaluated: the percentage of emergence, the index of emergence velocity, the average time of germination, the height of the seedling and the number of leaves. The data obtained was submitted to variance analysis, and the averages were grouped by the Scott and Knott (1974) test. The genetic diversity was studied according to the Tocher grouping method, based on the Mahalanobis distance (Dii) and canonical variables. The fifteen populations are divergent among themselves and the Rio Branco Estirão do Veado, Rio Branco Onofre and Igarapé Agua Boa populations are indicated to have hybridization with other populations due to the high divergence, as well as the rates of emergence and vigor of the seedlings. The height of the seedlings, percentage and speed of emergence, are those that most indicate genetic divergence. The measuring techniques of genetic divergence, canonical variables Mahalanobis distances are useful and corroborating in the evaluation of genetic divergence of the camu-camu plant.

In Amazonia, the potential for camu-camu is in its use in the preparation of foods like juices, sweets and ferments (Rodrigues et al., 2004;Teixeira et al., 2004;Chirinos et al., 2010;Akter, Oh, Eun, & Ahmed, 2011).It also constitutes a raw material for the cosmetic, chemical, pharmacological industries, food preservation and production of aerated beverages (Correa, 2000;Yuyama, 2011).Thus, the production and the utilization of the fruit appear to be viable alternatives in regional development, as a means of aggregating value from the natural resources available in the region (Welter et al., 2011;Chagas et al., 2015).
The principal function of genetic improvement of the camu-camu is selecting the genotypes which maximize yield from the first phases of their development (Pinedo, Linares, Mendoza, & Anguiz, 2004;Yuyama & Valente, 2011).Genetic divergence is one of the most important parameters evaluated by improvers of the plant in the initial phase of a genetic improvement programme, because, adequately explored, it may accelerate the genetic progress of particular characteristics (Negreiros et al., 2008) The seeds were manually separated from the fruit and the residual pulp, by friction in a fine grade sieve.After removal, the seeds were washed in running water and treated with sodium hypochlorite solution at 10% for five minutes.The seeds were not submitted to artificial drying.Thereafter, the seeds were stored for 15 days in cold chamber at 10 ºC in polypropylene sacks, and the relative humidity of the air was maintained at 50%.Subsequently, they were planted in seedbeds containing sand substrate and sawdust in 1:1 proportions.The seedbed was placed in a nebulization chamber, with irrigation intervals at four times per day, for a period of 10 minutes.The seeds were distributed at a distance of 2 cm between rows and 1 cm between them in line, and at 1.5 cm of depth.The evaluation criteria adopted was counting as the epicotyl appeared 1.5 cm above the surface of the substrate, in emergence stage.The experimental delineation was entirely random, with fifteen treatments (indigenous populations) and fifteen repetitions (subsamples), considering thirty seeds per subsample as an experimental unit, totalling 450 seeds per treatment.
At 40 days after planting, the experiment was finalized as two populations had already presented 100% emergence.The percentage of emergence was evaluated, the speed emergence of seedling (SES, index) (Maguire, 1962), the average time of emergence (Yuyama, Mendes, & Valente, 2011), the height of the plant shoots (cm) and the number of leaves emitted.The emergence percentage data were transformed in square root of arcsine "x/100" and the SES in square root "x + 0.5" (Gotelli & Ellison, 2011).

arcosine√x/100
(1) The data from the rest of the variables was not transformed.The SES was established from the emergence test, with daily evaluations being done upon the emergence of the first plants up until the 40th day.
The data was submitted to variance analysis in order to verify the existence of genetic variability among the populations, being that their averages were grouped in accordance with the Scott and Knott (1974) test, at 5% probability.Thereafter, multivariated analyses were used, applying the grouping and canonical variable techniques with the assistance of INFOGEN software, version 2013 (Balzarini & Di Rienzo, 2013).
In the grouping technique, Mahalanobis generalized distance was utilized (D2ii) (Mahalanobis, 1936) as a dissimilarity measure.In the group delimitations, the Tocher optimization method was used, adopting the criteria that the average of the measurements of genetic divergence within each group aught to be less than the average distances between groups (Cruz, Regazzi, & Carneiro, 2004).
Additionally, the relative contributions of the characteristics to genetic divergence was quantified by Mahalanobis generalized distances, utilizing the criteria proposed by Singh (1981), analyzed with the assistance of GENES software, version 2005 (Cruz, 2008).

Results
By univariated variance analysis, there were significant differences between the population averages (p < 0.01), through testing of F, for all the evaluated characteristics (Table 2), indicating at least divergence among the populations.
Based on the grouping of averages, it was verified that only two populations (RJI and RAN) presented percentages of emergence (EP) below 50%, while the population average of emergence had been 73.8%.The BRB EV and IAB populations obtained 100% EP, indicating that these materials are promissory for future work in improvement.The populations which presented the highest indices of emergence velocity, plant shoot height and number of leaves were those that obtained values above 80% emergence, with the stand out populations being the BRB EV, RB LM and IAB.The shortest emergence times were registered in the RB LM and BRB ON populations with values less than 33 days (Table 2).Note.* : Averages with the same letter, in column, belong to the same Scott-Knott grouping at 5% probability; **: Significant in F testing at 1% probability; CV = variation coefficient %.
With a basis in the relative magnitude of D2ii values, the formation of seven distinct groups was verified using the Tocher grouping method, being that the largest concentration of populations was in the first group (Table 3).

Discussion
According to Tekrony and Egli (1991), the vigour of plants, observed in the field by the ability of the seed to emerge and grow rapidly and vigorously, is a factor which can influence the productivity of cultures.According to Pinedo, Linares, Mendoza, and Anguiz, (2004), part of genetic improvement of the camu-camu is related to the selection of genotypes based on their rapidity, percentage of emergence and vigour of plant shoots, among other agronomic characteristics.
According to Cruz, Regazzi, and Carneiro (2004), when the first two canonical variables are above 80% of total variation, their utilization is satisfactory in the study of genetic divergence by way of the evaluation of the graphic dispersal of the scores in relation to the canonical variables (CV1 and CV2).This constant structure, through the formation of group coincidence in the utilization of complementary methods for morpho-agronomic characteristics, generates greater confidence in the results (Sudré et al., 2005;Oliveira, Ferreira, & Santos, 2007;Negreiros, Bergo, Miqueloni, & Lunz, 2013).Thus, the BRB AB and BRB AT populations that had to be classified as one group (Group 4), with the Tocher method, were classified by canonical variables within group III, together with the BRB EV, BRB ON and IAB populations.However, the IPI population, isolated in the Tocher method, was classified, by canonical variables, within group II, together with the RAN and RJI populations (Figure 2).
In group III, the populations of the lower Branco river BRB AB, AT, EV and ON, are interconnected by the same hydrographical region, which probably is related to the high values registered in the initial plant characteristics, representing a region of great potential for obtaining promissory genotypes for future improvement works.
The plant shoot height variable was the most important in distinguishing the populations in the first canonical variable (CV1) (68.31%), followed by the percentage of emergence and the average time of emergence, being that the contribution of these three characteristics to the genetic diversity among the 15 populations is confirmed, with a basis in the criteria proposed by Singh (1981) (Table 4).Thus contribution was verified in descending order as follows: plant shoot height, percentage of emergence, average time of emergence and SES.The number of leaves presented the lowest estimate of genetic diversity between the populations (S.j), not being important for the evaluation of genetic divergence between populations.19.9 1.56 Note.S.j = S is the average relative importance for each variable; j = for the study of genetic diversity.
The percentage of emergence and plant shoot height characteristics contributed to 86.24% of divergence, which could be the initial agronomic parameters to be considered in the selection of genotypes in future works in genetic improvement in the species, bearing the values observed in mind.

Conclusions
There is genetic divergence among the fifteen populations and the BRB EV, BRB ON and IAB populations can be indicated for hybridization with other populations due to high genetic divergence, rate of emergence and vigour of the plant shoots.
The height of the plant shoots, percentage of emergence and the ATE are the characteristics of highest contribution to genetic divergence identified in the camu-camu plant.
The canonical variables and the Mahalanobis distance are useful and complementary in the evaluation of genetic divergence of the camu-camu plant. in . In relation to the germination and emergence of red-wine colou he samples (n) cks, maintaine ed a code with

Table 1 .
Natural camu-camu populations prospected in the State of Roraima in different locations, municipalities and Hydrographical Regions

Table 2 .
Average values for emergence of plants (EP, %), speed emergence of seedling (SES, index), average time of emergence (ATE, days), height of plants (HP, cm) and number of leaves (NL) obtained for 15 indigenous populations of the camu-camu plant

Table 4 .
Estimates of relative contribution of each characteristic (S.j) to the genetic divergence between the camu-camu plant populations, based on the partition of the total D 2