Lecythis pisonis Cambess ( sapucaia ) : seed morphometric traits and germinative response

This study aimed to test a possible correlation between germination and the variables weight and density of seeds of Lecythis pisonis (sapucaia), as well as raise issues which may contribute to know germination in this species. For this, we followed up the soaking and germination process of intact and scariied seeds of L. pisonis under controlled conditions (26°C, with average irradiance of 34.6 mmol.m.s and in the dark). In another trial, seeds of sapucaia were individually characterized with regard to weight and density, and, then, sown in plastic boxes illed with soil and kept in a natural environment (average irradiance ±12.5 mmol.m.s at 12:00 p.m.), with temperatures ranging from 15 to 32°C. The germination (shoot emergence) of each seed was followed up within the period from September 2009 to January 2010. The results suggest that the coat does not represent a restrictive factor for soaking and germination of sapucaia. Germination increased proportionally to the increase in seed density. Based on these results, we recommend the use of seeds with densities > 0.9 g.mL, in order to ensure a high percentage of seedlings.


Introduction
Tree regeneration in lowland tropical forests is mainly done by seeds (RAICH; KHOON, 1990), thus, studies on germination may help us understand the ecophysiology and the distribution of species within a community, and they also provide important clues concerning seed technology and species management.Lecythis pisonis Cambess (Lecythidaceae), commonly named "sapucaia" in Brazil, is a tall (up to 35 m) and wide (1.5 m) tree when growing in isolation.The species is either classiied as early secondary (CARVALHO, 2006) or shade-tolerant climax and it occurs naturally in the Atlantic Forest and in the Amazon, being endemic to Brazil (CARVALHO et al., 2006).
The seed of sapucaia (1.7-2.6 cm wide and 3.6-4.3cm long) is angular, grooved, and it has a white pulpy aril at the end opposite to the micropyle (PAGOTTO, 1982;CARVALHO, 2006;BRAGA et al., 2007).According to Floriano (2004), the aril must be removed to overcome seed dormancy, while Carvalho (2006) reports that seeds of sapucaia require no pre-treatment for germination.Regarding longevity, the seeds are classiied as orthodox (desiccation-tolerant) and they can be stored at 5ºC or -18ºC without viability loss (CARVALHO, 2006;CARVALHO et al., 2006).The emergence of seedlings of sapucaia has been reported to take place between 30 and 90 days after sowing (CARVALHO, 2006).
According to Braga et al. (2007), although seeds of Lecythis pisonis are relatively uniform in length, the sample of seeds analysed by the autors was heterogeneous in terms of width, thickness, and weight.Thus, it is possible to see that seeds of L. pisonis may vary in terms of morphometric traits even within the same sample, and this variations may be related to changes in the seed physiological traits, such as germinability and germination rate (ANDERSSON, 1996;SCARPA;VALIO, 2008).This study aimed to test a possible relation between seed morphometric traits and germination of seeds of sapucaia, thus contributing to projects focused on species propagation, as well as raising issues on the germination process.

Materials and Methods
Fruits were collected from a population of 6 L. pisonis Cambess individuals growing in Rio Claro,São Paulo,Brazil (22º23'46.08"S and 47º32'47.75"W) in September and October 2009.The fruits were transported to the laboratory and left on a bench under two 40 W luorescent tubes at room temperature (22-28°C), and the seeds were, then, removed from these fruits whose operculum spontaneously detached up to 10 days after harvest.Fruits containing seeds whose opercula had already dropped were also collected.After detachment of the operculum, the seeds were manually extracted and the aril was removed, then, the seeds were placed in plastic trays at room temperature, where they remained for up to 8 days before being used in trials.
In the irst trial (trial I), a sample of 72 seeds was randomly removed from the batch and it was divided into 2 groups with 36 seeds each.To test the possible inluence of the seed coat, seeds from a group were mechanically scariied in a small area (about 0.5 cm 2 ) on the mid-section of the seed using an emery, while the other group of seeds were left intact.Each seed was, then, weighed on a precise weighing scale of two decimal points, wrapped in ilter paper kept saturated with distilled water, and put into six 12 cm x 12 cm x 3.5 cm germination boxes (Gerbox), with 6 seeds per Gerbox.The boxes were placed at 26°C, either under luorescent tubes (34.6 mmol.m -2 .s - ) or in the dark in growth cabinets.Under dark conditions, black boxes were used.The effect of white light was analyzed with regard to possible photoblastic behaviour of seeds, since no information on that issue was found in the scientiic literature.Each seed was daily removed from the Gerbox, inspected for radicle protrusion, weighed, and returned to the same box.Water uptake by seeds (I) was determined by means of the formula: I = [(m tn -m t0 )/ m t0 ].100, where m tn = seed fresh weight at time n and m t0 = seed fresh weight at time zero.As we evaluated the variance normality and homogeneity, the mean I values and the respective standard deviations and interval conidence (a = 0.05) were calculated.
Other trial (trial II) was carried out to evaluate the relation between the emergence of seedlings and seed Demo Version, http://www.verydoc.comand http://www.verypdf.combetween treatments disappeared after that period.It was also observed that the soaking curve of scariied seeds stabilised after 264 h (11 days), while the curve of intact seeds showed a steady, albeit slow, increase throughout the trial (Figure 1).Under the experimental conditions adopted (full water availability and temperature @ 26°C), germination (radicle protrusion) occurred between 15 and 31 days (average of 19.5 ± 1.5 days) after sowing.Within the scariied batch, only 1 seed germinated after 25 days of soaking.The seeds used in the experiments weighed 4.8 ± 0.2 g per seed (fresh weight, FW).The FW of newly collected seeds decreased from 6.2 ± 0.3 g to 4.5 ± 0.2 g after the seeds were left for 8 days on the lab bench at room temperature (22-28°C) and relative air humidity @ 75%, thus corresponding to a moisture loss around 27%.
In trial II, we observed that the seed density showed a low (R 2 = 0.3) correlation with FW and it was not correlated (R 2 = 0.006) with V (data not shown).The sprouting of seeds with densities ranging from 0.3 gmL -1 to 0.70 gmL -1 was null, while in seeds with densities > 0.70 g.mL -1 the proportion of germinated seeds related to N (number of seeds per class) increased according to seed density (Figure 2a).Considering the seed sample size (n = 124), the distribution of densities differed from normal and showed a bi-modal pattern, with a lower number of seeds in the density classes ranging from Demo Version, http://www.verydoc.comand http://www.verypdf.com0.301 to 0.700 g.mL -1 and a large amount of seeds with densities ranging from 0.701 to 1.100 g.mL -1 (Figure 2a).The relation between shoot emergence and fresh weight (FW) was not clear, with seeds from the intermediate FW classes (3.01-6.00g) sprouting proportionally more than those from the heavier class (above 6.01 g) (Figure 2b).The frequency distribution of FW did not differ from normal (K-S, p > 0.2), and it exhibits a left-skewed and platicurtic pattern.
In trial II, where seeds were sown directly in containers with substrate, the average time to shoot emergence was 55 ± 3.3 days.A determination coeficient (R 2 ) of 0.004 was found between germination proportion and fresh weight, while a R 2 = 0.959 was observed for the relation between germination and seed density from 0.601 g.mL -1 .Non-germinated seeds at the end of trials (123 days) were visually inspected and all of them showed deterioration signs.

Discussion
Seeds of Lecythis pisonis were soaked relatively fast in the irst 24 h, in accordance to Pagotto (1982), who reported that hydration of the seeds of sapucaia stabilises at around 3% after this time, probably due to the "extremely hard" seed coat.In this article, we report that the soaking of intact seeds steadily increased throughout the trial period (nearly 22 days) and it attained values > 40% before radicle protrusion.Scariication caused a slight increase in the soaking time course, but it had no effect on the inal seed hydration percentage, showing that, contrary to Pagotto (1982), but in accordance to Carvalho (2006), the seed coat was not the main limiting factor for soaking the seeds of sapucaia, and no pretreatment for germination is required for the species.However, seeds of L. pisonis from different batches or harvests may differ to one another in terms of seed coat resistance to water uptake, as reported for some species (LACERDA et al., 2004).
In this study, we found a fresh weight value of 5.28 g/seed -1 , while Pagotto (1982) and Braga et al. (2007) used seeds which weighed 4.71 g and 8.65 g, respectively.The FW of freshly collected seeds of sapucaia decreased nearly 27% after being left under room conditions, thus, both the time between seed collection and measurement and the storage conditions may inluence on the FW values.According to Carvalho et al. (2006), the decrease of seed moisture from 20.8% to 6.2% did not affect the germination of L. pisonis seeds, which was classiied as orthodox (desiccation-tolerant).It has been reported that the germination (shoot emergence) capacity of seeds of sapucaia ranges from 51% to 68% (CARVALHO, 2006), and it is only moderate (LORENZI, 1992).
Around 72% of the seeds used in this study showed FW values (weight per seed) ranging from 3.37 to 7.36 g, while Pagotto (1982) and Braga et al. (2007) reported the highest frequency of seeds in the ranges from 3.1 to 5.0 g and from 7.0 to 9.9 g, respectively.The volume of individual seeds of sapucaia was correlated with FW, mainly in seeds where germinability was either medium or high, and showed relatively high densities (weight volume -1 ).The germination (shoot emergence) of seeds of sapucaia with density ≥ 0.901g.L -1 was > 60% and it was null at density classes < 0.701 g.L -1 .The relation between seed density and germination capacity is known, and the results found here are similar, for example, to those found by Krieg and Bartee (1975), who reported that the germination of seeds of Gossypium hirsutum L. is positively related to seed density.These authors also reported that the cotton seed quality depends more on seed density than on seed size or weight, and that density is closely related to the embryo maturity status, being more independent of genotype than other seed trait.
Therefore, for obtaining a relatively high percentage of seedlings of sapucaia, we may recommend to sow seeds whose density is > 0.901 gmL -1 , and their germination can reach values > 60%.Seeds of sapucaia with FW < 3.01 g did not germinate, whilst seeds with FW ranging from 3.01 g to 6.0 g produced the largest number of seedlings.However, the relation between shoot emergence and FW was not as evident as it was with regard to the relation between emergence and seed density.Thus, the FW variable, taken in isolation, is not a good indicator of physiological seed quality in L. pisonis, although fresh weight is a variable easier to measure than density.Taking into account the dependence on variation of the seed morphometric traits, the results obtained here need to be corroborated by other regional evaluations in Demo Version, http://www.verydoc.comand http://www.verypdf.comorder to draw strong conclusions, with potential to be applied to a wider range of seed morphotypes.

FIGURE 2 :
FIGURE 2: Frequency distribution of seed density (A) and fresh weight (B) of Lecythis pisonis.N = 124.Open bars = number of sampled seeds; shaded bars = number of germinated seeds.