COMPARISONS BETWEEN TWO ECONOMICALLY VALUABLE FOREST SPECIES ( Eucalyptus grandis AND Pinus taeda ) IN RELATION TO SEED BEHAVIOUR UNDER CONTROLLED DETERIORATION 1

1 Recebido em 03.02.2009 e aceito para publicação em 25.08.2010. 2 Faculty of Agricultural Engineering, UNICAMP, Campinas, Brazil. E-mail: <jussara.fantinatti@gmail.com>. 3 Plant Protection Agency, Sao Paulo State, Campinas, Brazil. E-mail: <usberti@cati.sp.gov.br>. ABSTRACT – The objectives of this work were to analyze seed behaviour under controlled deterioration and estimate viability equations for forest species Eucalyptus grandis and Pinus taeda. Desired moisture content levels were achieved from initial values after either rehydration over water or drying over silica gel, both at 25 oC. Seed sub samples with 8 moisture contents each for E. grandis (1.2 to 18.1%, initial value of 11.3%) and P. taeda (1.5 to 19.5%, initial value of 12.9%) were sealed in laminate aluminium-foil packets and stored in incubators maintained at 40, 50 and 65 oC. The seeds from these species exhibited true orthodox and sub-orthodox storage behaviour, respectively, however E. grandis showed higher seed storability, probably due to a different seed chemical composition. Lowest moisture content limits estimated for application of the viability equations at 65 oC were 4.9 and 4.1 mc for E. grandis and P. taeda, on equilibrium with ±20% RH. The viability equation estimated quantified the response of seed longevity to storage environment well with KE = 9.661 and 8.838; CW = 6.467 and 5.981; CH = 0.03498 and 0.10340; CQ = 0.0002330 and 0.0005476, for E. grandis and P. taeda, respectively. .


INTRODUCTION
Eucalyptus spp.and Pinus spp.are the main sources for supplying raw material for the production of short and long fiber cellulose and mechanical processing and have been established around 3,800,000 ha of brazilian reforested areas.
Seed storability is affected mainly by moisture content (mc) and storage temperature (T).According to Harrington (1963), mc is the most important factor in seed longevity and the viability period could be doubled after reductions of either 1% in mc or 5.6°C in T. Roberts (1973) classified the seeds in storage as orthodox, showing a medium life around 3-15 years, with seed longevity inversely proportional to T and mc whereas recalcitrant seeds have short viability and more difficult conservation.The orthodox seed behavior can be investigated thereby reducing both moisture content and temperature, and their longevity can be described in a quantifiable and predictable way Bonner (1990) proposed a specific classification for forest species: true orthodox, which tolerates desiccation at low T and maintains seed viability for long periods in dry conditions (Eucalyptus, Pinus and Acacia); sub-orthodox, which could be stored in the same conditions above for the maximum of 6 years, presenting high lipid content, small size and fine seed coats; recalcitrant temperate, which are sensitive to desiccation but could be stored at T around freezing (Quercus spp.) and recalcitrant tropical, presenting big seeds and being sensitive to desiccation and freezing (coconut, cocoa and mango).
Several equations have been proposed to relate storage conditions with seed viability, aiming at predicting seed survival as well as seed sensibility to mc and T. The viability equation presented by Ellis and Roberts (1980) ) allows estimates of any seed lot viability during storage at different T and mc and includes five constants, as follows: one quantifying the initial seed lot quality (K i ), one for each species (K E ), one indicating the logarithmic response of seed longevity to mc (C W ) and two constants for linear and quadratic T (C H and C Q , respectively).
The main objectives of this work were (1) analyzing E. grandis W.Hill ex Maiden and P. taeda L. seed behavior under controlled deterioration and (2) estimating viability equations.E. grandis physical purity analysis was performed using a 5-g-working sample, which was separated in pure seeds, other seeds and inert matter.The 1,000seed weight determination for both species was carried out using eight 100-pure seed sub samples (ISTA, 2006).

MATERIAL AND METHODS
Lipid content was detected in 2-g-ground seed samples, placed in Whatman paper cartridges (80 mm length x 33 mm of internal diameter); the extraction was performed with ether of petroleum in a Soxhlet apparatus for 5 h and the amount of lipids was determined gravimetrically (AOAC, 1995).Activity of water was carried out at 25 ± 0.3°C on three replicates of ground seeds for each mc, using the chilled-mirror dew point technique.
Seed mc were adjusted to eight levels from initial values, using desiccators with silica gel or through rehydration over water (4 cm), both at 25°C.Seed sub samples were weighed periodically for controlling the amount of removed or absorbed water, until desired mc had been achieved.After a 4 day-period for stabilization at 4°C in sealed containers, seed mc (%, f.wt.) was determined at 130-133°C during 2 hours for each mc (ISTA, 2006), using three 5 g sub samples of intact (E.grandis) and ground seeds (P.taeda) and the values recorded are listed in Table 1.Comparisons between two economically...
Seeds were sealed in laminated paper (polyester / aluminium / low density polyethylene, with total thickness of 65 µm) with a minimum of 10 x 0.6 g sub samples for E. grandis and 10 x 200 seed sub samples for P. taeda for each combination mc / T and stored at 40, 50 and 65 (± 0.5°C), until complete survival curves were achieved (Table 2).
E. grandis seed germination tests were performed using 4 x 0.15 g sub samples; for P. taeda, a previous dormancy releasing through stratification was proceeded (5°C, 28 days), using 8 x 25 seed sub samples.Seed sub samples were placed over a moistened filter paper in 11 x 11 cm plastic boxes and kept in an incubator under alternate temperature regime (20-30°C, 16-8 hours, 40 W white light at 30°C); seedling counts were recorded at 7 and 14 days (E.grandis) and at 7 and the final one at 28 days (P.taeda).Seeds were considered germinated according to the criterion of the radicle protrusion (SOUZA and CARDOSO, 2000).
Seeds with mc below 6% were previously rehydrated at 25°C during 48 hours over a nylon screen in a plastic box with 1.5 cm of distilled water on the bottom, aiming at avoiding predictable damages due to fast rehydration.
The origin of the survival curves on the Y-axis doesn't converge to the same point, requiring a statistical adjustment in order to get the same value (K i ) with the smallest possible deviation.The germination percentage values were plotted in X-Y graphs and regression lines were fitted to probit values through Y-axis scale transformation.The residual deviances for survival curves at different T were analyzed to give the best fit to the viability equation, using the multiple regression analysis within the generalized linear model (HAY et al., 2003).
The specific constants of each species were achieved using decimal logarithms of sigma (frequency of seed deaths distribution in time or the number of days required for seed germination drop by one probit) and mc for each combination mc / T. Seed germination with or without rehydration was analyzed through a randomized statistical design.

RESULTS AND DISCUSSION
Seed physical pure percentages and 1,000-seed weights detected were 19.0% and 0.226 g (E.grandis) and 100.0%and 25.0 g (P.taeda).Extraction relationship or fruit / seed relationship is the amount of capsules required to produce one kg of mixed seeds and presented a medium rate of 15:1 for E. grandis.Taking into account those values, the numbers of pure seeds / kg were estimated in 841,643 and 40,000 for E. grandis and P. taeda, respectively, which is similar to the result reported by Vieira and Diniz, 1995 (± 700,000 / kg, E. grandis).According to Gentil and Ferreira (2002), the 1,000-seed weight is affected by seed size uniformity, foundation plants and environmental factors.

FANTINATTI, J.B. e USBERTI R.
Seed lipid contents recorded in E. grandis and P. taeda were 24.2 and 52.3%, which were higher than those achieved for Dalbergia nigra and Dimorphandra mollis (7.0% and 14%), respectively (CHAVES and USBERTI, 2004) E. grandis seeds lost water faster over silica gel after 48-h-drying, E. grandis and P. taeda presented seed mc around 6.8 and 7.4% (Figure 1), probably due to the different lipid contents (BROOKER et al., 1992).Seed mc for both species presented an inverse relationship with drying period.The drying process over silica gel revealed positive results for the species and the lowest mc were obtained without damages to the seeds.
No statistical differences were recorded for E. grandis seed germination among mc (Table 3); however the highest and lowest percentages were obtained with 7.5 / 11.3% mc (no rehydration) and with 1.2% mc (with rehydration), emphasizing the need for rehydration before germination test for mc lower than 6%.Lowest P. taeda seed germination percentages were recorded with 8.0 and 1.5% mc, unrelated to rehydration however seeds with 12.9% mc presented a significant increase in germination after rehydration.Bradford (1995) reported that seed germination normally increases after rehydration.
Seeds of P. taeda presented a lower equilibrium relative humidity (erh) than E. grandis, when stored in similar conditions.Oily seeds present erh lower than starchy ones, when exposed to similar environments, since they don't absorb or absorb less water, due to being hydrophobic.
Seed longevity increased logarithmically according to reduction of mc until all weak bound water had been removed.The inflection points of the desorption isotherms for E. grandis and P. taeda (4.9% mc, erh at 22.9% and 4.1% mc, erh at 21.2%, Figure 2) mean the discontinuity of the inverse relationship between longevity and mc or the lowest mc limits for application of the viability  Comparisons between two economically... equation at 65°C.Those values have already been estimated for another species with variable results, as follows: 2.4% mc for peanuts (USBERTI and GOMES, 1998); 6.2% mc for peas (ELLIS et al., 1989); 2.9 and 4.0% mc for D. nigra and D. mollis (CHAVES and USBERTI, 2004) and 3.6% mc for cottonseed (USBERTI et al., 2006).
Probit seed survival curves for the species after storage at 40, 50 and 65°C are displayed in Figure 3, showing the effects of mc and T on seed storability.When mc and T were reduced, predictable seed longevity increases were observed.Ellis (1984) reported that seed survival curves always present the same format, differing among them just in relation to the deterioration period, but presenting the same standard deviation.In the present study the sigmoid shape of the survival curves was strongly affected by temperature and mc increase especially for P. taeda seeds.
Relationship differences between mc and longevity were observed and P. taeda seed sensitivity to mc was higher than E. grandis.Tompsett (1986) worked with two arboreal species and verified that Ulmus carpinifolia Borkh.seed was significantly more sensitive to mc than Terminalia brassii Exell.
Seeds of E. grandis stored at 40°C with 7.5% mc had a sigma of 90 days, while for P. taeda with 8.0% mc that value was reduced to 5 days (Figure 3).The combination of highest T and mc resulted in seed longevity reductions, confirming the tremendous effect of mc on seed longevity.E. grandis seeds stored at 50°C with 7.5% mc presented a sigma of 40 days, while for P. taeda with 8.0% mc that value was reduced to 2 days (Figure 3).In seed samples with 4.7 and 4.0% mc, the calculated sigma values were 55 and 4 days, respectively.Almeida et al., 2005, working with three Pinus species, reported that P. taeda seed germination was not affected after a 2-minute-storage at d" 50°C.
Seeds of E. grandis stored at 65C with 1.2 and 7.5% mc presented sigma values of 40 and 1 days, respectively; however in P. taeda with 1.5% mc that time was 3 days (Figure 3).For seed germination to reach values lower than 5%, with 3.0 and 4.0% mc, 96 and 3.25 days were required, respectively.P. taeda seed germination reduction was higher than for E. grandis throughout the experiment, since oily seeds deteriorate faster than starchy ones.
Seed mc and sigma values after storage at 40, 50 and 65°C are presented in Table 4. E. grandis seeds with 18.1% mc stored at 40 and 50°C presented sigma values of 0.4 and 0.1 day; in P. taeda with 17.0% mc stored at 40 and 50°C, those values decreased to 0.009 and 0.002 day.
Constraining 24 survival curves to the same origin point on the Y-axis yielded different Ki values for E. grandis and P. taeda (1.351 and 1.819; 90 and 98% seed germination).Significant effects for log mc and T were observed while non-significant ones were calculated for the interaction log mc / T (F = 0.47 P. taeda; 3.08 E. grandis).
The constants K E and C W estimated for E. grandis (9.661 and 6.467) were different from those calculated for P. taeda (8.838 and 5.981), confirming the observation done by Ellis (1984) that for cereal seeds those values are higher than oily seeds, indicating the need for drying those seeds to lower mc levels.The constants that reflect seed sensitivity to T (C H and C Q ) were 0.03498 and 0.0002330 (E.grandis) and 0.10340 and 0.0005476 (P.taeda).
The viability equation constants for E. grandis appear to be close to that reported for cottonseed (K E = 9.240, C W = 5.190, C H = 0.03965 and C Q = 0.000426) (USBERTI et al., 2006).An inverse relationship was observed between mc and seed longevity for both species.
Figure 4 shows the logarithmic relationship between seed mc and sigma at 40, 50 and 65°C and reveals the parallelism among the straight lines, due to different T used.Dickie et al. (1990) reported that the viability equation could be applied from -13 to 90°C and also observed a negative relationship between C W values and seed oil content, among different species.Medeiros et al. (1998), using K E = 7.5498 and C W = 3.76 and adopting C H = 0.0329 and C Q = 0.000478 as universal constants, estimated the sigma value of 1,167 years for the species Astronium urundeuva (Fr.All.)Engl.stored at -20°C with mc in equilibrium at 15°C and 15% RH.
Using the viability equations and adopting 4°C (domestic refrigerator) and 7.5 and 8.0% mc (erh at 40 and 45%), achieved in a 2-day drying period over silica gel, it is expected that E. grandis and P. taeda seeds would present sigma values of 19.9 and 2.9 years.That different performance could be due to the different chemical composition of those seeds, since P. taeda presented higher lipid content.Seed moisture content (log %, f.wt.)Sigma (log)

CONCLUSIONS
. P. taeda and E. grandis presented, respectively, sub-and true orthodox seed behaviour to storage.
. P. taeda revealed erh and seed storability lower than E. grandis.
. Lowest limits for application of the viability equation at 65°C were 4.9 and 4.1% mc for E. grandis and P. taeda, respectively.

Figure 4 -
Figure 4 -Relationship between the logarithm of seed mc and the logarithm of sigma for E. grandis (A, on the left) and P. taeda (B, on the right), stored at 40, 50 and 65°C.Solid lines represent slopes, which are constrained to one common value.The dotted lines show different trends at 65°C, below the lower mc limit for application of the viability equation, pointed out by the arrows.Figura 4 -Relações entre o logaritmo de grau de umidade da semente e o logaritmo de sigma para de E. grandis (A, à esquerda) e P. taeda (B, à direita), após o armazenamento a 40, 50 e 65 °C.Linhas sólidas representam curvas, após serem direcionadas a um ponto comum.As linhas pontilhas mostram diferentes tendências a 65 °C, abaixo do limite inferior de grau de umidade para a aplicação da equação de viabilidade, indicados indicado pelas setas.

Table 1 -
Seed moisture content analyzed for the species.

Table 3 -
Seed germination (%) of E. grandis and P. taeda after drying over silica gel.