To design conservation strategies, the extent of plant richness of tropical forests needs to be characterized in terms of their seed longevity. In this study, we examined the potential seed longevity, that is, storage ex situ, of species from south-eastern Mexico: Chamaedorea glaucifolia, Cymbopetalum baillonii, Magnolia mexicana, Nectandra coriacea, and Ternstroemia tepezapote. Immediately after collection, seeds were stored at different temperatures (≤23℃). We evaluated seed germination after different storage durations. Seed water content (WC) was determined for each period. Seed desiccation sensitivity was determined as WC50, which is the WC at which the initial seed viability decreases to 50%; further, the time required to reach WC50 was also determined. Subsequently, we analyzed the relations between seed functional traits with other morphological and functional traits, along with the weather characteristics of their respective habitat. All of the studied species had short-lived seeds; they exhibited desiccation sensitivity after storage with differences across the species. Additionally, C. baillonii exhibited differences in seed desiccation sensitivity across 2 years of seed collection. Interaction was observed between storage time and storage temperature: Seeds exhibited less deterioration at 15℃ in C. glaucifolia and C. baillonii and at 5℃ in M. mexicana and N. coriacea. Seed storage behavior is discussed in this article. Finally, a relationship determined between germination traits, and seed WC, embryo size, endosperm amount, and rain and temperature patterns in the month of seed dispersal explained the limited longevity of the studied species.
Introduction
Plants with short-lived desiccation-sensitive seeds form an important biological group in tropical forests. They account for ∼15% to 19% of global plant species richness (Wyse & Dickie, 2017). However, in tropical environments, they constitute ∼50% (Tweddle, Dickie, Baskin, & Baskin, 2003). Generally, these seeds are large and have a low seed coat ratio (ratio of endocarp and seed coat mass to dispersal unit mass) and high water content (WC) at dispersal (Vázquez-Yanes & Orozco-Segovia, 1993; Hamilton, Offord, Cuneo, & Deseo, 2013). Further, they maintain a high metabolic rate even after their dispersal from the mother plant (Berjak & Pammenter, 2008), so they generally germinate at a fast rate. These reasons explain their short potential longevity (lifespan under optimal environment storage conditions; Vázquez-Yanes & Orozco-Segovia, 1993). Therefore, they are also classified as recalcitrant seeds, or they may be classified as intermediate seeds if they can survive considerable dehydration, but fail to survive conventional subfreezing storage temperature (Hong & Ellis, 1996). Storage difficulties coupled with potential environmental modifications due to climatic change and land use turnover make these species the central concern in terms of the conservation and restoration of tropical forests (O’Brien, Philipson, Tay, & Hector, 2013). Thus, it is necessary to evaluate the potential longevity and degree of seed desiccation sensitivity in tropical species as a first step toward defining their optimal ex situ management and susceptibility to habitat change.
Once short-lived desiccation-sensitive seeds are dispersed, their initial WC begins to decrease. At a certain point, the water deficit of the seeds triggers physiological, structural, and molecular damages at the cell level (Farrant, Berjak, & Pammenter, 1985; Obroucheva, Sinkevich, & Lityagina, 2016). These damages are reflected by a loss of vigor, which is a property that is expressed by the rate and the uniformity of seed germination (Rajjou et al., 2012). In this manner, the desiccation sensitivity of the seed can be identified with the determination of its “critical water content” or WC50 (King & Roberts, 1979) as well as the time elapsed to reach this value. The WC50 is the WC at which the initial seed viability decreases to 50% (Hill, Edwards, & Franks, 2012). Recalcitrant seeds have a WC50 in the range 25% to 20% and intermediate seeds ranged between 10% and 5% (Hong & Ellis, 1996). In comparison, orthodox seeds can reach a WC < 5% without a reduction of their initial viability (Hong & Ellis, 1996). Conversely, variability in WC50 is related to the morphological and physiological traits of seeds (Hill et al., 2012) and with the weather characteristics of the habitat (Dussert et al., 2000; Rodríguez, Orozco-Segovia, Sánchez-Coronado, & Vázquez-Yanes, 2000). Moreover, WC50 can vary according to the dehydration rate of seeds. A slow dehydration rate allows seeds to remain in a range of WC that allows germination metabolism to occur, but this causes an accumulation of metabolic and structural damage that increases desiccation sensitivity. Therefore, WC50 will be high during a fast dehydration rate (Farrant et al., 1985; Berjak & Pammenter, 2008). During storage, seed dehydration rate is slow compared with the fast dehydration techniques (e.g., using silica gel or an air fan). However, slow drying can provide us information regarding what occurs in the habitat and where slow drying conditions are present (Vázquez-Yanes & Orozco-Segovia, 1994).
The dehydration rate can be modified by external factors, such as relative humidity and temperature (Berjak & Pammenter, 2008). Reduction of the temperature slows down the metabolic activity, prevents germination, and reduces rate of water loss and cell damage of seeds (McDonald, 2004). Nevertheless, the high WC and metabolic activity of short-lived seeds render them sensitive to chilling damage even at temperatures above 0℃, that is, arrest of the enzymatic reactions and structural damage of cell membranes (Tommasi, Paciolla, Concetta de Pinto, & De Gara, 2006). Further, if the temperature drops to 0℃, freezing damage can occur (Hong & Ellis, 1996). Thus, the recommended temperature for storage of tropical recalcitrant seeds is between the optimum for germination and the temperature at which no chilling damage occurs (≥10℃; Hong & Ellis, 1996). Additionally, these seeds can exhibit interannual variability in chilling sensitiveness (Berjak & Pammenter, 2008).
Mexico has tremendous plant diversity, but knowledge regarding the seed biology of most of the wild tropical species is extremely limited (Vázquez-Yanes, Batis-Muñoz, Alcocer-Silva, Gual-Díaz, & Sánchez-Dirzo, 2001). Determination of seed longevity and the desiccation sensitivity of wild species can increase the probability of tropical forest conservation (Hamilton et al., 2013; Wyse & Dickie, 2017) and being able to determine the effects of future climatic changes on the habitat of various species (O’Brien et al., 2013). In our research, we evaluated the seed longevity of five species from tropical forest in south-eastern Mexico. This research concentrates on the following questions: (a) What are the effects of storage time on seed germination? (b) What are the effects of storage temperature on seed longevity? (c) Does the desiccation sensitivity of seeds vary between years of collection? (d) What was the storage behavior of the studied species? (e) How are seed traits related to seed longevity, that is, WC50 and the time required to reach WC50 and other functional and morphological traits of seeds, in relation to the environmental factors of the habitat of the studied species?
Material and Methods
Seed Collection and Study Site
Fruit collection was done during the dispersal seasons of 2015 and 2016, in two localities with tropical forest, in south-eastern Mexico. One of them was the tropical rain forest at the UNAM Tropical Biology Station, localized in San Andrés Tuxtla (18°34′5″ N, 95°04′26″ W; 155 m asl). The second locality was the transition region between the tropical rain forest and the dry forest in Ocozocoautla, Chiapas, within the confluence area of the Central Depression of Chiapas and the North Mountains situated in this state (16°51′18″ N, 93°23′47″ W, 904 m asl). This site constitutes a part of the buffer area of the El Ocote Biosphere Reserve. The annual mean precipitation is 4,725 mm for Los Tuxtlas and 1,100 mm for El Ocote. The annual mean temperature is 24℃ for Los Tuxtlas and 23.4℃ for El Ocote (Instituto Nacional de Estadística, Geografía e Informática, 2003; Soto & Gama, 1997; Gutiérrez-García & Ricker, 2011).
Study Species
The goals of this research were addressed with five subcanopy species: Chamaedorea glaucifolia H.Wendl. (Arecaceae), Cymbopetalum baillonii R.E.Fr. (Annonaceae), Magnolia mexicana DC. (Magnoliaceae), Nectandra coriacea (Sw.) Griseb (Lauraceae), and Ternstroemia tepezapote Cham. & Schltdl. (Pentaphylacaceae). The taxonomic status of these species is in accordance with The Plant List (2013). These species are shade tolerant and inhabit mature forests (Standley & Steyermark, 1946, 1949; Coates & Estrada, 1988; Becerra-Vázquez, Ramírez-Marcial, & Holz, 2011). Additionally, these species have local utility for the human settlements and potential economic value (Escobar-Ocampo & Ochoa-Gaona, 2007). All these species, except the understory palm C. glaucifolia that has a seed covered by the fruit pericarp (Corner, 1976) as its dispersal unit (henceforth, we will consider it as a seed), are trees. We collected seeds of C. baillonii in Los Tuxtlas, Veracruz, and seeds of C. glaucifolia, M. mexicana, N. coriacea, and T. tepezapote in El Ocote. Seeds characteristics and other biological and ecological traits of the studied species are presented in Figure 1 and Table 1.
Figure 1.
Fruits (upper row) and seeds (bottom row) of the studied species. (a and b) Chamaedorea glaucifolia, (c) and d) Cymbopetalum baillonii, (e and f) Magnolia mexicana, (g and h) Nectandra coriacea, and (i and j) Ternstroemia tepezapote. Image credit: A. G. Becerra-Vázquez.
Table 1.
The Ecological Traits of the Studied Species, Environmental Traits of Their Habitat, and Seed Morphological Traits and Dates of Seed Collection and Season of Seed Dispersal.
Processing of Fruits and Seeds
We collected fruits of the studied species during the dispersal season of 2016 except for C. glaucifolia which seeds were collected in 2015. Additionally, in 2016, we collected seeds of C. baillonii and M. mexicana during the dispersal season of 2015. The dates of collection are shown in Table 1. We collected fruits directly from at least 10 mature trees. Immediately after collection, fruits were deposited in either plastic containers or black plastic bags and covered with a soil layer taken from the study area to avoid seed dehydration. After 2 days, fruits were taken from the recipients in a dark room (22 ± 0.9℃, 50 ± 3.2% RH). Seeds inside fleshy fruits were cleaned in the laboratory. Because fruits of M. mexicana are dry and woody, we placed them on a table in a dark room until fruit dehiscence occurred. Seeds with an aryl were cleaned by gentle rubbing on a fine steel mesh.
Morphological Traits of Seeds
We measured the length (L), width (W), and thickness (T) of recently collected seeds (RC-seeds, n = 30) with an electronic vernier caliper (accuracy = 0.01 mm). Subsequently, seed volume (V) was calculated with the formula for obtaining the volume of an ellipsoid (Cerdà & García-Fayos, 2002) as follows: V = 1.333 × π × (L/2) × (W/2) × (T/2).
We measured the fresh weight of individual seeds (FWt1) with an electronic analytical balance (model A-200DS, precision 0.001 g, Fisher Scientific, Fairlawn, NJ). Subsequently, individual seeds were dried in an oven (model 107801, Boekel Industries, Inc. Philadelphia, PA) at 80℃ for 48 hr to avoid seed combustion, due to the low ignition point of seed lipids. Seeds were weighed again (DWt1), and the initial water content (WCt1 of RC-seeds) was calculated. The seed WC was determined on dry basis (WCdb) as follows: WCt1 = ((FWt1 – DWt1)/DWt1) × 100 (Equation 1).
Seed Storage
Seeds were stored in closed glass jars with three replicates for each species. Germination of stored seeds (St-seeds) was evaluated after different storage times depending on the species (storage times, t2 … tn). For C. baillonii and M. mexicana, we also evaluated the effect of the year of seed collection: 2015 (2015-S) and 2016 (2016-S) fruiting seasons. The jars were stored in laboratory conditions (23 ± 0.5℃, 50 ± 2.2% RH). To evaluate the effect of storage temperature, we maintained additional jars in growth chambers with controlled temperature (40 ± 5% RH, Labline Instruments Inc., Melrose Park, IL) at 15℃ and at 5℃. In addition, C. glaucifolia and C. baillonii 2015-S were stored at 10℃. Seeds of C. glaucifolia were stored for 30, 101, and 426 days, seeds of C. baillonii for 30, 180, and 360 days (also 730 days for 2015 seeds), seeds of M. mexicana were stored for 25 and 210 days (and for 90 days only for 2015-S), seeds of N coriacea for 35, 90, and 180 days, and seeds of T. tepezapote for 16, 41, 251, and 365 days. All species differed in t2 … tn, according to the seed collection dates and the number of collected seeds.
Prior to storage, seeds were treated with fungicide (Interguzan 30-30, pentachlorinenitrebenzene, and terametiltiuram disulphide). Subsequently, the jars were closed and sealed with a plastic film and placed in the storage site. The jars were reviewed and aerated periodically to prevent fungal contamination or avoid the increase in HR inside the jars due to the seed WC loss. After each storage time (t2 . . . tn), a sample of seeds were sown on agar, and another sample (n ≥ 4) was utilized to determine the WC (WC t2 … tn). For this purpose, Equation (1) was used, with the seed FW and DW data for each t2 … tn (FWt2 … tn and DWt2 … tn, respectively).
Seed Germination
Before (RC-seeds) and after storage (St-seeds), the seeds were sown on plates of 1% agar (10 g/L agar/water; Bioxon, Becton Dickinson de México S.A. de C.V., México). The agar medium was placed in transparent plastic boxes (12 × 16 × 5.5 cm). Before sowing, seeds were disinfected with a 1% sodium hypochlorite solution and, subsequently, with a 0.2% fungicide solution (Interguzan 30-30, pentachlorinenitrebenzene and terametiltiuram disulphide). All the sown boxes (30 seeds per box) were placed in growth chambers (25℃, 12/12 h photoperiod). Germination took place when the radicle protruded. Germination was registered every third day until no germination took place. Seeds that did not germinate were fully covered with fungal infection or were rotten. We used three replications for species for each germination essay.
Data Analysis
To compare the mean values for the different morphological traits of seeds (V, FW, DW, and WC), considering all the years of collection, we applied the Kruskal–Wallis test, because the assumptions required for analysis of variance (ANOVA) were not met with the available data. Tukey and Kramer’s test constituted the post hoc comparison. These analyses were conducted with R software, version 3.2.3 (R Core Team, 2016).
For each species, we obtained the germination parameters relative to each germination test of RC-seeds and St-seeds. First, the cumulative germination percentages were arcsine transformed and fitted to the exponential sigmoid curve y = a / [1 + b(−cx)] using the Table Curve 2D software, version 5.01 (AISN Software, Chicago, IL). All the fitted curves had R2 ≥ 0.9 and p ≤ .01. From the fitted curve, we obtained the lag time (LT), maximum germination rate (MGR; i.e., maximum first derivative of the sigmoid curve), and mean germination time (MGT; i.e., time for maximum germination rate). These variables, along with the final germination (FG), were the germination parameters.
The mean values of the germination parameters of RC-seeds and St-seeds were compared, considering each storage time and each storage temperature (temperature). For this, we applied ANOVA tests, followed by the Tukey’s test for post hoc comparisons. When the variance analysis assumptions were not met, the Kruskal–Wallis and post hoc Tukey and Kramer’s tests were applied. To evaluate if there was any interaction between storage time and temperature, we employed a two-way ANOVA, and for C. baillonii and M. mexicana, a three-way ANOVA was conducted considering the year of seed collection. These analyses were conducted with R. Additionally, to determine the effect of temperature on seed germination (for 2015-S in C. glaucifolia and M. mexicana and for 2016-S in N. coriacea and T. tepezapote and for both years in C. baillonii), a multiple regression analysis with temperature and storage time as factors and the same intercept value (FG before storage or t1) was performed. We applied the square root or logarithm transformation to satisfy the assumptions of the regression analysis. Subsequently, ANOVA test was employed to determine the statistical significance of each slope values. This analysis was conducted with Statgraphics Centurion XVI version 16.1.03 (Statpoint Technologies, Inc.).
To evaluate seed longevity and desiccation sensitivity between species, we determined both WC50, that is WC when the initial viability decreases to 50%, and the time required to reach WC50 (in days), respectively. To determine WC50, the values of seed viability (FG) and seed WC for each storage time (t2 . . . tn) were fitted to the functions indicated in Appendix B. In the case of C. glaucifolia, we determined the WC50 from the fitted curve (FG vs. WC), because the seed FG achieved after 14 months of storage was higher than 50%. In addition, the time of storage at which seeds reach WC50 was calculated using a regression analysis (WC vs. storage time). All fittings were done with Table Curve 2D. To enable comparisons of WC50 and the time required to reach WC50 values between species, we selected the seeds from the same year and stored at the same temperature. For this, we considered 2016-S (2015-S in C. glaucifolia) stored in laboratory conditions (23 ± 0.5℃, 50 ± 2.2% RH, and only for N. coriacea, to avoid seed germination during storage, these seeds were stored at 15 ± 0.5℃, 40 ± 5% RH). The mean values of WC50 and the time required to reach WC50 were compared with ANOVA and followed for the Tukey’s post hoc test. Also, we made comparisons between years of seed collection for both C. baillonii and M. mexicana (2015-S and 2016-S, respectively). Therefore, we applied the t-test or Wilcoxon rank sum test in cases of no normality and or homocestadicity of the data. Analyses were performed with R.
Finally, we performed a principal component analysis to explore the relationships between species in terms of (a) morphological traits, as seed DW, volume, initial WC, relative amount of endosperm, and embryo size, (b) functional traits, as seed WC50, time required to reach WC50, and initial values of LT, MGR, and MGT, and (c) ecological and environmental traits, as the season of seed dispersal and both mean maximum temperature and total precipitation for the month of seed collection (Table 1). Before the analysis, all variables were normalized. For this analysis, we included data for 2016-S of all species and 2015-S for C. glaucifolia. The analysis was also done with R.
Results
Morphological Traits of Seeds
The values for each morphological trait of the seeds of studied species are presented in Table 2. The volume ranged from 95.7 ± 17.0 to 386.9 ± 87.6 mm3, FW from 0.16 ± 0.026 to 0.53 ± 0.101 g, and DW from 0.11 ± 0.019 g to 0.34 ± 0.075 g. C. glaucifolia had the smallest and lightest seeds, while C. baillonii had the largest and the heaviest seeds in terms of FW and DW. Seeds of M. mexicana had the lowest WC value (19.6 ± 5.63%). For the other species, WC was > 36%; seeds of T. tepezapote had the highest value (75.3 ± 13.81%). Seeds of C. bailloni collected in 2015-S were significantly heavier in FW than those from 2016-S (0.62 ± 0.057 and 0.53 ± 0.101 g, respectively).
Table 2.
Seed Morphological Traits (n ≥ 30) of the Studied Species ().
Seed Storage
Before seed storage (t1), final germination (FG) ranged between 67% in T. tepezapote to 94% in C. glaucifolia (Table 3). The LT ranged from 8 days in N. coriacea to 26 days in T. tepezapote (Table 4). MGR ranged from 1.1% day−1 in C. glaucifolia to 5.4% day−1 in C. baillonii (2016-S; Table 3). MGT ranged from 22 days in C. baillonii (2016-S) to 89 days in the seeds of C. glaucifolia (Table 4).
Table 3.
Final Germination and Maximum Germination Rate of Seeds () After Different Storage Times or Durations (t2 . . . tn) in Different Storage Temperatures (℃) for Five Species From Tropical Forests of Mexico.
Table 4.
Lag Time and Mean Germination Time of Seeds () After Different Storage Times or Durations (t2 . . . tn) in Different Storage Temperatures (℃) for Five Species From Tropical Forests of Mexico.
Table 5.
Results of the Multiway ANOVA on Effects of Year (Y), Storage Time (Sti), and Storage Temperature (ST) on Germination of C. baillonii and M. mexicana Seeds.
After storage (St-seeds), both storage time and temperature exerted a significant effect on all germination parameters (FG, LT, MGR, and MGT) and in all species, except C. glaucifolia (Tables 3Table 4.Table 5. to 6). As the storage time increased, both FG and MRG decreased, and LT and MGT increased. There was a negative relation between the duration of storage and seed viability of all species and all storage temperatures (Figure 2). In contrast, in C. glaucifolia, only storage time had a significant effect on the MGT and MGR (Table 6). In this species, the MGT of St-seeds decreased with time, and the MGR increased after 101 days of storage, but after 426 days, the MGR decreased.
Table 6.
Results of the Two-Way ANOVA on Effects of Storage Time (Sti) and Storage Temperature (ST) on Germination of C. glaucifolia, N. coriacea, and T. tepezapote Seeds.
Figure 2.
Final germination (%, ± SE, n = three replications) of seeds of the studied species after different storage times (t2 . . . tn) at different storage temperatures. The evaluation was done for seeds collected in 2015 (Chamaedorea glaucifolia and Magnolia mexicana), 2016 (Nectandra coriacea and Ternstroemia tepezapote), and those collected in both years, Cymbopetalum baillonii. Seeds were stored in closed glass jars at 23℃ (room temperature, black square symbol), 15℃ (white square), 10℃ (up, white triangle), or 5℃ (down, gray triangle). The values of multiple linear regression analysis are presented in Appendix A.
Results of multiple linear regression showed that slopes of linear regression between FG and storage time differed significantly between temperatures for each species, as occurred in all species except in T. tepezapote (Figure 2). This means that the storage temperature had an effect in seed longevity: Lower slope values indicate less seed deterioration (Tables 3 and 4, Figure 2). The St-seeds at 15℃ showed slower deterioration than at 23℃, 10℃, and 5℃ in C. glaucifolia, F(3, 4) = 7.04, p < .05, and in C. baillonii for both years of seed collection, F(2, 3) = 11.13, p < .05 for 2015-S and F(2, 3) = 28.54, p < .01 for 2016-S. The St-seeds at 5℃ showed slower deterioration than at 23℃ and 15℃ in M. mexicana, F(2, 3) = 17.78, p < .01, and at 15℃ in N. coriacea, F(1, 2) = 7.15, p < .05. Although slopes did not differ significantly for T. tepezapote, F(1, 3) = 1.99, p > .05, after 251 days, germination only occurred in seeds stored at 5℃ (Table 3). Summary of multiple regression analysis is shown in Appendix A.
The multiway ANOVA tests applied to the mean values of all parameters confirm in some cases an interaction between the factors year, storage time, and temperature (Tables 5 and 6). Individually, the year of seed collection exerted a significant effect on the St-MGT in C. baillonii (Table 5). In this species, RC-seeds collected in 2016 had a shorter MGT than RC-seeds collected in 2015, but in 2015-S seeds, the St-MGT was shorter compared with RC-seeds (Table 4). Conversely, in C. baillonii, we found a triple interaction between the collection year, storage time, and temperature (Table 5). That is, 2016-S seeds stored for 180 days at 23℃ showed the longest St-MGT; whereas, 2015-S did not exhibit significant differences (Table 4). On the other hand, the collection year had a double interaction with both storage time and temperature for all germination parameters of C. baillonii seeds, except St-LG (Table 5). After 180 days of storage at 15℃, the FG of 2016-S decreased. Also, after 30 days of storage at 23℃ for both years, the MGT decreased in 2015-S, but in 2016-S, it increased. Finally, only 2016-S showed differences in St-LT and St-MGR in terms of temperature of storage (Tables 3 and 4).
An interaction between storage time and temperature also was observed for St-FG, St-LT, and St-MGT in seeds of N. coriacea and T. tepezapote (Table 6). Seeds of N. coriacea stored for 180 days at 5℃ had a higher FG and shorter LT and MGT than seeds stored at 15℃, while in T. tepezapote, germination occurred only in seeds stored for 251 days at 5℃ (Tables 3 and 4).
We found significant differences in WC50 values between species, ANOVA, F(4, 10) = 329.1, p < .0001. Seeds of C. glaucifolia had the lowest WC50, 4.6 ± 0.08% (Figure 3). C. baillonii and M. mexicana showed no difference between them in WC50, with 12.7 ± 1.25% and 11.7 ± 0.29%, respectively (Figure 3). Seeds of N. coriacea and T. tepezapote had the highest WC50, with 37.9 ± 1.96% and 59.1 ± 1.87%, respectively (Figure 3). In addition, the time required to reach WC50 varied with the species, ANOVA, F(3, 8) = 272.8, p < .0001. Seeds of M. mexicana required the shortest time, 9.1 ± 0.70 days, followed by T. tepezapote seeds that required 17.7 ± 1.46 days (Figure 3). Seeds of C. baillonii and N. coriacea required 88.4 ± 9.27 and 150.9 ± 3.78 days, respectively (Figure 3). After 426 days, the seeds of C. glaucifolia showed a reduction in their FG from 100% to 57%; thus, they had the longest WC50 (Figure 3). Functions used for the determination of WC50 and time to reach WC50 are shown in Appendix B.
Figure 3.
The determinations of the critical WC (WC50, Panel A) and the time at which the seeds reached WC50 (B) for the seeds of the species studied. The final mean seed germination ( ± SE, n = three replications) was expressed as a relative percentage with respect to the initial seed germination before storage (germination at t1 = 100%). Seed water content was expressed in a dry weight basis. The arrow on the x-axis indicates the WC50 value of seeds (A) or the time required to reach WC50 (B) for each species (abbreviated names in bold and italicized letters).
We found significant differences in WC50 between the 2 years for C. baillonii (t = − 20.55, p < .0001). The 2015-S seeds had lower WC50, compared with 2016-S seeds, with 4.8 ± 0.07%; but there was no difference in the time required to reach WC50 (t = 1.3098, p > .05). No differences were found in M. mexicana between 2015-S and 2016-S with respect to its WC50 (t = 1.7132, p > .05) and the time to reach WC50 (W = 0.0, p > .05).
Relationships Between Biological, Ecological, and Environmental Traits
The principal component analysis showed that Components 1, 2, and 3 accounted for 94% of the total variation (Figure 4, Appendix C). Component 1 explained 38% of the variation and was represented for the positive loadings of seed DW, MGR, volume, and mean maximum temperature in the month of seed collection, followed by negative loadings of seed MGT and LT (Figure 4, Appendix C). Thus, large seeds germinate fast and are dispersed in the hottest months. Component 2 explained 34% of the variation and was represented for positive loadings in seed FG, the time required to reach WC50, relative amount of endosperm, dispersal time, maximum mean temperature, and total precipitation in the month of seed collection, followed by negative loadings of seed WC, WC50, and embryo size (Figure 4, Appendix C). Therefore, seeds with low WC, small embryos, and an abundance of endosperm tended to have less desiccation sensitivity, high viability (final germination), and were dispersed in wet months with high maximum temperatures. Component 3, which explained 21% of the variation, had positive loadings in seed LT and relative amount of endosperm, followed by negative loadings in seed WC, embryo size, dispersal time, and total precipitation in the month of seed collection (Figure 4, Appendix C). Thus, seeds with low WC, abundance of endosperm, small embryo, and those dispersed in dry seasons take more time to germinate.
Figure 4.
Relations between ecological, morphological, and functional, as well as environmental traits of the studied species, seeds, and habitat (tropical forest of Chiapas and Veracruz). Eigenvalues and eigenvectors are showed in Appendix C. FG = final germination; LT = lag time; MGR = mean germination rate; MGT = mean germination time; DW = seed dry weight; WC = seed water content dry basis; WC50 = WC at which viability decreased by 50% from the initial viability; TWC50 = time to reach WC50; Endosperm = relative amount of endosperm in the seed; Embryo = embryo size; Volume = seed volume; PP-MSC = total precipitation in the month of seed collection; Temp-MSC = mean maximum temperature in the month of seed collection; Dispersal = season of seed dispersal.
Discussion
All species had short-lived desiccation-sensitive seeds, because their vigor and viability decreased after storage. These functional traits, along with morphological and physiological traits, were in accordance to those reported for tropical species with short-lived desiccation-sensitive seeds (Pritchard et al., 2004; Hamilton et al., 2013), but we found variation in seed longevity between them. Indeed, we found that seed longevity was related to other functional and ecological traits, along with the prevailing weather conditions at the time of seed dispersal. Seed WC and internal structure are related to longevity (Hong & Ellis, 1996; Hill, Edwards, & Franks, 2010). Moreover, the amount of precipitation and temperature influence the seed development (Finch-Savage & Farrant, 1997).
Seeds of both T. tepezapote and N. coriacea had WC50 > 30%, which signifies that both species might be recalcitrant, but they differ greatly in terms of the time required to reach WC50 (17 and 150 days, respectively). However, WC50 was determined in T. tepezapote seeds stored at 23℃, as was done in the other species studied, while we pointed out that N. coriacea seeds were stored at 15℃. This difference in storage temperature was because in 2014 seeds of N. coriacea with 78.8 ± 2.93% germination did not germinate after 90 days of storage at room environment (23 ± 0.5℃, 50 ± 2.2% RH; A. G. Becerra-Vázquez, personal observation, January, 2015). Thus, a storage temperature of 15℃ may have entailed a longer time to reach WC50. Therefore, the time required to reach WC50 for seeds of N. coriacea at 23℃ might be less than 90 days, while at 15℃ seeds of N. coriacea had a FG of 13% after 180 days. Regardless of this, the WC50 values of both species are consistent with those reported by Hong and Ellis (1996) for tropical recalcitrant seeds. In case of T. tepezapote, a closely related species T. brasiliensis has seeds with ecological longevity below 60 days (Pires, Cardoso, Joly, & Rodrigues, 2009). Inside the tropical forest, temperature could be almost constant above and beneath the litter (Vázquez-Yanes & Orozco-Segovia, 1994), closer to 23℃ than to 15℃. Among the other Nectandra species, some are classified as recalcitrant (de Carvalho, Davide, Silva, & Carvalho, 2008).
Even though M. mexicana seeds had low WC50 (11.7%), the short longevity of its seeds might be related to the anatomical and functional traits of both seeds and fruits. A small embryo is a trait of the family Magnoliaceae (Niembro-Rocas, 1989), and morphological dormancy is reported for species of Magnolia (Royal Botanic Gardens Kew, 2017). Nevertheless, in our study, seeds of M. mexicana that did not germinate were rotten. Thus, its small embryos might be highly susceptible to dehydration, without entering into dormancy state, despite the large size of the surrounding endosperm (Alcántara-Flores, 2002). On the other hand, seeds of M. mexicana had lower WC50, as in C. glaucifolia and C. baillonii, but M. mexicana seeds took shorter time to reach WC50 (9 days) compared with both species (> 80 days). Therefore, M. mexicana seeds might be classified as recalcitrant. Seeds of M. ovata, a species closely related to M. mexicana (Figlar & Nooteboom, 2004), loose water rapidly during dehydration in silica gel (∼10% from initial 100%), and after this dehydration, the FG decrease from 84% to 19% (José, Da Silva, Davide, Melo, & Toorop, 2011). Pupim et al. (2009) found that M. ovata seeds have a relatively low WC (WCdb = 30% calculated from WCfb, according to Caddick (2005)). This species has a dehiscent dry fruit that exposes seeds to drying before dispersal, thus rendering them highly susceptible to dehydration (José et al., 2011). This was observed in field in M. mexicana.
Seeds of C. baillonii had low WC50 (12%) and required 80 days to reach WC50. Therefore, its seeds might be recalcitrant. Seeds of C. baillonii have been reported to be desiccation-sensitive (Rodríguez et al., 2000). As pointed out earlier, seeds of M. mexicana and C. baillonii exhibited no difference in their WC50; but the fact that seeds of C. baillonii required a longer time to reach WC50 than M. mexicana is interesting. Both species have small embryos and abundant endosperm (Niembro-Rocas, 1989). However, they differ in seed anatomy, because the seed coat of C. baillonii has an inner tegument with multiple folds, and these folds extend into the fissures present in the endosperm tissue, that is, ruminate endosperm (Niembro-Rocas, 1989). Also, C. baillonii seeds were larger than those of M. mexicana. However, a single factor such as seed size cannot predict longevity in desiccation-sensitive seeds (Hill et al., 2012), influencing partially the dehydration rate of the seed, as it occurs in some tropical species (Hill et al., 2010). Large seeds have a great seed surface to mass volume ratio (Cleri, 2016); thus, they dehydrate at a slower rate than small seeds. Therefore, seed structure and size might explain the differences in seed longevity of C. baillonii.
We found annual variation in desiccation sensitivity in seeds of C. baillonii: 2015-S had the lowest WC50 compared with 2016-S. Weather conditions have influence on seed development. Indeed, RC-seeds from 2015 had higher FW than 2016. Seeds produced in different years can exhibit variation in their morphological and functional traits, as is the case of some tropical species (Sánchez-Coronado et al., 2007; Lamarca et al., 2016). Thus, weather conditions might have a similar effect during the development of the seeds of C.baillonii, because 2014 (the development year for seeds collected in 2015) had higher monthly mean maximum temperatures compared with 2015 (climatic data from Torre CONAGUA-SMN—CONANP—IBUNAM, Comisión Nacional del Agua, 2017). On the other hand, we found that seed vigor for C. baillonii increased after 30 days of storage for seeds collected in 2015 but not in 2016. This result agrees with that of Rodríguez et al. (2000), which found that a previous mild dehydration of the seeds of C. baillonii increased germination rate and final germination. The improving of germination velocity with the mild dehydration of seeds before germination is also found in other tropical species. (Eggers, Erdey, Pammenter, & Berjak, 2007; Rodríguez et al., 2000). During mild dehydration, seeds might end the seed maturation phase (maturation and dryness; Vertucci & Farrant, 1995) and maintain their active metabolism; so, if after this period they are placed in optimal germination conditions, seed germination rate might increase, as it occurs in recalcitrant seeds of Avicennia marina (Farrant et al., 1985).
Seeds of C. glaucifolia had the lowest value of WC50 (4.6%) and required the longest time to WC50 (∼426 days). This seems to be closely related to the values reported for intermediate seeds (Hong & Ellis, 1996), so C. glaucifolia might have this storage behavior. Seeds of C. elegans lost >50% of their initial viability when WC of seed reach 16%; while in the less sensitive C. microspadix seeds, the loss of >50% of initial viability occurs when seeds reach WC of 7% (Carpenter & Ostmark, 1994). Therefore, seeds of C. glaucifolia had a desiccation response similar to that of C. microspadix. Conversely, seeds of C. glaucifolia are the smallest compared with the other studied species; however, they might have a high dehydration rate corresponding to their high surface/volume ratio (Cleri, 2016). A structural trait in seed of C. glaucifolia might be related to its longer time required to reach WC50. The fruit and seed tissues (endocarp and hard endosperm) that surround its embryo (Corner, 1976) isolate it more than the seed coat does in the other studied species. Internal differences in water distribution along with physical changes in seed structure might explain the unexpected desiccation patterns in seeds, as is the case with other rainforest species (Hill et al., 2010). Finally, morphophysiological dormancy is common in palms (Baskin & Baskin, 2014), such as Chamaedorea spp (Carpenter & Ostmark, 1994). The removal of dormancy might require a seed life span longer than in quiescent seeds that have a faster germination.
Effect of Temperature in Seed Longevity
The longevity of tropical desiccation-sensitive seeds can be extended by storage at ≥10℃ (Hong & Ellis, 1996). In this study, longevity of C. glaucifolia and C. baillonii seeds was longer at 15℃ than at 23℃, 10℃, and 5℃. Nevertheless, seeds of M. mexicana, N. coriacea, and T. tepezapote exhibited slower aging at 5℃ than at ≥15℃. Seeds of these three species have lipid reserves (Niembro-Rocas, 1989), like temperate red oaks with seeds that have a high lipid content (17.54 ± 4.43%, n = 26 species; data included from Bonner & Vozzo, 1987; Xia, Seal, Chen, Zhou, & Pritchard, 2010) and relatively longer viability than white oaks. Tropical species that can also be found in subtropical and even temperate habitats, such as Calophyllum brasiliense and Persea americana, can prolong their longevity at temperatures below 10℃ (Gálvez-Cendegui, Peñaloza, Oyanedel, & Castro, 2017; Nery, Prudente, Alvarenga, Paiva, & Nery, 2017). Moreover, differences in lipid composition in seeds are related to differences in the chilling sensitivity of several species of Cuphea, a tropical genus with intermediate seeds (Crane, Miller, Van Roekel, & Walters, 2003). Thus, to clarify this variation, further research must include a biochemical analysis of seeds that considers phylogenetic affinity and the geographic distribution of species.
Seeds of the studied species, except C. glaucifolia, are probably recalcitrant. Since we did not follow the protocol to determine seed storage behavior (e.g., Hong & Ellis, 1996), further research is needed that includes the storage of seeds dehydrated to a specific WC and placed at temperatures above and below freezing. Moreover, a detailed determination of the presence of seed dormancy is required, as in C. glaucifolia, because it demonstrated a reduction in MGT after storage, which could indicate dormancy removal. The presence of dormancy could mask the degree of desiccation sensitivity in tropical seeds (Rodríguez et al., 2000).
Implications for Conservation
The current deterioration of tropical forest requires the implementation of conservation strategies, but highly threatened species, such as those with desiccation sensitive seeds, clearly require special attention compared with other members of the plant community. Thus, knowledge about desiccation sensitivity in seeds is an essential step before designing in situ conservation and restoration programs for species with short-lived seeds (e.g., seedlings from seed in nurseries, direct seeding). In this study, we discovered that all species had short-lived desiccation-sensitive seeds. Therefore, this seed trait must be considered for conservation and restoration strategies of the species’ habitat. In the same manner, the propagation and storage of seeds of these species must take into account their limited longevity and the temperatures with which the longevity can be extended.
Acknowledgments
The authors thank Posgrado en Ciencias Biológicas for the academic and scientific support and Los Tuxtlas Biological Station for providing logistical support. The authors also thank Ana Elena Mendoza Ochoa for the resources granted, M. Esther Sánchez-Coronado for statistical and methodological support, as well as Amílcar Ovidio Gómez Pérez and Isaías Landa Trujillo for technical and field assistance.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by DGAPA PAPIIT UNAM (No. IN 205715) and Consejo Nacional de Ciencia y Tecnología (No. 221015).
References
Appendices
Appendices
Appendix A.
Results of Multiple Linear Regressions Between Final Germination, Storage Time (Sti) and Storage Temperature (ST) for the Studied Species.
Appendix B.
Functions Used for the Determination of WC50 and Time to Reach WC50, Expressed as the Relation Between the Seed’s Water Content (WC) and Seed Viability (FG) and WC and Storage Time (Sti).
Appendix C.
Eigenvalues and Eigenvectors of Principal Component Analysis Done With Morphological and Functional Seed Traits Along With Ecological and Environmental Traits of the Habitat of the Studied Species.
