Seed Longevity of Five Tropical Species From South-Eastern Mexico: Changes in Seed Germination During Storage

Abstract

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.

Keywords

native plants seed longevity storage seed desiccation sensitivity tropical forest

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.

Species	Seed collection date	Season of seed dispersal^a	Precipitation^b (mm)	Max. Temp.^b (℃)	Embryo size^c	Endosperm amount
Cg	October 2015	Late rainy season***	206.8	26.9	Small⁴	High⁴
Cb	April 2016	Late dry season**	20.8	28.8	Small^1,2	High^1,2
Mm	March 2016	Mid dry season*	0	27.1	Small^1,3	High^1,3
Nc	September 2016	Late rainy season*	104.4	27.8	Large^,2	Absent^1,2
Tt	January 2016	Early dry season*	10.7	25.7	Large^,2	Low^1,2

Note. Cg = Chamaedorea glaucifolia; Cb = Cymbopetalum baillonii; Mm = Magnolia mexicana; Nc = Nectandra coriacea; Tt = Ternstroemia tepezapote; Precipitation = total precipitation in the month of seed collection; Max. Temp. = maximum mean temperature in the month of seed collection; L, W, T = seed length, width, and thickness.

Annual rain pattern in the collection sites (Instituto Nacional de Estadística, Geografía e Informática, 2003; Gutiérrez & Ricker, 2011). Dispersal dates from Becerra-Vázquez et al. (2011)*, Rodríguez et al. (2000)** and personal observations (A. G. Becerra Vázquez)*** in the study site.

Total precipitation and mean of maximum temperature for the date of seed collections from Station 7319 and Torre CONAGUA-SMN – CONANP – IBUNAM (Comisión Nacional del Agua, 2017).

Embryo and endosperm proportions are relative to seed size (¹Corner, 1976; ²Niembro-Rocas, 1989; ³Alcántara-Flores, 2002; ⁴Baskin & Baskin, 2014) and corroborated by personal observations (A. G. Becerra-Vázquez).

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 (FW_t1) 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 (DW_t1), and the initial water content (WC_t1 of RC-seeds) was calculated. The seed WC was determined on dry basis (WC_db) as follows: WC_t1 = ((FW_t1 – DW_t1)/DW_t1) × 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 (FW_{t2 … tn} and DW_{t2 … 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^(−c ^x ⁾] 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 ( $\bar{x} \pm SE$ ).

Species	Year	L × W × T (mm)	Volume (mm³)	FW (g)	DW (g)	WC_db (%)
Chamaedorea glaucifolia	2015	6.1, 5.5, 5.3 (±0.45, 0.30, 0.29)	95.7 ± 17.07e	0.16 ± 0.026e	0.11 ± 0.019d	36.3 ± 2.36c
Cymbopetalum baillonii	2015	15.4, 8.7, 5.5 (±1.27, 0.69, 0.65)	396.6 ± 64.57a	0.62 ± 0.057a	0.37 ± 0.057a	68.6 ± 14.51ab
	2016	14.9, 8.5, 5.7 (±1.27, 0.65, 0.75)	386.9 ± 87.69a	0.53 ± 0.101b	0.34 ± 0.075a	56.3 ± 16.26b
Magnolia mexicana	2015	10.4, 8.4, 4.5 (0.69, 1.37, 0.39)	211.6 ± 38.07c	0.20 ± 0.038d	0.16 ± 0.027c	21.9 ± 4.05d
	2016	10.5, 8.2, 4.6 (±0.74, 0.86, 0.53)	212.9 ± 42.05c	0.22 ± 0.039d	0.18 ± 0.030c	19.6 ± 5.63d
Nectandra coriacea	2016	9.0, 8.0, 7.8 (±0.82, 0.67, 0.63)	302.4 ± 73.02b	0.39 ± 0.098bc	0.25 ± 0.063b	56.5 ± 5.84b
Ternstroemia tepezapote	2016	10.6, 6.4, 4.7 (±0.95, 0.68, 0.77)	167.8 ± 26.6d	0.21 ± 0.04d	0.12 ± 0.021d	75.3 ± 13.81a
			F_{(6, 224)} = 219.2	χ²= 435.04	χ²= 461.21	χ²= 493.7

Note. L = seed length; W = seed width; T = seed thickness; Volume = seed volume; FW = fresh weight; DW = dry weight; WC_db = water content dry basis. Lowercase letters indicate statistical differences inside each column (Tukey’s test or Tukey and Kramer’s test, p < .0001). Year of seed collection for species is indicated.

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 ( $\bar{x} \pm SE$ ) After Different Storage Times or Durations (t2 . . . tn) in Different Storage Temperatures (℃) for Five Species From Tropical Forests of Mexico.

Storage time	Final germination (%)				Maximum germination rate (% × day⁻¹)
(days)	23℃	15℃	10℃	5℃	23℃	15℃	10℃	5℃
Chamaedorea glaucifolia (2015)
	F = 18.66, df = 11.13, p < .0001				F = 4.206, df = 11.23, p < .01
0 (t1)	94.4 ± 4.0a				1.1 ± 0.05c
33	92.2 ± 2.23a	88.8 ± 2.22a	81.1 ± 1.11ab	95.5 ± 2.22a	2.2 ± 0.17abc	2.4 ± 0.07abc	2.1 ± .037abc	2.4 ± 0.04abc
101	74.4 ± 1.11abc	85.5 ± 1.11ab	62.2 ± 9.68bcd	NE	3.3 ± 0.23ab	3.9 ± 0.93a	2.7 ± 0.39abc	NE
426	54.4 ± 2.93cde	63.3 ± 5.77bcd	27.7 ± 8.01e	43.3 ± 5.77de	1.6 ± 0.32bc	2.4 ± 0.52abc	0.7 ± 0.21c	1.8 ± 0.42bc
Cymbopetalum baillonii (2015)
	F = 20.12, df = 7.16, p < .0001				F = 0.898, df = 6.13, p > .05
0 (t1)	84.2 ± 1.92a				4.4 ± 0.36
30	64.4 ± 1.11ab	67.7 ± 2.93a	66.6 ± 8.81a	NE	4.7 ± 0.40	3.6 ± 0.38	3.0 ± 0.42	NE
180	12.2 ± 8.88c	63.3 ± 9.62ab	33.3 ± 8.38bc	NE	3.1 ± 0.63	8.0 ± 4.55	3.7 ± 0.84	NE
730	NG	5.5 ± 2.22c	NG	NE	NG	ND	NG	NE
Cymbopetalum baillonii (2016)
	F = 105.1, df = 8.18, p < .0001				F = 18.03, df = 6.14, p < .0001
0 (t1)	72.5 ± 7.40a				5.4 ± 0.41a
30	82.2 ± 2.93a	85.5 ± 2.93a	NE	91.1 ± 4.00a	5.1 ± 0.15bc	4.6 ± 0.36c	NE	7.6 ± 0.69ab
180	9.9 ± 3.33cd	52.2 ± 7.77b	NE	22.2 ± 2.93c	2.5 ± 0.98cd	3.1 ± 0.18cd	NE	1.5 ± 0.21d
365	3.3 ± 1.92d	8.8 ± 1.11cd	NE	ND	ND	ND	NE	ND
Magnolia mexicana (2015)
	F = 22.72, df = 3.8, p < .001				F = 0.996, df = 3.8, p > .05
0 (t1)	88.8 ± 2.93a				1.92 ± 0.30
25	18.0 ± 3.80b	35.0 ± 12.58b	NE	NE	0.8 ± 0.06	1.2 ± 0.47	NE	NE
90	NG	NG	NE	17.1 ± 4.36b	NG	NG	NE	1.0 ± 0.78
210	NG	NG	NE	NG	NG	NG	NE	NG
Magnolia mexicana (2016)
	F = 47.52, df = 3.8, p < .0001				F = 3.108, df = 2.6, p > .05
0 (t1)	77.7 ± 4.84a				1.9 ± 0.09
25	2.2 ± 2.2c	40.0 ± 6.93b	NE	55.5 ± 2.93b	ND	1.4 ± 0.29	NE	2.6 ± 0.50
210	NG	NG	NE	NG	NG	NG	NE	NG
Nectandra coriacea (2016)
	F = 47.31, df = 6.14, p < .0001				F = 4.523, df = 6.14, p < .01
0 (t1)	82.2 ± 4.84abc				1.5 ± 0.16ab
35	NE	100.0 ± 0.00a	NE	96.6 ± 3.33ab	NE	1.3 ± 0.08b	NE	1.4 ± 0.13ab
90	NE	64.4 ± 4.44c	NE	76.6 ± 5.09bc	NE	1.3 ± 1.16b	NE	2.4 ± 0.21a
180	NE	13.3 ± 1.92d	NE	64.4 ± 6.18c	NE	1.8 ± 0.38ab	NE	2.3 ± 0.26ab
Ternstroemia tepezapote (2016)
	F = 11.25, df = 5.12, p < .001				F = 7.23, df = 5.12, p < .01
0 (t1)	67.7 ± 5.55a				1.6 ± 0.11ab
16	23.3 ± 3.84c	54.4 ± 11.11ab	NE	NE	0.7 ± 0.02bc	1.8 ± 0.37a	NE	NE
41	27.7 ± 1.11bc	53.3 ± 3.33ab	NE	NE	1.3 ± 0.15abc	1.7 ± 0.10a	NE	NE
251	NG	NG	NE	18.8 ± 5.87c	NG	NG	NE	0.6 ± 0.17c

Note. NG = no germination, ND: no determined; NE = no evaluated. Lowercase letters indicate statistical differences (Tukey’s test after ANOVA, p < .05, or Tukey and Kramer’s test after Kruskal-Wallis, p < .05). Comparisons were done individually for each germination parameter for each species and for each year of seed collection of C. baillonii and M. mexicana.

Table 4.

Lag Time and Mean Germination Time of Seeds ( $\bar{x} \pm SE$ ) After Different Storage Times or Durations (t2 . . . tn) in Different Storage Temperatures (℃) for Five Species From Tropical Forests of Mexico.

Storage time	Lag time (days)				Mean germination time (days)
(days)	23℃	15℃	10℃	5℃	23℃	15℃	10℃	5℃
Chamaedorea glaucifolia (2015)
	F = 3.407, df = 11.23, p < .01				χ²= 32.134, df = 11, p < .001
0 (t1)	22.6 ± 2.61ab				89.7 ± 2.71c
33	22.3 ± 1.17ab	25.0 ± 1.79a	26.0 ± 4.52a	26.1 ± 0.65a	55.4 ± 0.67a	55.4 ± 0.67a	57.1 ± 1.91a	61.4 ± 2.61b
101	14.0 ± 1.80ab	18.2 ± 3.25ab	18.0 ± 2.31ab	NE	30.2 ± 0.52a	34.7 ± 0.53a	36.0 ± 0.64a	NE
426	11.9 ± 0.98b	18.7 ± 3.55ab	25.0 ± 8.02ab	23.3 ± 2.48ab	40.9 ± 4.29a	44.9 ± 0.56a	53.4 ± 6.56a	44.3 ± 0.98a
Cymbopetalum baillonii (2015)
	F = 1.964, df = 6.13, p > .05				F = 4.611, df = 6.13, p < .05
0 (t1)	19.6 ± 0.46				35.6 ± 0.93a
30	16.4 ± 1.48	14.0 ± 1.40	11.4 ± 1.15	NE	27.1 ± 1.38b	28.7 ± 0.79ab	28.4 ± 0.70ab	NE
180	20.0 ± 3.27	23.0 ± 5.98	18.4 ± 0.71	NE	26.4 ± 1.00b	34.2 ± 2.69ab	26.7 ± 0.82b	NE
730	NG	ND	NG	NE	NG	ND	NG	NE
Cymbopetalum baillonii (2016)
	F = 12.7, df = 6.14, p < .0001				F = 10.96, df = 6.14, p < .001
0 (t1)	11.5 ± 0.71a				22.1 ± 1.45cd
30	10.6 ± 0.93a	7.5 ± 1.29a	NE	9.8 ± 1.37a	23.0 ± 0.68bcd	22.0 ± 0.25bcd	NE	20.2 ± 0.44d
180	23.4 ± 2.50b	13.3 ± 0.25a	NE	11.8 ± 1.26a	30.1 ± 0.76a	26.6 ± 1.42ab	NE	25.4 ± 2.08abc
365	ND	ND	NE	ND	ND	ND	NE	ND
Magnolia mexicana (2015)
	F = 11.91, df = 3.8, p < .01				F = 2.815, df = 3.8, p > .05
0 (t1)	18.56 ± 5.78c				54.18 ± 3.13
25	58.2 ± 5.34a	34.3 ± 3.07bc	NE	NE	71.6 ± 4.62	58.0 ± 2.20	NE	NE
90	NG	NG	NE	42.2 ± 4.47ab	NG	NG	NE	69.0 ± 8.06
210	NG	NG	NE	NG	NG	NG	NE	NG
Magnolia mexicana (2016)
	F = 9.396, df = 2.6, p < .05				F = 4.353, df = 2.6, p > .05
0 (t1)	25.5 ± 1.46a				57.2 ± 0.55
25	ND	36.3 ± 2.27b	NE	34.1 ± 1.76b	ND	60.4 ± 1.20	NE	54.1 ± 2.23
210	NG	NG	NE	NG	NG	NG	NE	NG
Nectandra coriacea (2016)
	F = 107.5, df = 6.14, p < .0001				F = 7.603, df = 6.14, p < .001
0 (t1)	8.7 ± 0.83c				43.6 ± 1.45abc
35	NE	7.1 ± 0.01c	NE	6.3 ± 0.79c	NE	56.9 ± 4.74a	NE	53.0 ± 8.37ab
90	NE	13.0 ± 2.16bc	NE	13.0 ± 1.63bc	NE	48.2 ± 0.81abc	NE	36.0 ± 0.81c
180	NE	52.8 ± 2.20a	NE	16.8 ± 1.94b	NE	60.9 ± 0.73a	NE	37.4 ± 2.20bc
Ternstroemia tepezapote (2016)
	F = 10.7, df = 5.12, p < .001				F = 8.85, df = 5.12, p < .01
0 (t1)	26.4 ± 1.13c				49.9 ± 1.65b
16	15.5 ± 1.12c	16.8 ± 0.67bc	NE	NE	42.8 ± 2.25b	40.4 ± 2.33b	NE	NE
41	29.6 ± 1.82ab	18.4 ± 2.05bc	NE	NE	48.3 ± 0.69b	41.6 ± 1.36b	NE	NE
251	NG	NG	NE	38.1 ± 8.52a	NG	NG	NE	66.5 ± 7.50a

Note. NG = no germination; ND = no determined, NE = no evaluated. Lowercase letters indicate statistical differences (Tukey’s test after ANOVA, p < .05, or Tukey and Kramer’s test after Kruskal-Wallis, p < .05). Comparisons were done individually for each germination parameter, for each species and for each year of seed collection of C. baillonii and M. mexicana.

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.

	Cymbopetalum baillonii				Magnolia mexicana
Source of
Variation	MS	F	p	df	MS	F	p	df
Final germination (%)
Y	2	0.024	.8774	1	101	0.945	.3456	1
Sti	11525	135.99	.0001	4	7053	65.822	.0001	2
ST	1121	13.223	.0001	3	2447	22.842	.0001	2
Y × Sti	892	10.526	.001	2	32	0.301	.5911	1
Y × ST	187	2.203	.1469	1	327	3.05	.0999	1
Sti × ST	849	10.023	.0001	4
Y × ST × Sti	30	0.35	.5582	1
Residuals	2881			34	107			16
Lag time (days)
Y	259.84	17.55	.0002	1	44.5	1.135	.3015	1
Sti	201.93	13.639	.0001	2	1084.2	27.669	.0001	2
ST	53.28	3.598	.0262	3	595	15.186	.001	2
Y × Sti	9.7	0.655	.5275	2	12.5	0.318	.5799	1
Y × ST	62.87	4.246	.0490	1
Sti × ST	18.4	1.243	.3136	3
Y × ST × Sti	55.12	3.723	.0642	1
Residuals	14.81			27	39.2			17
Mean germination time (days)
Y	336.2	55.362	.0001	1	145.32	3.567	.0761	1
Sti	64.2	10.575	.001	2	146.51	3.596	.0498	2
ST	16.9	2.791	.0595	3	208.11	5.108	.0183	2
Y × Sti	88.7	14.609	.0001	2	2.44	0.06	.8096	1
Y × ST	64.7	10.658	.0029	1
Sti × ST	7.8	1.29	.2980	3
Y × ST × Sti	26	4.274	.0484	1
Residuals	6.1			27	40.74			17
Maximum germination rate (% × day⁻¹)
Y	0.07	0.012	.9149	1	2.5731	5.522	.0311	1
Sti	21.46	3.725	.0372	2	0.5744	1.233	.3162	2
ST	3.74	0.648	.5908	3	1.7047	3.659	.0477	2
Y × Sti	34.08	5.915	.0074	2	0.0104	0.022	.8829	1
Y × ST	3.27	0.567	.4580	1
Sti × ST	10.96	1.903	.1530	3
Y × ST × Sti	8.3	1.44	.2405	1
Residuals	5.76			27	0.4659			17

Note. Interactions that were statistically-significant (p values) are indicated with values in bold. MS = mean square; df = degrees of freedom.

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 3 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.

	Chamaedorea glaucifolia				Nectandra coriacea				Ternstroemia tepezapote
Source of
variation	MS	F	P	df	MS	F	p	df	MS	F	p	df
Final germination (%)
Sti	757.6	1.795	.185	2	3623	68.47	.0001	3	1215.2	11.25	.001	3
ST	440	1.043	.39	3	1867	35.29	.0001	2	2408.3	22.294	.001	1
Sti × ST	51	0.121	.887	2	1141	21.57	.0001	1	23.1	0.214	.6517	1
Residuals	422			27	53			14	108			12
Lag time (days)
Sti	81.31	2.878	.0736	2	10.819	93.1	.0001	3	0.644	15.027	.001	3
ST	70.85	2.508	.0801	3	5.538	47.66	.0001	2	0.1218	2.841	.1176	1
Sti × ST	0.96	0.034	.9667	2	4.797	41.28	.0001	1	0.2379	5.551	.0363	1
Residuals	28.25			27	0.116			14	0.0429			12
Mean germination time (days)
Sti	3579	75.362	.0001	2	0.0745	4.688	.0181	3	0.14816	13.581	.001	3
ST	47	0.992	.412	3	0.3809	23.959	.001	2	0.03263	2.991	.1093	1
Sti × ST	3	0.07	.933	2	0.0604	3.799	.0480	1	0.00582	0.533	.4793	1
Residuals	47			27	0.0159			14	0.01091			12
Maximum germination rate (% × day⁻¹)
Sti	7.56	14.571	.0001	2	0.5823	3.894	.0324	3	0.6616	6.298	.0082	3
ST	1.34	2.582	.0741	3	1.5822	10.58	.0057	2	1.4499	13.802	.0029	1
Sti × ST	0.124	0.238	.7897	2	0.3648	2.439	.1233	1	0.3679	3.502	.0858	1
Residuals	0.519			27	0.1496			14	0.1051			12

Note. Interactions that were statistically significant (p values) are indicated with values in bold. MS = mean square; df = degrees of freedom.

Figure 2.

Final germination (%, $\bar{x}$ ± 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 ( $\bar{x}$ ± 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 WC_50. 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 (WC_db = 30% calculated from WC_fb, 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.

Footnotes

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).

Appendices

References

Alcántara-Flores, E. (2002). Estructura e histoquímica de las partes vegetales usadas en medicina tradicional de Talauma mexicana [Structure and histochemistry of plant parts used in traditional medicine of Talauma mexicana] (Bachelor Thesis). Mexico City, Mexico: Facultad de Ciencias, Universidad Nacional Autónoma de México.

Baskin

J. M.

Baskin

C. C.

(2014) What kind of seed dormancy might palms have? Seed Science Research 24: 17–22.

Becerra-Vázquez

A. G.

Ramírez-Marcial

Holz

S. C.

(2011) Tropical dry forest restoration in Chiapas, Mexico, and basic knowledge for native tree species: Phenology, seed germination and seedling growth. In: Newton

A. C.

Tejedor

(eds) Principles and practice of forest landscape restoration: Case studies from the drylands of Latin America, Gland, Switzerland: International Union for Conservation of Nature, pp. 148–149.

Berjak

Pammenter

N. W.

(2008) From Avicennia to Zizania: Seed recalcitrance in perspective. Annals of Botany 101: 213–228.

Bonner, F. T., & Vozzo, J. A. (1987). Seed biology and technology of Quercus (Gen. Tech. Rep. SO-66). New Orleans, LA: USDA Forest Service.

Caddick, L. (2005). Moisture isotherms and conversion of indices used to describe moisture relations in stored grain. http://sgrl.csiro.au/storage/moisture/conversion.html (accessed January 2008).

Carpenter

W. J.

Ostmark

E. R.

(1994) Temperature and desiccation affect the germination of Chamaedorea palm seed. Proceedings of the Florida State Horticultural Society 107: 183–186.

Cerdà

Garcıa-Fayos

(2002) The influence of seed size and shape on their removal by water erosion. Catena 48: 293–301.

Cleri

(2016) The physics of living systems, Switzerland: Springer International Publishing.

10.

Coates

Estrada

(1988) Frugivory and seed dispersal in Cymbopetalum baillonii (Annonaceae) at Los Tuxtlas, Mexico. Journal of Tropical Ecology 4: 157–172.

11.

Comisión Nacional del Agua. (2017). Online access tool for climatological information gathered for the stations of the Servicio Meteorológico Nacional de México. Retrieved from http://smn.cna.gob.mx/es/climatologia/informacion-climatologica.

12.

Corner

E. J. H.

(1976) The seeds of dicotyledons Vol. 1, Cambridge, UK: Cambridge University Press.

13.

Crane

Miller

A. L.

Van Roekel

J. W.

Walters

(2003) Triacylglycerols determine the unusual storage physiology of Cuphea seed. Planta 217: 699–708.

14.

de Carvalho

L. R.

Davide

A. C.

da Silva

E. A. A.

de Carvalho

M. L. M.

(2008) Classificação de sementes de espécies florestais dos gêneros Nectandra e Ocotea (Lauraceae) quanto ao comportamento no armazenamento [Classification of seeds of forest species of the genus Nectandra and Ocotea (Lauraceae) regarding the storage behavior]. Revista Brasileira de Sementes 30: 1–9.

15.

Dussert

Chabrillange

Engelmann

Anthony

Louarn

Hamon

(2000) Relationship between seed desiccation sensitivity, seed water content at maturity and climatic characteristics of native environments of nine Coffea L. species. Seed Science Research 10: 293–300.

16.

Eggers

Erdey

Pammenter

N. W.

Berjak

(2007) Storage and germination responses of recalcitrant seeds subjected to mild dehydration. In: Adkins

Ashmore

Navie

S. C.

(eds) Seeds: Biology, development and ecology, Wallinford, UK: CABI, pp. 85–92.

17.

Escobar-Ocampo

Ochoa-Gaona

(2007) Estructura y composición florística de la vegetación del Parque Educativo Laguna Bélgica, Chiapas, México [Structure and floristic composition of the vegetation of the Laguna Bélgica Educational Park, Chiapas, Mexico]. Revista Mexicana de Biodiversidad 78: 391–419.

18.

Farrant

J. M.

Berjak

Pammenter

N. W.

(1985) The effect of drying rate on viability retention of recalcitrant propagules of Avicennia marina. South African Journal of Botany 51: 432–438.

19.

Figlar

R. B.

Nooteboom

H. P.

(2004) Notes on Magnoliaceae IV. Blumea 49: 87–100.

20.

Finch-Savage

W. E.

Farrant

J. M.

(1997) The development of desiccation-sensitive seeds in Quercus robur L.: Reserve accumulation and plant growth regulators. Seed Science Research 7: 35–39.

21.

Gálvez-Cendegui

Peñaloza

Oyanedel

Castro

(2017) Storage, size and vigor of ‘Esther’ avocado seeds (Persea americana Mill.). Ciencia e Investigación Agraria 44: 1–6.

22.

Gutiérrez-García

Ricker

(2011) Climate and climate change in the region of Los Tuxtlas (Veracruz, Mexico): A statistical analysis. Atmósfera 24: 347–373.

23.

Hamilton

K. N.

Offord

C. A.

Cuneo

Deseo

M. A.

(2013) A comparative study of seed morphology in relation to desiccation tolerance and other physiological responses in 71 Eastern Australian rainforest species. Plant Species Biology 28: 51–62.

24.

Hill

J. P.

Edwards

Franks

P. J.

(2010) How long does it take for different seeds to dry? Functional Plant Biology 37: 575–583.

25.

Hill

J. P.

Edwards

Franks

P. J.

(2012) Size is not everything for desiccation-sensitive seeds. Journal of Ecology 100: 1131–1140.

26.

Hong

T. D.

Ellis

R. H.

(1996) A protocol to determine seed storage behavior, Rome, Italy: International Plant Genetic Resources Institute.

27.

Instituto Nacional de Estadística, Geografía e Informática (2003) (2003) Cuaderno estadístico municipal Ocozocoautla de Espinoza, Estado de Chiapas [Municipal statistical notebook Ocozocoautla de Espinoza, State of Chiapas], Mexico: Instituto Nacional de Estadística, Geografía e Informática.

28.

José

A. C.

Da Silva

E. A. A.

Davide

A. C.

Melo

A. J. S.

Toorop

P. E.

(2011) Effects of drying rate and storage time on Magnolia ovata Spreng. seed viability. Seed Science and Technology 39: 425–434.

29.

King

M. W.

Roberts

E. H.

(1979) The storage of recalcitrant seeds: Achievements and possible approaches, Rome, Italy: International Board for Plant Genetic Resources.

30.

Lamarca

E. V.

de Camargo

M. B. P.

Teixeira de

S. P.

da Silva

E. A. A.

Faria

J. M. R.

Barbedo

C. J.

(2016) Variations in desiccation tolerance in seeds of Eugenia pyriformis: Dispersal at different stages of maturation. Revista Ciência Agronômica 47: 118–126.

31.

McDonald

M. B.

(2004) Orthodox seed deterioration and its repair. In: Benech-Arnold

R. L.

Sánchez

R. A.

(eds) Handbook of seed physiology: Applications to agriculture, New York, NY: Food Products Press, pp. 273–304.

32.

Nery

F. C.

Prudente

D. O.

Alvarenga

A. A.

Paiva

Nery

M. C.

(2017) Storage of Calophyllum brasiliense Cambess. seeds. Brazilian Journal of Biology 77: 431–436.

33.

Niembro-Rocas

(1989) Semillas de plantas leñosas: Morfología comparada [Seeds of woody plants: Comparative morphology], Mexico City, Mexico: Limusa.

34.

O’Brien

M. J.

Philipson

C. D.

Tay

Hector

(2013) The influence of variable rainfall frequency on germination and early growth of shade-tolerant dipterocarp seedlings in Borneo. PLoS One 8: e70287.

35.

Obroucheva

Sinkevich

Lityagina

(2016) Physiological aspects of seed recalcitrance: A case study on the tree Aesculus hippocastanum. Tree Physiology 36: 1127–1150.

36.

Pires

L. A.

Cardoso

V. J. M.

Joly

C. A.

Rodrigues

R. R.

(2009) Germination of Ternstroemia brasiliensis Cambess. (Pentaphylacaceae) from a restinga forest. Acta Botanica Brasilica 23: 57–66.

37.

Pritchard

H. W.

Daws

M. I.

Fletcher

B. J.

Gaméné

C. S.

Msanga

H. P.

Omondi

(2004) Ecological correlates of seed desiccation tolerance in tropical African dryland trees. American Journal of Botany 91: 863–870.

38.

Pupim

T. L.

Novembre

A. D. da L.C.

Brancalion

P. H. S.

de Moraes

M. H. D.

Mondo

V. H. V.

Labonia

V. D. de S.

(2009) Conservação de sementes de Magnolia ovata St. Hil [Preservation of seeds of Magnolia ovata St. Hil]. Revista Brasileira de Sementes 31: 96–105.

39.

R Core Team (2016) (2016) R: A language and environment for statistical computing, Vienna, Austria: R Foundation for Statistical Computing.

40.

Rajjou

Duval

Gallardo

Catusse

Bally

Job

(2012) Seed germination and vigor. Annual Review of Plant Biology 63: 507–533.

41.

Rodríguez

M. D. C.

Orozco-Segovia

Sánchez-Coronado

M. E.

Vázquez-Yanes

(2000) Seed germination of six mature neotropical rain forest species in response to dehydration. Tree Physiology 20: 693–699.

42.

Royal Botanic Gardens Kew. (2017). Seed information database (SID). Version 7.1. Retrieved from http://data.kew.org/sid/.

43.

Sánchez-Coronado

M. E.

Coates

Castro-Colina

Gamboa-de Buen

Paez-Valencia

Barradas

Orozco-Segovia

(2007) Improving seed germination and seedling growth of Omphalea oleifera (Euphorbiaceae) for restoration projects in tropical rain forests. Forest Ecology and Management 243: 144–155.

44.

Soto

Gama

(1997) Climas. In: González-Soriano

Dirzo

Vogt

R. C.

(eds) Historia natural de Los Tuxtlas [Natural history of Los Tuxtlas], Mexico City, Mexico: Instituto de Biología, Universidad Nacional Autónoma de México, pp. 7–23.

45.

Standley

P. C.

Steyermark

J. A.

(1946) Flora of Guatemala Vol 24, Part IV, Chicago, IL: Fieldiana Botany, Chicago Natural History Museum.

46.

Standley

P. C.

Steyermark

J. A.

(1949) Flora of Guatemala Vol 24, Part VI, Chicago, IL: Fieldiana Botany, Chicago Natural History Museum.

47.

Tommasi

Paciolla

Concetta de Pinto

De Gara

(2006) Effects of storage temperature on viability, germination and antioxidant metabolism in Ginkgo biloba L. seeds. Plant Physiology and Biochemistry 44: 359–368.

48.

Tweddle

J. C.

Dickie

J. B.

Baskin

C. C.

Baskin

J. M.

(2003) Ecological aspects of seed desiccation sensitivity. Journal of Ecology 91: 294–304.

49.

Vázquez-Yanes

Batis-Muñoz

A. I.

Alcocer-Silva

M. I.

Gual-Díaz

Sánchez-Dirzo

(2001) Árboles y arbustos nativos potencialmente valiosos para la restauración ecológica y la reforestación [Native trees and shrubs potentially valuable for ecological restoration and reforestation], Mexico City, Mexico: Instituto de Ecología, Universidad Nacional Autónoma de México.

50.

Vázquez-Yanes

Orozco-Segovia

(1993) Patterns of seed longevity and germination in the tropical rain forest. Annual Review of Ecology and Systematics 24: 69–87.

51.

Vázquez-Yanes

Orozco-Segovia

(1994) Signals for seeds to sense and respond to gaps. In: Caldwell

M. M.

Pearcy

R. W.

(eds) Exploitation of environmental heterogeneity by plants: Ecophysiological processes above and below ground, New York, NY: Academic Press, pp. 209–236.

52.

Vertucci

C. W.

Farrant

J. M.

(1995) Acquisition and loss of desiccation tolerance. In: Kigel

Galili

(eds) Seed development and germination, New York, NY: Marcel Dekker, pp. 237–271.

53.

Wyse

S. V.

Dickie

J. B.

(2017) Predicting the global incidence of seed desiccation sensitivity. Journal of Ecology 105: 1082–1093.

54.

Xia

Seal

C. E.

Chen

W. Y.

Zhou

Z. K.

Pritchard

H. W.

(2010) Fruit oil contents of the genus Quercus (Fagaceae): A comparative study on acorns of subgenus Quercus and the Asian subgenus Cyclobalanopsis. Seed Science and Technology 38: 136–145.