Thermal Effects on Early Life Stages of Leptocereus (Cactaceae) Species from Cuban Seasonally Dry Tropical Forests

Abstract

Rising temperatures are among the most predictable outcomes of climate change, and cacti are particularly vulnerable at the germination stage. We tested seeds of ten Cuban Leptocereus species from coastal and inland habitats under five temperature regimes to evaluate germination responses, thermal buffering capacity for optimal germination, photoblastic behavior, recovery after heat stress, and seedling vigor. Germination declined sharply with increasing temperature, revealing minimal thermal buffering capacity for optimal germination. All species exhibited positively photoblastic seeds, while recovery and the degree of physiological dormancy varied among taxa. Except for one taxon, most displayed partial dormancy that could stagger germination over time. Seedling vigor was not affected by high temperatures in the same way in all species. Overall, our findings suggest that climate warming will further constrain the germination niche of Leptocereus, underscoring the importance of conservation measures such as ex situ propagation.

1. Introduction

The potential impact of climate change on plant populations is a central concern for conservation [1,2,3], including early life stages [4] and species adapted to arid environments [5]. Projected increases in temperature and drought [6] were once thought to favor the spread of dry- and heat-tolerant plant species such as cacti. However, the tolerance of adult cacti to high temperature and water stress is absent in the early life stages, when seeds and seedlings are especially vulnerable [7,8,9]. Any habitat warming on these early stages could compromise population persistence by reducing germination and establishment, both critical for medium- and long-term survival [10,11,12]. Alarmingly, projections suggest that 60–90% of cactus species may be at risk under future climate scenarios [5].

The effects of habitat warming on cactus seed germination have been widely studied e.g., [9,13,14,15,16], but results are inconsistent [17]. Germination may be unaffected [14,16], enhanced [9,16], or reduced [13,14,16]. Such contrasting responses may reflect interspecific differences in thermal buffering capacity for optimal germination (TBCog). We defined TBCog as the difference between the seed’s optimal germination temperature and the mean seasonal temperature during its peak germination period. The thermal buffering capacity of germination referred by Seal et al. [16] covers the range between suboptimal and supraoptimal germination temperatures relative to the mean temperature of the wettest quarter and differs from TBCog in that we are only considering the range of temperatures between the optimal germination temperature and the mean temperature of the wettest quarter. In practice, species from warmer habitats tend to have narrower TBCog, making them more sensitive to temperature increases, as demonstrated by Sentinella et al. [18] and Barrios et al. [17].

In cacti, optimal germination temperatures are typically 5–10 °C above the seasonal mean, while the margin for tropical species is only 1–3 °C or less [16,17]. Although one might expect temperate species to have lower optimal germination temperatures than tropical ones, evidence indicates that both, tropical and temperate species germinate optimally around 30 °C [17]. This suggests that temperate species possess greater thermal buffering capacity for optimal germination, whereas species from warm climates (already germinating near their upper limit) are far more vulnerable to future warming. Importantly, tropical cacti remain underrepresented in germination studies [17].

The West Indies host around 116 native cactus species, 84% of them endemic [19]. Cuba is the diversity hotspot of the region, with 49 native species, ~75% of which are endemic [20]. The West Indies species occupy diverse ecosystems, but about half of them are restricted to coastal zones [20], potentially increasing their exposure and vulnerability to climate change. Despite this, knowledge of Cuban cactus germination ecology (both inland and coastal) remains extremely limited [21,22].

Here, we investigate three questions using ten endemic Cuban species of Leptocereus: (1) Do coastal species have narrower thermal buffering capacity for optimal germination than inland species? (2) What TBCog values can be expected for Leptocereus under projected 2070 climate scenarios? (3) What do recovery capacity and seedling vigor reveal about adaptability to warming? To address these questions, we determined optimal germination temperature (Tₒ), combined field-measured temperature data with climate projections, and evaluated photoblastic response, recovery after heat stress, and seedling vigor. We hypothesized that: (1) coastal Leptocereus species would have narrower TBCog but similar germinability compared with inland species; (2) TBCog would approach zero under projected climate scenarios; and (3) recovery and seedling vigor would reveal high adaptability and tolerance to elevated temperatures.

2. Materials and Methods

2.1. Study Species: Seed Collection and Processing

Species of the genus Leptocereus are threatened [23], endemic Cuban cacti with diverse growth forms, including arborescent, erect, scandent, and prostrate shrubs [24]. They occur in both semi-arid coastal and dry inland karstic hill habitats (Table 1) characterized by seasonally dry tropical forests [24]. Rainfall occurs during summer and drought happens during winter. Vegetation of coastal habitats is formed by dry evergreen forest with an annual rainfall average of 700–1200 mm. The vegetation of karstic hills is semi-deciduous tropical forest with an annual rainfall average of 1400–1700 mm [25]. During the summer months in the study sites, the maximum temperatures typically range from 35 °C and 40 °C [19,22].

Ten collection sites for ten species of Leptocereus were visited between 2015 and 2017 (Table 1). At least seven fruits were collected from different individuals of each species, except for L. assurgens, from which only three individuals could be sampled. Fruits were carefully opened with a knife, and the seeds were extracted and rinsed under running water through a sieve until all mucilage was removed. The seeds were then air-dried at room temperature in the shade and stored in paper bags until use. Collections were made mainly during the rainy season (May–September), except for L. scopulophilus, which was collected during the dry season (January). Germination trials were initiated within 0–4 months of collection (Table 1).

Table 1. Habitat characteristics of Leptocereus species. m a.s.l., meters above sea level; Pm, precipitation of the wettest quarter [33]. SCD: seed collection date (month/year); SSD: seed sowing date (month/year).

Species	Locality, Province	Habitat	m a.s.l.	Pm (mm)	SCD	SSD
L. santamarinae	Playa Herradura, Las Tunas	Coastal	2	377	Aug/17	Sep/17
L. sylvestris	Sendero el Guafe, Granma	Coastal	18	387	Sep/15	Nov/15
L. arboreus	Desembocadura del río Yaguanabo, Cienfuegos	Coastal	26	510	Sep/15	Nov/15
L. wrightii	Boca de Jaruco, Mayabeque	Coastal	29	479	May/17	May/17
L. maxonii	Mesa de la tinta, Guantánamo	Inland karstic hill	394	501	Jun/16	Oct/16
L. carinatus	Sierra del Chorrillo, Camagüey	Inland karstic hill	244	627	Sep/17	Sep/17
L. scopulophilus	Pan de Matanzas, Matanzas	Inland karstic hill	212	617	Jan/17	Jan/17
L. leonii	Sierra de Anafe, Artemisa	Inland karstic hill	147	575	May/16	May/16
L. assurgens	Chichones del Indio, Viñales, Pinar del Río	Inland karstic hill	348	703	Jun/16	Oct/16
L. ekmanii	Sierra de Guanes, Pinar del Río	Inland karstic hill	49	611	Jun/17	Jun/17

Table 1. Habitat characteristics of Leptocereus species. m a.s.l., meters above sea level; Pm, precipitation of the wettest quarter [33]. SCD: seed collection date (month/year); SSD: seed sowing date (month/year).

Species	Locality, Province	Habitat	m a.s.l.	Pm (mm)	SCD	SSD
L. santamarinae	Playa Herradura, Las Tunas	Coastal	2	377	Aug/17	Sep/17
L. sylvestris	Sendero el Guafe, Granma	Coastal	18	387	Sep/15	Nov/15
L. arboreus	Desembocadura del río Yaguanabo, Cienfuegos	Coastal	26	510	Sep/15	Nov/15
L. wrightii	Boca de Jaruco, Mayabeque	Coastal	29	479	May/17	May/17
L. maxonii	Mesa de la tinta, Guantánamo	Inland karstic hill	394	501	Jun/16	Oct/16
L. carinatus	Sierra del Chorrillo, Camagüey	Inland karstic hill	244	627	Sep/17	Sep/17
L. scopulophilus	Pan de Matanzas, Matanzas	Inland karstic hill	212	617	Jan/17	Jan/17
L. leonii	Sierra de Anafe, Artemisa	Inland karstic hill	147	575	May/16	May/16
L. assurgens	Chichones del Indio, Viñales, Pinar del Río	Inland karstic hill	348	703	Jun/16	Oct/16
L. ekmanii	Sierra de Guanes, Pinar del Río	Inland karstic hill	49	611	Jun/17	Jun/17

2.2. Treatments

For all species, the effects of five temperature regimes were tested: one constant (25 °C) and four alternating (25/30 °C, 25/35 °C, 25/40 °C, and 25/45 °C), under two light conditions (continuous darkness and light/darkness). For L. assurgens and L. ekmanii, seed germination performance was evaluated under only two temperature regimes due to the limited number of seeds available. Each treatment consisted of seven replicates of 25 seeds, sown in Petri dishes (60 × 15 mm) containing 1% agar as substrate.

Prior to sowing, seeds were disinfected following this sequence: 1 min in 0.2% sodium lauryl sulfate (SLS) detergent, rinse with distilled water, 1 min in 95% ethanol, rinse with distilled water, 5 min in 1% sodium hypochlorite, rinse with distilled water, 1 min in 95% ethanol, and a final rinse with distilled water.

Petri dishes were placed in growth chambers (FRIOCEL 111L, Germany) with an 8 h photoperiod provided by white, fluorescent lamps (40 μmol m⁻² s⁻¹, 400–700 nm). Darkness was achieved by covering the dishes with double aluminum foil and placing them at 25 °C. Alternating temperature regimes (25/30 °C, 25/35 °C, 25/40 °C, and 25/45 °C) were programmed in growth chambers as follows: 12 h at 25 °C (darkness), 8 h at the higher temperature (light), and 4 h of transition between temperatures (darkness). Thus, mean temperatures for these treatments were 27.08, 29.16, 31.25, and 33.33 °C, respectively. High temperatures were included based on the conditions of Cuban Leptocereus habitats [22]. Specifically, the highest alternating temperatures (25/40 °C, and 25/45 °C) were included based on temperatures recorded in Cuban open habitats [19] and the reported ceiling temperatures for several Caribbean cacti, which range from 39.4 to 47.2 °C [16].

Germination was recorded daily, except for the dark treatment, which was evaluated only on day 28. Seeds were considered germinated when the radicle emerged. After four weeks, 20 seedlings were randomly selected, blotted with filter paper to remove excess water, and weighed on a Sartorius analytical balance (precision 0.0001 g). Seedling mass was measured only for the 25 °C and 25/35 °C treatments, except for L. assurgens and L. wrightii, where seedlings from the 25/30 °C treatment were measured instead of those from 25 °C.

After four weeks, seeds that had not germinated were rinsed thoroughly with distilled water and re-incubated for another four weeks at 25 °C under light (40 μmol m⁻² s⁻¹, 400–700 nm). Germination recovery was calculated following [26] as: Recovery = [(a − b)(c − b)] × 100; where a is the number of seeds germinated by day 56, b is the number germinated by day 28, and c is the number of seeds per Petri dish (25). Seeds that remained ungerminated at day 56 were subjected to a cut test to determine viability: seeds with a firm white embryo were classified as viable, while those with a soft or gray embryo were classified as non-viable [26].

To assess germination, three indices were measured:

(1)

Minimum germination time (T_min), which is calculated as the day at which the first germination occurred [22].

(2)

Mean germination time (MGT), an index of germination speed expressed in days [27]. MGT was calculated as: $MGT={\displaystyle \frac{\sum n_{i·}{t}_i}{\sum n_i}}$; where n_i is the number of seeds that germinated at time i, and t_i is the elapsed time from the start of the experiment to the i^th observation [28].

(3)

Germinability (G), the final percentage of germinated seeds at the end of the experiment (day 28) [22].

In this study, we calculated the Relative Light Germination (RLG) index [29] to evaluate the light requirement for germination. The index was calculated as: RLG = Gl/(Gd + Gl), where Gl is the germination percentage in light, and Gd is the germination percentage in darkness. Both Gd and Gl values were obtained from the mean germination percentages at 25 °C. The RGL index ranges from (germination only in darkness) to 1 (germination only in light). Species with RLG > 0.75 are considered light-dependent (positively photoblastic), those with RLG < 0.25 are light-repellent (negatively photoblastic), and values between 0.25 and 0.75 indicate light-indifferent behavior [30].

2.3. Measurement and Projection of Temperatures at Collection Sites

A HOBO U23 Pro v2 data logger (Bourne, MA, USA) was installed at nine of the ten collection sites, recording temperature (±0.1 °C) every 30 min. These records covered a minimum of two years. Future projections were generated using two climatic variables: mean temperature of the wettest quarter and maximum temperature of the warmest month derived from climate change scenarios for 2070 across ten general circulation models (GCMs): BCC-CSM1-1, CCSM4, CNRM-CM5, HadGEM2-ES, MIROC5, MPI-ESM-LR, MRI-CGCM3, GISS-E2-R, NorESM1-M, and IPSL-CM5A-LR [31]. For each GCM, two radiative forcing scenarios (2.6 and 8.5 W m⁻²) were considered, representing the lower and upper extremes of possible greenhouse gas concentrations, with 8.5 reflecting the worst-case scenario [32].

To estimate the temperature increase at each site, the current baseline temperature from WorldClim 2.0 [33,34] was subtracted from the projected 2070 temperature in each GCM. The mean of these differences was then added to the average values recorded by the data loggers at each location. For La Mesa de La Tinta, where no logger was available, mean temperature of the wettest quarter and maximum temperature of the warmest month were obtained directly from WorldClim 2.0 [33,34].

2.4. Data Analysis

Each germination index was analyzed using a best-fit model that included temperature, species, and their interaction as predictor variables. For germinability and recovery, Generalized Linear Models (GLMs) with a quasibinomial error distribution and a logit link function were fitted. T_min and MGT were analyzed with GLMs assuming a Gamma error distribution and a logarithmic link function. The significance of main effects and interactions was assessed with Type II analysis of variance (ANOVA) using likelihood ratio tests (LRTs). Post hoc comparisons were performed with Tukey’s test on estimated marginal means (EMMs), with a significance level of α = 0.05.

Seedling mass was analyzed using a simple linear model (lm), with temperature as the independent variable. Pairwise comparisons of estimated means between the two temperature levels were conducted for each species. Significance was evaluated with t-tests, applying a Bonferroni correction for multiple comparisons.

All analyses were performed in R (version 2025.05.1) using the emmeans (version 1.11.2) and car (version 3.1.3) packages [35]. Figures were generated in SigmaPlot 10.0 and refined in Adobe Photoshop CS3.

3. Results

3.1. Germination Speed (T_min and MGT)

In most species, germination under light began between the first and second week after sowing (Figure 1). Only L. wrightii and L. ekmanii germinated after the second week, as did L. assurgens at 25/35 °C (

Table 2. Germination indices of Leptocereus seeds after four weeks. Values are means ± standard deviation. Lowercase letters indicate differences among temperatures within a species, while capital letters indicate differences among species within a temperature. Different letters denote significant differences at p < 0.05. T_min: minimum germination time required for seeds to begin germination; MGT: mean germination time; G: germinability; GRL: relative light germination index.

Species	Cond.	Temp (°C)	Tmin (Days)	MGT (Days)	G (%)	GRL
L. santamarinae	darkness	25			10.28 ± 7.25	0.89
	light	25	6.71 ± 0.76 aD	9.74 ± 0.33 aE	99.42 ± 1.51 a
		25/30	6.57 ± 0.53 aE	9.52 ± 0.27 aG	98.85 ± 1.95 a
		25/35	9.14 ±0.69 bF	13.02 ± 1.11 bD	32.00 ± 11.31 b
L. sylvestris	darkness	25			2.8 ± 3.8	0.95
	light	25	10.42 ± 0.53 aC	16.05 ± 1.79 aC	55.42 ± 10.43 a
		25/30	10.42 ± 0.78 aCD	17.02 ± 1.88 aCD	42.85 ± 6.41 a
		25/35	14.14 ± 5.49 bBCD	21.76 ± 3.84 bA	19.42 ± 10.18 b
L. arboreus	darkness	25			0	1.00
	light	25	10.57 ± 0.78 aC	15.95 ± 0.94 aC	71.43 ± 9.36 a
		25/30	10.71 ± 0.49 aBC	15.27 ± 1.57 aDE	69.71 ± 11.74 a
		25/35	17.00 ± 5.62 bABC	19.76 ± 4.94 bAB	8.57 ± 7.81 b
L. wrightii	darkness	25			0	1.00
	light	25	14.14 ± 2.27 aAB	19.60 ± 1.73 aAB	66.28 ± 11.51 a
		25/30	18.28 ± 3.09 bA	23.71 ± 2.19 bA	53.14 ± 11.25 a
		25/35	25.00 ± 2.83 cA	25.00 ± 2.83 bA	1.14 ± 1.95 b
L. maxonii	darkness	25			4.0 ± 4.0	0.96
	light	25	10.28 ± 0.75 aC	15.78 ± 0.96 bC	90.28 ± 6.87 a
		25/30	9.85 ± 1.46 aCD	13.88 ± 0.63 aEF	94.85 ± 5.01 a
		25/35	13.71 ± 2.56 bCDE	19.95 ± 1.50 cAB	36.57 ± 10.18 b
L. carinatus	darkness	25			0	1.00
	light	25	10.14 ± 0.38 aC	13.63 ± 0.65 aD	93.14 ± 11.24 a
		25/30	10.00 ± 0.00 aCD	12.40 ± 0.35 aF	94.85 ± 1.95 a
		25/35	10.57 ± 0.53 aEF	13.71 ± 1.18 aCD	56.71 ± 7.09 b
L. scopulophilus	darkness	25			0	1.00
	light	25	10.14 ± 2.03 aC	15.87 ± 1.57 aC	66.28 ± 13.63 a
		25/30	9.00 ± 1.15 aDE	16.47 ± 1.19 aD	73.14 ± 10.76 a
		25/35	12.42 ± 1.51 bDE	16.82 ± 1.85 aBC	20.00 ± 6.92 b
L. leonii	darkness	25			0	1.00
	light	25	11.28 ± 2.49 aBC	17.30 ± 1.41 aBC	42.28 ± 5.09 a
		25/30	11.85 ± 1.34 aBC	18.18 ± 2.49 aBC	29.14 ± 6.81 ab
		25/35	13.42 ± 3.05 aCDE	19.92 ± 2.19 bAB	21.71 ± 9.76 b
L. assurgens	darkness	25/30			0	1.00
		25/30	14.00 ± 2.16 aAB	19.83 ± 2.43 aB	28.00 ± 11.07 a
		25/35	15.57 ± 2.99 aBCD	19.92 ± 1.59 aAB	14.86 ± 9.15 b
L. ekmanii	light	25	17.00 ± 3.61 aA	21.47 ± 1.09 aA	25.71 ± 8.60 a
		25/35	18.57 ± 2.94 aAB	22.68 ± 2.16 aA	11.43 ± 7.80 b

). T_min differed significantly among species (D = 10.50; df = 9; p < 0.0001), temperatures (D = 2.29; df = 2; p < 0.0001), and their interaction (D = 1.16; df = 16; p < 0.0001), indicating that the effect of temperature on T_min varied across species. T_min was delayed by 0.6–3.9 days in the 25/35 °C treatment, with the greatest effect observed in the coastal species L. wrightii and L. arboreus (Table 2).

Mean germination time (MGT) also differed significantly among species (D = 7.90; df = 9; p < 0.0001), temperatures (D = 1.06; df = 2; p < 0.0001), and their interaction (D = 0.76; df = 16; p < 0.0001) (Table 2). The coastal species L. wrightii and L. assurgens, together with L. ekmanii from inland karstic hills, exhibited the slowest germination rates (Table 2).

3.2. Germinability

Final germination percentage differed significantly among species (D = 37.23; df = 9; p < 0.0001), temperatures (D = 38.63; df = 2; p < 0.0001), and their interaction (D = 8.99; df = 16; p < 0.0001) (Table 2), indicating that the effect of temperature on germination varied across species. Treatments at 25 °C and 25/30 °C were optimal for all species (Table 2). No seeds germinated at 25/45 °C, and only four species germinated at 25/40 °C, with values below 10% (Figure 1). At 25/35 °C, L. santamarinae (67.4%) and L. wrightii (65.1%) showed the greatest reduction in germination compared to 25 °C (Table 2). In contrast, L. carinatus exhibited the highest germinability at 25/35 °C and was the least affected by temperature increase (Table 2). Only three species germinated in darkness, at <15%, indicating that all species studied are positively photoblastic (Table 2).

3.3. Recovery

All species recovered germination when seeds previously exposed to 25/35 °C, 25/40 °C, and 25/45 °C were transferred to 25 °C (Figure 2). Significant differences in recovery across temperatures (D = 32.99; df = 7; p < 0.0001) were observed in six species, half of which showed greater recovery after exposure to 25/45 °C (Figure 2A). Leptocereus maxonii and L. wrightii exhibited the highest recovery rates, except at 25/45 °C, where L. santamarinae and L. sylvestris showed comparable values, while L. wrightii was not evaluated (Figure 2B). Across species, most seeds that failed to germinate by week 8 remained viable (

Table 3. Viability (%) of Leptocereus seeds that did not germinate after 56 days. Values marked with an asterisk indicate viability at day 28. NE, not evaluated.

Species	Alive Seeds (%)
	Darkness	Light
	25 °C	25 °C	25/30 °C	25/35 °C	25/40 °C	25/45 °C
L. santamarinae	87.50	NE	NE	91.83	94.05	95.65
L. sylvestris	74.11	64.38	63.54	73.56	79.31	80.32
L. arboreus	88.57	67.44	80.43	90.56	97.10	90.22
L. wrightii	47.05	100	13.33	70.00	66.67	NE
L. maxonii	100	87.50 *	55.56 *	91.67	97.56	92.59
L. carinatus	93.12	100	100	100	100	86.76
L. scopulophilus	87.35	77.96	67.29	58.06	87.60	85.03
L. leonii	96.26 *	94.23	85.36	92.85	90.47	90.00
L. assurgens	100	NE	99.06 *	99.18 *	NE	NE
L. ekmanii	NE	85.71	NE	100	NE	NE

).

Recovery from seeds previously kept in darkness also differed among species. Neither L. arboreus nor L. leonii recovered after four weeks in darkness, and only 0.6% of L. scopulophilus seeds germinated. In contrast, L. wrightii and L. maxonii showed the highest recovery rates, similar to their performance under light conditions (Figure 3).

3.4. Seedling Vigor

Twenty-eight days after sowing, seedling mass ranged from 10.5–41.3 mg at 25 °C and 7.9–44.7 mg at 25/35 °C. The highest values for both treatments were recorded in L. sylvestris. Four species showed no significant differences in seedling mass between treatments. In L. maxonii, seedlings grown at 25/35 °C had significantly higher mass than those at 25 °C (Figure 4). Conversely, seedlings of L. leonii and L. assurgens grown at 25 °C and 25/30 °C had greater mass than those at 25/35 °C (Figure 4).

3.5. Thermal Buffering Capacity for Optimal Germination (TBCog)

Six of the ten Leptocereus species (L. santamarinae, L. sylvestris, L. arboreus, L. wrightii, L. scopulophilus, and L. ekmanii) currently lack a thermal buffering capacity for optimal germination (TBCog) (Figure 5). Average wet-season temperatures were higher in coastal habitats than in inland karstic hills (Figure 5). In coastal sites, TBCog was exceeded by +1.25 °C (±0.95–1.88 °C). In inland karstic hills, the mean TBCog was only 0.82 °C, and was surpassed in L. scopulophilus (+0.16 °C) and L. ekmanii (+0.10 °C). L. maxonii exhibited the highest TBCog, at 2.58 °C (Figure 5).

Climate models predict that, by 2070, mean wet-season temperatures in Leptocereus habitats will increase by 0.9–2.6 °C under RCP 2.6 and 1.2–4.2 °C under RCP 8.5. The greatest increases are projected for Pan de Matanzas and the mouth of the Yaguanabo River, habitats of L. scopulophilus and L. arboreus, respectively. Currently, maximum temperatures during the warmest month range from 30.3–39.7 °C, with projected increases of 0.8–2.0 °C under RCP 2.6 and 2.1–4.2 °C under RCP 8.5 (Figure 5). By 2070, models predict that under RCP 2.6, coastal habitats during the wettest quarter will exceed average temperatures of 29.2 °C (equivalent to the 25/35 °C treatment). Under RCP 8.5, only two sites (Mesa de La Tinta and Chichones del Indio) are expected to remain below this threshold (Figure 5).

4. Discussion

4.1. Germination and Thermal Buffering Capacity for Optimal Germination

Our results are consistent with the rapid germination pattern reported for most cacti, which typically occurs within the first few days [36,37,38] or weeks after sowing [21,39,40] under optimal temperatures. This pattern is associated with seed traits such as a permeable seed coat that promotes rapid imbibition and the absence of physiological dormancy in the portion of seed lots that germinate quickly [17].

We identified 25 °C and 25/30 °C as optimal germination temperatures, as both treatments yielded the highest germinability and germination speeds (Table 2). This agrees with the general pattern for cactus seeds, which germinate optimally between 20–30 °C [17]. However, three species (L. sylvestris, L. leonii, and L. wrightii) showed a tendency toward higher germinability at 25 °C, although differences were not statistically significant. On the other hand, although L. assurgens and L. ekmanii appear to have the same germination behavior as the other species evaluated (Table 2), interpretations related to their thermal sensitivity and TBCog (discussed in the following paragraphs) should be considered with caution. This is because only a small number of seeds could be collected for these species, which meant that they could only be evaluated at two temperatures.

Our findings support the hypothesis that coastal Leptocereus species exhibit narrower TBCog values than inland species. Indeed, coastal species inhabit areas where mean wet-season temperatures are up to ~2 °C above the optimal germination temperature (Figure 5), suggesting that their germination niche is narrower. Germination of these species likely occurs in microsites with lower-than-average temperatures for this season. According to Seal et al. [16], species with this thermal profile are more vulnerable to climate warming than species whose optimal germination temperatures exceed mean seasonal temperatures.

Regardless of habitat, our results consistently revealed little or no TBCog across all Leptocereus species. This finding aligns with the reduced populations and low natural recruitment observed in most Cuban species of the genus, independent of direct anthropogenic causes [41,42,43]. Several studies have reported the scarcity of juveniles in L. nudiflorus populations [44,45,46], a pattern typical of the genus in Cuba, except for L. scopulophilus [19], but even in this species, its largest population is distributed over less than a quarter of a hectare [47]. Moreover, the low germination percentages of most species at 29.2 °C (25/35 °C treatment) and under slight decreases in water potential [22] reinforce their vulnerability to the warming and drought projected for the Caribbean [48]. In fact, according to Barrios et al. [22] temperatures in the soil under direct solar radiation can reach up to 41 °C. Subsequent year-long monitoring of soil and air temperatures in the Yaguanabo coastal dry forest recorded even higher soil temperatures, reaching up to 53.9 °C, while maximum air temperatures reached 38.1 °C. The average difference between soil and air temperatures during the hottest hours of the day was 8.6 °C (with maximums of 15 °C) in the rainy season, increasing to 11.9 °C (with maximums of 22.5 °C) in the dry season [19]. Consequently, the germination niche of these species is already restricted and is likely to become narrower under future climate scenarios, especially under RCP 8.5, where most habitats are projected to exceed the 25/35 °C threshold during the wettest quarter by 2070, with maximum monthly temperatures surpassing 35 °C (Figure 5).

Contrary to expectations, coastal species, despite occupying the warmest habitats during the wet season (close to the 25/35 °C average), were generally the most negatively affected by 25/35 °C treatments (Figure 1). This result is consistent with evidence showing that species from warm ecosystems do not necessarily respond positively to further warming in terms of germination [17,18]. An interesting case was L. carinatus, which was least affected by 25/35 °C (Table 2). This species inhabits inland karstic hills in one of the cooler study sites. Its relatively higher tolerance may relate to phylogenetic affinities, as it is the closest Cuban species to Leptocereus from Hispaniola–Puerto Rico [24], where L. paniculatus shows an optimal germination temperature of 30.3 °C [16].

The alternating temperature of 25/35 °C has been evaluated in 45 cactus species, 39 of them in non-dormant seed lots (e.g., [40,49,50,51,52,53,54,55]). In contrast to our findings in Leptocereus, most species (28) germinated at >80% under this regime. Only nine species (including Rhipsalis, Selenicereus, Pilosocereus, Echinopsis, Melocactus, and Cereus) showed similarly reduced germination. Overall, our study highlights the high sensitivity of Cuban Leptocereus seeds to average temperatures near 30 °C or higher [21,22], in contrast to the many cacti capable of germinating at ≥30 °C (e.g., [15,40,50,53,55,56,57,58]), and even at 40 °C in some columnar taxa [59].

4.2. Recovery and Seed Dormancy

Despite the strong inhibitory effects of high temperatures on germination, all species exhibited recovery when seeds were transferred back to favorable conditions, indicating the absence of lethal heat damage in most cases. In some taxa (L. maxonii and L. wrightii), recovery rates were particularly high, even after exposure to 25/40 °C and 25/45 °C. This capacity for recovery, coupled with the relatively high proportion of viable ungerminated seeds, suggests a form of physiological dormancy that could function as a bet-hedging mechanism, spreading germination over time [17]. Such traits may partially buffer populations against interannual variability in rainfall and temperature, though the degree of protection may be insufficient under chronic warming.

Seeds that failed to germinate at 25/35, 25/40, and 25/45 °C were thermally inhibited [60], resuming germination once temperatures were reduced. This confirms the adaptability and resilience of Leptocereus seeds, which remain viable under restrictive temperatures, consistent with Barrios et al. [22] and with evidence from other cacti capable of germinating even after exposure to 50–110 °C (e.g., [61,62,63,64,65,66,67,68,69,70]).

With the exception of L. santamarinae, all species exhibited some degree of physiological dormancy (Figure 1; Table 2), consistent with earlier reports for L. scopulophilus [21,71]. Dormancy in Cuban Leptocereus may provide an adaptive advantage by spreading germination over time, reducing intraspecific competition among seedlings, and enabling at least transitory soil seed banks. Dormancy could also prevent premature germination during brief winter rains associated with cold fronts, which last only a few days and may otherwise cause early seedling mortality [72]. Although rapid germination is common in Cactoideae, physiological dormancy, long considered the only dormancy class in the family [17,73,74], may be more widespread than expected. Indeed, Willis et al. [74] found that 65% of 153 cactus species exhibit physiological dormancy. Yet, studies on dormancy in cacti remain scarce (e.g., [75,76,77]).

Few explicitly classify ungerminated seeds as dormant [78,79], complicating species-level comparisons. For conservation purposes, further work is needed to determine how dormancy is broken in Cuban Leptocereus. Given their fleshy, indehiscent fruits, dormancy could be naturally alleviated through mechanical or chemical scarification during gut passage by dispersers. In the laboratory, chemical scarification (e.g., H₂SO₄, HCl, NaClO, H₂O₂) or mechanical scarification (sandpaper) could provide effective protocols for propagation of threatened Cuban species [23].

The seed recovery and partial dormancy observed in Leptocereus highlight their adaptive value in highly variable environments. In arid and seasonally dry ecosystems, where rainfall is scarce and unpredictable, the timing of germination often determines seedling survival [7,17]. Dormancy enables a portion of seeds to remain viable but ungerminated during unfavorable season or years [80], reducing the risk of recruitment failure through a bet-hedging strategy [81]. Likewise, recovery after exposure to supraoptimal temperatures shows that seeds can withstand transient heat stress and germinate once conditions improve [17]. These traits may buffer populations against short-term climatic variability and extreme events, which are expected to become more frequent under climate change. Yet their effectiveness depends on the temporal scale of stress. While dormancy and recovery enhance resilience over short to medium timescales, they are unlikely to offset the chronic increases in temperature projected for coming decades, when habitats may exceed species’ thermal thresholds.

4.3. Seed Photoblastism

The photoblastism observed in Leptocereus seeds is consistent with reports from other Cactoideae, including members of Pachycereeae [40,82] and Stenocereinae [83], where positively photoblastic seeds are common. This trait is critical for seed survival and the maintenance of soil seed banks [84,85,86]. Indeed, positive photoblastism has been reported for more than 250 cactus species [17,82,83,87], in line with evidence for soil seed banks in cacti [88,89,90]. Moreover, Leptocereus seeds exhibited varying degrees of skotodormancy, similar to results reported for Cacteae [77,91], where dark conditions induced secondary dormancy through loss of light sensitivity [92], without loss of viability. Skotodormancy may therefore represent an adaptive strategy for survival in unpredictable environments [73,93,94].

4.4. Seedling Vigor

Seedling vigor comparisons between temperatures revealed the three response types commonly observed in similar studies. Most species showed no significant differences in seedling mass between treatments, consistent with findings for Selenicereus sp. [95], Opuntia ficus-indica [96], and Echinopsis candicans and Gymnocalycium mostii [14]. This suggests that seedlings may tolerate higher temperatures better than seeds, as also observed in L. maxonii, where mass increased at higher temperatures, paralleling results for Pilosocereus robinii at 25/35 °C [54]. Conversely, L. leonii and L. assurgens exhibited reduced seedling vigor at the highest temperature tested (Figure 4). This observation is similar to findings regarding Cereus fernambucensis [97] and suggests that temperatures higher than those evaluated could negatively affect their establishment. Although studies in natural microenvironments are required to validate this assumption, several studies on cacti have shown that the high temperatures reached in open areas of ecosystems can increase seed mortality [98] and significantly reduce seedling establishment [21,99,100,101].

5. Conclusions

Cuban Leptocereus species may be facing an increasingly unfavorable outlook for sustaining populations through sexual reproduction. For most taxa, successful germination will depend on seeds locating microsites with cooler conditions than those currently available. Ongoing climate change, characterized by rising temperatures and reduced rainfall, is likely to further constrain germination success. Even so, the presence of dormancy, the capacity to recover after thermal and water stress, and the probable orthodox storage behavior of seeds may enable Leptocereus to persist in the soil seed bank until more favorable conditions arise.

The results presented here have important implications for the conservation of Leptocereus. The lack of thermal buffering capacity for optimal germination in most species, combined with reduced germination under warming, indicates that natural regeneration could be increasingly constrained under future climates. This poses a serious threat to species restricted to coastal areas, where mean wet-season temperatures already exceed or approach germination optima. Conservation strategies must therefore anticipate a likely reduction in recruitment from seed in the wild and consider complementary interventions.