Introduction

Interspecific variations among congeneric species are directly related to their evolution and ability to occupy different habitats or ecological niches (Singhakumara et al. 2003; Rossatto et al. 2013). Studies with congeneric species minimized the effect of phylogenetic differences and make clearer the ecological filters that limit the distribution and occurrence of a plant species (Bevill and Louda 1999; Lloyd et al. 2002; Swenson et al. 2006; Zhang et al. 2010). The genus Inga has approximately 300 species and is one of the most abundant and diversified in the plant kingdom (Valencia et al. 1994; Steege et al. 2013; Nicholls et al. 2015; Wiggins et al. 2016). In Brazil, it is represented by about 144 species, of which 72% are restricted to the Brazilian Atlantic Rainforest (Mata 2009). Among Inga spp., two are noteworthy for the contrast of distribution on Brazilian territory. Inga marginata and I. subnuda are evergreen N-fixing tree species in initial stages of forest regeneration (Aidar et al. 2003; Simonelli et al. 2010; Marchiori et al. 2016), and they are very important for forest restoration (Ferreira et al. 2007; Ramos Silva and Corrêa 2008) and agroforestry systems (Souza et al. 2012). Both species are usually classified as pioneers or early successional (Ramos Silva and Corrêa 2008; Rolim et al. 2006; Guaratini et al. 2008). However, I. marginata has a wide geographic distribution and occurs in the Cerrado (Brazilian savanna), Amazonian Forest, Pantanal, and Brazilian Atlantic Rainforests (Teixeira et al. 2008; Siminski et al. 2011; Silva et al. 2016), whereas I. subnuda is endemic to the Brazilian Atlantic Forest (Peterle et al. 2015; Silva et al. 2015), with high occurrence in restinga areas (Zamith and Scarano 2004; Souza et al. 2008; Peterle et al. 2015; Silva et al. 2015) and lowland and montane rainforests (Peterle et al. 2015).

The light gradients inside forest ecosystems influence the colonization and spatial distribution of species and position occupied by each species in the ecological succession (Swaine and Whitmore 1988; Bazzaz and Pickett 1980; Valladares et al. 2000; Poorter et al. 2006; Valladares and Niinemets 2008). The pioneer or early successional trees usually have characteristics similar to those of sun plants, whereas trees that regenerate in the late successional stages, have the typical characteristics of shade plants in their initial growth phase (seedlings and saplings) (Bazzaz and Pickett 1980; Swaine and Whitmore 1988; Poorter et al. 2006). The adjustments to light changes during forest succession are related to changes in leaf and whole-plant carbon balance (Givnish 1988; Aleric and Kirkman 2005). Low light availability compromises the net carbon gain, and light levels above the processing capacity of plants can damage the photosynthetic apparatus (Lüttge 2008). In both situations, impairments to the growth (Kitajima 1994; Aleric and Kirkman 2005; Lusk and Jorgensen 2013) and survival (Kitajima 1994; Ashton et al. 1995; Goodale et al. 2014) of seedlings and saplings are observed. In addition, inside a forest, seedlings and saplings sometimes grow under the shade of larger trees and may be subjected to a high incidence of light in the gaps (Palow and Oberbauer 2009). The ability of seedlings and saplings of different tree species to tolerate low light levels inside a forest is a very important feature that drives forest dynamics and structure (Valladares and Niinemets 2008). However, the ability to tolerate high light is important for plants to colonize gaps or open areas (Davies 1998).

Plants are very plastic organisms, and several morphological, anatomical, and physiological characteristics contribute to the ability of plants to cope with multiple environmental factors (Chapin et al. 1987) and colonize different ecosystems. Plants adapted to the shade environments of forest understories have higher chlorophyll content and lower chlorophyll a/b ratio (Boardman 1977; Givnish 1988; Valladares and Niinemets 2008). In addition, shade plants have a higher apparent quantum yield (α), lower dark respiration rate (Rd), lower light compensation point (LCP), as well as lower maximum gross photosynthetic rate in saturating light (Pgmax) than sun plants (Givnish 1988; Lee 1996; Yano and Terashima 2004; Pearcy 2007; Hallik et al. 2012; Yang et al. 2013). Sun plants have thicker leaves, higher carotenoid content (Carot), and greater stomatal density when compared to shade plants (Strauss-Debenedetti and Bazzaz 1996). Plants adapted to dry environments exhibit some morphological and physiological characteristics similar to those of plants adapted to high light environments (Chaves et al. 2002; Lüttge 2008), such as mechanisms to dissipate the high excitation states of energy during the photochemistry of photosynthesis. Although analysis of physiological and anatomical changes may elucidate the adaptive strategies of plants in different ecosystems, biomass allocation can also be an excellent descriptor of the ability of plants to cope with multiple environmental limiting factors (Gleeson and Tilman 1992). For example, Markesteijn and Poorter (2009) compared the seedling root morphology and biomass allocation of 62 tropical tree species and reported that the average values of leaf area (LA), specific leaf area (SLA), leaf mass fraction (LMF), and leaf area ratio (LAR) were significant higher in the species of moist rather than dry forests; the average values of root mass fraction (RMF) were significantly higher in the seedlings of dry forest trees. Thus, congeneric species are expected occupy the same position in the ecological succession but have different geographical distributions (moist versus dry tropical forests), similar physiological characteristics in relation to shade tolerance, and different biomass allocation attributes related to the capacity to colonize drought-prone ecosystems.

We considered that I. marginata and I. subnuda have similar positions in the forest succession and different geographic distributions and conducted this study to compare their physiological growth and biomass allocation responses to contrasting light environments. We tested the hypothesis that these two congeneric species have similar photosynthetic and growth responses to contrasting light availability but different biomass allocations.

Materials and methods

Study site, plant materials, and experimental set-up

The experiment was performed at Universidade Estadual de Santa Cruz (UESC), located in Ilhéus, Bahia, Brazil (39°13′59″O; 14°45′15″S). The seedlings were produced in the Instituto Floresta Viva nursery, Serra Grande, Uruçuca, Bahia, Brazil (39°7′60″W, 14°31′0″S). At 120 days after germination, 64 plants of each species were transferred to UESC and transplanted in 1.5-L plastic bags containing the forest soil as the substrate. After 1 week of acclimatization, 54 plants of each species were placed in PVC tubes of 100 mm diameter and 250 mm height. A randomized block design was used, where a single repetition of each light environment was assigned to each block (a total of three treatments and three blocks). In each block, six plants of each species were placed, with a total of 18 plants per light environment. At 80 days after the beginning of the experiment, supplemental fertilization was performed, that is, 50 mL of the following solution was applied to each plant: 10 L of water, 26.34 g of urea, 40 g of monoammonium phosphate (MAP), and 10.68 g of potassium chloride (KCl) as a source of the basic nutrients nitrogen, phosphorus, and potassium, respectively. The experiment had a total duration of 105 days.

The plants were grown at three levels of light availability. The light environments were established using black screens arranged on wooden structures of 1 m width × 1 m length × 0.80 m height. Each light environment was repeated three times. The photosynthetic active radiation (PAR) was monitored in each treatment by using S-LIA-M003 light radiation sensors coupled to Hobo Micro Station Data Logger H21-000 (Onset, USA). Temperature (T) and relative air humidity (RH) were monitored throughout the experiment with an S-THB-M002 sensor coupled to the data logger. The air vapor pressure deficit (VPD) was calculated from the values of T and RH. The station was programmed to collect PAR data every second and store the average values every 10 min. The average daily PAR was calculated from the data stored every 10 min, after considering the time of sunrise and sunset. The mean PAR values throughout the experiment in the three light environments were approximately 36 (L1), 15 (L2), and 6 (L3) mol photons m−2 day−1. During the experiment, the daily mean values of T and VPD were 24.5 °C and 0.63 kPa, and the maximum daily values were 37.6 °C and 3.2 kPa, respectively.

Plant growth and biomass allocation

At the beginning of the experiment, ten plants of each species were used to measure the height (H), stem diameter (D), number of leaves (LN), leaf area (LA), and dry mass of roots (RDM), stems (SDM), leaves (LDM), and their total (TDM). The LA was measured with an LI-3100 area meter (Li-Cor, Inc. Lincoln, Nebraska, USA). The biomass was measured after washing the plants, separating the roots, stems, and leaves and storing them in properly labeled paper bags, and drying all the plant tissues in a forced circulation oven at 50 °C to constant dry mass. At 105 days after the start of the experiment, evaluation of D, H, and LN as well as biomass and LA was performed in all plants of the experiment, followed by the same procedures performed at the beginning of the experiment. From the biomass values, we calculated the following variables: specific leaf area (SLA = LA/LDM), leaf area ratio (LAR = LA/TDM) and stem (SMF = SDM/TDM), leaf (LMF = LDM/TDM) and root (RMF = RDM/TDM) mass fraction. The mean relative growth rate in mass (RGRM) and mean net assimilation rate (NAR) were calculated following Hunt (1990): RGRM = ([1nTDM105 − 1nTDM0]/105) and NAR = [(TDM105 − TDM0)/(LA105 − LA0)] × [(lnLA105 − lnLA0)/105]; where 105 is the total time in days from the beginning to the end of the experiment. Based on the first and last measurements of D and H we calculated the relative growth rates in diameter (RGRD) and height (RGRH) following the same procedure described above for RGRM.

Light–response curves

The light–response curves were created at the end of the experiment by analysis of leaf gas exchange with a portable LI6400 photosynthesis measurement system (Li-Cor, USA) equipped with a 6400-02B RedBlue light source. Measurements were always obtained between 08:00 and 11:00 a.m. from the third fully expanded leaf from the apex of three individuals of each species per block/light environment. The equipment was programmed to provide instantaneous PAR values of 1200, 1000, 800, 600, 400, 200, 100, 50, 25, 10, 5 and 0 μmol photons m−2 s−1 at a chamber temperature of 27 °C and CO2 concentration (394.92 ± 2.69 μmol mol−1). Then, six models were tested for curve fitting (de Lobo et al. 2013). From the lowest value of the sum of the squares and the highest value of R2, we selected the model proposed by Jassby and Platt (1976):

$${\text{Pn}} = {\text{Pg}}{\max} \times \tan{h} \;(\phi(I_{0} ) \times {\text{PAR/Pg}}{\max} )-{\text{Rd,}}$$
(1)
$${\text{LCP}} = \tan h \, ({\text{Rd/}}\phi(I_{0} ) \times {\text{Pg}}{\max} ) /\phi(I_{0} ),$$
(2)
$${\text{LSP}} = \tan{h} \;((( ( {\text{Pg}}{\max} - {\text{Rd)}} \times 0.85) + {\text{Rd}}) / {\text{Pg}}{\max} ) \times {\text{Pg}}{\max} /\phi(I_{0} ).$$
(3)

On the basis of this model, the following parameters were obtained: maximum net photosynthetic rate (Pn), maximum gross photosynthetic rate (Pgmax), apparent quantum efficiency (α), dark respiration rate (Rd), light compensation point (LCP) and light saturation point (LSP).

Pigment content

The contents of chlorophyll a (Chl a), b (Chl b) and carotenoids (Carot) and chlorophyll a/b ratio (Chl a/b) were measured for comparing the acclimation capacity of the light-collecting systems of the two species in the three environments (Boardman 1977; Givnish 1988; Valladares and Niinemets 2008). The total chlorophyll content, in addition to indicating the acclimatization of leaves to light availability, is related to nitrogen and magnesium contents in the leaves and, consequently, to the acquisition of these mineral nutrients (Lambers et al. 2008). For chlorophyll extraction, the third fully expanded leaf of four seedlings per species/block/treatment was used. Chlorophyll measurements were obtained from the same leaves in which photosynthesis were measured. Five leaf discs, whose sum of areas is equal to 1 cm2, were removed from the leaves and immersed in tubes containing 3 mL of dimethylsulfoxide (DMSO) and saturated with 5% CaCO3 (Hiscox and Israelstam 1979). The tubes containing the samples were wrapped in foil for protection from light and maintained for 24 h at room temperature (25 °C). Then, absorbance of the extracts was measured at wavelengths 480, 649, and 665 nm by using a dual beam spectrophotometer (Mesu Lab Enterprise Co, Ltda., China). The contents of chlorophylls and carotenoids were calculated on the basis of area by using the equations proposed by Wellburn (1994).

Statistical analysis

We performed analysis of variance (two-way ANOVA), followed by Tukey’s mean comparison test at 5% level of significance by using statistical software R. A randomized block design was used in a 2 × 3 × 3 factorial scheme with two species, three light environments, three blocks, and six plants per species/block for growth and biomass allocation variables, three plants per species/block for light–response curves, and four plants per species/block for pigment contents.

Results

All the seedlings of the two species survived the different light levels, and no type of pathogen injury or insect attack was observed during the experiment. Independent of light conditions, significant differences between species were observed for RGRM, RGRD, RMF, SMF, LMF, LAR, LN, SLA, Chla + b, Chla/Chlb ratio, and Carot (Table 1). Non-significant differences were observed between species for all light-response curve parameters. Among variables with significant differences between species, the average values of RGRM, RGRD, LMF, LAR, SLA, and Chla + b were higher for I. subnuda than for I. marginata. Conversely, the average values of RMF, SMF, LN, Chla/Chlb ratio, and Carot were higher for I. marginata than for I. subnuda. Independent of the light environment, the average values of LMF, LAR, and SLA were about 16%, 36%, and 34%, respectively, higher for I. subnuda than for I. marginata.

Table 1 Mean values (± se) and significance levels of two-way ANOVA for net assimilation rate (NAR, mg cm2 day−1), relative growth rate in mass (RGRM, mg g−1 day−1), height (RGRH cm cm1 day−1), and diameter (RGRD mm mm−1 day−1), root mass fraction (RMF), stem mass fraction (SMF), leaf mass fraction (LMF), leaf area ratio (LAR cm2 g−1), number of leaves (LN), specific leaf area (SLA cm2 g−1), maximum gross photosynthetic rate (Pgmax, μmol CO2 m−2 s−1), dark respiration rate (Rd, μmol CO2 m−2 s−1), apparent quantum efficiency (α, μmol CO2 μmol photons−1), light compensation point (LCP, μmol photons m−2 s−1), light saturation point (LSP, μmol photons m−2 s−1), chlorophyll a (Chla, mg m−2), chlorophyll b (Chlb, mg m−2), total (Chla + b, mg m−2), Chla/Chlb ratio, carotenoids (Carot, mg m−2) of young plants of Inga spp. subjected to contrasting light conditions for 105 days
Full size table

The light environments significantly influenced NAR, RGRM, RGRH, RGRD, RMF, LMF, LAR, SLA, Rd, LCP, Chla, and Chlb (Table 1). With the increments in light availability, and independent of the species, NAR, RGRM, RGRD, RMF, and Rd increased but RGRH, LMF, LAR, Chla, and Chlb decreased (Table 2). From the highest (L1) to lowest (L3) light conditions, for both species, decreases in order of 27% and 51% were observed for RGRM and NAR, respectively. Conversely, an increase in the order of 78% was observed from L1 to L3 for LAR (Table 2).

Table 2 Mean values (± se) of for net assimilation rate (NAR, mg cm2 day−1), relative growth rate in mass (RGRM, mg g−1 day−1), height (RGRH cm cm1 day−1) and diameter (RGRD mm mm day−1), leaf mass fraction (LMF), root mass fraction (RMF), dark respiration rate (Rd, μmol CO2 m−2 s−1), and contents of chlorophyll a (Chla, mg m−2) and chlorophyll b (Chlb, mg m−2) of young plants of Inga spp. subjected to contrasting light conditions for 105 days
Full size table

Interactions between species and light environments were observed only for LAR, SLA, and LCP (Table 1). The average values of LAR were higher for I. subnuda than for I. marginata in all the light environments. A proportional decrease was observed for I. marginata with increased light availability, but no significant difference for this variable between L2 and L3 was detected for I. subnuda (Table 3). From L1 to L3, increases in LAR in the order of 85% and 73% were observed for I. marginata and I. subnuda, respectively. A similar trend was observed for SLA. From L1 to L3, increases in SLA in the order of 82% and 73% were observed for I. marginata and I. subnuda, respectively. The average values of LCP increased proportionally to the light availability for I. marginata; however, after a significant increase in LCP was observed between L3 and L2, a significant decrease was observed from L2 to L1 for I. subnuda.

Table 3 Mean values (± se) of specific leaf area (SLA cm2 g−1), leaf area ratio (LAR, cm2 g−1), and light compensation point (LCP, μmol photons m−2 s−1) of young plants of Inga spp. subjected to contrasting light environments for 105 days
Full size table

Discussion

Morphological and physiological adjustments such as those that occur in I. marginata and I. subnuda are inherent to the process of acclimatization of plants to contrasting light environments (Valladares and Niinemets 2008). Among the changes, we highlighted the decrease in LAR and increase in NAR with an increase in light availability. The decrease in LAR promotes a decrease in the transpiration demand, whereas the increase in NAR reflects the changes in carbon balance at leaf and plant scales and contributes to higher RGRM. However, the lower values of NAR and RGRM in L2 and L3 indicate impairment of growth of the young plants of I. marginata and I. subnuda below 15 mol photons m−2 day−1. The higher RGRH and lower RGRD under low light availability are recurrent in forest species as a shade escape strategy (Davidson et al. 2002; Yang et al. 2013; Amissah et al. 2015). Higher RGRH, associated with the maintenance of positive photosynthetic rates, indicates that both species can respond to a gap-opening situation (Matsubara et al. 2008). However, the investment in height occurs at the expense of a smaller increment in diameter, as exhibited by the lower values of RGRD in L2 and L3 for both species; this makes them more susceptible to mechanical damage.

With respect to biomass allocation, I. marginata allocated more carbon to roots than I. subnuda did, which may indicate a larger storage capacity that would allow seedlings to survive in limited light environments and have reserves to respond rapidly when light availability increases (DeLucia et al. 1998; Saldanã-Acosta et al. 2009; Yang et al. 2013). However, higher values of LMR can be considered a strategy of I. subnuda to increase the carbon gain (Souza and Válio 2003; Feijó et al. 2009), leading to the higher values of RGRM observed for this species under all light conditions. Plants growing in environments with greater light availability usually have increased leaf temperature and are subjected to a higher evaporative demand of the atmosphere, which leads to higher transpiration rates (Lenhard et al. 2013). Higher values of RMF, and lower values of LAR, under high light availability, as found in I. marginata, can be considered a strategy to maintain higher water supply and hydration of plants (Poorter and Markesteijn 2008; Markesteijn and Poorter 2009; Cortina et al. 2013). However, the higher allocation of biomass to leaves (higher LAR values), associated with the increase in SLA, in environments with low light availability, as found in I. subnuda, indicates a greater investment of carbon to capture light (Santelices et al. 2014; Tang et al. 2015). Thus, the higher mean values of RMF and lower LAR and SLA verified for I. marginata indicate different strategies for occupancy of similar niches for the two species in different ecosystems. I. subnuda invests in light capture (higher LAR and SLA), and I. marginata invests in storage (higher RMF). In addition, the lower values of LAR and SLA and higher values of RMF in I. marginata than in I. subnuda indicate a greater capacity to cope with environments with limited water (Poorter and Markesteijn 2008; Markesteijn and Poorter 2009), which is consistent with its wide geographic distribution (Teixeira et al. 2008; Siminski et al. 2011; Silva et al. 2016).

Under low light availability, plants tend to produce leaves whose epidermis is thinner, with larger surface area per unit mass and more spongy structure (Givnish 1988; Gong et al. 2016). The inherent high values of leaf mass per area, the inverse of SLA, are related to the occurrence of a plant species in dry and nutrient-poor habitats (Wright et al. 2002). The significantly higher values of SLA in I. subnuda than in I. marginata in all the light environments indicate that this species is adapted to environments where water is not a limiting factor; in contrast, I. marginata has the leaf characteristics of environments prone to water deficit. The comparison of SLA between the species is related to the geographic distribution: I. subnuda occurs in the moist forests of the Brazilian Atlantic coast (Peterle et al. 2015; Silva et al. 2015), and I. marginata has a widespread distribution in the central Brazilian ecosystems (Teixeira et al. 2008; Siminski et al. 2011; Silva et al. 2016). The results of the congeneric species of Inga, that is, I. marginata exhibited higher RMF, lower LAR, and lower SLA than I. subnuda, are consistent with those of other studies that compared seedlings and adult trees from forest and savanna formations (Hoffmann and Franco 2003).

At the leaf scale, the carbon balance of both I. marginata and I. subnuda was minimally affected by the light levels because only LCP and Rd were altered by changes in light availability. The decrease in LCP may be related to a significant decrease in Rd when plants grow under limited light conditions. Plants acclimated to shade generally have a lower concentration of photosynthesis-related proteins and enzymes, which reduces their respiratory cost (Griffin et al. 2004). The decrease in metabolic rates, as exhibited by the low values of Rd, associated with the changes in SLA in low light environments, is a typical response of plants to low light availability (Boardman 1977; Givnish 1988; Valladares and Niinemets 2008). In addition, under low irradiance conditions, lower LCP values increase the photosynthetic performance, minimizing CO2 loss by improving carbon gain potential (Walters and Reich 1999; Santos et al. 2012; Yang et al. 2013). However, the absence of significant differences between light environments for Pgmax and LSP indicates that both species were not able to reach the maximum potential photosynthesis values throughout the experiment under light conditions below 15 mol photons m−2 day−1. Thus, based on the data for leaf gas exchange and growth, we can conclude that the responses of both species are typical of trees that cannot tolerate shade in the phase of seedlings and saplings and need a great amount of light to complete the life cycle (Poorter 1999; Davidson et al. 2002; Souza and Válio 2003; Singhakumara et al. 2003; Kelly 2009; Yang et al. 2013).

Changes in RGRM in response to changes in light availability are directly related to both changes in NAR and LAR (Poorter and van der Werf 1998; Poorter 1999). In our study, the decreases in NAR in low light were associated with the limited capacity of both species to reach the maximum potential photosynthesis throughout the experiment under light conditions below 15 mol photons m−2 day−1 (as the mean values of Pgmax and LSP did not change with changes in light availability). However, both species showed increased LMF and LAR and decreased RMF and Rd from the highest (L1) to the lowest (L3) light conditions. A greater dependence of RGRM to NAR at high light and to LAR at low light is well documented in the literature, and it is dependent on changes in the carbon balance at leaf and whole-plant scales (Poorter and van der Werf 1998). For both I. marginata and I. subnuda, the low values of Rd at leaf scale and changes in biomass allocation, i.e., greater LAR and lower RMF, at whole-plant scale may be the main mechanisms of adjustment to low light availability. As discussed above, it was not sufficient to maintain high values of RGRM, and low growth rates below 15 mol photons m−2 day−1 clearly indicate that I. marginata and I. subnuda are typical light-demanding species.

At high light availability, usually, the leaves have a lower concentration of chlorophyll, which decreases their light absorption capacity (Valladares and Niinemets 2008). This pattern was not observed in our study, where, independent of the species, the contents of Chla and Chlb remained constant between treatments. In this scenario, the dissipation of energy becomes extremely important, reducing the risk of photoinhibition. Although the contents of Chla and Chlb were similar among treatments, Chla/Chlb ratio was higher in I. marginata than in I. subnuda under all light conditions. Equivalent results were found in Dalbergia miscolobium (savanna species), whose leaves presented higher Chla/Chlb ratio than Dalbergia nigra (Atlantic rain forest species) under full sunlight (Barros et al. 2012). This alteration in the proportion of chlorophyll contents has already been interpreted as an indication of higher proportion of reaction center to light-harvesting complex, which favors higher rates of photochemical work (Demmig-Adams 1998). However, the increase in Carot content in both species when exposed to the L1 environment indicates that, under this condition, the leaves have greater capacity to dissipate the excess energy of excitation. In this case, the increased Carot content protects the chloroplast membrane from photo-oxidation by eliminating singlet oxygen and extinguishing the triplet state of chlorophyll (Demmig-Adams 1990; Tang et al. 2015). In addition, the higher mean values of Carot in I. marginata than in I. subnuda indicate that this species has more effective mechanisms to dissipate the high excitation states of energy in photosynthesis, which can be an advantage under conditions in which plants need to cope with high light and water deficit (Chaves et al. 2002).

In conclusion, our hypothesis was partially supported because the two species had similar photosynthetic responses to changes in light availability, but differences in growth and biomass allocation. The absence of significant differences between the light environments with respect to Pgmax and LSP indicates that both species were not able to reach the maximum potential photosynthesis values at 15 and 6 mol photons m−2 day−1, demonstrating that they are typical light-demanding species. The higher RGRM of I. subnuda is associated with its higher allocation of biomass to light capture, as shown by the higher values of LMF and LAR. Conversely, the higher values of RMF for I. marginata in contrast to the higher values of LMF, LAR, and SLA for I. subnuda indicate that I. marginata should be better adapted to environments where water could be a limiting factor, which is consistent with its wide geographic distribution.

Author contribution statement

HASP, ACD and MSM designed the experiment. HASP, GSC and ACD performed the experiment. HASP and ACS performed statistical analyses. HASP, ACD, MCS, and MSM wrote the manuscript.