Contrasting water, dry matter and air contents distinguish orthophylls, sclerophylls and succophylls (leaf succulents)

Differences in leaf texture (hardness, thickness) distinguish orthophylls (soft leaves), sclerophylls (hard leaves) and (semi)succophylls (water-storing leaves). Texture is controlled by dry matter, water and air contents. Our aim was to a) identify the best index of succulence, b) assess how these three components vary with leaf type, and c) derive bounds for these properties among the four main leaf-texture classes. Eight contrasting species from the Namib Desert, South Africa were assessed for their leaf area (A), thickness (z), dry mass (D), saturated water content (Q), and relative volume of dry matter, water and air to derive various indices of leaf texture. Q/A (= QV•z), where QV is saturated water storage per unit volume of leaf, is an ideal index of succulence. Specific leaf area (SLA) is more suitable as an index of hardness (SLA-1 = D/A) but only among non-succulents. Rising leaf specific gravity among sclero-orthophylls is due to replacement of air by dry matter but water among succophylls. Collation of 13 worldwide studies showed that orthophylls can be distinguished by Q/A ≤ 0.45 mg water mm-2 leaf surface from succophylls with ≥ 0.9, such that there is a divergent relationship among plants regarding their water-storing properties. Semi-succophylls can be defined as having a Q/A > 0.45 to < 0.9, and sclerophylls can be separated from orthophylls by a SLA ≤ 10 mm2 mg-1 dry mass. The distribution of these leaf texture classes may vary greatly within, and especially between, local floras.

Methods
Fieldwork

Leaves were collected from eight species growing wild at Groenriviersond, 500 km north of Cape Town, South Africa (30º 51' S, 17º 34' E). The species were selected to cover the full range of textures among perennials at the study site [specific leaf area (SLA) ranged from 2 to 20 mm2 mg-1, Lamont and Lamont 2000]. They were Pteronia onobromoides (Asteraceae, shrub to 50 cm tall, hard-leaved), Salvia lanceolata (Lamiaceae, shrub to 1 m tall, soft-leaved), Eriocephalus africanus (Asteraceae, shrub to 1 m tall, soft-leaved), Stoeberia utilis (Aizoaceae, syn. Mesembryanthemaceae, ground creeper, succulent), Ruschia fugitans (Aizoacae, syn. Mesembryanthemaceae, ground creeper, large-leaved succulent), Zygophyllum morgsana (Zygophyllaceae, shrub to 50 cm, semi-succulent), Othonna cylindrica (Asteraceae, shrub to 40 cm, succulent) and Senecio aff. sarcoides (Asteraceae, undershrub, small-leaved succulent). Nomenclature is as given in Eccles et al. (1999) and, from hereon, only the genus names are used. Leaves of all species were iso(bi)lateral and sessile (except Salvia).  As this is a shrubland, all species were growing in the open so that differences in microclimate would have had no role in affecting the results. They varied from apparently sclerophyllous to highly succulent. On a water mass content per unit volume basis, Qv, two species were in the range 40-50%, three were 60-70%, and three were 80-95% (Lamont and Lamont 2000). Thus, the water-storing properties of the eight species studied formed a well-defined gradient that proved ideal for testing the hypotheses outlined here.

The study site lies in the southern portion of the Namib Desert. The vegetation is part of the succulent karoo and consists of clumps of climbers to woody shrubs up to 2 m tall (Eccles et al. 2001). The soil is red aeolian sand overlying an impenetrable silcrete hardpan at about 2 m depth. Rainfall was 79 mm in the year of the study although fog and dew are regular occurrences (Fradera-Soler et al. 2021).

Laboratory work

Current season’s mature twigs (100–150 mm long) were removed from side branches of 6–8 plants of each species by cutting under water predawn. They were kept with their ends in water at 17.5–20.5ºC and covered with plastic bags for 1–4 days in the laboratory to promote full hydration. They were then recut under water and their pressure-volume relations determined following the protocol of Radford and Lamont (1992). The balancing pressure was achieved with a digital pressure chamber, model 1003, PMS Instruments, Corvallis, OR, USA. In order to obtain turgid (saturated) mass as needed for this study, wet weight values of twigs were extrapolated to Y  = 0, i.e., full turgidity. Ten mature, full-sized leaves were removed from other stems, and these plus the original supporting twigs used were weighed, frozen at -16ºC to rupture the cells and hasten drying, dried at 72ºC for 48 h and reweighed. From this, turgid mass of the twigs was used to obtain leaf turgid mass (60–95% of total mass for individual twigs).

Midpoint thickness of 10 leaves from three plants was determined with callipers. Projected area (A) was obtained by placing 30 leaves or more diagonally on the conveyor belt of an area meter (Li-Cor 3000, Lincoln, NK, USA). Adjustments were made for the shape of leaves and their volume (V) determined geometrically (Lamont et al. 2015): five were cylindrical (V = p/4z•A) where z was diameter, two were laminate (V = z•A where z was thickness) and one was subulate (V = mean z•A), all lacking midribs. SLA [A/D = (DV•z)-1, where D = dry leaf mass and DV = dry leaf density on a volume basis, Witkowski and Lamont 1991] was adjusted for leaf shape in the same way, and DV and QV (dry matter and saturated water mass per unit leaf volume) were based on these measurements.

Volume of dry matter was determined by removing all, and only, mature leaves from six twigs, bulking and macerating after oven-drying as above to pass through a 1.1 mm mesh, then twice through a 0.3 mm mesh. The powder was then moistened with a wetting agent (1% Tween 20) in distilled water to form a thick paste. A cork borer (internal diameter 3.58 mm) was pushed into the paste, to produce an initial firm cylinder 20–40 mm in length. It was then placed on a paper tissue to absorb water over plastic sheeting on a fibrocement base. An iron rod of diameter 3.50 mm was pushed into the borer and tapped with a small hammer about 30 times, until water no longer squeezed out of the bottom. The pressure applied was up to 5.1 kg cm-2 but it was usually about 2.1 kg cm-2. The cylinder of compressed paste was forced out with the rod, and the ends cut with a razor blade as required to produce a perfect cylinder and its length and width determined with callipers. 3–5 cylinders were obtained per species. They were dried at 65ºC for 40 h and kept in a desiccator until weighing.

Properties assessed

Knowing Dv (D/V) and volume of dry matter per unit dry mass (VD/D), the contribution of dry matter volume – essentially cell walls, but including protein and most solutes) – to total volume [(VD/D)(D/V) = VD/V = FD] and air, Fa = (1 – FD) were calculated. Thus, colloidal protein and other non-soluble substances were put with structure rather than cytoplasm or vacuole when estimating volume fractions (as in Roderick et al. 1999b). Some solutes may not have been adsorbed or held back by the cell-wall components during compression, but, even if some were lost, their contribution to volume would be negligible (< 0.01% of dry matter according to our estimates). Formulae for specific gravity (r) based on either non-air leaf volume (rQ+D) or total leaf volume (rl) (see Table 1) were as given in Roderick et al. (1999a).

Z Wang et al. (2022) fitted a single relationship between SLA and Q/D for over 3,000 species distributed throughout most of the world’s vegetation types, especially in China and North America, that did not appear to include succophylls (confirmed in a later figure). We therefore decided to compare Q/D with SLA (A/D) (as did they among many other structural and ecophysioloical properties of the species), whose slope produces Q/A – our preferred index of succulence, with all data sets which did include succulents that we could locate: ours and 12 others (Table 3). These were identified by feeding in the keywords: water content, succulent, SLA, into Google Scholar®. This yielded about 200 papers that were inspected to see if they provided data on Q/M and SLA, or they could be obtained from leaf dry matter (LDMC) or water content on a turgid mass basis. LDMC = D/(D + Q) that was inverted and 1 subtracted to give Q/D. Q/(Q + D) values were inverted, 1 subtracted, and re-inverted to give Q/D. Some data were obtained by measuring the length of bars in figures with callipers (to 0.05 mm) and converted to the indices of interest using the scale on the axis. Some data sets had to be rejected as their units were clearly incorrect and the solution was not obvious, while some listed papers proved impossible to obtain.

Half the papers used fresh mass to calculate Q that is misleading as it should be based on turgid mass if it is to be treated as a property of the plant rather than as a proximate response to the conditions of the day. No attempts to at least harvest the plants predawn were evident (see above for how turgid mass can be obtained, including extrapolation from pressure-volume curves). In this case, Q/D was multiplied by 1.1 (i.e., water content increased by 10%) to adjust for the error unless it was grown in hydroponics or water content exceeded 90% when any error would be small. For similar reasons, plants subjected to drought stress treatments were ignored.  In passing, we note that many species had leaves with a QM of only 50% or less and wonder how successful the claims of turgid mass were.

叶片质地（硬度、厚度）的差异可区分软质叶（orthophylls）、硬叶（sclerophylls）和（半）肉质叶（(semi)succophylls，储水叶片）。叶片质地受干物质、水分和空气含量调控。本研究的目标为：a）确定最优的肉质性指标；b）评估上述三种组分如何随叶片类型变化；c）推导四类主要叶片质地类群中这些性状的取值范围。本研究对采自南非纳米布沙漠的8个对比物种进行了测定，测量了叶面积（A）、厚度（z）、干质量（D）、饱和含水量（Q）以及干物质、水分和空气的相对体积，以计算多种叶片质地指标。其中，Q/A（= QV·z，其中QV为单位叶体积的饱和储水量）是理想的肉质性指标。比叶面积（SLA）更适合作为硬度指标（SLA⁻¹ = D/A），但该关系仅适用于非肉质植物。硬-软质叶的叶片比重随干物质取代空气而上升，而肉质叶的比重则随水分变化。对全球13项研究的汇总分析显示，软质叶的Q/A ≤ 0.45 mg水分·mm⁻²叶表面积，肉质叶的Q/A ≥ 0.9，由此可见植物的储水特性存在分化关系。（半）肉质叶可定义为Q/A介于0.45~0.9之间的类群，硬叶与软质叶可通过SLA ≤ 10 mm²·mg⁻¹干质量进行区分。这些叶片质地类群的分布在区域植物区系内部，尤其是不同区系之间差异显著。

## 研究方法
### 野外工作
本研究于南非开普敦以北500 km的Groenriviersond（30°51′S，17°34′E）采集8种野生植物的叶片。所选物种覆盖了研究区域多年生植物的全部质地范围[比叶面积（SLA）范围为2~20 mm²·mg⁻¹，Lamont和Lamont 2000]，分别为：*Pteronia onobromoides*（菊科，灌木，株高可达50 cm，硬叶）、*Salvia lanceolata*（唇形科，灌木，株高可达1 m，软叶）、*Eriocephalus africanus*（菊科，灌木，株高可达1 m，软叶）、*Stoeberia utilis*（番杏科，同物异名：Aizoaceae，原Mesembryanthemaceae，匍匐草本，肉质）、*Ruschia fugitans*（番杏科，同物异名：Mesembryanthemaceae，匍匐草本，大叶肉质）、*Zygophyllum morgsana*（蒺藜科，灌木，株高可达50 cm，半肉质）、*Othonna cylindrica*（菊科，灌木，株高可达40 cm，肉质）以及*Senecio aff. sarcoides*（菊科，亚灌木，小叶肉质）。物种命名遵循Eccles等（1999）的标准，后文仅使用属名进行指代。所有物种的叶片均为等面（两面型）叶且无柄（除*Salvia*外）。由于研究区域为灌丛地，所有物种均生长于开阔生境，因此微气候差异不会对实验结果产生影响。供试物种的叶片质地从典型硬叶到高度肉质呈连续梯度分布。按单位体积含水量计算，2个物种的Qv（单位体积饱和储水量）为40%~50%，3个为60%~70%，剩余3个为80%~95%（Lamont和Lamont 2000）。因此，本研究中8个物种的储水特性形成了清晰的梯度，非常适合验证本文提出的假说。

研究区域位于纳米布沙漠南部，植被属于肉质卡鲁（succulent karoo）植被区，涵盖从藤本到株高可达2 m的木本灌丛（Eccles等2001）。土壤为红色风成沙，下方约2 m处为不透水的硅质硬磐。研究当年降雨量为79 mm，但该区域常年有雾和露水出现（Fradera-Soler等2021）。

### 实验室工作
于黎明前从每个物种的6~8株植物的侧枝上剪取当季成熟的嫩枝（长100~150 mm），剪取过程始终保持切口浸入水中。将嫩枝末端浸入水中，置于17.5~20.5℃的实验室环境中，并用塑料袋覆盖1~4天以促进充分水合。随后再次将切口浸入水中重新剪切，按照Radford和Lamont（1992）的方法测定其压力-体积曲线。使用美国俄勒冈州科瓦利斯市PMS Instruments公司生产的型号为1003的数字压力室获得平衡压力。为获取本研究所需的膨润（饱和）质量，将嫩枝的鲜重外推至水势Y=0的状态，即完全膨润状态。从其余茎秆上摘取10片成熟完整的叶片，连同原支撑嫩枝一同称重，随后在-16℃下冷冻以破裂细胞并加速干燥，再于72℃下烘干48 h后再次称重。通过该步骤得到的嫩枝膨润质量可推算叶片膨润质量（占单嫩枝总质量的60%~95%）。

使用游标卡尺测定3株植物的10片叶片的中点厚度。将30片及以上叶片斜置于叶面积仪（Li-Cor 3000，美国内布拉斯加州林肯市）的传送带上，获取叶片投影面积（A）。根据Lamont等（2015）的方法，通过几何计算对叶片形状和体积（V）进行校正：其中5片为圆柱形（V=π/4·z·A，z为直径），2片为扁平形（V=z·A，z为厚度），1片为锥形（V=平均z·A），所有叶片均无中脉。比叶面积（SLA）[A/D=(DV·z)⁻¹，其中D为叶片干质量，DV为单位体积干叶密度，Witkowski和Lamont 1991]按照相同方式针对叶片形状进行校正，DV和QV（单位叶体积的干物质和饱和水质量）均基于上述测量结果计算。

干物质体积的测定方法为：从6根嫩枝上摘取全部成熟叶片，按照前述方法烘干后混合研磨，依次通过1.1 mm和0.3 mm筛网。将粉末用含有1%吐温20（Tween 20）的蒸馏水润湿，制成浓稠的糊状。使用内径3.58 mm的软木打孔器将糊状物料压制成初始长度20~40 mm的紧实圆柱。将圆柱置于铺有塑料膜的纤维水泥底座的纸巾上吸水，随后将直径3.50 mm的铁棒插入打孔器，用小锤轻敲约30次，直至底部不再有水挤出。施加的压力最高可达5.1 kg·cm⁻²，通常约为2.1 kg·cm⁻²。用铁棒将压实的糊状圆柱推出，必要时用剃须刀片切割两端以获得完美的圆柱，再用游标卡尺测量其长度和直径。每个物种可获得3~5个圆柱样品，将其在65℃下烘干40 h后置于干燥器中保存至称重。

### 测定的性状
已知Dv（D/V）和单位干质量的干物质体积（VD/D），可计算干物质体积（主要为细胞壁，也包含蛋白质和大部分溶质）占总体积的比例[(VD/D)(D/V)=VD/V=FD]以及空气体积占比Fa=1-FD。因此，在估算体积分数时，胶体蛋白质和其他不溶性物质被归类为结构组分，而非细胞质或液泡（如Roderick等1999b所述）。部分溶质在压缩过程中可能未被细胞壁组分吸附或截留，但即使存在少量流失，其对体积的贡献可忽略不计（据本研究估算，不足干物质的0.01%）。基于非空气叶体积（rQ+D）或总叶体积（rl）的比重（r）计算公式详见Roderick等（1999a）（见表1）。

Wang等（2022）针对全球多数植被类型（尤其中国和北美）的3000余种植物建立了SLA与Q/D的单一回归关系，但该数据集似乎未包含肉质叶类群（后续图表验证了这一点）。因此，本研究决定将Q/D与SLA（A/D）进行对比（该团队此前也曾针对物种的多种结构和生理生态性状开展类似分析），其斜率即可得到Q/A——本研究优选的肉质性指标。本研究收集了所有可获取的包含肉质植物的数据集，包括本团队的数据集以及另外12项已发表数据集（见表3）。通过在谷歌学术（"Google Scholar®"）中以"water content, succulent, SLA"为关键词检索获取相关文献，共得到约200篇论文，筛选其中提供Q/M与SLA数据，或可通过叶干物质含量（leaf dry matter content, LDMC）或膨润质量基准含水量推导得到相关数据的文献。叶干物质含量LDMC=D/(D+Q)，取其倒数并减1即可得到Q/D；Q/(Q+D)值取倒数、减1后再次取倒数，也可得到Q/D。部分数据通过游标卡尺测量图表中柱形的高度（精度至0.05 mm），并结合坐标轴刻度换算得到目标指标。部分数据集因单位明显错误且无法修正而被剔除，部分已列出的文献也无法获取。

半数文献使用鲜质量计算Q，这一做法存在误导性：若将含水量作为植物的固有性状而非当日环境的即时响应，则Q应基于膨润质量计算。此类文献均未尝试至少在黎明前采集样品（详见前文获取膨润质量的方法，包括通过压力-体积曲线外推）。针对此类文献，将Q/D乘以1.1（即含水量提升10%）以校正误差，但若植物为水培种植或含水量超过90%，则误差可忽略不计。出于类似原因，本研究排除了经历干旱胁迫处理的植物。顺带一提，本研究注意到许多物种的叶片QM仅为50%或更低，不禁质疑其膨润质量测定结果的可靠性。