Floral longevity (FL), defined as the length of time that a flower remains open and functional, influences the processes of pollen removal and pollination (Primack, 1985; Ashman and Schoen, 1994). This functional trait varies greatly among species and is an important contributor to increased reproductive success because a longer flowering period can allow plants to attract more pollinators. However, a trade-off may exist between the benefit of increased pollination success and the cost of floral maintenance (Rathcke, 2003). Researchers have suggest that FL is affected by many biotic (Rathcke, 2003; Giblin, 2005; Weber and Goodwillie, 2012) and abiotic factors (Vespirini and Pacini, 2005; Arroyo et al., 2013; Jorgensen and Arathi, 2013). For example, longer flower life spans might be associated with less frequent pollination (Giblin, 2005; Weber and Goodwillie, 2012). Longevity is also improved in plants growing at higher elevations or lower temperatures, or in areas where soil moisture is high (Vespirini and Pacini, 2005; Arroyo et al., 2013; Jorgensen and Arathi, 2013).

Leaf dry mass per unit area (LMA) is a central trait in ecology (Poorter et al., 2009), which shows strong correlations with a suite of important leaf functional traits across a diverse group of species (Small, 1972; Reich et al., 1997; Wright et al., 2004). Longer lived leaves have higher LMA, which influences leaf construction cost and physiological functions such as water transport and use (Fu et al., 2012). A high LMA is typical for stress tolerant species (Wright et al., 2004; Poorter et al., 2009). A direct consequence of greater LMA in leaves with higher lifespan is reduced photosynthetic rate per mass investment. High LMA is thus typically associated with greater drought tolerance (Wright et al., 2004; Poorter et al., 2009). Unlike leaves, flowers do not contribute much to carbon assimilation and are relatively short lived but still may transpire significant amounts of water (Roddy and Dawson, 2012; Teixido and Valladares, 2014). On the contrary, they often experience desiccating conditions that would lead to wilting and prevent pollen dispersal. Therefore, they must maintain water balance and turgor during flowering to attract pollinators. Although longer-lived flowers may increase the opportunities for pollination and reproductive success, especially in habitats where pollinators are in shorter supply, however the attendant costs to those plants are uncertain. This phenomenon of life span versus construction costs has already been described for leaves. In fact, the “leaf economic spectrum” occurs across all major terrestrial plant groups, where a higher (LMA) is invested in longer-lived leaves (Wright et al., 2004; Sack et al., 2013; John et al., 2017). The role of water transport traits in flower evolution is complex, with phylogeny an important determinant of flower hydraulic characteristics (e.g., vein density) that has evolved independently of leaves (Roddy et al., 2013). In addition, the benefits associated with longevity are also very different in flowers when compared to leaves. Thus, it is important to understand whether flowers exhibit trait linkages with longevity similar to those found for leaves.

An adequate water supply is needed during all periods of floral display, including bud expansion, flower opening, nectar production, and the maintenance of turgor in floral organs (Mohan Ram and Rao, 1984; Patino and Grace, 2002; Tsukaguchi et al., 2003; van Doorn and van Meeteren, 2003; Galen, 2005). Flower maintenance may require a considerable amount of water (Roddy and Dawson, 2012) and, under dry conditions, those plants can lose more water from their flowers than from the leaves (Lambrecht, 2013). Therefore, understanding the physiological mechanisms of floral water transport and water relations may provide new insights into the evolution of flowers (Galen, 2000; Chapotin et al., 2003; Feild et al., 2009a,b; Roddy et al., 2016). Even though water has an essential role throughout the floral lifespan, few studies have focused on water relations in flowers (Feild et al., 2009b; Lambrecht et al., 2011; Roddy et al., 2016). Contrasting results in different species have indicated that some flowers are phloem-hydrated (Trolinder et al., 1993; Chapotin et al., 2003). Other studies, such as that involving Magnolia grandiflora, has shown that the giant flowers of that species are hydrated by the xylem (Feild et al., 2009a,b). However, quantitative investigations have been lacking about water relations in flowers from different species, even though they can vary greatly in floral characteristics such as longevity, size, color, and shape.

Orchids, an important group of plants both economically and ecologically, are well-known for their ornamental flowers. In addition, many species show enhanced drought tolerance due to their epiphytic growth habit (Zhang et al., 2012). The genus Paphiopedilum exhibits wide variations in floral life span (from 15 to 60 days) as well as high diversity in morphology and physiology (Karasawa and Saito, 1982; Guan et al., 2011; Zhang et al., 2011). This makes it an ideal system for studying potential functional associations between FL and flower physiology. Moreover, this genus has a well-studied phylogeny that can facilitate phylogenetically based data analyses (Cox et al., 1997) and the interpretation of patterns of functional trait evolution.

Here, we examined FL, floral dry mass per unit area (FMA), and pressure–volume traits using the flowers from 11 Paphiopedilum species growing in southwestern China. Our aim was to identify important functional associations between flower life span and water relations. Specifically, we hypothesized that, similar to leaves, flowers exhibit a positive correlation between cost (in terms of mass per unit area) and longevity, such that flowers from species with greater FL would also have higher FMA values. We also examined whether floral traits are correlated with leaf traits or they are independent of each other due to different selective pressures they experienced.

Within our sample group of 11 species of Paphiopedilum, we found large interspecific diversity in leaf dry mass per unit area, leaf area, flower dry mass per unit area, floral area, FL, turgor loss point, and bulk modulus of elasticity (Table 1). Furthermore, significant relationships were found among traits associated with FL, FMA, and flower maintenance. Whereas, FL was positively correlated with FMA (P = 0.01), but not with FA (P = 0.74) (Figures 3, 4). Even after phylogeny was considered, FL was still correlated with FMA (P = 0.03) (Figure 3). Longevity was correlated negatively with π_tlp (P = 0.02) and positively with ε (P < 0.001) (Figure 5). These correlations of FL with π_tlp and ε remained significant before and after accounting for the phylogenetic relationships (π_tlp : P = 0.04), ε: P < 0.001).

We observed no correlation between FA and LA (P = 0.87), and FA was still not correlated with LA after considering phylogeny (P = 0.47) (Figure 6). Likewise, phenotypically and phylogenetically, FMA showed no relationship with LMA (Pearson correlation: P = 0.53, phylogenetically independent contrast correlation: P = 0.27) (Figure 6).

Our results solidly support the hypothesis that FL is tightly coupled to the cost of flower maintenance in terms of flower dry mass per unit area and drought tolerance. This study provides novel insight into the physiological mechanism for maintaining FL in species of Paphiopedilum, and our data reveal important functional associations related to the maintenance of longer-lived flowers based on water relations in Paphiopedilum.

Our findings support the hypothesis that FMA is positively correlated with FL, and they provide strong evidence that drought tolerance coincides with FL in Paphiopedilum. Species with different flower life spans also vary in their flower water relations. These contrasting strategies among species are vital to their survival because different capacities for water maintenance must mean different costs. Although researchers are beginning to explore these processes in vegetative organs, such as leaves and stems (Fu et al., 2012; Simonin et al., 2012; Carins-Murphy et al., 2014), little attention has been given to their significance in reproductive organs. Further investigations into the costs of sustaining FL in orchids under a range of ecological conditions and amid various growth forms are required if we are to understand how this lineage of species has become so successful.

F-PZ, S-BZ, and HH designed this study; F-PZ, Y-JY, Q-YY, and WZ conducted the experiments; F-PZ performed the analysis and wrote the manuscript; and TB and G-YH helped to evaluate and edit the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.