Cladophora Kützing (1843) is one of the largest green-algal genera and has a worldwide distribution. Within the family Cladophoraceae, the genus Cladophora is characterized by its simple thallus architecture with branched, uniseriate filaments of multinucleate cells. It is a heterogeneous and confusing assemblage of species with varied appearances. Because most morphological characteristics used to define species, such as branching pattern, color, cell dimensions, overlap among described species, vary with age and environment, taxonomic identification within the genus has challenged phycologists for decades (van den Hoek, 1963, 1982; Söderström, 1963, 1965; Dodds and Gudder, 1992). More than 600 species have been attributed to the genus over the past 200 years. While more than 400 species were turned out to be synonymous ones (van den Hoek, 1963, 1982; van den Hoek C and Chihara, 2000). At the present, there are only 196 species accepted taxonomically (Guiry, 2018).

Cladophora gracilis was first reported by Griffiths ex Harvey (1834) as Conferva gracilis, and then transferred to Cladophora by Kützing (1845). This species was widely distributed around the world (Adams, 1907; Taylor, 1957, 1960; Yoshida, 1998; Lee and Kang, 2001). Van den Hoek (1963) listed it as a synonym of C. sericea. For seaweed classification, a number of characteristics such as thallus height, color and cell size are dependent on plant maturity and environmental conditions (Mathieson et al., 1981). Due to various morphology characteristics, C. gracilis had been reported many varieties (Taylor, 1957). Hence, in nature, the extensive morphology plasticity makes it problematic to delineate the species. With the development of DNA sequencing, the difficulty of morphological taxonomy motivated our use of molecular techniques to assist classification. 18S rDNA and nuclear ribosomal DNA internal transcribed spacer regions (ITS1 and ITS2) had been employed in a range of phycological studies (Bakker et al., 1992, 1994; Zechman et al., 1994; Goff et al., 1994). In present study, we aimed to clarify the extremely extensive phenotypic plasticity of C. gracilis based on molecular analysis, and tried to explore the availability of it DNA barcode.

As a heterogeneous assembles of species, the taxonomy of Cladophora is always problematic. Our morphological observation and molecular classification strongly revealed the widely intraspecific plasticity, and prompted us to reconsider the reliability of the taxonomy criteria.

One of the most valuable taxonomic criteria in genus Cladophora is its branching system. While the nine samples of C. gracilis showed highly variable branching system, making it difficult to classify. Some specimens including AST2010015 and AST2010028 are characterized by a distinctly sparse ramification, the thalli being reduced to main axes. These old plants’ reduced branching system interfered with the accurate taxonomy, and then molecular assistance was resorted. Additionally, specimen AST2010027 may be confused with C. catenata due to the sparse, irregular and wide-angled branches as well as the small plants (2–5 cm) (Figs 1p-q). This specimen was collected at sea beach, exposed to fierce wave action. The small plants entangled together and floated with seawater, under which circumstance they developed into irregular branches. While they can be distinguished from each other by cell size, with cell diameter 200–500 μm of C. catenata, obviously larger than that of C. gracilis. Branching pattern can also be variable. Branches from main axes were fastigiated, verticillate for AST2010017, alternate for AST2009001, AST2010004 and AST2010005, opposite for AST2010015 and AST2010028. For most samples, the ultimate branches were secund, such as AST2009001, AST2010004, AST2010005, AST2010010 and AST2010029. Besides, ultimate branch systems were often more or less incurved or recurved, such as AST2009001. Generally, most specimens had 1–3 branches per node, with the exception of sample AST2010017. There were up to 5 branches per node, forming fasciculate terminal branch-systems, which belonged to the feature of C. fascicularis. This kind of interspecific overlapped characteristics usually leads to incorrect taxonomy result. From what has been discussed above, it can be drown that branching system may vary widely among intraspecific populations and should not be considered as stable taxonomic criterion.

Thallus color as well as height was directly visible traits of plants, while both of which were also highly variable. Among the surveyed samples, thallus color varied from dark green to light green. Especially, for AST2010010, there is obviously distinct color within the same individual. Part of the plant grew at shaded localities, its robust filaments showing dark green habit. While for other parts exposed to direct sunlight, its relatively thin filaments exhibited light green (Figs 1i-k). This conspicuous contrast weakened the reliability of thallus color as taxonomic character. Besides, the height of full grown plants was variable too, ranging from 5 cm to 20 cm. Likewise, this character was influenced by environmental conditions. Exposure to wave action promoted the formation of short, compact tufts (van den Hoek, 1982). Thus, as stated above, the sample AST2010027 formed short, spherical tufts. For most other samples attaching at rigid rocks of sheltered water pools, such as AST2010004, AST2010005 and AST2010029, filaments length can reached to 20 cm. We can conclude from above that thallus habits including color and height exist extensive intraspecific plasticity for C. gracilis, and care must be taken when these characteristics are used for distinction of the species.

As one of the most important characteristics, diameter of main axes, branches and ultimate branches are also highly variable. Of course, the distinction between slender species, such as C. flexuosa and very robust species such as C. catenata is obvious, but the existence of a series of samples with intermediate sizes makes the use of this characteristic problematic (Table 2). For the reported C. gracilis, the diameter of main axes is 100–140 μm, that of branches is 70–125 μm, and that of ultimate branches is 50–75 μm (Ding and Luan, 2013). Yet from Table 2, we can see that not all the samples corresponded to that criterion. Obviously, delineating species is difficult because most easily observed morphological traits are highly variable. This extensive phenotypic variety makes molecular approach play a significant role in classification. Analysis of both the conserved 18S rDNA and fast-evolving ITS sequences demonstrated that the nine samples belonged to the same evolution entity. They can be applied as DNA barcode for accurate anchoring this species.

It is clear that there is extensive intraspecific heterogeneity in traditional measured morphological traits, making it difficult to develop a reliable method of distinguishing species. In view of the difficulty in morphology taxonomy of Cladophora, molecular data has potential feasibility assisting traditional classification methodology.

We are grateful to Saren Gaowa, Wei Zhang, Tao Wang, Chunmei Bao for their kind laboratory technical assistance, and the members of marine biological museum of IOCAS for their help.